Hans J. Herrmann, Lucas Böttcher - Computational Statistical

Index
ab-initio molecular dynamics, 230 bond percolation, 22, 94, see percolation
ab-initio simulation, 222 bootstrap percolation, 34
absorbing phase, 127, see directed percolation Boris Lubachevsky, 194
acceptance probability, 74, see Markov chain Born rule, 223
adaptive mesh, 174, see Adaptive Born–Oppenheimer approximation, 226
Particle-Particle-Particle-Mesh (AP3 M) bosons, 224
method boundary conditions, 83
Adaptive Particle-Particle-Particle-Mesh (AP3 M) box counting method, 36, see fractal dimension
method, 172 Box–Muller trick, 13, see transformation method
additive RNG, 4, see random-number generator broad histogram method, 101, see histogram
(RNG) method
adjacency matrix, 50, see network Brownian motion, 3, 42, see random walk
Amdahl’s law, see parallelization techniques, 126 burning method, 19–20, see percolation
Aneesur Rahman, 186
angular momentum, 152, 165, see classical
canonical coordinates, 55, see statistical mechanics
mechanics
canonical ensemble, 57, see physical ensemble
angular velocity, 165, see classical mechanics
canonical pressure ensemble, 57, 58, see physical
anomalous diffusion, 46, see random walk
ensemble
antiferromagnetic, 60
canonical velocity distribution, 182
Antonio Coniglio, 96
Car–Parrinello method, 231
autocorrelation function, 86
cellular automata, 133, 215
classical mechanics, 149
ballistic deposition, 141
Barabási–Albert network, 52, see network cloud-in-cell, 172, see Particle-Mesh (PM) method
Bellman-Ford-Moore algorithm, 54, see cluster algorithm, 96
shortest-path length cluster-size distribution, 25, 27, 32
Benoit B. Mandelbrot, 34 clustering coefficient, 52, see network
Berendsen barostat, 185 clustering instability, 198
Berendsen thermostat, 178 collision angle, 192
Berni Alder, 150 commutator, 224
betweenness centrality, 53, see network complementary error function, 171
Binder cumulant, 89, see statistical mechanics congruential RNG (multiplicative), 4–6, see
binomial distribution, 43 random-number generator (RNG)
bipartite, 116 Coniglio–Klein clusters, 96
body-fixed coordinate system, 164, see classical conservation law, 151, see classical mechanics
mechanics constraint force, 161
Boltzmann distribution, 76, see statistical contact dynamics, 208
mechanics contact force, 209
Boltzmann equation, 215 contact process, 127, see non-equilibrium process
Boltzmann machine, 114, see machine learning contact time, 152
252
Index 253
contrastive divergence, 117, see restricted dynamic load sharing, 125, see parallelization
Boltzmann machine techniques
Control parameter, 59, see statistical mechanics dynamical critical exponent, 86
convolution theorem, 172
coordination number, 22, see degree Eden model, 140
corrections to scaling, 90, see scaling edge list, 50, see network
correlation function, 29, 42, 64 edges, 50, see network
correlation length, 30, 65, 87, 106 Edsger W. Dijkstra, 53
Correlation test, 10, see random-number generator Edwards–Wilkinson (EW) equation, 142, see
(RNG) irreversible growth
Coulomb friction, 190 effective mass, 195
Coulomb graph, 208, 211 elastic collision, 187
covering, 36, see fractal dimension electric field, 175
Creutz algorithm, 81, 137, see statistical mechanics ellipsoidal particles, 202
critical exponent, 30, 62, 64, 110, 129 energy, 78, see statistical mechanics
critical phenomena, 17 energy conservation, 152, see conservation law
critical point, 31, 59, 105 ensemble average, 56, 57, 67, 70, see statistical
critical region, 30–31 mechanics
critical slowing down, 86, 96 ensemble method, 40, see fractal dimension
critical temperature, 60, see Ising model, 84 entropic sampling, 102
cross product, 165, see classical mechanics entropy, 58, see statistical mechanics
cube test, 5, 9, see random-number generator equipartition theorem, 177, 184
(RNG) Erdõs–Rényi network, 52, see network
ergodicity, 75, see Markov chain
D2Q9 model, 217, see lattice Boltzmann method ergodicity hypothesis, 56, see statistical mechanics
damped oscillator, 188 Ernst Ising, 60
Daniel T. Gillespie, 130 Euler angles, 166, 167, see classical mechanics
data collapse, 31, 43, 66, 88, see scaling Euler’s equations, 165, see classical mechanics
David Landau, 61, see statistical mechanics event-driven molecular dynamics, 192
decimation, 107, see renormalization group Ewald sum, 170
degree, 22, 51, see coordination number, see exchange operator, 223
network exponential distribution, 13, see transformation
degree distribution, 51, see network method
degrees of freedom, 163, see classical mechanics
Denis J. Evans, 166 Family–Vicsek scaling, 138
density functional theory (DFT), 228 farming, 122, see parallelization techniques
Derrick H. Lehmer, 4 fast Fourier transform (FFT), 173, 174
detailed balance, 76, 77, 79, 96, see Markov chain Fereydoon Family, 138
dielectric medium, 175 fermions, 224
Dietrich Stauffer, 27, 28 Fernando Alonso-Marroquin, 205
diffusion equation, 45, see random walk ferromagnetic, 60
diffusion-limited aggregation, 142, see irreversible finite size scaling, 87
growth finite-size effects, 37
Dijkstra algorithm, 54, see shortest-path length first-order phase transition, see phase transition, 92
Direct simulation Monte Carlo (DSMC), 219 fixed point, 105, 108
directed network, 50, see network flat histogram method, 102, see histogram method
directed percolation, 127, see non-equilibrium fluctuating phase, 127, see directed percolation
process fluctuation-dissipation theorem, 63, see Ising
discrete element method (DEM), 190 model, 111
discrete Fourier transform (DFT), 174 fluid flow, 215
dissipation, 187 forward Euler method, 155, see integration method
Dissipative Particle Dynamics (DPD), 220 Fourier transform, 173
domain decomposition, 125, 159, see fractal dimension, 33–35
parallelization techniques friction angle, 190
Donald E. Knuth, 159 friction coefficient, 190
drop, see interface, 84 fugacity, 48
254 Index
Game of Life, 136, see cellular automata interface, 83

Gaussian charge distribution, 171 interface width, 84
Gaussian distribution, 13, see transformation internal energy, 58, see statistical mechanics
method irreversible growth, 138
Gaussian quadrature, 217 Ising model, 59, see statistical mechanics
Gaussian-type orbitals, 225
Gene M. Amdahl, 125 Jacobian, 110
general gradient approximation (GGA), 229 Jean Jacques Moreau, 209
generalized dimensions, 38, see multifractality Jian-Sheng Wang, 102
generating function, 48, 94 Joan Adler, 34
Geoffrey Hinton, 112 John H. Conway, 136
George E. P. Box, 13 John Hammersley, 17
George Marsaglia, 5 John Hopfield, 114
George Pólya, 45 Josiah W. Gibbs, 57
Gibbs free energy, 58, see statistical mechanics Julia M. Yeomans, 220
Gillespie algorithm, 130, see kinetic Monte Carlo
Glauber dynamics (heat bath Monte Carlo), 80, 96, Kadhar–Parisi–Zhang (KPZ) equation, 142, see
114, see statistical mechanics irreversible growth
gold colloids, 35, 36, see fractal dimension Kasteleyn–Fortuin theorem, 95
gradient descent, 115, 117, see machine learning Kawasaki dynamics, 81, 83, see statistical
Graeme Bird, 219 mechanics
grand canonical ensemble, 57, see physical kinetic energy, 150, see classical mechanics
ensemble kinetic gelation, 31
gravitational interactions, 175 kinetic growth walk, 48, see self-avoiding walk
Green’s function, 172 kinetic Monte Carlo, 130
ground state, 226 kink-jump algorithm, 48, see self-avoiding walk
growth exponent, 138, see irreversible growth Knudsen number, 219
Kohn–Sham approximation, 228
Hamilton equations, 151, see classical mechanics Kohn–Sham potential, 228
Hamilton operator, 223 Kullback–Leibler divergence, 115, see relative
Hamiltonian, 56, see statistical mechanics entropy
Hans C. Andersen, 184 Kurt Binder, 89
hard-sphere potential, 151 Kyozi Kawasaki, 81, see statistical mechanics
Harry E. Stanley, 98
Hebbian learning, 113, see machine learning lab-fixed system, 164, see classical mechanics
Heisenberg model, 98, see n-vector model Lagged Fibonacci RNG (additive), 7–8, see
Hellmann–Feynman theorem, 229 random-number generator (RNG)
Helmholtz free energy, 58, see statistical mechanics Lagrange multiplier, 161
Hénon map, 39, see multifractality Laplace Gillespie algorithm, 132, see Gillespie
Herman Berendsen, 179 algorithm
Hertzian contact, 188 Lars Onsager, 60, 175
Hilbert space, 222 lattice Boltzmann method (LBM), 215
histogram method, 100 lattice gas automata, 215
Hohenberg–Kohn theorems, 227 leapfrog method, 154, see integration method
Hooke’s law, 189 Lennard–Jones potential, 151
Hopfield network, 112, see machine learning Leo P. Kadanoff, 106
Hoshen–Kopelman algorithm, 24, 26 Leonhard M. Sander, 143
hyperscaling, 65, see statistical mechanics linear correlation function, 86
linked-cell method, 159
importance sampling, 68, 74, see Monte Carlo Liouville’s theorem, 56, 182
inclined plane, 190 Liouville’s theorems, see statistical mechanics
in-degree, 51, see degree local-density approximation (LDA), 228
inelastic collapse, 198 long-range potential, 158, 169, 173
inelastic collision, 187 lookup table, 78, 157, 192
inertia tensor, 165, see classical mechanics Loup Verlet, 154
instantaneous temperature, 177 Lu Jeu Sham, 228
integration method, 153 Lubachevsky method, 194
Index 255
Lucilla de Arcangelis, 144, 145 non-smooth mechanics, 208

Ludwig Boltzmann, 55 normal restitution coefficient, 197
Lyapunov exponent, 155 Nosé–Hoover thermostat, 179
numerical error, 154
M(RT)2 algorithm, 77, 96, see statistical mechanics
machine learning, 112 optimization, 119
magnetic susceptibility, 62, 88, see Ising model order parameter, 23, 24, 60, see statistical
magnetization, 61, see Ising model mechanics, 127
majority rule, 106 out-degree, 51, see degree
many-body wave function, 223
Marian Boguñá, 131 (pseudo) random-number generator (RNG), 4
Markov chain, 74, see Monte Carlo PageRank, 43, see random walk
Marsaglia’s “Diehard“ tests, 10 parallelization techniques, 122
Marsaglia’s theorem, 5, see congruential RNG Parrinello–Rahman barostat, 186
master equation, 75, see Markov chain Particle-Mesh (PM) method, 171
maximal period, 4, see random-number generator Particle-Particle-Particle-Mesh (P3 M) method, 172
(RNG) partition function, 48, 58, 94, 106, 182
Maxwell–Boltzmann distribution, 178, 216 Paul Ehrenfest, 59
McCulloch-Pitts neuron, 112, see machine learning Paul Erdõs, 46
mean-square displacement, 45 Paul J. Flory, 47
measure, 38, see multifractality Paul Peter Ewald, 169
Mersenne Twister, 9, see random-number Pauli exclusion principle, 224
generator (RNG) Paulo Murilo Castro de Oliveira, 101
Mersenne prime, 4–5 Per Lötstedt, 208
Mervin E. Muller, 13 percolation, 17
Michael Creutz, 82 percolation threshold, 19, 22, see percolation
Michele Parrinello, 231 permutation, 120
microcanonical ensemble, 57, 58, see physical persisting contact, 210
ensemble Phase, 223
microcanonical Monte Carlo, 82, see Creutz phase space, 55, see statistical mechanics
algorithm phase transition, 58, see statistical mechanics
Minkowski cow, 206 physical ensemble, 57, see statistical mechanics
modulo (mod) operator, 3–4 Pierre C. Hohenberg, 226
moment of inertia, 164, see classical mechanics Pieter W. Kasteleyn, 94
momentum, 152, see classical mechanics pivot algorithm, 49, see self-avoiding walk
Monte Carlo, 67 plastic deformation, 189
multi canonical Monte Carlo, 102 Poisson brackets, 56, see statistical mechanics
multi-particle collision dynamics (MPC), 218, see Poisson equation, 172
stochastic rotation dynamics (SRD) Poisson process, 131
multi-spin coding, 122, see parallelization polygon, 205
techniques polymer, 47
multifractality, 37, see fractal dimension potential energy, 150, see classical mechanics
multiplicative RNG, 4, see random-number Potts model, 93, see statistical mechanics
generator (RNG) pseudo random numbers, 3
Pulay forces, 225
n-vector model, 98
Navier–Stokes equations, 215 Q2R, 82, see Creutz algorithm
nearest-grid point, 172, see Particle-Mesh (PM) quantum mechanics, 222
method quaternions, 167
network (graph), 50, 94
Newton’s equations of motion, 151, see classical radius of gyration, 40
mechanics random deposition, 138
nodes, 50, see network random deposition with surface diffusion, 139
non-equilibrium process, 127 random numbers, 3
non linear correlation function, 85 random walk, 42, 45
non-Markovian Gillespie algorithm, 131, see Raoul Kopelman, 25
Gillespie algorithm Raymond Kapral, 219
256 Index
reaction-field method, 175 spreading rate, 128, see contact process

regular diffusion, 45, see random walk square test, 5, 9, see random-number generator
regular network, 51, see network (RNG)
rejection method, 15, 48, see self-avoiding walk statistical mechanics, 55
relative entropy, 115, see Kullback–Leibler Stefan Luding, 199
divergence Stephen Wolfram, 135
Renfrey B. Potts, 93 stochastic method, 183
renormalization group, 105 stochastic rotation dynamics (SRD), 218
reptation algorithm, 48, see self-avoiding walk strange attractor, 39, see multifractality
response function, 62, see statistical mechanics Sundman time transformation, 180
restitution coefficient, 187 superellipsoid, 204
restricted Boltzmann machine, 116, see Boltzmann supervised learning, 112, see machine learning
machine susceptible-infected-susceptible (SIS) model,
restricted solid on solid model (RSOS), 140 128–129
Robert Brown, 42 Swendsen–Wang algorithm, 96, see cluster
Robert C. Tausworthe, 7 algorithm
Robert D. Carmichael, 4
Robert Swendsen, 111 Tamás Vicsek, 138
Roberto Car, 231 tangential restitution coefficient, 196
rotating particles, 198 TC model, 199
rotation group, 164, see classical mechanics thermal equilibrium, 57, see statistical mechanics
roughening transition, see interface, 84 thermal inertia, 180
Roy J. Glauber, 79 Thomas Witten, 143
time average, 55, see statistical mechanics
sandbox method, 39, see multifractality time-dependent Schrödinger equation, 223
Sauro Succi, 217 time-independent Schrödinger equation, 223
scale invariance, 106, see renormalization group torque, 163, 164, 212, see classical mechanics
scaling, 27, 91 transformation method, 12
scaling function, 27, 32, see scaling transition probability, 74, see Markov chain
scaling relations, 65, see statistical mechanics trapezoidal rule, 69
Scott Kirkpatrick, 120 traveling salesman problem, 120
Sean McNamara, 198 triangle, 52, see network
second moment, 28, 31
second-order phase transition, 59, 90, see phase umbrella sampling, 103, see histogram method
transition unbiased estimator, 70
seed, 4, 8, see random-number generator (RNG) uncorrelated configurations, 87
self-avoiding walk, 47, see random walk undirected network, 50, see network
self-similarity, 34, see fractal dimension universality, 24
shortest-path kinetic Monte Carlo algorithm, 132, universality classes, see statistical mechanics, 66
see Gillespie algorithm unsupervised learning, 112, see machine learning
shortest-path length, 33, 52, see network upper critical dimension, 65, see statistical
Shuichi Nosé, 180 mechanics
Sierpinski triangle, 35–37, see fractal dimension
sigmoid function, 114 vectorization, 124, see parallelization techniques
Signorini graph, 208, 211 velocity rescaling, 178
Signorini problem, 208 Verlet method, 154, 164, see integration method
simulated annealing, 119, see optimization Verlet table, 158
site percolation, 22, see percolation volatile fractal, 42
Slater determinant, 224 von Kárman vortex street, 218
small-world network, 52, see network
smoothed particle hydrodynamics, 221 Walter Kohn, 228
Sol–gel transition, 18 Wang–Landau method, 102, see histogram method
specific heat, 63, see Ising model water molecule, 161
sphere-packing problem, 71, see Monte Carlo Watts–Strogatz network, 52, see network
spherical model, 98, see n-vector model Wilhelm Lenz, 59
spherpolygon, 205 William G. Hoover, 179
spontaneous magnetization, 61 William Klein, 96
Index 257
Wolff algorithm, 97, see cluster algorithm XY model, 98, see n-vector model
woodpecker toy, 208
wrapping probability, 21, see percolation Zenon paradox, 199
Zierler Trinomial condition, 8, see Lagged
XOR function, 122 Fibonacci RNG (additive)
Computational Statistical Physics
Providing a detailed and pedagogical account of the rapidly growing field of com-
putational statistical physics, this book covers both the theoretical foundations of
equilibrium and nonequilibrium statistical physics and modern, computational appli-
cations such as percolation, random walks, magnetic systems, machine learning
dynamics, and spreading processes on complex networks. A detailed discussion of
molecular dynamics simulations, a topic of great importance in biophysics and phys-
ical chemistry, is also included. The accessible and self-contained approach adopted
by the authors makes this book suitable for teaching courses at the graduate level, and
numerous worked examples and end-of-chapter problems allow students to test their
progress and understanding.
Lucas Böttcher is Assistant Professor of Computational Social Science at Frankfurt

School of Finance and Management and Research Scientist at UCLA’s Department
of Computational Medicine. His research areas include statistical physics, applied
mathematics, complex systems science, and computational physics. He is interested
in the application of concepts and models from statistical physics to other disciplines,
including biology, ecology, and sociology.
Hans J. Herrmann is Directeur de Recherche at CNRS in Paris, visiting Professor at the

Federal University of Ceará in Brazil and Emeritus of ETH Zürich. He is a Guggen-
heim Fellow (1986), a member of the Brazilian Academy of Science, and a Fellow
of the American Physical Society. He has been the recipient of numerous prestigious
awards, including the Max-Planck Prize (2002), the Gentner-Kastler Prize (2005),
the Aneesur Rahman Prize (2018), and an IBM Faculty Award (2009), and he has
received an ERC Advanced Grant (2012). He is Managing Editor of International
Journal of Modern Physics C and Granular Matter. He has coauthored more than 700
publications and coedited 13 books.
Computational Statistical Physics
LUC AS B ÖT TC HER
Frankfurt School of Finance and Management, University of California, Los Angeles
HANS J. HERRMANN
Centre National de la Recherche Scientifique (CNRS), Federal University of Ceará, ETH Zürich
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India
103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467
Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.
www.cambridge.org
Information on this title: www.cambridge.org/9781108841429
DOI: 10.1017/9781108882316

c Lucas Böttcher and Hans J. Herrmann 2021
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2021
Printed in the United Kingdom by TJ Books Limited, Padstow Cornwall
A catalogue record for this publication is available from the British Library.
ISBN 978-1-108-84142-9 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Dedicated to all students who attended our lectures on Computational Physics.
— Lucas Böttcher and Hans J. Herrmann
I also dedicate this book to my great-grandparents Doris and Gerhard.
— Lucas Böttcher
Contents
Preface page xi
What is Computational Physics? xiii
Part I Stochastic Methods 1
1 Random Numbers 3
1.1 Definition of Random Numbers 3
1.2 Congruential RNG (Multiplicative) 4
1.3 Lagged Fibonacci RNG (Additive) 7
1.4 Available Libraries 8
1.5 How Good is an RNG? 9
1.6 Nonuniform Distributions 11
2 Random-Geometrical Models 17
2.1 Percolation 17
2.2 The Sol–Gel Transition 17
2.3 The Percolation Model 19
2.4 Fractals 34
2.5 Walks 42
2.6 Complex Networks 50
3 Equilibrium Systems 55
3.1 Classical Statistical Mechanics 55
3.2 Ising Model 59
4 Monte Carlo Methods 67

4.1 Computation of Integrals 67
4.2 Integration Errors 68
4.3 Hard Spheres in a Box 71
4.4 Markov Chains 74
4.5 M(RT)2 Algorithm 76
4.6 Glauber Dynamics (Heat Bath Dynamics) 79
4.7 Binary Mixtures and Kawasaki Dynamics 80
4.8 Creutz Algorithm 81
vii
viii Contents
4.9 Boundary Conditions 83

4.10 Application to Interfaces 83
5 Phase Transitions 85
5.1 Temporal Correlations 85
5.2 Decorrelated Configurations 87
5.3 Finite-Size Scaling 87
5.4 Binder Cumulant 89
5.5 First-Order Transitions 90
6 Cluster Algorithms 93
6.1 Potts Model 93
6.2 The Kasteleyn and Fortuin Theorem 94
6.3 Coniglio–Klein Clusters 95
6.4 Swendsen–Wang Algorithm 96
6.5 Wolff Algorithm 97
6.6 Continuous Degrees of Freedom: The n-Vector Model 98
7 Histogram Methods 100

7.1 Broad Histogram Method 101
7.2 Flat Histogram Method 102
7.3 Umbrella Sampling 103
8 Renormalization Group 105

8.1 Real Space Renormalization 105
8.2 Renormalization and Free Energy 105
8.3 Majority Rule 106
8.4 Decimation of the One-Dimensional Ising Model 107
8.5 Generalization 109
8.6 Monte Carlo Renormalization Group 111
9 Learning and Optimizing 112

9.1 Hopfield Network 112
9.2 Boltzmann Machine Learning 114
9.3 Simulated Annealing 119
10 Parallelization 122
10.1 Multispin Coding 122
10.2 Vectorization 124
10.3 Domain Decomposition 125
11 Nonequilibrium Systems 127

11.1 Directed Percolation and Gillespie Algorithms 127
11.2 Cellular Automata 133
11.3 Irreversible Growth 137
Contents ix
Part II Molecular Dynamics 147
12 Basic Molecular Dynamics 149

12.1 Introduction 149
12.2 Equations of Motion 150
12.3 Contact Time 152
12.4 Verlet Method 153
12.5 Leapfrog Method 154
13 Optimizing Molecular Dynamics 157

13.1 Verlet Tables 158
13.2 Linked-Cell Method 158
14 Dynamics of Composed Particles 161

14.1 Lagrange Multipliers 161
14.2 Rigid Bodies 163
15 Long-Range Potentials 169

15.1 Ewald Summation 169
15.2 Particle-Mesh Method 171
15.3 Reaction Field Method 175
16 Canonical Ensemble 177

16.1 Velocity Rescaling 178
16.2 Constraint Method 179
16.3 Nosé–Hoover Thermostat 179
16.4 Stochastic Method 183
16.5 Constant Pressure 184
16.6 Parrinello–Rahman Barostat 185
17 Inelastic Collisions in Molecular Dynamics 187

17.1 Restitution Coefficient 187
17.2 Plastic Deformation 188
17.3 Coulomb Friction and Discrete Element Method 190
18 Event-Driven Molecular Dynamics 192

18.1 Event-Driven Procedure 192
18.2 Lubachevsky Method 194
18.3 Collision with Perfect Slip 195
18.4 Collision with Rotation 196
18.5 Inelastic Collisions 197
18.6 Inelastic Collapse 198
19 Nonspherical Particles 202

19.1 Ellipsoidal Particles 202
19.2 Polygons 205
19.3 Spheropolygons 205
x Contents
20 Contact Dynamics 208

20.1 One-Dimensional Contact 208
20.2 Generalization to N Particles 212
21 Discrete Fluid Models 215

21.1 Lattice Gas Automata 215
21.2 Lattice Boltzmann Method 215
21.3 Stochastic Rotation Dynamics 218
21.4 Direct Simulation Monte Carlo 219
21.5 Dissipative Particle Dynamics 220
21.6 Smoothed Particle Hydrodynamics 221
22 Ab Initio Simulations 222

22.2 Implementation of Wave Functions 225
22.3 Born–Oppenheimer Approximation 225
22.4 Hohenberg–Kohn Theorems 226
22.5 Kohn–Sham Approximation 228
22.6 Hellmann–Feynman Theorem 229
22.7 Car–Parrinello Method 230
References 232
Index 252
Preface
This textbook was born from the lecture notes of a course given first at the University
of Stuttgart (1999–2006) as a core subject of the bachelor curriculum in computational
physics and then at ETH Zurich as part of the bachelor and master curriculum in
physics and computational science and engineering. Over the years, the course was
continuously modified according to current scientific findings and breakthroughs in
the fields of statistical physics and complex systems science.
The book is divided into two parts: Stochastic Methods (Part I) and Molecular
Dynamics (Part II). In Part I, we cover computational approaches to studying per-
colation, random walks, spin systems, and complex networks. We include examples
and exercises to illustrate that computational methods are invaluable in obtaining
further insights into certain systems when analytical approaches are not feasible. In
the context of Boltzmann machines, we also describe recent developments in con-
necting statistical physics concepts to machine learning. We conclude Part I with
a discussion of computational methods to study nonequilibrium systems and high-
light applications in modeling epidemic spreading, opinion dynamics, and irreversible
growth.
In Part II, we focus on molecular dynamics and establish a connection between the
study of microscopic particle interactions and their emerging macroscopic properties
that can be analyzed statistically. We provide an overview of different simulation tech-
niques that enable the reader to computationally study a broad range of interaction
potentials and particle shapes. We discuss thermostat and barostat methods to simu-
late molecular interactions in canonical temperature and pressure ensembles. For the
simulation of rigid particles, we cover the method of contact dynamics. To account
for quantum-mechanical effects in molecular-dynamics simulations, we conclude with
outlining the basic concepts of density functional theory and the Car–Parrinello
method.
Throughout the book, we point to applications of the discussed mathematical and
computational methods in different fields besides physics, including engineering, social
sciences, and biology. In addition to providing a solid theoretical background in com-
putational statistical physics, the aim of the book is to inspire the reader to develop
their own simulation codes and computational experiments. We therefore include a
carefully chosen selection of detailed exercises, which will enable the reader to imple-
ment and better understand the discussed theory and algorithms. For the interested
reader, we also include side notes and information boxes that explore certain topics in
greater depth.
xi
xii Preface
We assume that the reader is familiar with at least one programming language and
has some basic knowledge in classical mechanics, electrodynamics, statistical physics,
and numerical mathematics.
This book benefited from the comments of many colleagues. In particular, we
thank Stefan Luding, Malte Henkel, Dirk Kadau, Lorenz Müller, Marco-Andrea
Buchmann, Nicola Offeddu, Marcel Thielmann, Madis Ollikainen, Alisha Föry, Josh
LeClair, Fernando Alonso-Marroquin, Yupeng Jiang, and Thomas Asikis for their
feedback. We also thank Giovanni Balduzzi for his support with performing differ-
ent simulations. Moreover, we are grateful to the numerous scientists in the fields of
computational and statistical physics for contributing photos that we included in bio-
graphical panels to highlight their work. Historically, a gender imbalance has been
very much ingrained in computational physics, but fortunately the situation is starting
to change.
What is Computational Physics?
Computational physics is the study and implementation of numerical algorithms to

solve problems in physics by means of computers. As we will see throughout the
book, finding numerical solutions to a given problem is useful because there are only
very few systems that can be solved analytically. In particular, computational physics
methods are used to simulate many-body particle systems. The simulated “virtual real-
ity” is sometimes referred to as the third branch of physics (between experiment and
theory) [1].
The analysis and visualization of large data sets that are generated numerically and
experimentally is also part of computational physics but will not be treated in the
present book.
Computational physics plays an important role in the following fields:
• Computational fluid dynamics (CFD): solving and analyzing problems that involve
fluid flows
• Classical phase transitions: percolation, critical phenomena
• Solid state physics (quantum mechanics)
• High-energy physics/particle physics: in particular, lattice quantum chromodynam-
ics (“lattice QCD”)
• Astrophysics: many-body simulations of stars, galaxies, etc.
• Geophysics and solid mechanics: earthquake simulations, fracture, rupture, crack
propagation, etc.
• Agent-based modeling (and interdisciplinary applications): complex networks in
biology, economy, social sciences, and other disciplines
xiii
PART I
STOCH ASTIC METHODS

Random Numbers 1
Random numbers (RNs) are important for scientific simulations. We will see that they
are used in many different applications, including the following:
• To simulate random events in experimental data (e. g., radioactive decay)

• To simulate thermal fluctuations (e. g., Brownian motion)
• To incorporate a lack of detailed knowledge (e. g., traffic or stock market simula-
tions)
• To test the stability of a system with respect to perturbations
• To perform random sampling
1.1 Definition of Random Numbers
Random numbers are a sequence of numbers in random order. The probability that a
given number occurs next in such a sequence is always the same. Physical systems can
produce random events, for example, in electronic circuits (electronic flicker noise) and
in systems where quantum effects play an important role (e. g., radioactive decay or
photon emissions in semiconductors). However, physical random numbers are usually
bad in the sense that they are often correlated and not reproducible.
Generating random numbers with algorithms is also a bit problematic because
computers are completely deterministic, while randomness should be nondetermin-
istic. We therefore must content ourselves with generating so-called pseudo-random
numbers, which are computed with deterministic algorithms based on strongly non-
linear functions. These numbers should follow a well-defined distribution and should
have long periods. Furthermore, we should be able to compute them quickly and in a
reproducible way.
A very important function for generating pseudo-random numbers is the modulo
operator mod (% in C++). It determines the remainder of a division of one integer
number by another one.
Given two numbers a (dividend) and n (divisor), we write a modulo n or a mod n,
which represents the remainder of dividing a by n. For the mathematical definition of
the modulo operator, we consider the integers a, q, and r. We then express a as
a = nq + r , (1.1)
3
4 Random Numbers
where r (0 ≤ r < |n|) is the remainder (i. e., the result of a mod n). One useful property
of the mod operator for generating RNs is that it is a strongly nonlinear function.
We distinguish between two classes of pseudo-random number generators (RNGs):
multiplicative and additive RNGs.
• Multiplicative RNGs are simpler and faster to program and are based on integers.
• Additive RNGs are more difficult to implement and are based on binary variables.
In the following sections, we describe these RNGs in detail and outline different
methods that allow us to examine the quality of random sequences.
1.2 Congruential RNG (Multiplicative)
The simplest form of a congruential RNG [2, 3] was proposed by Lehmer (see
Figure 1.1). It is based on the mod operator.
Congruential RNG
To define the congruential RNG, we choose two integer numbers, c and p, and a seed value x0 such that
c, p, x0 ∈ Z. We then generate the sequence xi ∈ Z, i ∈ N iteratively according to
xi = (cxi−1 ) mod p . (1.2)
This iteration generates random numbers in the interval [0, p − 1]1 . To transform a
random number xi ∈ [0, p − 1] into a normalized random number zi ∈ [0, 1), we simply
divide xi by p and obtain
xi
0 ≤ zi = < 1, (1.3)
p
where zi is a rational number (i. e., zi ∈ Q). The random numbers zi are homogeneously
distributed, which means that every number between 0 and 1 is equally probable.
Figure 1.1 Derrick H. Lehmer Since all integers are smaller than p, the sequence must repeat after maximally (p−1)
(1905–1991) was an American iterations. Thus, the maximal period of the RNG defined by eq. (1.2) is (p − 1). If we
mathematician who worked on number pick the seed value x0 = 0, the sequence remains at this value. Therefore, x0 = 0 cannot
theory and theory of computation. be used as a seed of the described congruential RNG.
In 1910, the American mathematician Robert D. Carmichael [4] proved that the
maximal period of a congruential RNG can be obtained if p is a Mersenne prime
number and if it is the smallest integer number that satisfies
c p−1 mod p = 1 . (1.4)
1 Throughout the book, we adopt the notation that closed square brackets [ ] in intervals are equivalent
to “≤” and “≥” and parentheses ( ) correspond to “<” and “>,” respectively. Thus, the interval [0, 1]
corresponds to 0 ≤ x ≤ 1, x ∈ R and (0, 1) means 0 < x < 1, x ∈ R.
1.2 Congruential RNG (Multiplicative) 5
A Mersenne number is defined as Mn = 2n − 1 with n ∈ N. If the number M is prime,

it is referred to as Mersenne prime. As of May 2021, only 51 Mersenne primes were
known. The largest has 24,862,048 digits and was discovered by Patrick Laroche in
December 2018 within the Great Internet Mersenne Prime Search (GIMPS) [5]. In
1988, Park and Miller [6] proposed the following numbers to generate the maximal
period of congruential RNGs, here, in pseudo-code:
c o n s t i n t p =2147483647;
c o n s t i n t c =16807;
Figure 1.2 Cube test: plot of consecutive
random numbers xi , xi+1 , xi+2 with 15
i n t rnd =42; / / s e e d
clearly visible planes (“RANDU” algorithm
with c = 65539, p = 231 , x0 = 1).
rnd =( c ∗ rnd)%p;
p r i n t rnd;
The number p = 2147483648 is a Mersenne prime and corresponds to the maximal

integer of 32 bits: 231 − 1.
To assess the homogeneity of numbers generated by an RNG, we can either plot two
consecutive pseudo-random numbers (xi , xi+1 ) (the so-called square test) or employ
a “cube test” for three consecutive numbers (xi , xi+1 , xi+2 ). In Figure 1.2, we show
an example of an RNG that clearly fails the cube test. All numbers that are gen-
erated by this particular RNG lie on just 15 parallel planes, indicating correlation
effects that may lead to undesired effects in simulations of stochastic processes. In
1968, George Marsaglia (see Figure 1.3) showed that all congruential RNGs as
defined in eq. (1.2) produce pseudo-random numbers that lie on equally spaced hyper- Figure 1.3 George Marsaglia
planes [8]. The distance between these planes decreases with the length of the period. (1924–2011) made many contributions
For the interested reader, we briefly summarize the corresponding theorem on the to improving and testing RNGs [7].
following page.
Additional Information: Marsaglia’s Theorem

Let {zi } = {xi /p} (i ∈ N) denote normalized numbers of an RNG sequence, and let π1 = (z1 , . . . , zn ),
π2 = (z2 , . . . , zn+1 ), π3 = (z3 , . . . , zn+2 ), . . . be the points in the unit n-cube formed by n succes-
sive numbers zi . Marsaglia showed that all points π1 , π2 , . . . lie on parallel hyperplanes [8]. Formally, if
a1 , a2 , . . . , an ∈ Z is any choice of integers such that
a1 + a2 c + a3 c2 + · · · + an cn−1 ≡ 0 mod p,
then all the points π1 , π2 , . . . will lie in the set of parallel hyperplanes defined by the equations
a1 y1 + a2 y2 + · · · + an yn = 0, ±1, ±2, . . . , yi ∈ R, 1 ≤ i ≤ n.
6 Random Numbers
Additional Information: Marsaglia’s Theorem (cont.)

There are at most
|a1 | + |a2 | + · · · + |an |
of these hyperplanes that intersect the unit n-cube. Note that there is always a choice of a1 , . . . , an such
that all of the points fall in fewer than (n!p)1/n hyperplanes. The theorem can be proven in four steps:
Step 1: If
a1 + a2 c + a3 c2 + · · · + an cn−1 ≡ 0 mod p,
one can prove that
a1 zi + a2 zi+1 + · · · + an zi+n−1
is an integer for every i.
Step 2: The point πi = (zi , zi+1 , . . . , zi+n−1 ) must lie in one of the hyperplanes
a1 y1 + a2 y2 + · · · + an yn = 0, ±1, ±2, . . . , yi ∈ R, 1 ≤ i ≤ n.
Step 3: The number of hyperplanes of the above type, which intersect the unit n-cube, is at most
|a1 | + |a2 | + · · · + |an | .
Step 4: For every multiplier c and modulus p, there is a set of integers a1 , . . . , an (not all zero) such that
a1 + a2 c + a3 c2 + · · · + an cn−1 ≡ 0 mod p
and
|a1 | + |a2 | + · · · + |an | ≤ (n!p)1/n .
This is of course only the outline of the proof. The exact details are described in Marsaglia (1968) [8]. In a
similarway, it is possible to show that for congruential RNGs the distance between the planes must be larger
p
than n
.
Exercise: Congruential RNG

Task 1
Write a program that generates random numbers using a congruential random number generator. First, use
small values for c (< 50) and p (< 50), for example, c = 3 and p = 31.
• Check for correlations using the square test. That is, create a plot of two consecutive random numbers
(xi , xi+1 ). (What is the maximum number of random numbers that have to be created until you can see
all possible lines/planes for this specific random number generator?)
• Create a corresponding 3D plot for the cube test.
• Do the same for other random number generators (at least one!), for example, by changing c and p. You
may also compare your results to those obtained with C++ built-in generators, such as rand() and
drand48().
1.3 Lagged Fibonacci RNG (Additive) 7
Exercise: Congruential RNG (cont.)

Task 2
Generate a homogeneous distribution of random points inside a circle. How should the coordinates r and φ
be chosen using uniformly distributed random numbers?
Task 3
Test your RNG for different c and p using the χ2 test:
• Divide the range of random numbers into k bins. That is, divide the range into discrete intervals of the same
size, so that the probability of a random number lying in interval i is pi = 1/k.
• Using each RNG, generate at least one sequence of n numbers. For each sequence, determine Ni , the
number of random numbers in interval i (choose n such that all npi ≥ 5).
• Compute the χ2 value for one specific sequence s of random numbers

k
(Ni − npi )2
χ2s = .
i=1
npi
Use the table from Knuth [9] to check if the random numbers are uniformly distributed and compare the
quality of different RNGs.
• Calculate χ2s for different sequences (i.e., different seeds of the RNG). You can then plot the cumulative
probability for the χ2 in comparison to the theoretically expected values (values again from the table of
Ref. [9]).
1.3 Lagged Fibonacci RNG (Additive)
A slightly more complicated RNG is the lagged Fibonacci algorithm proposed by

Robert C. Tausworthe (see Figure 1.4) in 1965 [10]. Lagged Fibonacci-type generators
can achieve extremely long periods and allow us to make some predictions about their
underlying correlation effects.
To define lagged Fibonacci RNGs, we start with a sequence of binary numbers
xi ∈ {0, 1} (1 ≤ i ≤ b). The next bit in our sequence xb+1 is

xb+1 = xb+1− j mod 2 , (1.5)
j∈J

where J ⊂ [1, . . . , b]. The sum j∈J xb+1− j includes only a subset of all the other
bits, so the new bit could, for instance, simply be based on the first and third bits, Figure 1.4 Robert C. Tausworthe is a
xb+1 = (x1 + x3 )mod 2 (or any other subset). We now focus on the properties of retired senior research engineer at the Jet
RNGs that are defined by eq. (1.5) and consider a two-element lagged Fibonacci Propulsion Laboratory, California Institute
generator. of Technology.
8 Random Numbers
Two-Element Lagged Fibonacci RNG

Let c, d ∈ {1, . . . , b} with d ≤ c. The RNG sequence elements xb+1 are recursively generated according to
xb+1 = (xb+1−c + xb+1−d ) mod 2 .
We immediately see that we need some initial sequence of at least c bits to start from
(a so-called seed sequence). Except for all bits equal to zero, any other seed configu-
ration can be used. One possibility is to use a seed sequence that was generated by a
congruential RNG.
The maximum period of sequences that are generated by the outlined two-element
lagged Fibonacci RNG is 2c −1. As for congruential RNGs, there are conditions for the
choice of the parameters c and d to obtain the maximum period. For lagged Fibonacci
generators, c and d must satisfy the Zierler–Trinomial condition, which states that the
trinomial
T c,d (z) = 1 + zc + zd , (1.6)
where z is a binary number, cannot be factorized into subpolynomials (i.e., is prim-
itive). A possible choice of numbers that satisfy the Zierler condition is (c, d) =
(250, 103). The generator is named after Kirkpatrick (see Figure 9.10) and Stoll, who
proposed these numbers in 1981 [11].
Some examples of known pairs (c, d) follow:
(c, d)
(250, 103) Kirkpatrick–Stoll (1981) [11]
(1.7)
(132049, 54454) J. R. Heringa et al. (1992) [12]
(6972593, 3037958) R. P. Brent et al. (2003) [13].
We may use one of the following methods to convert the obtained binary sequences to
natural numbers (e.g., 32 bit unsigned integers):
• Running 32 lagged Fibonacci generators in parallel (this can be done very effi-
ciently): The problem with this method is the initialization, because all 32 initial
sequences have to be uncorrelated. The quality of the initial sequences has a major
impact on the quality of the produced random numbers.
• Extracting a 32-bit-long part from the sequence: This method is relatively slow
because for each random number we have to generate 32 new elements in the binary
sequence.
1.4 Available Libraries
In general, we should always make sure that we use a high-quality RNG. For some
purposes it might be sufficient to use drand48 (in C/C++). If your compiler supports
1.5 How Good is an RNG? 9
the C++11 standard (or above), there are different implementations already avail-
able in the “random” library. For example, linear congruential RNGs are available
by calling minstd_rand from the linear_congruential_engine class. A useful general-
purpose RNG is the so-called Mersenne twister [14], which was developed in 1997 by
Makoto Matsumoto and Takuji Nishimura. The Mersenne twister has some struc-
tural similarities to lagged Fibonacci RNGs and belongs to the class of generalized
feedback shift register algorithms. The name “Mersenne twister” was chosen because
its period length is a Mersenne prime. In C++, we can use the Mersenne twister to
generate uniformly distributed random numbers as follows:
# i n c l u d e <i o s t r e a m >
# i n c l u d e <random>
u s i n g namespace s t d ;
i n t main ( )
{
random_device rd;
mt19937 mt_rng ( rd ( ) ) ;
u n i f o r m _ r e a l _ d i s t r i b u t i o n <double > u ( 0 . 0 , 1 . 0 ) ;
double rnd = u ( mt_rng ) ;
c o u t << rnd << e n d l ;
return 0;
}
In the above code listing, we use the random_device as seed for the mt19937 generator.
This generator has a very long period of 219937 − 1. For the sake of reproducibility of
numerical results, it is recommended to store all seeds in a file. In PYTHON, no further
efforts are required to use the Mersenne twister because the mt19937 is the core RNG
in all PYTHON distributions.
1.5 How Good is an RNG?
There are many possibilities to test how “random” a certain RNG sequence is. Possible
tests for a given sequence {si }, i ∈ N include the following:
1. Square test (see Section 1.2 for details)

2. Cube test (see Section 1.2 for details)
10 Random Numbers
3. Average value: the arithmetic mean of all numbers in the sequence {si } should cor-
respond to the analytical mean value. Let us assume here that the numbers si are
rescaled to lie in the interval si ∈ [0, 1). The arithmetic mean should then be
1
N
1
s̄ = lim si = . (1.8)
N→∞ N i=1 2
The more numbers that are averaged, the better 21 will be approximated.
4. Fluctuation of the mean value (χ2 test): the distribution around the mean value
should behave like a Gaussian distribution.
5. Spectral analysis (Fourier analysis): Let {si } denote values of a function. It is possi-
ble to perform a Fourier transform of such a function by means of the fast Fourier
transform (FFT; see details in Section 15.2.1). If the frequency distribution corre-
sponds to white noise (uniform distribution), the randomness is good; otherwise,
peaks will show up (resonances).
6. Correlation test: Analysis of correlations such as
si si+d − s2i , (1.9)
for different values of d.
Of course, this list is not complete. Many other tests can be used to check the quality
of RNG sequences.
Probably the most famous set of RNG tests is the Marsaglia’s “Diehard” set. These
Diehard tests are a battery of statistical tests for measuring the quality of a set of
random numbers. They were developed over many years and published for the first
time by Marsaglia on a CD-ROM with random numbers in 1995 [15]. Marsaglia’s
tests were inspired by different applications, and each can measure different types of
correlations.
Additional Information: Marsaglia’s “Diehard” Tests
• Birthday spacings: If random points are chosen in a large interval, the spacings between points should be
asymptotically Poisson distributed. The name stems from the birthday paradox.1
• Overlapping permutations: When analyzing five consecutive random numbers, the 120 possible orderings
should occur with equal probability.
• Ranks of matrices: A number of bits of some number of random numbers is formed into a matrix over {0,1}.
The rank of this matrix is then determined, and the ranks are counted.
• Monkey test: Sequences of some number of bits are taken as words, and the number of overlapping words
in a stream is counted. The number of words not appearing should follow a known distribution. The name
is based on the infinite monkey theorem.2
• Parking lot test: Randomly place unit circles in a 100 × 100 square. If the circle overlaps an existing one,
try again. After 12,000 tries, the number of successfully “parked” circles should follow a certain normal
distribution.
Additional Information: Marsaglia’s “Diehard” Tests (cont.)
• Minimum distance test: Find the minimum distance of 8,000 uniformly randomly placed points in a
10,000 × 10,000 square. The square of this distance should be exponentially distributed with a certain
mean.
• Random spheres test: Put 4,000 randomly chosen points in a cube of side length 1,000. Now a sphere is
placed on every point with a radius corresponding to the minimum distance to another point. The smallest
sphere’s volume should then be exponentially distributed.
• Squeeze test: 231 is multiplied by random floats in [0, 1) until 1 is reached. After 100,000 repetitions, the
number of floats needed to reach 1 should follow a certain distribution.
• Overlapping sums test: Sequences of 100 consecutive floats are summed up in a very long sequence of
random floats in [0, 1). The sums should be normally distributed.
• Runs test: Ascending and descending runs in a long sequence of random floats in [0, 1) are counted. The
counts should follow a certain distribution.
• Craps test: 200,000 games of craps3 are played. The number of wins and the number of throws per game
should follow a certain distribution.
1
The birthday paradox states that the probability of two randomly chosen persons having the same birthday in a group of 23 (or more) people
is more than 50 percent. For 57 or more people, the probability is already larger than 99 percent. Finally, for at least 366 people, the probability
is exactly 100 percent. This is not paradoxical in a logical sense; it is called a paradox nevertheless since intuition would suggest probabilities
much lower than 50 percent.
2
The infinite monkey theorem states that a monkey hitting keys at random on a typewriter keyboard for an infinite amount of time will almost
surely (i.e., with probability 1) write a certain text, such as the complete works of William Shakespeare.
3
A dice game.
1.6 Nonuniform Distributions
Thus far, we have only considered uniform distributions of pseudo-random numbers.

The congruential and lagged Fibonacci RNGs produce numbers that can easily be
mapped to the interval [0, 1) or any other interval by simple shifts and multiplications.
However, if we want to generate random numbers that are distributed according to
a certain distribution (e. g., a Gaussian distribution), the algorithms presented so far
are not able to do so. However, we may employ techniques that permit us to trans-
form uniform pseudo-random numbers to other distributions. There are essentially
two different ways to perform this transformation:
• We can apply transformation methods if the target distribution is known analytically,

is integrable, and the resulting expression is invertible.
• However, if the target distribution is not known analytically or if it cannot be
analytically integrated and inverted, we have to use the so-called rejection method.
These methods are explained in the following sections.

12 Random Numbers
1.6.1 Transformation Method

For a certain class of distributions, we can generate pseudo-random numbers from
uniformly distributed random numbers by applying a mathematical transformation.
The transformation method works particularly nicely for the most common distri-
butions (exponential and normal distributions). While the transformation is rather
straightforward, it is not always feasible – depending on the analytical form of the
distribution.
For a positive continuous random variable z and probability density function (PDF)
f (z), the probability that z lies between z and z + dz is f (z) dz. Thus, the probability that
z takes on values between a and b is
b
Pr(a ≤ z ≤ b) = f (z) dz. (1.10)
a
The corresponding cumulative distribution function (CDF) is

z
F(z) = f (z ) dz . (1.11)
0
The idea behind the transformation method is to find the equivalence between the
CDFs of a uniform distribution fu and the distribution of interest. The PDF of the
uniform distribution with support [0,1] is
⎧
⎪
⎪
⎪
⎨ 1 , for z ∈ [0, 1] ,
fu (z) = ⎪
⎪ (1.12)
⎪
⎩ 0 , otherwise .
Now we want to obtain random numbers that are distributed according to a second
PDF f (y). If we compare the areas of integration (i.e., CDFs), we find
z y

z= fu (z ) dz = f (y ) dy , (1.13)
0 0
where z is a uniformly distributed random variable and y a random variable distributed

according to the desired distribution f (y). The first equality in eq. (1.13) follows from
the definition of the uniform distribution in eq. (1.12) and we impose the second equal-
ity to find a suitable change of variables such that the CDFs of fu and f are the same.
Inverting F(y) leads to
y = F −1 (z) . (1.14)
This shows that a transformation between the two distributions can be found if and
only if
y
1. the integral F(y) = 0 f (y ) dy can be solved analytically in a closed form (i.e., f (y)
must be integrable),
2. there exists a closed-form analytical expression of the inverse of z = F(y) such that
y = F −1 (z) (i. e., F(y) must be invertible).
Of course, these conditions can be overcome to a certain extent by precomputing/tabu-

lating and inverting F(y) numerically if the integral is well behaved (i.e., nonsingular).
Then, it is possible to transform the uniform numbers numerically.
We are now going to demonstrate this method for the two most commonly used
distributions: the exponential distribution and the Gaussian distribution. We will see
that already in the case of a Gaussian distribution, quite some work is required to
perform such a transformation.
The Exponential Distribution

The exponential distribution is
f (y) = ke−yk . (1.15)
By applying the area equality of eq. (1.13), we find

y

z= ke−y k dy , (1.16)
0
and thus Figure 1.5 George E. P. Box (1919–2013)

z= −e−y k |y0 =1−e −yk
. (1.17) was a statistician working at different
research institutions, including Princeton
Solving for y yields University and the University of
1 Wisconsin–Madison.
y = − ln(1 − z) , (1.18)
k
so that for each homogeneously distributed random number z we get an exponentially
distributed random number y through eq. (1.18).
The Gaussian Distribution

Methods for generating normally distributed random numbers are very useful because
there are many applications and examples where such numbers are needed.
The Gaussian or normal distribution is
1 y2
f (y) = √ e− 2σ2 , (1.19)
2πσ2
where σ represents the standard deviation of the distribution. Unfortunately, we can
only solve its integral between −∞ and ∞ analytically:
∞ Figure 1.6 Mervin E. Muller (1928–2018)
1 y 2
√ e− 2σ2 dy = 1 . (1.20) was a computer scientist working at the
2πσ2 −∞
Ohio State University. Photograph
However, Box and Muller [16] (see Figures 1.5 and 1.6) have introduced the following courtesy The Ohio State University.
elegant trick to circumvent this restriction. We take two (uncorrelated) uniform ran-
dom variables z1 and z2 . We apply the area equality of eq. (1.13) again, but this time
we write it as a product of the two random variables
14 Random Numbers
y1 y2
y 2
y 2
1 − 2σ 1
dy 1 e− 2σ dy 2
1 2
z1 · z 2 = √ e · √
−∞ 2πσ2 −∞ 2πσ2
y2 y1 (1.21)
1 − y 21 +y 22
= 2
e 2σ dy1 dy2 .
−∞ −∞ 2πσ
We now solve the integral by transforming the variables y 1 and y 2 into polar
coordinates
r2 = y21 + y22 , (1.22)

y1
tan φ = , (1.23)
y2
with
dy 1 dy 2 = r dr dφ . (1.24)
We apply these coordinate transformations to eq. (1.21) to obtain

φ r
1 r 2
z1 · z2 = 2
e− 2σ2 r dr dφ (1.25)
2πσ 0 0
r
φ 2
−r 2
= e 2σ r dr (1.26)
2πσ2 0
φ 2
− r2
= · σ 2
1 − e 2σ , (1.27)
2πσ2
giving
y2 +y2

1 y1
· 1 − e− 2σ2 .
1 2
z1 · z2 = arctan (1.28)
2π y2

≡z1 ≡z2
By separating these two terms (and associating them to z1 and z2 , respectively), we can
invert this function and find
y21 + y22 = −2σ2 ln(1 − z2 ) , (1.29)

y1 sin(2πz1 )
= tan(2πz1 ) = . (1.30)
y2 cos(2πz1 )
One solution of these two coupled equations is

y1 = −2σ2 ln(1 − z2 ) sin(2πz1 ) , (1.31)

y2 = −2σ2 ln(1 − z2 ) cos(2πz1 ) . (1.32)
Thus, by using two uniformly distributed random numbers z1 and z2 , we obtain

(through Eqs. (1.31) and (1.32)) two normally distributed random numbers y1 and y2 .
1.6.2 Random Points on a Sphere

Another application of the transformation method is to generate points that are dis-
tributed uniformly at random over the surface of a sphere with radius R. We denote
the coordinates of the points by coordinates (φ, θ) where φ ∈ [0, 2π) and θ ∈ [0, π].
An overly simplified attempt would be to generate two uniformly distributed ran-

dom numbers u, v ∈ [0, 1] and then set φ = 2πu and θ = vπ. We show the resulting
distribution of points in the top panel of Figure 1.7. It is evident that there is a larger
concentration of points in the pole regions when compared to the remaining parts of
the sphere. The reason is that we are not properly accounting for the surface element
dA = R2 sin(θ)dφdθ.
To obtain points that are distributed uniformly at random over a sphere surface, we
have to apply the transformation method to the distribution of θ:
θ
1
v= sin(θ ) dθ . (1.33)
2 0
Figure 1.7 Distribution of random points
The prefactor 1/2 guarantees that the cumulative distribution function is normalized on a sphere.
to unity for θ = π. Based on eq. (1.33), we find θ = arccos(2v − 1) and as in the first
attempt we use φ = 2πu. The points are now uniformly distributed over the sphere’s
surface as we show in the bottom panel of Figure 1.7.
1.6.3 The Rejection Method

We have seen in Section 1.6.1 that there are two conditions that a function has to
satisfy to apply the transformation method: integrability and invertability. If either of
these conditions is not satisfied, there exists no analytical method to obtain random
numbers for such a distribution. It is important to note that this is particularly relevant
for experimentally obtained data, where no analytical description is available. In such
a case, we have to resort to a numerical method (rejection method) to obtain arbitrarily
distributed random numbers.
Let f (y) be the distribution from which we would like to obtain random numbers. A
necessary condition for the rejection method to work is that f (y) is bounded and only
exists over a finite domain. That is, f (y) < A for y ∈ [0, B], with A, B ∈ R and A, B < ∞.
We then define a box with edge lengths B and A (see Figure 1.8).
Illustration of the rejection method. Sample points are placed within the box and rejected if they lie above the curve Figure 1.8
(i.e, inside the blue-shaded area), and accepted otherwise.
16 Random Numbers
We now generate two pseudo-random variables z1 and z2 with z1 , z2 ∈ [0, 1). If we

consider the point with coordinates (Bz1 , Az2 ), we see that it surely lies within the
defined box. If this point lies above the curve describing the distribution f (y) (i. e.,
Az2 > f (Bz1 )), the point is rejected (hence the name of the method). Otherwise y = Bz1
is retained as a random number that is distributed according to f (y).
In principle, the method works quite well. It can be improved considering the
following points:
• It is desirable to have a good guess for the upper bound. Obviously, the better the
guess, the less points are rejected. In the above description of the algorithm, we have
assumed a rectangular box. This is, however, not a necessary condition. In principle,
one can take any domain on which homogeneously distributed random numbers are
easily generated.
• There is a method to make the rejection method faster (but also more complicated).
We may use N boxes to cover f (y) and define the individual box with side length Ai
and bi = Bi+1 − Bi for l ≤ i ≤ N. Then, the approximation of f (y) is much better (this
is related to the definition of the Riemann integral).
Random-Geometrical Models 2
2.1 Percolation
In materials science and chemistry, percolation refers to the movement or filtering of

fluids through porous media. The term stems from a Latin word and is still common
in Italian.1
A very simple and basic model of such a process was first introduced by Broadbent
and Hammersley (see Figure 2.1) in 1957 [17].
While the original idea was to model fluid motion through a porous material (e. g.,
a container filled with glass beads), it was found that the model has many other
applications. Furthermore, it was observed that the model possesses some interesting
universal features characteristic of so-called critical phenomena.2
Applications include:
• Porous media: for example, used in the oil industry and as a model for the pollution
of soils
• Sol-gel transitions
• “Mixtures” of conductors and insulators: finding the point at which a conducting
material becomes insulating
• Spreading of (forest) fires [18] Figure 2.1 John Hammersley
• Spreading of epidemics and computer viruses [19, 20] (1920–2004) was a British
• Crashing of stock markets (D. Sornette [21]) mathematician who made substantial
• Landslide election victories (S. Galam [22]) contributions to probability theory.
• Recognition of pathogenic antigens by T-cells (A. Perelson [23])
2.2 The Sol–Gel Transition
The formation of gelatine is quite special from a chemical and physical point of view.
Initially, it is a fluid containing many small monomers (emulsion) which is referred
to as sol. If we place the sol in a fridge, it becomes a gel. The process taking place
1 ita: percolare: 1 Passare attraverso. ~ filtrare. 2 Far filtrare. (1 pass through ~ filter, 2 make sth. filter).
2 The notion of critical phenomena describes the physics associated with critical points and phase
transitions and is discussed later in this book.
17
18 Random-Geometrical Models
t
Figure 2.2 Sol–Gel transition. (a) Monomers are dispersed in the liquid (sol). (b) When the liquid is cooled, the monomers start
polymerizing and growing. (c) Once a huge macromolecule (which stretches across the whole container) has formed,
the sol–gel transition occurs at the gel time tG .
t
Figure 2.3 Viscosity and elastic modulus for the system acrylamide-bis-acrylamide as a function of time. At the gelation time tG ,
the viscosity diverges to infinity, and the elastic modulus, previously at zero, increases to a finite value. The figure is
taken from Ref. [24].
is schematically illustrated in Figure 2.2. Upon cooling, monomers start polymer-

izing and polymers start growing. At some point in time, one molecule becomes
sufficiently long to span from one side of the container to the other, which is when
the so-called percolation transition occurs. The polymerization and growth process is
experimentally accessible; one can, for instance, measure the elastic modulus or the
viscosity as a function of time. After a characteristic time (“gel time”) tG , the elas-
tic modulus suddenly increases from zero to a finite value. Similarly, the viscosity
increases and diverges at tG ; this is reflected in experimental findings (see Figure 2.3).
2.3 The Percolation Model
Modeling the process of gelation is surprisingly simple [25]. Let us assume that we
have a square lattice of side length L (e. g., L = 40) such that every cell can either be
occupied or empty. We fill each cell with probability p. That is, for each cell, we create
a random number z ∈ [0, 1) and compare it to p. If z < p, the cell will be marked as
occupied; otherwise, the cell remains empty. This can be done for different values of p
(see Figure 2.4). For small values of p, most cells are empty, whereas for big values of
p, most cells are occupied. If two nearest neighbors are occupied, one says that they
are connected and a group of connected sites is called a cluster. There exists a critical
probability p = pc (pc = 0.592 . . . for an infinite lattice) where, for the first time, a fully
connected cluster of cells will span the box (such a configuration is called a percolating
cluster). Although pc is uniquely defined for infinite lattice sizes, we can also find it for
finite sizes; in that case, pc is the average probability at which a percolating cluster
occurs. The critical probability is referred to as the percolation threshold.
2.3.1 The Burning Method

In order to calculate the critical probability associated with a lattice, we need a method
to detect if a spanning cluster exists in a given system (such as that shown in right panel
of Figure 2.4).
The burning method [26] not only provides us with a Boolean feedback (yes/no), but
also calculates the minimal path length (the minimal distance between opposite sides,
following only connected sites). The name of the method stems from its implementa-
tion. Imagine a grid with occupied and unoccupied sites; an occupied site represents
a tree while an unoccupied site is simply empty space. If we start a fire at the very top
of our grid, all trees in the first row will start to light up as they are in the immediate
Percolation on a square lattice with side length L = 40 for different values of p. For p < pc , there are many small Figure 2.4
clusters, while for p > pc , one big spanning cluster is present.
t
Figure 2.5 The burning method. A fire is started in the first line. At every iteration, the fire from a burning tree lights up occupied
neighboring cells. The algorithm ends if all reachable trees are burned or if the burning method has reached the
bottom line.
vicinity of the fire. Obviously, this forest fire will spread and nearest neighbors of the
trees in the first row will soon catch on fire as well. Thus, in the second iterative step
all trees immediately adjacent to burning trees get torched.
Clearly, the iterative method only comes to an end if the fire has reached the bot-
tom or if the fire runs out of fuel (i. e., no unburned trees neighboring a burning site)
and consequently dies out. If the inferno has reached the bottom, we can determine
the length of the shortest path from the iteration counter; that is, the number of itera-
tions needed to reach the other side defines the shortest-path length of the percolating
cluster.
The algorithm is as follows:
Burning Method
1. Label occupied sites by a one and unoccupied ones by zero.

2. Label all occupied cells in the top line with the marker t = 2.
3. Iteration step t + 1:
(a) Go through all the cells which have label t.
(b) For each t-labeled cell:
i. Check if any nearest neighbor (on the square lattice: North, East, South, West) is occupied and not
burning (label is 1).
ii. Assign any such neighbors the label t + 1.
Burning Method (cont.)
4. Repeat step 2 (with t → t + 1) until either there are no neighbors to burn anymore or the bottom
line has been reached. In the latter case, the latest label minus 1 defines the shortest-path length.
A graphical representation of the burning algorithm on the square lattice is given in

Figure 2.5. In the shown example, the shortest-path length is 22 − 1 = 21 units.
Exercise: Percolation and Forest Fire

Task 1
Write a program to simulate percolation with occupation probability p on a square lattice of size L × L.
Consider different lattice sizes and illustrate the resulting lattice with occupied and unoccupied sites.
Task 2
Write a program for a simple model for forest fires. This can be used to check if a sample contains a spanning
cluster (i.e., if it percolates):
• For a set of parameters L and p, create a square lattice, like in the previous task (representing the for-
est). Color the occupied sites with green (representing trees) or assign a number to them that represents
unburned forest.
• In the first time step burn all the trees in the first row (change the color to red or assign a number that
represents burning trees).
• In the next time step, burn all the trees neighboring the burning trees (red ones) and burn out the
previously burning trees (change their color to black or assign a number that represents burned trees).
• Repeat until the fire stops.
• Could you find a spanning cluster?
• You can now also measure the following quantities:
– the shortest-path length (i.e., the number of time steps that the fire needs to reach the other side).
– the lifetime of the fire defined as the total number of time steps needed before the fire stops.
2.3.2 The Percolation Threshold

When we carry out simulations of the percolation model, we observe that the prob-
ability of obtaining a spanning cluster or “wrapping probability” W(p) depends on
the occupation probability p (see Figure 2.6). When we also vary the lattice size, we
see that the transition from low- to high-wrapping probabilities W(p) becomes more
abrupt with increasing lattice size. If we choose very large lattices, the wrapping prob-
ability will start to resemble a step function. The occupation probability at which we
t
Figure 2.6 We show the wrapping probability W(p) for site percolation on the square lattice as a function of the occupation
probability for different lattice sizes along with a step function for comparison. A more detailed analysis reveals that
the percolation threshold is pc ≈ 0.592746 [27].
t
(a) Square lattice (b) Triangular lattice (c) Martini lattice (d) (4,6,12) lattice (e) Honeycomb lattice
Figure 2.7 Examples of different lattices.
observe the increasingly abrupt transition is the percolation threshold pc . When we

carry out a sequence of simulations with increasing lattice sizes, we can determine the
percolation threshold on the square lattice to be approximately pc = 0.59274 . . . [27].
The percolation threshold is a characteristic value for a given type of lattice. We find
different percolation thresholds depending on the type of lattice under consideration.
We show values for some lattice types in Table 2.1.
Percolation can be performed not only for lattice sites (site percolation) but also
for the corresponding bonds by randomly occupying them (bond percolation). For
a given lattice, the number of neighbors may be different for bond percolation when
compared to site percolation. This difference will be also reflected in the correspond-
ing bond percolation threshold.3 Looking at the different lattices in Figure 2.7 and
Table 2.1, it becomes clear that thresholds decrease with the number of neighbors
(also called “coordination number” or “degree”). For example, the honeycomb lat-
tice (see Figure 2.7e) has the highest threshold in two dimensions. Furthermore, the
threshold of some lattices such as the triangular lattice (see Figure 2.7b) can be calcu-
lated analytically [28] (denoted by an asterisk * in Table 2.1). Intuitively, it is clear that
3 Please note that unless otherwise stated, we subsequently use the term percolation instead of site perco-
lation for the sake of brevity. The general behavior of bond percolation is qualitatively similar to that of
site percolation.
Table 2.1 Percolation thresholds for different lattices for site

percolation and bond percolation
Lattice Site Bond
Square 0.59275 [29] 1/2* [30]

Triangular 1/2* [30] 2 sin (π/18)* [30]
Honeycomb 0.696 [29] 1-2 sin (π/18)* [31]
Cubic (body-centered) 0.246 [32] 0.180 [33]
Cubic (face-centered) 0.198 [32] 0.120 [33]
Cubic (simple) 0.312 [32] 0.249 [33]
Diamond 0.430 [34] 0.390 [34]
4-Hypercubic 0.197 [35] 0.160 [35]
5-Hypercubic 0.141 [35] 0.118 [35]
6-Hypercubic 0.109 [35] 0.094 [35]
7-Hypercubic 0.089 [35] 0.079 [35]
Note: Numbers with an asterisk (*) can be calculated analytically.
the threshold should depend on the geometry of the lattice; thinking of the burning
algorithm, the speed at which the “fire” spreads depends on the spatial configuration
of “trees.”
2.3.3 The Order Parameter

We have determined the percolation threshold in the previous section and outlined
that the transition between spanning and nonspanning configurations occurs exactly
at pc for infinitely large systems. Let us now consider probabilities p > pc , where we
will always find a spanning cluster (for sufficiently large systems). Naturally, we may
ask ourselves how many of the sites belong to the spanning cluster (which is also the
biggest one). More precisely, we can define the fraction of sites which belong to the
largest cluster P(p) as a function of the occupation probability p. We call this quantity
the order parameter (for its dependence on p, see Figure 2.8).
Obviously, for p < pc the order parameter is zero, as there are no spanning clusters.
For p > pc the order parameter approaches a finite value and increases with p. When
analyzing the behavior of the order parameter, we find that P(p) behaves like a power
law in the region close to the percolation threshold
P(p) ∝ (p − pc )β , (2.1)
with a dimension-dependent exponent β. This exponent neither depends on the type of

lattice nor on the detailed connectivity rule (bond or site percolation) and is therefore
tFigure 2.8 Order parameter P(p) as a function of the occupation probability p for site percolation on the square lattice. The
system exhibits a “critical” behavior at pc . Starting at pc , the order parameter grows as P(p) ∝ (p − pc )β , where
β = 365
in 2D and β ≈ 0.41 in 3D.
considered “universal.” We will discuss this concept in more detail in the following
sections.
Exercise: Percolation Threshold and Order Parameter

Consider the percolation model and determine the probability to find a spanning cluster, shortest-path length,
and the lifetime of the fire. You may proceed as follows:
• For a given number of lattice sites N and occupation probability p make several measurements (e.g.,
1,000) using different seeds for the random number generator, and compute the fraction of samples with a
spanning cluster, the average shortest-path length (when percolation occurs), and the lifetime of the fire.
• Repeat the same for different values of p. Plot the results against p.
• Change N and compare the results. (For each N you should obtain a curve.)
• Find the value of the threshold probability pc for which the system changes its behavior.
Please consider the following point:
• Choosing a site to be occupied with probability p can be simply done in the following way: generate a
random number between 0 and 1. If it is smaller than p set the site as occupied.
2.3.4 The Cluster-Size Distribution

After having investigated the percolation threshold and the fraction of sites in
the spanning cluster, the natural extension would be to identify all the clusters
(see Figure 2.9). An efficient algorithm is necessary to perform such a task. In fact,
Figure 2.9 Percolation on a square lattice
there are several such algorithms; the most popular (and for this purpose most effi-
with L = 40 for p = 0.4. The different
cient) algorithm is the Hoshen–Kopelman algorithm [36], which was developed in
shades of gray indicate lattice sites that
1976 by Joseph Hoshen and Raoul Kopelman (see Figure 2.10).
belong to different clusters.
We represent the percolation configuration as a matrix Ni j that can have values of

0 (site is unoccupied) and 1 (site is occupied). Furthermore, let k be a running index
labeling the clusters in Ni j . We also introduce an array Mk to keep track of the mass
of cluster k (the number of sites belonging to the cluster k). We start the algorithm
by setting k = 2 (since 0 and 1 are already taken) and search for the first occupied
site in Ni j beginning in the upper left corner of the lattice. We then add this site to
the array Mk=2 = 1 and set the entry in Ni j to k (so it is branded as pertaining to the
cluster k).
We then go over all lattice sites Ni j , line by line as a type writer, and try to detect
whether an occupied site belongs to an already known cluster or a new one. We comb
through the lattice from top-left to bottom-right; the criteria are rather simple:
Figure 2.10 Raoul Kopelman is a
• If a site is occupied and the top and left neighbors are empty, we have found a new professor at the University of Michigan.
cluster and we set k to k + 1, Ni j = k and Mk = 1.
• If only one of the sites (top or left) has the value k0 (i.e., is part of a cluster), we
increase the corresponding value in the array, Mk0 , by one (setting Mk0 to Mk0 + 1).
We label the new site accordingly. That is, we set Ni j = k0 .
• If both neighboring sites are occupied with k1 and k2 , respectively (assuming k1
k2 ) – meaning that they are already part of a cluster – we choose one of them (e. g.,
k1 ). We set the matrix entry to the chosen value, Ni j = k1 and increase the array
value not only by one but also by the whole number of sites already in the second
cluster (here k2 ), Mk1 → Mk1 + Mk2 + 1. Of course we have to mark the second array
Mk2 in some way so that we know that its cluster size has been transferred over to
Mk1 which we do by setting it to −k1 . We have thus branded Mk2 in such a way that
we immediately recognize that it does not serve as a counter anymore (as a cluster
cannot consist of a negative number of sites). Furthermore, should we encounter
an occupied site neighboring a k2 site, we can have a look at Mk2 to see that we are
actually dealing with cluster k1 (revealing the “true” cluster number).
The last point is crucial, as we usually deal with a number of sites that are marked in
such a way that we have to first recursively detect which cluster they pertain to before
carrying out the algorithm any further. The recursive detection stops as soon as we
have found a k0 with Mk0 ≥ 0 (i. e., a “true” cluster number).
Once all the sites Ni j have been visited, the algorithm ends up with a number l of
clusters, where l is smaller than the largest obtained cluster index kmax . The only thing
left to do is to construct a histogram of the different cluster sizes. This is done by
looping through all the clusters k ∈ {2, . . . , kmax } while skipping negative Mk .
We visualize the Hoshen–Kopelman algorithm in Figure 2.11. The algorithm is very
efficient because it visits every site only once and it scales linearly with the number of
sites.
Once we have run the algorithm for a given lattice (or collection of lattices and taken
the average), we can calculate the cluster-size distribution n s (i.e., the number of clus-
ters of size s per site) as a function of the occupation probability p. These results are
illustrated in Figure 2.12, where the first graph (Figure 2.12a) represents the behavior
t
Figure 2.11 The Hoshen–Kopelman algorithm applied to a percolation configuration on the square lattice. The numbers denote
the running cluster variable k.
Hoshen–Kopelman Algorithm
1. k = 2, Mk = 1
2. For all i, j of Ni j
(a) If top and left are empty (or nonexistent) k → k + 1, Ni j = k, Mk = 1.
(b) If one is occupied with k0 then Ni j = k0 , Mk0 → Mk0 + 1.
(c) If both are occupied with k1 and k2 (and k1 k2 ) then choose one, for example, k1 and Ni j → k1 ,
Mk1 → Mk1 + Mk2 + 1, Mk2 = −k1 .
(d) If both are occupied with k1 , Ni j = k1 , Mk1 → Mk1 + 1.
(e) If any of the ks considered has a negative mass Mk , find the original cluster they reference to and use
its cluster number and weight instead by using while(Mk < 0) k = −Mk .
3. For k ∈ {2, . . . , kmax }, if Mk > 0 then N(Mk ) → N(Mk ) + 1.
4. For all cluster sizes s, determine n s = N(s)/N where n s is the desired cluster-size distribution.
for subcritical occupation probabilities (p < pc ), the second graph (Figure 2.12b)
shows the behavior at the critical occupation probability (p = pc ), and the third graph
(Figure 2.12c) depicts the behavior for overcritical occupation probabilities (p > pc ).
In the subcritical regime (p < pc ), we find that n s (p) obeys an exponential function
multiplied with a power law. In contrast, in the overcritical region (p > pc ), we observe
a distribution exhibiting an exponential decay but with an argument that is stretched
with a power s1−1/d .
We observe in Figure 2.12b that n s (p = pc ) follows as a straight line in a double
logarithmic plot, implying a power law behavior (i.e., n s (p = pc ) ∝ s−τ ). The exponent
Cluster-size distribution of percolation on a square lattice for (a) p < pc , (b) p = pc , and (c) p > pc . The gray Figure 2.12
solid line in the center panel has slope −τ = −187/91 ≈ −2.05.
τ depends on the system dimension and is 187 91 in 2D while numerical results indicate
that it is 2.1892(1) in 3D [35]). The bounds for τ are 2 ≤ τ ≤ 5/2.
We summarize the behavior of the cluster-size distribution n s in the three different
regions:
⎧
⎪
⎪
⎪ s−θ eas , p < pc ,
⎪
⎪
⎪
⎨ −τ
n s (p) ∝ ⎪
⎪ s , p = pc , (2.2)
⎪
⎪
⎪
⎪
⎩ e−bs 1−1/d
, p > pc .
We now compare the distributions to the cluster-size distribution at pc and may plot a
rescaled distribution
ñ s (p) = n s (p)/n s (p = pc ) ,
which is defined in all three regions of eq. (2.2). A plot of ñ s (p) as a function of (p −
pc )sσ is shown in Figure 2.13. Choosing the right value of σ, we see that all data for Figure 2.13 Scaling behavior of the
different s and p fall on one single curve. This is called data collapse. In 2D, σ = 36/91 cluster-size distribution for percolation on
and in 3D the exponent σ = 0.445(10) [33]. the square lattice. Original plot made by
We notice that the cluster-size distributions are all described by the following scaling Prof. Dietrich Stauffer.
relation
n s (p) = s−τ ± [(p − pc )sσ ], (2.3)
with the scaling functions4 ± , where the subscript ± stands for p > pc (+) and
p < pc (−), respectively. This scaling law was first described by Dietrich Stauffer (see
Figure 2.14).
4 Scaling function: A function of two variables that can be expressed as a function of one single variable.
Additional Information: Moments of a Distribution

The nth moment of a distribution P(x) is defined as

μn = xn P(x) dx .
the 0th moment of a normalized distribution is simply μ0 = 1, and the first moment is
Therefore,
μ1 = xP(x) dx = E(x) (the expectation value). Accordingly, μ2 is the standard deviation of the
distribution around the expectation value. For a discretely valued argument s, the integral becomes a sum.
The second moment of the cluster-size distribution n s (p) is

2
χ= s ns . (2.4)
s
The prime ( ) indicates the exclusion of the largest cluster in the sum (as the largest
Figure 2.14 Dietrich Stauffer cluster would make χ infinite at p > pc for N → ∞). We find
(1943–2019) was a professor for χ ∝ |p − pc |−γ , (2.5)
theoretical physics at the University of
Cologne. with γ = 43/18 in 2D and γ ≈ 1.8357 in 3D [37]. The second moment is a very strong
indicator of pc , as we can see a very clear divergence around pc .
In the gray box below, we show that the scaling exponents we have seen so far are
related by [38]
3−τ
γ= . (2.6)
σ
An overview of the different exponents is given in Ref. [18].
Additional Information: Scaling Relation for the Moments of the Cluster-Size Distribution

For the second moment of the cluster-size distribution, we have s ’ s2 n s ∝ |p − pc |−γ with γ =

(3 − τ)/σ (see eq. (2.6)). For general moments of the cluster-size distribution, we find s ’ sk n s ∝
|p − pc |(τ−1−k)/σ using eq. (2.3) [38]:
∞ ∞
s ns =
k
s n s ds =
k
sk−τ ± [(p − pc )sσ ] ds
s 0 0
±∞
= σ−1 sk−τ+1 z−1 ± (z) dz (2.7)
0

= σ |p − pc |(τ−1−k)/σ |z|(1+k−τ)/σ z−1 ± (z) dz ,
−1
where we substituted z = (p − pc )sσ in the third step. Integrals above pc run from 0 to +∞ and below
pc from 0 to −∞. For large ±z, the scaling function ± (z) decays exponentially. We have thus confirmed

s ’ sk n s ∝ |p − pc |(τ−1−k)/σ .
Exercise: Cluster-Size Distribution

Implement the Hoshen–Kopelman algorithm for a square lattice of size L × L and occupation probability p.
• The algorithm determines which sites belong to each cluster. Each cluster gets a number k (starting from
k = 2 to distinguish from “untouched” sites). The size (mass) M(k) of a cluster is stored.
• Here, we will go through the system from left to right and top to bottom.
• There are three possible situations for the top and left neighboring sites of an occupied site:
– both are unoccupied: new cluster (k → k + 1, M(k) = 1).
– one is occupied or both are occupied with same cluster (Cluster k0 ): site belongs to cluster k0 , M(k0 )
increased by 1.
– both are occupied with two different clusters (k1 and k2 ): clusters meet each other and have to be
united to one cluster (e.g., M(k1 ) → M(k1 )+ M(k2 )+1). Make an annotation that, for example,
cluster k2 belongs to cluster k1 (M(k2 ) = −k1 ). (Note: Typically the larger k-value is chosen to be
added to the smaller one.)
• For determining the “true” cluster number in cases when M(k) is negative, it is useful to implement a
function (recursive or while loop). Note that in all cases mentioned above you first have to determine the
“true” cluster numbers k0 , k1 , k2 .
• To get the cluster-size distribution you have to count the number of clusters of each size after reaching the
end of the system.
• Plot the cluster-size distribution for different values of p.
Please consider the following points:
• The cluster-size distribution n s is normalized per site, that is, the total number of clusters of each size
divided by total number of sites (L2 ). (For this reason, n s < 1.)
• Depending on the value of p (below, at, or above pc ) a different behavior of the cluster-size distribution is
observed.
2.3.5 Correlation Length

To further analyze percolation clusters, we now consider the correlation function c(r)
that describes connectivity correlations as a function of the radial distance r. For one
cluster (e. g., the largest cluster), we can use density correlations since, by definition,
all sites in a cluster are connected. Otherwise, if we want to study correlation effects
between multiple clusters, we should focus on connectivity correlations. Large values
of the correlation function c(r) mean that two quantities strongly influence each other,
whereas zero means that they are uncorrelated.
Numerically, we can obtain the correlation function c(r) by placing concentric cir-
cles or spheres around a site in the center region of the cluster. The correlation function
is defined as

Γ d
2
c(r) = [M(r + Δr) − M(r)] , (2.8)
2πd/2 rd−1 Δr
where M(r) (mass of the cluster) denotes the number of occupied sites in a d-
dimensional hypersphere of radius r. The prefactor in eq. (2.8) is the inverse of the
volume of a d-dimensional annulus between r and r + Δr, and Γ(·) is the gamma func-
tion. We see that c(r) is the number of occupied sites within an annulus of thickness Δr
at a distance r from the center and normalized by the volume of the annulus. We apply
this method to the largest percolation cluster for different occupation probabilities p.
When we compute c(r) for a given cluster (see Figure 2.15), we find that the correlation
function usually decreases exponentially with distance r, eventually with an offset C.
Mathematically, this means

r
c(r) ∝ C + exp − , (2.9)
ξ
where the constant C is equal to the order parameter of percolation, P(p), and thus
vanishes for p < pc . The newly introduced quantity ξ is the correlation length. It
describes the typical length scale over which the correlation function decays. Below
pc , the correlation length ξ is proportional to the radius of a typical cluster.
Figure 2.15 Illustration of how to
When we analyze the dependence of the correlation length ξ on the occupation
numerically determine the correlation
probability p (see Figure 2.16), we find that it diverges at the critical occupation
function c(r) of a percolation cluster. At
probability pc [39],
a given radius r, we count the number of
⎧
occupied sites within an annulus of ⎪
⎪
⎪
⎨ 3, in 2D ,
4
thickness Δr. Note that the origin ξ(p) ∝ |p − pc |−ν where ν = ⎪ (2.10)
⎪
⎪
⎩ 0.8751(11) ,
(r = 0) must belong to the cluster. in 3D .
Apparently, the assumption of an exponential behavior is no longer valid at pc . In fact,

at pc the correlation function follows a power law [33]
⎧
⎪
⎪
⎪
⎨ 24 , in 2D ,
5
c(r) ∝ r−(d−2+η) with η=⎪
⎪ (2.11)
⎪
⎩ −0.046(8) , in 3D .
The exponents ν and η in Eqs. (2.10) and (2.11) are again universal critical exponents.
2.3.6 Finite Size Effects

Figure 2.16 Singular behavior of ξ with
We encounter problems when the system size L is smaller than the correlation length
respect to p in an infinite system.
ξ. Instead of the singularity that was mentioned in eq. (2.10), the correlation length
ξ only takes on finite values (see Figure 2.17). We consider the values p1 and p2 that
are defined by correlation lengths ξ(p1 ) and ξ(p2 ) that are of the order of the linear
system size L. The region between p1 and p2 is called the critical region. We cannot
trust any quantity obtained numerically within this critical region. Based on
L = ξ(p1 ) ∝ (p1 − pc )−ν (2.12)
and
p1 − p2 ≈ 2(p1 − pc ) ,
we find that the critical region shrinks with system size like
1
p1 − p2 ∝ L− ν . (2.13)
In the limit L → ∞, the critical region will vanish. We can obviously not realize such
a limit on a computer. If we identify our numerically obtained effective critical occu-
pation probability peff (L) with p1 (ξ(p1 ) ≈ L), we will always obtain peff (L) < pc . We
assume that we are close enough to pc such that the deviation from the true pc is Figure 2.17 The finite size of the system
proportional to L−1/ν . We thus find leads to a cutoff in the correlation length
1 (orange solid line). The critical region is
peff (L) = pc 1 − aL− ν , (2.14) defined as the interval that is confined by
the two dashed black vertical lines.
where a > 0 is a constant. The best we can do at this point is using the data acquired
for finite system sizes and extrapolate to the values of the infinite system. To do so,
we can use eq. (2.14), plot peff as a function of L−1/ν , and extrapolate the data to the
point at which the vertical axis is crossed (corresponding to the limit L−1/ν → 0). This
method enables us to find the critical occupation probability pc .
2.3.7 Finite Size Scaling

In the neighborhood of the percolation threshold pc , we find that the dependence of
the second moment of the cluster-size distribution χ (see eq. (2.4)) on p and L can be
reduced to a one variable function. This is a consequence of a general scaling law that
exists around all critical points.
By plotting χ around pc against the occupation probability p for several values of
L, we obtain curves that differ strongly around the critical point. The most important
difference is the height of the peaks which is found to grow as a power law with the
linear system size L. Using this observation and the fact that the critical region shrinks
like L−1/ν as shown in eq. (2.13), we can rescale the horizontal and vertical axis of the
left panel of Figure 2.18 to get a data collapse as shown in the corresponding right
panel. Such a data collapse can be also observed in related models such as kinetic
gelation (see Figure 2.19).5
Let us try to comprehend what this data collapse means. We have a function χ
that was originally a function of two parameters (the occupation probability p and
the system size L) and that now behaves as though it was a one-parameter func-
tion. Mathematically, the data collapse of χ as shown in Figures 2.18 and 2.19 can
be written as
5 In kinetic gelation, bonds are randomly occupied so as to form cross-linked chains in order to describe
gelation through additive polymerization [40].
t
(a) (b)
Figure 2.18 For site percolation on a square lattice with side length L, we show (left) the system-size dependence of the second
moment of the cluster-size distribution (eq. (2.4)) and (right) the corresponding finite-size scaling (eq. (2.15)). The
number of samples is 104 . We normalized χ(p, L) by max p [χ(p, L = 1024)] in the left panel.
γ 1
χ(p, L) = L ν ℵχ (p − pc )L ν , (2.15)
where ℵχ is called a scaling function of χ. This is an example of the finite size scaling
first proposed in Ref. [41].
When we approach the critical occupation probability pc , the scaling function ℵχ
approaches a constant and we find that the peak χmax depends on the system size, so
γ
χmax (L) ∝ L ν . (2.16)
This can be verified by plotting the maxima of the data of the left panel of Figure 2.18
as a function of L. These expressions are reminiscent of those previously introduced
in the context of the cluster-size distribution in Section 2.3.4. We recall that a scaling
function is a function that combines two variables (in this context p, L), which can be
expressed in terms of only one variable (the so-called scaling field).
Based on eq. (2.5), we know that χ(p) ∝ |p − pc |−γ in the vicinity of pc . Thus, at pc
and in the limit L → ∞, we find
Figure 2.19 Finite size scaling of χ for
γ 1
the kinetic gelation model on a cubic lim χ(p, L) = lim L ν ℵχ (p − pc )L ν ∝ |p − pc |−γ . (2.17)
L→∞ L→∞
lattice. We observe a data collapse when
plotting χL−γ/ν against |p−p c | 1/ν
pc L . This implies that ℵχ (x) ∝ |x|−γ for x → ∞.
The right curve corresponds to p < pc
and the left one to p > pc . The straight
2.3.8 Size Dependence of the Order Parameter
lines have a slope of −γ (see the critical
exponents of χ for infinite systems,
We are now going to consider the order parameter P(p) at the critical occupation
eq. (2.5)). The figure is taken from
probability pc . We denote the size of the largest cluster by s∞ and the side length of
Ref. [40].
the lattice by L. When we plot s∞ against L, we notice that there is a power law at work,
s∞ ∝ Ldf , (2.18)
where the exponent df depends on the dimension of the system. We illustrate this
behavior in Figure 2.20. For two-dimensional lattices, we find df = 91/48 and for three
dimensions we find df = 2.5226(1) [42]. The exponent df is called the fractal dimension,
which will be explained in more detail in Section 2.4.
The size of the largest cluster is proportional to the product of system size Ld
and P(p) (i. e., the probability of a lattice site to belong to the largest cluster).
That is,
s∞ ∝ Ld P(p) . (2.19)
We now combine Eqs. (2.19) and (2.10) and obtain
s∞ ∝ Ld P(p) ∝ Ld |p − pc |β ∝ Ld L−β/ν , (2.20)
where we used that the correlation length is of the order of L for a finite system in
the vicinity of pc . Based on Eqs. (2.18) and (2.20), we can establish the following Figure 2.20 Log-log plot of the size
connection: dependence of the order parameter in a
β square lattice at the percolation threshold
df = d − . (2.21) pc ≈ 0.5972. The slope of the gray solid
ν
line is the fractal dimension
Analogous to eq. (2.15), the order parameter P also satisfies a finite-size scaling
df = 91/48 ≈ 1.9.
relation for p > pc
β
1
P(p, L) = L− ν ℵP (p − pc )L ν . (2.22)
At pc we obtain
β
P ∝ L− ν . (2.23)
When we combine all of this, we find

β
− ν +d
s∞ ∝ PL ∝ L
d
∝ Ldf ,
in agreement with Eqs. (2.20) and (2.21).
2.3.9 The Shortest Path

We have outlined how to decide whether there exists a spanning cluster by means of the
burning algorithm, and we have determined and analyzed the cluster-size distribution Figure 2.21 To illustrate the
with the Hoshen–Kopelman algorithm. One unanswered question, however, is how shortest-path dependence on the system
the shortest-path length of a spanning cluster behaves as a function of the system size. size, we show the Euclidean distance r
Simulations show that the shortest-path length of a spanning cluster satisfies as a function of the shortest-path length
for four-dimensional site (top curve)
∝ Ldmin , and bond (bottom curve) percolation.
Both quantities are related through
which is also a power law of the linear lattice size L. The exponent dmin depends on the
r ∝ 1/dmin . The figure is taken from
dimension [44]:
⎧ Ref. [43].
⎪
⎪
⎪ 1.13077(2) , in 2D ,
⎪
⎪
⎪
⎨
dmin = ⎪ ⎪ 1.3756(6) , in 3D , (2.24)
⎪
⎪
⎪
⎪
⎩ 1.607(5) , in 4D .
We can determine dmin for site percolation and bond percolation as well as for differ-
ent lattice types. The quantity dmin is called the fractal dimension of the shortest path, a
concept which we will discuss in Section 2.4. In Figure 2.21, we show the shortest path
dependence on the system size for site and bond percolation for a four-dimensional
system.
2.3.10 Bootstrap Percolation
Many variants of percolation have been invented to describe different physical phe-
nomena. One of those is bootstrap percolation [45–47], which has been studied by
Joan Adler (see Figure 2.22). In the canonical bootstrap model, sites are initially occu-
Figure 2.22 Joan Adler is a professor pied randomly (with a given probability p) as in the percolation model. Next, all sites
emeritus at the Technion in Israel. that do not have at least m occupied neighbors are removed. We then repeat this last
step until the resulting lattice configuration does not change anymore. In the case of
m = 1, all isolated sites are removed and for m = 2 all “dangling” sites are removed.
This does not affect pc . If we, however, choose m ≥ 3, the percolation threshold pc and
the nature of the transition are strongly dependent on the lattice. On some lattices and
for specific values of m, the order parameter even jumps at the transition (first order
transition).
Bootstrap percolation has many applications in solid state physics and fluid flow in
porous media. We show an example of bootstrap percolation in Figure 2.23.
2.4 Fractals
2.4.1 Self-Similarity
Fractals are objects that exhibit self-similarity. That is, certain patterns occur repeat-
edly at different spatial scales. Benoit B. Mandelbrot (see Figure 2.24) coined the term
fractal [48]. He also introduced the concept of fractal dimension to characterize fractal
objects. Before taking a closer look at this concept, it may be useful to consider some
examples of self-similarity. In a nutshell, we could define an object to be self-similar
if it is built up of smaller copies of itself. Such objects occur both in mathematics and
nature. Let us discuss a few examples.
The Sierpinski Triangle

Figure 2.24 Benoit B. Mandelbrot The Sierpinski triangle is a mathematical object that is constructed by an iterative
(1924–2010) coined the term fractal. He application of a simple operation. It was first described by the Polish mathemati-
worked at IBM and different academic cian Wacław Franciszek Sierpiński (1882–1969). To construct a Sierpinski triangle,
institutions, including Harvard and Yale we subdivide a triangle into four sub-triangles and discard the center triangle
Universities. (see Figure 2.25). In the next step of the iteration, we subdivide each of the three
2.4 Fractals 35
The Sierpinski triangle is a self-similar mathematical object that is created iteratively. Figure 2.25
remaining triangles and remove each central triangle. This obviously produces an
object that is exactly self-similar in the limit of infinitely many iterations. When zoom-
ing in, we see the exact same image at all scales. The object becomes mathematically a
fractal in the limit of an infinite number of iterations.
Self-Similarity in Nature
Naturally occurring self-similar objects are usually only approximately self-similar.
What we mean by that is that we cannot zoom into natural patterns indefinitely often
and expect to find the same pattern at all scales. To better illustrate this, we consider a
tree. A tree has different branches, and the whole tree looks similar to a branch con-
nected to the tree trunk (see Figure 2.26). The branch itself resembles the smaller
branches attached to it and so on. Evidently, this breaks down after a few itera-
tions, when the leaves of the tree are reached. This example is based on a hierarchical
tree structure. However, natural self-similar patterns are not necessarily hierarchical
objects as illustrated by the following example.
Gold colloids were shown to arrange in fractals of fractal dimension 1.70 (the mean-
ing of this will soon be explained) by David Weitz in 1984 [49]. Colloidal gold is a
suspension of sub-micrometer-sized gold particles in a fluid (e. g. , water). These gold
colloids aggregate to fractal structures as we show in Figure 2.27. As in the case of
trees, gold colloids are also only approximately self-similar. If we zoom too deeply
into the colloidal structure, we will end up seeing individual gold particles and not
a colloidal structure (see top left panel of Figure 2.27). Moreover, the finite size of Figure 2.23 We show a freshly occupied
the aggregate colloids does not allow us to zoom out of the structure infinitely far. square lattice with periodic boundary
The diameter of gold particles and the finite size of aggregate colloids therefore define conditions having an occupation
lower and upper cutoffs, respectively. probability p = 0.55 (top) and the same
lattice after removing all “dangling” sites
with less than m = 2 occupied
2.4.2 Fractal Dimension: Mathematical Definition neighbors (bottom).
Keeping these examples in mind, we will now give the precise mathematical definition
of fractal dimension. To determine the fractal dimension of an object, we can use the
following (theoretical) procedure.
Consider all possible coverings of an object with d-dimensional spheres of radius
ri ≤ , where is the resolution. Let N (C) be the number of spheres used in the
covering C. Then, the volume of the covering is
Additional Information: Definition of Covering

A covering of an object X ⊆ Rd is a collection of sets whose union contains X as a subset. If we denote the
collection of sets by C = {Ai |i ∈ I}, we call C a covering of X if

X⊆ Ai .
i∈I
N
(C)
V (C) ∝ rid , (2.25)

i=1
where d is the dimension of the space into which the object is embedded.
We define V∗ as the volume of the covering that, among all coverings with the
smallest number of spheres, has minimal volume

V∗ = min min (V (C)) . (2.26)
V (C) N (C)
Figure 2.26 An example of the The fractal dimension of the object is then defined as
hierarchical branching pattern of a
baobab tree. log V∗ / d
df lim , (2.27)
→0 log (L/)
where L is the linear system dimension.

For most objects, we may simplify the definition of the fractal dimension in
eq. (2.27) in the following way: When the length of the object is stretched by a fac-
tor of a, its volume (or mass) grows by a factor of adf . We obtain this interpretation by
rewriting eq. (2.27) (in the limit → 0) as
V∗ L df
= , (2.28)
d
If we now consider a rescaling of L by a according to L → aL, we find that V∗ ∝ adf .
Thus, the volume V∗ scales as claimed.
As a simple example, we again consider the Sierpinski triangle (see Figure 2.28).
Stretching its sides by a factor of two evidently increases its area by a factor of
Figure 2.27 Gold colloids at different three. Using this observation and the aforementioned proportionality, we find that
scales [50]. the Sierpinski triangle has fractal dimension log(3)/ log(2) ≈ 1.585.
2.4.3 The Box Counting Method

The box counting method is a method to numerically determine the fractal dimen-
sion of an object. It is conceptually easy because the underlying idea is close to the
mathematical definition.
In the box counting method, we superimpose a grid with grid spacing on the
fractal object. We define the number of boxes in the grid that are not empty (contain a
2.4 Fractals 37
Rescaling the Sierpinski triangle. Stretching the length by a factor of two increases the black area by a factor of three. Figure 2.28
We apply the box counting algorithm to a percolation cluster on the square lattice at pc and plot the number of Figure 2.29
nonempty boxes N() against the inverse grid spacing −1 in a log-log scale. For = 1, we reach the grid spacing
of the underlying lattice (lower cutoff). For = L, one box covers the whole lattice of linear size L (upper cutoff).
The slope of the linear fit (blue solid line) corresponds to the measured fractal dimension.
part of the fractal) as N(). We do this for a large range of and plot N() as a function
of in a log-log plot. We show a typical result in Figure 2.29.
We recognize a region where the slope is constant in the plot. In this region,
the object is self-similar and the slope equals the fractal dimension. Outside this
self-similar regime, the finite resolution and finite size of the object destroy the self-
similarity. To convince ourselves that this indeed reproduces the fractal dimension
that we defined before, recall that defines a characteristic length scale, while N()
is proportional to the volume of the fractal object. Numerically, the box counting
method has rather pronounced finite-size effects, because at the border of an object
very sparsely occupied cells can have an important weight.
Multifractality
By expanding the box counting method, we may observe multifractality, a generaliza-
tion of the fractal dimension. Instead of only considering whether a box is completely
empty or not, we now also count how much mass of the fractal, that is, how many
occupied points or pixels there are in a certain box i; we will denote this number by
Ni (). We introduce the occupation density pi () = Ni ()/N, where N is the total num-
ber of occupied points across all boxes. Thus, pi () is the fraction of the mass of the

fractal contained in box i. Since pi () ≥ 0 for all i and i pi () = 1, the functions pi ()
form what is called in mathematics a measure.
Let us define the q-mass Mq of an object as

Mq = pi ()q . (2.29)
i
The q-mass varies with according to
Mq ∝ τq , (2.30)
where

log Mq
τq = lim . (2.31)
→0 log ()
Furthermore, the associated generalized dimensions dq = τq /(1 − q) are [51–53]

1 log Mq
dq = lim , (2.32)
1 − q →0 log ()
where d0 is the fractal dimension (df ) as defined before and d1 is the information dimen-
sion. In information theory, the mathematical structure of dq is similar to the Rényi
entropy.
Additional Information: Information Dimension
To evaluate the term Mq 1/(1−q) in the limit q → 1, we set q = 1 − q and consider the limit q → 0:
⎛ ⎞1 ⎛ ⎞1
⎜⎜⎜ 1−q ⎟⎟⎟ q ⎜⎜⎜ −q log(p ) ⎟⎟⎟ q
lim Mq 1−q = lim ⎜⎜⎝ pi ⎟⎟⎠ = lim ⎜⎜⎝ ⎟⎟⎠
1
pi e i
q→1 q →0 q →0
i i
⎡ ⎤1
⎢⎢⎢ ⎥⎥ q
= lim

⎢⎢⎣ pi 1 − q log(pi ) − O(q ) ⎥⎥⎥⎦
2
q →0
i
⎡ ⎤ q1 (2.33)
⎢⎢⎢ ⎥⎥
= lim

⎢
⎢⎣1 − q
pi log(pi )⎥⎥⎥⎦
q →0
i
⎡ ⎤
⎢⎢⎢ ⎥⎥
= exp ⎢⎢⎣− pi log(pi )⎥⎥⎥⎦ .
i
We find that
pi ln(pi )
lim dq = − lim i
. (2.34)
q→1 →0 log()
2.4 Fractals 39
The strange attractor arising from a solution of eq. (2.35) for a = 1.4 and b = 0.3 ((x0 , y0 ) = (0.6, 0.2)). Figure 2.30
Additional Information: Information Dimension (cont.)

That is, for q → 1, the generalized dimension dq corresponds to the Shannon entropy normalized by
log(). Because of this connection to information theory, the generalized dimension d1 is sometimes
referred to as information dimension. For q = 0, the dimension d0 is also called capacity dimension. A possi-
ble application of multifractal analysis is the characterization of the distribution of species in a certain habitat
(i.e., measuring species diversity) [54].
For simple fractals, dq does not depend on q and for multifractals dq decreases with
q. One example of a multifractal object is the attractor of the Hénon map, which
corresponds to the set of values (xn , yn ) that result from the iterations
⎛ ⎞ ⎛ ⎞
⎜⎜⎜ xn+1 ⎟⎟⎟ ⎜⎜⎜1 − ax2 + yn ⎟⎟⎟
⎜⎜⎜⎜ ⎟⎟ ⎜⎜ n ⎟⎟⎟⎟ ,
⎝y ⎟⎟⎠ = ⎜⎜⎝ bx ⎠ (2.35)
n+1 n
where n ∈ {1, 2, . . . }. This map was proposed by Michel Hénon. When iterating this
map of eq. (2.35) with certain a and b starting from any initial numbers (x0 , y0 ),
one ends up in an attractor, which for this specific case is called “strange attrac-
tor” because it is multifractal. We show an example for a = 1.4 and b = 0.3 in
Figure 2.30.
2.4.4 The Sandbox Method

We will now return to simple fractals and leave aside further considerations of multi-
fractality. The sandbox method is another technique that can be easily implemented to Figure 2.31 Illustration of the sandbox
numerically determine the fractal dimension of an approximately self-similar object. method for a percolation cluster on a
In general, it converges faster than the box counting method. We show an illustration square lattice: one measures the number
of this method in Figure 2.31. As in the case of the box counting method, we start with of filled pixels in increasingly large
a fractal object. First, we choose a site belonging to the fractal which is located in its concentrically placed boxes.
center region. Next, we place around this site a small box of size R on the fractal and
count the number of occupied sites (or pixels) N(R) in the box. We then successively
t
Figure 2.32 We apply the sandbox algorithm to one single percolation cluster at pc on a square lattice and plot the number of
nonempty sites N(R) against the linear box size R in log-log scale. The slope of the linear fit (blue solid line)
corresponds to the fractal dimension. To obtain a clearer linear dependence, one has to average each data point over
several clusters.
increase the box size R in small steps until we cover the whole picture with our box,
always storing N(R). We finally plot N(R) as a function of R in a log-log plot where the
fractal dimension is the corresponding slope (see Figure 2.32).
2.4.5 The Ensemble Method

Another possibility to measure the fractal dimension having many objects of different
sizes M is to determine their radius of gyration
)
1 2
Rg = ri − r j , (2.36)
M(M − 1) i j
where ri and r j are the positions of lattice sites i and j, respectively. We can now extract
the fractal dimension based on the proportionality
M ∝ Rdgf . (2.37)
This so-called ensemble method can, for instance, be used to determine the fractal
dimension from many clusters of a percolation configuration at the critical occupation
probability pc .
Exercise: Fractal Dimension of the Percolating Cluster

Compute the fractal dimension of a percolating cluster at pc and close to pc .
To find the percolating cluster you may use the burning method:
• The procedure is exactly the same as in the burning algorithm, except that only one occupied site is set on
fire.
2.4 Fractals 41
Exercise: Fractal Dimension of the Percolating Cluster (cont.)
• If the fire reaches the other side of the cluster, the system percolates and all the burned sites belong to the
percolating cluster. (Note: To find all sites of this cluster, you will have to burn further until the fire dies
out.)
• If the fire dies out without reaching the other side, the chosen starting site did not belong to the percolating
cluster. Choose another site from the first row. If the fire does not reach the other side starting from any
(filled) site in the first row, throw the sample away, make a new one, and try again.
To compute the fractal dimension of the obtained spanning cluster, assign to all sites that belong to the
spanning cluster a “1” and to all remaining sites a “0.”
Task 1: Box Counting Method
Once the percolating (or largest) cluster is found, measure the fractal dimension of the percolating cluster at
the critical point using the box counting algorithm:
1. Measure the number N of occupied cells for different grids. (A cell is considered to be occupied if it contains
at least one site of the percolating cluster.)
2. The finest grid, here called the “fine grid” is the original lattice (lower cutoff). Coarser grids consist of larger
cells which contain several cells of the finer one. A cell of a coarser grid is occupied if it contains at least one
occupied cell of the finer grid.
3. Plot N as a function of the inverse size of the cell, the slope of which is the fractal dimension.
Task 2: Sandbox Method
Measure the fractal dimension of a percolating cluster at the critical point using the sandbox algorithm:
1. Compute the mass N(R) (number of sites) of that part of the percolating cluster which lies within a square
box of size R that is located at the center of the system, where R ≤ L (L: system size of a square lattice).
(Note: For the sandbox method, the center of the boxes must be an occupied site.)
2. Vary R from small values R L (e.g., R = 3) to the maximal value R = L.
3. Plot N(R) as a function of R and measure the fractal dimension (the slope in the log-log plot).
2.4.6 The Correlation Function Method

In the correlation function method, we consider the spatial correlation function of
the density of an object. Such a function describes how the density is correlated over
various distances. Large correlations mean that two quantities strongly influence each
other, whereas no correlations (i. e., being equal to zero) would indicate that they are
uncorrelated.
Additional Information: Scaling of c(r) and g(r)

For percolation, we defined the correlation function c(r) according to eq. (2.8), which led to the scaling
c(r) ∝ r−(d−2+η) at the percolation threshold pc (see eq. (2.11)). In the definition of c(r), the origin
(r = 0) is part of the percolation cluster, whereas the definition of g(r) in eq. (2.38) is not based on this
assumption. The density correlation function g(r) can thus be seen as an average over c(r) [55, 56]. Using
df = d − β/ν (see eq. (2.21)) and the scaling relations of Section 3.2.4, we find that c(r) ∝ r−(d−2+η) ∝
r2(df −d) at pc , which is different from the scaling (2.39).
Let ρ(x) be the density of the object at point x. The correlation function of the
density at the origin and at a distance r is
g(r) = ρ(0) · ρ(r) . (2.38)
The angle brackets · denote a suitable averaging (for instance, over all pairs of points
separated by a distance r). For a fractal object, the correlation function g(r) decays like
a power law as [57]
g(r) ∝ rdf −d . (2.39)
Figure 2.33 Example of a volatile fractal.
Therefore, in a double-logarithmic plot of g(r) as a function of r, the slope of the
(visually) linear part is the fractal dimension minus the dimension of the space the
object is embedded into.
The correlation function method is mostly used, when determining the frac-
tal dimension experimentally using scattering techniques, since the dependence
of the scattered intensity is the Fourier transform of the density correlation
function g(r).
2.4.7 Volatile Fractals

As we have seen in Section 2.3.8, the spanning cluster of percolation at pc is fractal and
its fractal dimension is determined by the critical exponents of the order parameter
Figure 2.34 Robert Brown (1773–1858) and the correlation length as given in eq. (2.21). Percolation clusters are volatile. In a
was a Scottish botanist. volatile fractal, the cluster is redefined at each scale. That is, the spanning cluster of a
system of linear size L1 < L2 is not necessarily part of the spanning cluster in a system
of size L2 as we show in Figure 2.33.
2.5 Walks
2.5.1 Random Walks

Random walks are used to model many phenomena, including
• the motion of particles in liquids (Brownian motion, named after Robert Brown
(see Figure 2.34)) [58, 59],
• stock price fluctuations [60, 61],
2.5 Walks 43
(a) (b)
In the left panel, we show the binomial distribution (eq. (2.41)) for t ∈ {10, 102 , 103 } and the corresponding Figure 2.35
Gaussian approximation (eq. (2.47)). We observe a data collapse when plotting P(r, t)t1/2 against r/t1/2 in the
right panel.
• foraging behavior of animals [62], and

• polymer chains at the theta point [63, 64].
Moreover, random walks are also important in search algorithms such as the famous
PageRank algorithm [65]. We could continue this list with many more examples from
various disciplines [66]. For the interested reader, we refer to Refs. [67, 68].
Mathematically, a random walk is a stochastic process

t
X t = X0 + Zj , (2.40)
j=1
where Z1 , Z2 , . . . , Zt denote t independent random variables that can be scalars or vec-

tors. We set X0 = 0 and note that Xt can be defined either on a lattice (Zd ) or on a
continuous space Rd . In one dimension, one possibility is to consider binary random
variables Z j . That is, Z j = 1 with probability p and Z j = −1 with probability q = 1 − p.
Out of the t total steps, the walker moves l steps to the left (Z j = −1) and m steps to
the right (Z j = 1). Thus, after t steps, the walker is at position Xt = m − l = 2m − t.
The probability that the walker is at position Xt = 2m − t is distributed according to a
binomial distribution [68]

t m t−m
P (Xt = 2m − t) = p q . (2.41)
m
In Figure 2.35, we illustrate the binomial distribution of eq. (2.41) and the corre-
sponding Gaussian approximation (see eq. (2.47)). We observe a data collapse when
plotting P(r, t)t1/2 against r/t1/2 .
Additional Information: Properties of 1D Random Walk

Based on eq. (2.41), we can now build some intuition about the behavior of the considered 1D random walk by
computing the expectation value μr and variance σ2r of the walker position r = 2m − t. Note that we use
r instead of Xt for the sake of brevity. However, before doing so, we first have to determine the expectation
value μm and variance σ2m of m:
Additional Information: Properties of 1D random walk (cont.)
t
t m t−m
μm m = m p q
m=0
m
t
d t m t−m d (2.42)
=p p q = p (p + q)t
dp m=0 m dp
= tp (p + q)t = tp ,
and
σ2m m2 − m2 = m2 − (tp)2
t
t m t−m
= (m(m − 1) + m) p q − (tp)2
m=0
m
t
2 d
2
d t m t−m (2.43)
= p +p p q − (tp)2
dp2 dp m=0 m
2

2 d d
= p +p (p + q)t − (tp)2 = tp − tp2 = tpq .
dp2 dp
The relative width σm /μm decreases with t according to σm /μm ∝ t−1/2 . Now we are also able to deter-
mine the expectation value μr and the variance σr of the walker position r = 2m − t. Based on Eqs. (2.42)
and (2.43), we obtain
μr r = 2μm − t = t(p − q) , (2.44)
and
σ2r r2 − r2 = 4σ2m = 4tpq . (2.45)
If p q, there exists a bias in the average walking direction μr 0. However, we are more interested in
the undirected case where p = q = 1/2 such that μr = r = 0 and σ2r = r2 = t. In this case, the
mean-square displacement of the considered random walker is proportional to time t. In the limit of large t,
we use Stirling’s formula

1 1
ln t! = t + ln t − t + ln 2π + O t−1 , (2.46)
2 2
to approximate the binomial distribution of eq. (2.41) by a Gaussian [69]
2 r2
P(r, t) = √ e− 4Dt , (2.47)
4πDt
where D = 1/2 denotes the diffusion coefficient. Here we implicitly assumed that r ∈ {−t, . . . , t}
and t ∈ {1, 2, . . . } are integers such that the difference between two consecutive values of r or t is
Δr = Δt = 1.
2.5 Walks 45
(a) (b) (c)
Snapshots of three realizations of random walks (blue, orange, and green) that start at the center of the lattice. Figure 2.36
On a computer, we can implement a random walk on a lattice as follows:
Random Walk
• Initialize the walker at a certain lattice site.

• Select one of the possible neighbors uniformly at random and move the walker to the new lattice site.
• Repeat the previous step until a certain target length is reached.
It is also possible to simulate an off-lattice walk by randomly selecting an angle

φ ∼ U(0, 2π) instead of a certain neighboring lattice site. We show different realizations
of two-dimensional random walks at different times in Figure 2.36. The probability to
find the random walker at a distance (r, r + dr) from its starting point at t = 0 is
1 r2
P(r, t) dr = √ e− 4Dt dr , (2.48)
4πDt
where D is the diffusion constant. In d dimensions, the normalizing factor is (4πDt)−d/2
and P(r, t) is a solution of the diffusion equation
∂
P(r, t) = D∇2 P(r, t) , (2.49)
∂t
with initial condition P(r, 0) = δ(r). This means that a random walk corresponds to a
regular diffusion in the continuum limit. In dimensions d ≥ 2, the fractal dimension of
a random walk is df = 2 [68] and the mean-square displacement is proportional to t
(i. e., r2 ∝ t2/df ). Moreover, Mandelbrot conjectured that the fractal dimension of the
boundary of a planar random walk is 4/3. This was proven by Lawler, Schramm, and
Werner in 2000 [70]. Figure 2.37 George Pólya (1887–1985)
In 1921, George Pólya (see Figure 2.37) demonstrated that the probability of a ran- was a professor for mathematics at ETH
dom walker to return to the origin equals 1 only in one and two dimensions. Table 2.2 Zürich and at Stanford University.
summarizes some return probabilities for various dimensions [71]. In 1951, Dvoretzky Photograph courtesy ETH Library.
Table 2.2 The probability that a random walker will return

to the origin for different dimensions
Dimension Return probability
1 1
2 1
3 0.34053732955
4 0.19320167322
5 0.13517860982
6 0.10471549562
7 0.08584493411
8 0.07291264996
9 0.06344774965
10 0.05619753597
Note: From Ref. [71].
and Erdös (see Figure 2.38) identified the following asymptotic behavior of the number
of visited sites Ncov (t) at time t [72, 73]:
⎧√
⎪
⎪
⎪ t, d=1 ,
⎪
⎪
⎨
Figure 2.38 Paul Erdös (1913–1996) was Ncov (t) ∝ ⎪t/ ln t , d=2 , (2.50)
⎪
⎪
⎪
one of the most influential ⎪
⎩t , d>2 .
mathematicians of the twentieth century.
There also exists anomalous diffusion that is characterized by r2 ∝ t2/df with df 2.
For df < 2, the process is referred to as enhanced (or super) diffusion and for df >
2 as subdiffusion. One example of an enhanced diffusion phenomenon is the Lévy
flight [74] in which the random walker jumps from the current location to a location
r ≥ rmin > 0 that is distributed according to
p(r) ∝ 1/r1+κ , (2.51)
where κ > 0. In d dimensions, the integral over the function p(r) from rmin to ∞ can be
normalized if κ > d − 1. The resulting fractal dimension is df = κ for κ < 2 and df = 2
for κ ≥ 2 [68].
Exercise: Random Walks

Task 1: Simple Random Walk
• Generate a two-dimensional random walk consisting of N positions (i.e., N − 1 segments). Within each
step, the step size is fixed (length of the segments) but the angle has to be chosen at random.
• Measure the square of the end to end distance R2 (distance between position 1 and N).
2.5 Walks 47
Exercise: Random walks (cont.)
• Generate several configurations k ∈ {1, . . . , M} (random walks of same N, but different sequence of
random numbers) and compute the average of R2 as well as the estimated error Δ:
* +
1 2 1 2 2 , 2 2 + 2 , 1 2 2
M M
R =
2
Rk , Δ = R − R , R2 = R .
M k=1 M M k=1 k
(2.52)
Choose the number M of configurations such that the estimated error is below a desired value (e.g., 1
percent of the average value).
• Vary the number N of positions and check your result by comparing it to the analytical result.
Task 2: Chains of Spherical Particles

Now, consider chains consisting of N hard spherical particles of radius r in two dimensions (disks) or in three
dimensions (spheres). The particles are not allowed to overlap and neighboring particles in a chain touch each
other (self-avoiding random walk).

• Measure R2 and Δ (same as in task 1).
• Again, vary the number of particles N and compare the behavior to the result of task 1.

• R2 , R2 , etc. depend on the number of particles N.
• The chain of nonoverlapping particles is created in the way similar to the simple random walk, but it has
to be checked that there is no overlap with other particles. If an attempt is made which would create an
overlap with any particle in the chain, the entire chain must be rejected, and one has to start all over again.
2.5.2 Self-Avoiding Walks

An important extension of a random walk is the concept of walks that cannot inter-
sect their own trajectory (i. e., fulfill the condition of “excluded volume”). The most
prominent example is the so-called self-avoiding walk (SAW). In one dimension, the
walker would just move in one direction, and for dimensions d ≥ 4, the intersection
probability is vanishingly small, so we observe a regular random-walk behavior with
fractal dimension df = 2 [68]. In two and three dimensions, the fractal dimensions are
df = 4/3 and df ≈ 5/3, respectively [75–77]. The SAW was first introduced by Paul
Flory (see Figure 2.39) to describe polymers in a good solvent [75, 78]. Figure 2.39 Paul J. Flory (1910–1985)
For the SAW, all configurations of same chain length N have the same statistical made important contributions to the field
weight [79]: of polymer physics and chemistry. He was
awarded the Nobel Prize in 1974.
ΩN = μN N θ , (2.53)
where μ denotes a chemical potential and θ = 11/32 for all two-dimensional lat-
tices [76, 80]. The generating function is equivalent to a grand canonical partition
function (see Section 3.1.4)

Z(x) = ΩN xN = (μx)N N θ , (2.54)
N N
where x is the fugacity which corresponds to the statistical weight of adding one
element to the chain. The radius of convergence of Z(x) defines the critical fugacity
xc = 1/μ. The average chain length is
⎧
-- ⎪
⎪
⎪finite , if xc > 1 ,
∂ ln(Z) - NΩ ⎪
⎪
⎨
N = - = N N
=⎪
⎪ critical , if xc = 1 , (2.55)
∂x - x=1 N ΩN ⎪
⎪
⎪
⎩infinite , if x < 1 .
c
There also exist other models for non self-intersecting walks, like the kinetic growth
walk [81] that have different statistical properties than the SAW.
Creating SAWs numerically, which satisfy the statistical properties of Eqs. (2.53)–
(2.55) correctly is not completely trivial and we will present next a few such techniques.
The simplest method is the rejection method. In this method, we grow a chain of
length N by starting from one site and randomly move to one of the empty neighbors.
We repeat this procedure until the required length is reached. If we try to move to
an already occupied neighbor (self-intersection), we have to start all over again. This
method is straightforward to implement, but the rejection rate becomes extremely high
for longer walks. If we want to generate many samples of SAWs of a certain length N,
this simple method is computationally too expensive.
To overcome that sampling issue, other methods have been developed that make
the sampling of SAWs of length N more efficient. One method is the reptation6
algorithm [82]:
Reptation Algorithm
Start with any nonintersecting walk of length N.
1. Choose one end of the walk.
2. Place last monomer from chosen end in a random direction at the other end.
3. If there is no overlap, retain this configuration otherwise, reject the move.
4. Repeat above until the desired number of configurations is reached.
All retained configurations correspond to samples of an SAW of length N. Another

method that enables us to efficiently obtain samples of SAWs is the kink-jump
algorithm [83]:
6 The name derives from the term reptile because during the simulation, the chain moves back and forth
like a snake.
2.5 Walks 49
1’ 4’ 5 6 7’
1 2 3 4 7 8
1’’ 8’
An illustration of possible kink-jump algorithm updates in a chain that consists of eight particles. Blue monomers Figure 2.40
correspond to the initial configuration, whereas gray monomers indicate possible positions for kink-jump movements.
1 2 5 6
2’ 3 4 7 8
1’
An illustration of a possible pivot algorithm update in a chain that consists of eight particles. Blue monomers Figure 2.41
correspond to the initial configuration, whereas the gray monomers correspond to a subwalk that was rotated
counter-clockwise at an angle of 90◦ .
Kink-Jump Algorithm
1. Uniformly at random select one monomer.
2. If the monomer lies at the end of the chain or at a corner, determine the possible new local positions that
do not disrupt the chain and do not violate excluded volume condition.
3. If no new positions are available, keep the current chain.
4. If new positions are available, return to step 1 until the desired number of configurations is reached.
We illustrate possible kink-jump updates of monomer positions in Figure 2.40. Blue

monomers correspond to the initial configuration and gray monomers indicate possi-
ble positions for updates according to the kink-jump algorithm. Instead of changing
the location of only one monomer, we can also consider sub-chains. This brings us to
the pivot algorithm [84] as described in the beginning of the next page.
Examples of sub-chain transformations include reflections and rotations about
an axis. We show one example of a rotation transformation in Figure 2.41. In this
example, we initially select the monomer with label “3” and then rotate the subwalk
consisting of monomers “1” and “2” counterclockwise at an angle of 90◦ .
Pivot Algorithm
1. Uniformly at random select one monomer as pivot. The walk can be regarded as consisting of two subwalks
connected by the pivot.
2. Choose one of the two subwalks and apply a transformation that does not disrupt the chain.
3. If this produces no overlap, accept this move, otherwise reject it.
4. Return to step 2 and repeat until the desired number of configurations is reached.
2.6 Complex Networks
We can simulate random walks not only on regular lattices as in Section 2.5, but
also on more general networks that mimic the structure of complex physical and
social systems. Examples of systems that can be mathematically treated as networks
include power grids, railways, air traffic, gas pipes, highways, and the Internet. We
show two illustrations of networks in Figure 2.42. In the following sections, we pro-
vide an overview of concepts from network science, an interdisciplinary research field
that combines methods from graph theory, statistical physics, and computer science.
For further reading on network science, there exist many excellent textbooks [85–87].
2.6.1 Adjacency Matrix

Mathematically, a network G(V, E) is an ordered pair, where V and E are the corre-
sponding sets of nodes and edges. Connections between nodes are described by the
adjacency matrix A. We set the matrix element Auv = 1 if there exists an edge e ∈ E
between the node pair u, v ∈ V and otherwise we set Auv = 0. In a directed network
Auv can be different from Avu , whereas Auv = Avu for undirected networks. The adjacency
matrix of the (undirected) graph that we show in Figure 2.43 is
⎛ ⎞
⎜⎜⎜0 1 1 1 0 1 1⎟⎟⎟
⎜⎜⎜⎜ ⎟⎟
⎜⎜⎜1 0 1 0 1 1 0⎟⎟⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟
⎜⎜⎜ ⎟
⎜⎜⎜1 1 0 1 0 0 1⎟⎟⎟⎟⎟
⎜⎜⎜ ⎟⎟
A = ⎜⎜⎜1 0 1 0 1 0 0⎟⎟⎟⎟ . (2.56)
⎜⎜⎜ ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟
⎜⎜⎜0 1 0 1 0 1 0⎟⎟⎟
⎜⎜⎜ ⎟
⎜⎜⎜1 1 0 0 1 0 1⎟⎟⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟
⎝ ⎠
1 0 1 0 0 1 0
In this example, the adjacency matrix is sparse and it is computationally more efficient
to use edge lists instead of storing the complete matrix. For complete graphs where
all nodes are connected with each other, we have to store the whole adjacency matrix.
The elements (u, v) of the matrix powers Ak represent the number of paths of length k
from node u to node v. For an undirected network, the degree of node u is

ku = Auv = Avu . (2.57)
v∈V v∈V
For regular networks, all nodes have the same degree. An example of a regular network
is the square lattice with periodic boundary conditions. For a directed network, we
distinguish between in-degree (number of incoming edges) and out-degree (number of
outgoing edges). Mathematically, we denote the in-degree of node u by

kuin = Avu , (2.58)
v∈V
and the corresponding out-degree by

kuout = Auv . (2.59)
v∈V (a)
For undirected networks, summing over ku yields

ku = Auv = 2|E| , (2.60)
u∈V u∈V v∈V
where |E| denotes the number of edges of the network G(V, E). This equation implies
that every undirected network has an even number of nodes with an odd degree. If
there would be an odd number of nodes with odd degrees, the right-hand side of
eq. (2.60) would not be even (i.e., equal to 2|E|). This is also known as the handshak-
ing lemma. In a group (“network”) of handshaking people, an even number of people
(“nodes”) must shake an odd number of other people’s (“neighboring nodes”) hands.
For directed networks, we obtain

kuin = kuout = |E| . (2.61)
u∈V u∈V
2.6.2 Degree Distribution (b)
Networks can be characterized by their degree distribution P(k). For a square lat- Figure 2.42 An Apollonian network [88]
tice, the degree distribution is P(k) = δk4 (i.e., each node has degree k = 4). If the with 1096 nodes and mean degree
distribution follows a power law k ≈ 5.99 (a). A school friendship
P(k) ∝ k−γ , (2.62) network (based on an Add Health survey)
with 2539 nodes and mean degree
the network is called scale-free [91]. Examples of networks that are approximately
k= 8.24 (b). Both illustrations were
scale-free are the Internet and some social networks [92, 93]. One mechanism that
created with Gephi [89]; different colors
produces scale-free networks is preferential attachment where each new node is more
represent different communities.
likely to attach to existing nodes with high degree. Mathematically, each new node will
be attached to m ≤ m0 existing nodes and the attachment probability is proportional
to the number of edges of the existing nodes. Here, m0 is the initial number of nodes.
Networks that result from this type of preferential attachment are called Barabási–
Albert networks and their degree distribution is P(k) ∝ k−3 . We show an example of a
Barabási–Albert network in Figure 2.44.
Networks that are generated by a random process are also called random-graph
models [94, 95]. Another example of a random-graph model are Erdős–Rényi net-
works [96]. To generate them, we start with N isolated nodes and add new edges
between two uniformly at random selected nodes with probability p. The resulting
degree distribution is binomial:

N−1 k
P(k) = p (1 − p)N−1−k , (2.63)
k
where N is the number of nodes.
Watts–Strogatz networks [97] also belong to the class of random-graph models and
are generated by arranging nodes in a ring and connecting each node to its K nearest
Figure 2.43 An example of a graph with neighbors. The K/2 rightmost edges of each node are then rewired (i. e., reconnected)
seven nodes. with probability p (see Figure 2.45). The degree distribution of a Watts–Strogatz
network is [98]
(k,K)
f
K/2 (pK/2)k−K/2−l −pK/2
P(k) = (1 − p)l pK/2−l e , (2.64)
l=0
l (k − K/2 − l)!
where f (k, K) = min(k − K/2, K/2).
2.6.3 Small-World Networks

As a further network characteristic, we use dmin (s, t) to denote the shortest-path length
between a source node s and a target node t for each node pair s, t ∈ V. In a complete
graph, all nodes are connected and the shortest-path length between two nodes s and
t is dmin (s, t) = 1. Certain networks such as Watts–Strogatz networks (see Figure 2.45)
Figure 2.44 The degree distributions of a are so-called small-world networks. That is, the shortest-path length d between two
Barabási–Albert network. Black disks randomly chosen nodes satisfies
represent the degree distribution of a
d ∝ log(N) , (2.65)
single network realization (N=1600
nodes), and the gray solid line is the while the clustering coefficient C is significantly larger than in an equivalent random
corresponding analytic degree network. Mathematically, the local clustering coefficient c(u) of a node u with degree
distribution (power law with exponent ku ≥ 2 is the number of triangles T (G; u) that include that node divided by all possible
−3 [90]). The inset is an example of a triangles that could include u (i.e., the connected triples) [99]:
Barabási–Albert network with 100 2T (G; u)
nodes. The sizes of nodes scale with their c(u) = . (2.66)
ku (ku − 1)
betweenness centralities (see eq. (2.68)).
If the degree of a node is 0 or 1, the corresponding local clustering coefficient of that
node is 0. The mean local clustering coefficient of a network is
1
C(G) = c(u) . (2.67)
|V| u∈V
The interesting small world situation in which eq. (2.65) is valid (and the clustering
coefficient is high) occurs for intermediate values of p (see Figure 2.45). In scale-free
networks, one finds that d ∝ log log(N) [100]. These networks are said to have ultra
small-world properties.
Different regimes in the Watts–Strogatz model. Figure 2.45
To analyze networks with respect to their shortest-path characteristics, we use σ st

to denote the total number of shortest paths from s to t. The number of shortest paths
from s to t that traverse v ∈ V is σ st (v). The geodesic node betweenness centrality cB for
a node v is the sum (over all node pairs s, t in a network) of the number of shortest
paths from s to t that traverse v divided by the total number of shortest paths from s
to t [87]. Including a normalization factor, the geodesic node betweenness centrality is
defined as
σ st (v) 2
cB (v) = . (2.68)
σ (|V| − 1)(|V| − 2)
s,t∈V st

normalization factor
Note that σ st = 1 if s = t and that σ st (v) = 0 if v ∈ {s, t}. We define the geodesic node
betweenness of a network G as
1
C B (G) = cB (v) . (2.69)
|V| v∈V
That is, C B (G) is the mean geodesic node betweenness centrality over all nodes in a
network.
A large geodesic betweenness centrality cB (v) indicates that many shortest paths tra-
verse a particular node. This may help identifying important nodes in social, physical,
and communication networks [87, 101–103].
2.6.4 Dijkstra’s Algorithm

In the previous section, we defined the shortest path between two points on an arbi-
trary network G(V, E), where V is a set of nodes and E is a set of edges. With the Figure 2.46 Edsger W. Dijkstra
burning method (see Section 2.3.1), we have already seen an example of an algorithm (1930–2002) was a professor of
that identifies the shortest-path length (“burning time”) on a percolation cluster. For a computer science and mathematics at the
general graph G(V, E), Edsger Dijkstra (see Figure 2.46) proposed a greedy algorithm7 University of Texas at Austin and the
in 1959. Eindhoven University of Technology.
7 A greedy algorithm tries to find a global optimum through many local optimizations. At each step, the
algorithm makes the locally optimal choice.
Dijkstra’s Algorithm
1. We assign to every node a temporary distance value (zero for our initial node and infinity for all other
nodes).
2. We set the initial node as “burned” and mark all other nodes as “unburned.” All unburned nodes are stored
in a corresponding set.
3. For all unburned neighbors of the current node, we compute the corresponding distances. If the computed
distance is smaller than the current one, we assign the smaller value. Otherwise, we keep the current
value.
4. After having considered all of the neighbors of the current node, we mark the current node as “burned”
and remove it from the set of unburned nodes. We will not consider a burned node again.
5. We stop the algorithm if we reach the destination node or if the smallest tentative distance among the
nodes in the unvisited set is infinity.
6. Otherwise, we select the “unburned” node that is marked with the smallest distance, consider it as current
node, and return to step 3.
When combined with additional optimization techniques, the run time of Dijk-
stra’s algorithm is proportional to |E| + |V| log |V|. In addition to Dijkstra’s algorithm,
there also exists the Bellman–Ford–Moore algorithm that can also handle negative
weights [104].
Exercise: Analyzing Networks

In this exercise, we generate (i) Erdős–Rényi and (ii) Barabási–Albert networks and analyze some of their
properties.
(i) To build an Erdős–Rényi network, begin with a set of N isolated nodes and iterate over all possible node
pairs. Two nodes are connected with probability p.
(ii) Barabási–Albert networks are constructed via preferential attachment after starting from a connected
network with m0 nodes. Preferential attachment means that a new node is connected to m ≤ m0
existing nodes and the probability pi of being connected to an existing node i is proportional to the

degree ki of that node (i.e., pi = ki / j k j ).
Task 1: You can start with networks that have N = 100 nodes and later extend your analyses to networks
with 1000 and 10,000 nodes. Vary the connection probability p for Erdős-Rényi networks and the parameter
m for Barabási–Albert networks. What is the influence of these variations on the degree distribution?
Task 2: Implement Dijkstra’s algorithm and determine the average path length between two nodes for (i) and
(ii) as a function of the number of nodes N. What do you observe?
Equilibrium Systems 3
3.1 Classical Statistical Mechanics
In the first part of this chapter, our goal is to provide an overview of methods
to computationally study systems that can be described within the framework of
equilibrium statistical mechanics. We therefore first introduce the most important
concepts from classical statistical mechanics that enable us to mathematically capture
the macroscopic properties of interacting microscopic units. Historically, important
contributions to the microscopic formulation of thermodynamics were made by
Boltzmann and Gibbs [105, 106]. In particular, Gibbs’s notion of a statistical ensemble
enables us to interpret macroscopic physical quantities as averages over a large number
of different configurations of a system’s interacting units.
3.1.1 Phase Space
Let us consider a three-dimensional classical physical system that consists of N parti-

cles. We need a three-component vector to describe the position of a certain particle Figure 3.1 Ludwig Boltzmann
and another three-component vector to describe its momentum. Therefore, we denote (1844–1906) is one of the fathers of
the canonical coordinates and corresponding conjugate momenta of all N particles by statistical mechanics and formulated one
q1 , . . . , q3N and p1 , . . . , p3N , respectively. The 6N-dimensional space Γ that results from version of the ergodicity hypothesis.
the set of canonical coordinates defines the phase space. This concept has been intro-
duced by Ludwig Boltzmann (see Figure 3.1). The considered N particles could simply
be uncoupled harmonic oscillators. In this case, the phase space of each single particle
would look like the one we show in Figure 3.2. By keeping certain external parameters
such as temperature and pressure constant, we can measure a macroscopic physical
quantity by computing the time average

1 T
QT = lim Q(p(t), q(t)) dt (3.1)
T →∞ T 0
over different realizations of the underlying microscopic states. However, computing

the time average of a certain macroscopic quantity makes it necessary to determine the
time evolution of all microscopic states. If the number of involved particles is large, this
approach would be computationally infeasible. Instead of computing time averages, Figure 3.2 Orbits in phase space of an
another possibility is to consider an average over an ensemble of systems in different undamped (blue) and a damped (orange)
microstates (i. e., specific microscopic configurations of a certain system) under the harmonic oscillator.
same macroscopic conditions.
55
56 Equilibrium Systems
The assumption that all states in an ensemble are reached by the time evolution of
the corresponding system is referred to as ergodicity hypothesis. Next, we define the
ensemble average of a quantity Q(p, q) as

Q(p, q)ρ(p, q) dpdq
Q = , (3.2)
ρ(p, q) dpdq
where ρ (p, q) denotes the phase space density and dpdq is a shorthand notation for
dp3N dq3N .
3.1.2 Liouville Theorem

The dynamics of the considered N particles is described by their Hamiltonian H(p, q).
The equations of motion are
∂H ∂H
ṗi = − and q̇i = (i = 1, . . . , 3N) . (3.3)
∂qi ∂pi
Moreover, the temporal evolution of a phase space element of volume V and boundary
∂V is given by

∂
ρ dV + ρv dA = 0 , (3.4)
∂t V ∂V
where v = ( ṗ1 , . . . , ṗ3N , q̇1 , . . . , q̇3N ) is a generalized velocity vector. By applying the
divergence theorem to eq. (3.4), we find that ρ satisfies the continuity equation
∂ρ
+ ∇ · (ρv) = 0 , (3.5)
∂t
where ∇ = (∂/∂p1 , . . . , ∂/∂p3N , ∂/∂q1 , . . . , ∂q3N ). We can further simplify eq. (3.5) with
the help of the identity
3N
3N
∂q̇i ∂ ṗi ∂ ∂H ∂ ∂H
∇·v= + = − = 0. (3.6)
i=1
∂qi ∂pi ∂qi ∂pi ∂pi ∂qi
i=1
=0
We use Poisson brackets1 to rewrite eq. (3.5) and obtain Liouville’s Theorem,
∂ρ
= {H, ρ} , (3.8)
∂t
which describes the time evolution of the phase space density ρ.
1 The Poisson bracket is defined as

∂u ∂v ∂u ∂v

{u, v} = − = −{v, u}. (3.7)
i
∂qi ∂pi ∂pi ∂qi
3.1 Classical Statistical Mechanics 57
3.1.3 Thermal Equilibrium

In thermal equilibrium, the system reaches a steady state in which the distribution of
configurations is time independent. That is, the phase space density satisfies ∂ρ/∂t = 0.
Liouville’s theorem (eq. (3.8)) then leads to
v · ∇ρ = {H, ρ} = 0 . (3.9)
One possibility to satisfy this equation is to take a phase space density ρ that depends
on quantities such as energy or the number of particles, which are conserved during the
time evolution of the system. We may then use such a phase space density to replace
the time average (eq. (3.1)) by a corresponding ensemble average (eq. (3.2)).
In the subsequent sections, we also consider discrete configurations X for which we
define the ensemble average as
1
Q = Q(X)ρ(X) , (3.10)
Ω X

where Ω is the normalizing volume such that Ω−1 X ρ(X) = 1. With the help of ensem-
ble averages, systems can be described by macroscopic quantities such as temperature,
energy, and pressure.
3.1.4 Ensembles
We perform measurements to determine some observable quantities that character-
ize a given physical system while keeping other parameters constant. Mathematically,
such measurements can be seen as an “ensemble average” (Section 3.1.3), which is
based on a phase space density ρ that only depends on conserved quantities. However,
we should bear in mind that it is not possible to independently adjust the values of
all parameters. As an example, we may consider a classical gas. The dynamics of the
gas is described by perfect elastic collisions and therefore it is impossible to compress
the volume by keeping pressure and temperature unchanged at thermal equilibrium.
The system behaves differently depending on which quantities are kept constant. Some
quantities such as volume V and pressure p are conjugate to another. Either V can be
held constant or p, not both.
Other examples are energy E and temperature T , particle number N and chemical
potential μ, magnetization M and magnetic field H. Depending on the parameters held
constant, the system is described by a
• microcanonical ensemble: constant E, V, N,
• canonical ensemble: constant T , V, N,
• canonical pressure ensemble: constant T , p, N, Figure 3.3 Josiah W. Gibbs (1839–1903)
• grandcanonical ensemble: constant T , V, μ. introduced the concept of statistical
ensembles and derived the laws of
The notion of a physical ensemble was introduced by Josiah W. Gibbs (see Figure 3.3). thermodynamics from statistical
In the subsequent sections, we consider microcanonical and canonical systems. mechanics.
3.1.5 Microcanonical Ensemble

In a microcanonical ensemble, number of particles, volume, and energy are constant.
Any configuration X of the system has a constant energy E (X). The phase space
density is
1
ρ (X) = δ (H (X) − E) , (3.11)
Zmc
where Zmc is the partition function of the microcanonical ensemble:

Zmc = δ (H (X) − E) . (3.12)
X
3.1.6 Canonical Ensemble
Microcanonical ensembles are difficult to realize experimentally because every energy

exchange with the environment has to be suppressed. It is more common to deal with
systems exhibiting a fixed temperature T such as experiments at room temperature.
The corresponding ensemble is called canonical ensemble (see Figure 3.4).
At a given temperature T , the probability for a system to be in a certain configura-
Figure 3.4 In a canonical ensemble, we tion X with energy E (X) is
consider a system (system 1) that is
1 E (X)
coupled to a heat reservoir (system 2). ρ (X) = exp − , (3.13)
ZT kB T
The heat reservoir guarantees a constant
temperature. with the corresponding canonical partition function

E (X)
ZT = exp − . (3.14)
X
kB T
According to eq. (3.10), the ensemble average of a quantity Q in the canonical
ensemble is
1 − E(X)
Q = Q (X) e kB T . (3.15)
ZT X
3.1.7 Classifying Phase Transitions

Phase transitions are omnipresent in many physical, biological, and ecological sys-
tems [107]. Important examples of phase transitions include the solid–liquid and
liquid–gas transitions of water and other liquids. In Section 3.2, we will see that phase
transitions also occur in models of magnetism. In general, phase transitions occur in
systems in which underlying processes lead from a disordered to an ordered phase.
In an ordered phase, the constituents of the system (molecules, spins, etc.) are in a
state that exhibits some symmetry. In a thermodynamic system whose interactions
are defined by a Hamiltonian, phase transitions result from the competition between
internal energy U (or enthalpy H) and entropy S . To mathematically characterize this
competition, we use the Helmholtz free energy F = U − T S and the Gibbs free energy
G = H −T S as relevant thermodynamic potentials. The temperature T acts as a control
3.2 Ising Model 59
parameter to go from a disordered phase (“high temperature”) to an ordered phase

(“low temperature”). The transition occurs at a temperature T c . Based on a more his-
torical classification of Paul Ehrenfest, we say that a phase transition is of nth order
if [108]
m -- m --
∂ F -- ∂ F --
- = - (3.16)
∂T m V -T =T + ∂T m V -T =T −
c c
and
-- m --
∂m F -- ∂ F --
-- = - , (3.17)
∂V m T =T c+ ∂V m -T =T −
c
for m ≤ n − 1, and otherwise

-- n --
∂n F -- ∂ F --
-- - (3.18)
∂T n V T =T + ∂T n V -T =T −
c c
and
-- n --
∂n F -- ∂ F --
-- - . (3.19)
∂V n T =T c+ ∂V n -T =T −
c
We defined a phase transition of nth order based on the Helmholtz free energy F
(constant volume and temperature ensemble) because we will also use this potential
in our discussion of magnetic systems in Section 3.2. For systems at constant pressure
and temperature, it may be useful to use the Gibbs free energy G for which Eqs. (3.16)–
(3.19) can be used as well to define a phase transition of nth order. We again note that
the Ehrenfest classification is a more historical one. The majority of phase transitions
are of first and second order and modern classifications mainly distinguish between
these two cases. Second-order phase transitions with n > 1 are also called critical
points.
For other systems whose interactions are not defined by a Hamiltonian, it is possible
to classify phase transitions according to the behavior of the order parameter [109].
If the order parameter goes continuously to zero at T c , the phase transition is of sec-
ond (or higher) order. For discontinuous changes of the order parameter, the phase
transition is said to be of first order.
3.2 Ising Model
Because of its historical relevance for the study of phase transitions in statistical
physics and its broad applicability in many other fields, we now apply the outlined
terminology from Section 3.1 to the Ising model. Wilhelm Lenz proposed this model
to his doctoral student Ernst Ising (see Figure 3.5) to describe systems that are com-
posed of magnetic dipole moments. Each dipole moment can be in one of two states
(+1 or −1). The original goal was to model phase transitions in magnetic materials.
As part of his doctoral thesis in 1924, Ernst Ising showed that the one-dimensional
model exhibits no phase transition. For that reason, it was expected that this model
was of no particular use. Then in 1944, Lars Onsager (see Figure 15.4) derived the
partition function of the two-dimensional Ising model without magnetic field, finding
a critical point at a finite temperature T c [111] and later in 1949 published an equa-
tion to describe the temperature dependence of the magnetization. Onsager provided
no proof for his formula, but it is known that he derived it using Toeplitz determi-
nants [112]. A few years passed until a proof was established by Yang in 1952 [113].
Since then, the Ising model has been successfully applied to a large number of physical
and nonphysical problems such as magnetic systems, binary mixtures, and models of
opinion formation. Also when compared with experimental data, the Ising model has
been found to generally agree very well with observations made for certain magnetic
materials [114]. To date, no general analytical solution for the Ising model in three
dimensions is known. This is the reason why this mathematical model has been stud-
ied so intensively from a numerical perspective [115–119], using tools from statistical
Figure 3.5 Ernst Ising (1900–1998) was
physics – some of which we describe in this section.
a professor at Bradley University. Ising
We start our discussion of the properties of the Ising model by considering a two-
was born in Cologne and migrated to the
dimensional lattice with sites being in binary states σi ∈ {1, −1} that only interact with
United States. For more information
their nearest neighbors (see Figure 3.6). This restriction can be relaxed by letting the
about his biography, see Ref. [110].
sites also interact with the next-nearest neighbors and other extended neighborhoods.
The interaction of spins is described by the Hamiltonian

N
H ({σ}) = −J σi σ j − H σi , (3.20)
i, j i=1
where the first term denotes the interaction between all nearest neighbors (represented
by a sum over i, j), and the second term accounts for the interaction of each site with
an external magnetic field H, favoring the σi = +1 state if H > 0. In the ferromagnetic
case (J > 0), the energy is lowered by an amount of −J if two spins are parallel,
whereas the energy is increased by an amount of +J if two spins are antiparallel. For
H = 0, there exist two ground states at T = 0, namely, all spins up or all spins down.
If J < 0, the system is called antiferromagnetic, because antiparallel spins lead to an
energy minimization. No interaction occurs for J = 0. In the case of a ferromagnetic
Ising system, the first term in eq. (3.20) tries to create order by minimizing the overall
energy as a consequence of aligning spins in the same direction. The second term tends
to align the spins in the direction of the external field H. While the energy is lower in
Figure 3.6 An illustration of the an ordered state, thermal fluctuations tend to destroy the order by flipping single spins.
interaction of a magnetic dipole (black) For temperatures above a critical value T c , such fluctuations dominate the system
with its nearest neighbors (dark gray) on and there is no long-range alignment of spins observable anymore. However, the
a square lattice. domains of aligned spins grow as the temperature decreases below the critical
value. This transition, like any other phase transition (occurring in superconductors,
superfluids, and sol–gel systems) between ordered and disordered states, can be char-
acterized by an order parameter (see also Section 2.3.3). For the Ising model, the order
parameter is the spontaneous magnetization, which we define in the next section.
The value of the critical temperature depends on the underlying lattice. √ For
a two-
dimensional square lattice, the critical temperature (kB T c /J = 2/ ln 1 + 2 ≈ 2.269)
3.2 Ising Model 61
(a) (b)
(a) The magnetization M(T, H) for different fields H ≥ 0 (indicated by the arrow) as a function of T . The black Figure 3.8
solid line represents the spontaneous magnetization MS (T ) for H = 0 and should be interpreted in the sense that
limH → 0+ M(T, H). (b) Phase diagram with the first-order transition line (blue) that results from a sign change
of the external field.
has been computed analytically by Kramers and Wannier in 1941 [120] and Onsager
(see Figure 15.4) in 1944 [111]. In three dimensions, extensive Monte Carlo simulations
have been used by David Landau (see Figure 3.7) and collaborators to determine the
critical temperature to T c = 4.5115232(1) (J = kB = 1) [117]. In addition to the second-
order phase transition, there also exists a first-order transition in the Ising model. It
occurs below the critical temperature if the direction of the external magnetic field
changes. We illustrate this behavior in Figure 3.8. We observe a jump in the magne-
tization at H = 0 when we cross the first-order transition line. This jump is twice the
order parameter.
3.2.1 Order Parameter

The magnetization is
. / Figure 3.7 David Landau (University of
1
N
M (T, H) = σi , (3.21) Georgia) has made many high-precision
N i=1 Monte Carlo calculations.
and corresponds to the ensemble average of the mean value of all spins. If the external
field vanishes, the Hamiltonian is invariant under a simultaneous reversal of all spins.
In other words, a certain equilibrated configuration of spins would also have the same
energy if we would change the sign of every single spin.
Thus, the ensemble average defined by eq. (3.21) would not be a good measure of
the magnetization because it corresponds to an ensemble average over all possible
configurations. For every temperature, M (T ) vanishes because for every configuration
there exists a configuration of opposite sign which neutralizes it. Therefore, we use the
so-called spontaneous magnetization
. N /
1
MS (T ) = lim+ σi , (3.22)
H→0 N i=1
as the order parameter of the Ising model. In the definition of MS (T ), the symmetry
of the Ising model is broken by applying a vanishing positive field H that aligns the
t
Figure 3.9 We show the formation of magnetic domains in the Ising model for different temperatures. For larger temperatures,
the configurations are disordered due to thermal fluctuations. The simulations were performed on a square lattice
with 512 × 512 sites. Black means that spins are pointing upward, and white indicates spins that point downward.
spins in one direction. Another possibility of breaking the symmetry can be realized
by keeping the boundaries of the lattice in a certain state. This is, however, impracti-
cable if periodic boundaries are being used. In Figure 3.9, we illustrate typical domain
configurations for different temperatures. If T > T c , one observes that thermal fluctu-
ations lead to disordered configurations. On the other hand, magnetic domains form
for small enough temperatures. Similar to eq. (2.1) in the case of percolation, in the
vicinity of the critical temperature for T < T c , the spontaneous magnetization scales
as [109]
MS (T ) ∝ (T c − T )β , (3.23)
where β = 1/8 in two and β = 0.326419(3) in three dimensions [118]. For T = T c and
H → 0, we find [109]
M(T = T c , H) ∝ H 1/δ , (3.24)
where δ = 15 in two and δ = 4.78984(1) in three dimensions [118]. The exponents β

and δ are critical exponents and characterize, together with other exponents, the under-
lying phase transition. Different techniques such as series expansions, field-theoretic
methods, and very sophisticated Monte Carlo algorithms exist to determine critical
exponents and critical temperatures with high precision.
3.2.2 Response Functions

Response functions are the second derivatives of the free energy and describe the
sensitivity of the thermodynamic functions to fluctuations. The magnetic susceptibility
3.2 Ising Model 63
∂M (T, H)
χ (T ) = , (3.25)
∂H
quantifies the change of the magnetization M in response to an applied magnetic
field H. Fluctuations of the spontaneous magnetization can be used to determine the
susceptibility using the fluctuation-dissipation theorem:
N 0 1
χ (T ) = MS (T )2 − MS (T )2 . (3.26)
kB T
Analogously, the specific heat is connected to energy fluctuations via
∂ E 1 0 1
C(T ) = = E (T )2 − E (T )2 . (3.27)
∂T kB T 2
Additional Information: Proof of the Fluctuation-Dissipation Theorem for the

Susceptibility of the Ising Model
To derive eq. (3.26) for the Ising model, we use the definition of the spontaneous magnetization (see
eq. (3.22)), and plug it into eq. (3.25) to obtain
∂ M (T, H)
χ (T ) = lim+
H→0 ∂H
N E0 +H i=1
N
σi
∂ 1 {σ} i=1 σ i exp k B T
= lim+ N .
H→0 ∂H N E0 + H i=1 σi
exp
{σ}
kB T

=ZT (H)

Here we used the definition of the ensemble average (see eq. (3.15)) with E0 = J i, j σi σ j and the
canonical partition function of the Ising Hamiltonian ZT (H). Using the product rule yields
N 2 N
E0 +H i=1 σi
1 {σ} i=1 σ i exp kB T
χ (T ) = lim+
H→0 Nk B T ZT (H)
N 2
E0 +H i=1 σi
1 {σ}
N
i=1 σ i exp kB T
− 2 32
NkB T ZT (H)
N 0 1
= MS (T )2 − MS (T )2 ≥ 0 .
kB T
Equations (3.26) and (3.27) describe the variance of magnetization and energy,
respectively. Similar to the power law scaling of the spontaneous magnetization
defined in eq. (3.23), we find for the magnetic susceptibility in the vicinity of T c
χ (T ) ∝ |T c − T |−γ , (3.28)
t
(a) (b)
Figure 3.10 Schematic plots of the susceptibility χ(T ) and specific heat C(T ) as functions of temperature. In the
thermodynamic limit, both quantities diverge at the critical temperature T c .
where γ = 7/4 in two and γ = 1.23701(28) in three dimensions [117], and note the
analogy with eq. (2.5) in the case of percolation. Moreover, the specific heat exhibits
the following scaling:
C (T ) ∝ |T c − T |−α , (3.29)
where α = 0 in two2 and α = 0.11008(1) in three dimensions [118]. The temperature

dependence of susceptibility and specific heat is illustrated in Figure 3.10.
3.2.3 Correlation Length

The spin–spin correlation function is
G(r1 , r2 ; T, H) = σ1 σ2 − σ1 σ2 , (3.30)
where the vectors r1 and r2 point to lattice sites 1 and 2. If the system is translationally
and rotationally invariant, the correlation function only depends on r = |r1 − r2 |. At
the critical point, the correlation function decays as
G(r; T c , 0) ∝ r−(d−2+η) , (3.31)
where η is another critical exponent and d is the dimension of the system. This scal-
ing behavior is similar to the one of percolation (see eq. (2.11)). In two dimensions,
η = 1/4 and η = 0.036298(2) in three dimensions [118]. For temperatures differ-
ent from the critical temperature, the correlation function exhibits an exponential
decay
G(r; T, 0) ∝ r−ϑ e−r/ξ , (3.32)
2 The exponent of α = 0 corresponds here to a logarithmic decay since
|x|−s − 1
lim = −ln |x| .
s→0 s
However, numerically, it is difficult to decide if the exponent is zero or just has a small value [121].
3.2 Ising Model 65
Table 3.1 The critical exponents of the Ising model in different dimensions
Exponent d=2 d=3 d≥4
α 0 0.11008(1) [118] 0
β 1/8 0.32630(22) [117] 1
γ 7/4 1.237075(10) [118] 1
δ 15 4.78984(1) [118] 3
η 1/4 0.036298(2) [118] 0
ν 1 0.629971(4) [118] 1/2
where ξ defines the correlation length. In the two-dimensional Ising model, the expo-
nent ϑ equals 2 above and 1/2 below the transition point [121, 122]. In the vicinity of
T c , the correlation length ξ diverges according to
ξ(T ) ∝ |T c − T |−ν , (3.33)
with ν = 1 in two dimensions and ν = 0.629912(86) in three dimensions [117]. Note

the similarities with Eqs. (2.10) and (2.11) for percolation.
3.2.4 Critical Exponents and Universality

The aforementioned six critical exponents are connected by four scaling relations [109]
α + 2β + γ = 2 (Rushbrooke) , (3.34)
γ = β(δ − 1) (Widom) , (3.35)
γ = (2 − η)ν (Fisher) , (3.36)
2 − α = dν (Josephson) , (3.37)
that have been derived in the context of the phenomenological scaling theory for fer-
romagnetic systems [109, 123]. Due to these relations, the number of independent
exponents reduces to two. The Josephson law includes the spatial dimension d of
the system and thus defines a so-called hyperscaling relation. Such relations are not
valid above the upper critical dimension which is equal to dc = 4 for the Ising model.
Above the upper critical dimension, the critical exponents are equal to those obtained
with the mean-field approximation and are thus independent of dimension (see
Table 3.1).
The importance of critical exponents becomes more clear when we study different
systems exhibiting phase transitions. Critical control parameters such as T c in the Ising
model sensitively depend on the interaction details. However, critical exponents only
107
106
105
104
Δμ*/DΔρ*⏐Δρ*⏐8–1
103
102 3He
4He
10
Xe
CO2
1
H2O
10–1
10–2
10–2 10–1 1 10 102 103 104 105 106
X+X0
X0
t
Figure 3.11 Widom scaling function for five different gases. The scaling variable is defined as x = ΔT |Δρ|−1/β , and x0
depends on the amplitude B of the power law for the coexistence curve Δρ = BΔT β [121]. The figure is taken
from Ref. [125].
depend on fundamental system properties such as dimension or symmetries and are

therefore said to be universal. Based on these observations, the universality hypothesis
states that for equilibrium systems all critical phenomena can be reduced to a small
number of universality classes [124]. All systems belonging to a certain universality
class share the same critical exponents and the same scaling behavior near the crit-
ical point. Universality is not a mere theoretical concept, but can be also observed
experimentally.
In Figure 3.11, we show an example of five different fluids undergoing a liquid-
gas transition. All substances exhibit different interatomic interactions, and still we
observe a clear data collapse for the rescaled chemical potential.
Monte Carlo Methods 4
Monte Carlo methods are broadly applicable to different problems that are (i) impos-
sible to solve analytically and (ii) difficult to handle for other numerical techniques
due to the large computational complexity [126]. For example, it may be difficult to
numerically explore the complete phase space of a certain model. Instead, we could
apply appropriate random sampling techniques (i. e., Monte Carlo methods) and only
explore relevant parts of phase space. In this chapter, we focus on Monte Carlo algo-
rithms to numerically study the properties of the Ising model and other Hamiltonian
systems. The basic idea behind Monte Carlo is that random sampling of the phase
space, instead of averaging over all states, may be sufficient to compute ensemble aver-
ages of certain thermodynamic quantities. If the number of samples is large enough,
the computed estimate converges toward the correct value. The main steps of Monte
Carlo sampling are as follows.
Monte Carlo Sampling
1. Randomly choose a new configuration in phase space (using a Markov chain).

2. Accept or reject the new configuration, depending on the strategy used (e.g., Metropolis or Glauber
dynamics).
3. Compute the physical quantity and add it to the averaging procedure.
4. Repeat the previous steps.
4.1 Computation of Integrals
To compute the ensemble average

Q(p, q)ρ(p, q) dpdq
Q = (4.1)
ρ(p, q) dpdq
of a thermodynamic quantity Q (e. g., the energy of a gas), we have to approximate
a high-dimensional integral. We therefore first focus on Monte Carlo methods that
are applicable to integrals, particularly higher-dimensional integrals (we will see later
on that Monte Carlo is in fact the most efficient way to compute higher-dimensional
integrals).
Let us consider the integral of a function g(x) in an interval [a, b]. We may approx-
imate the integral by choosing N points xi (regularly or randomly) on the x-axis with
67
68 Monte Carlo Methods
t
(a) (b)
Figure 4.1 Examples of Monte Carlo sampling (orange dots) for a smooth function (a) and for a function with a singularity (b).
their corresponding values g(xi ), summing over these sampled results, and multiplying
the resulting expression with the length of the interval:

1
b N
g(x)dx ≈ (b − a) g(xi ). (4.2)
a N i=1
If we choose the points xi uniformly at random, the process is called simple sampling,
which works very well if g(x) is smooth.
But what if we cannot make the assumption that g(x) is smooth? Let us consider,
for instance, a function with a singularity at a certain value (see Figure 4.1). Due to
the limited number of points that we would usually choose around the singularity (in
simple sampling), we would not be able to really appreciate and reproduce its behavior
and eq. (4.2) would be a bad approximation. We thus need more precision which goes
beyond the possibilities of simple sampling. An improvement is provided by a sec-
ond function p(x) for which g(x)
p(x) is smooth. The sampling points are now distributed
according to p(x), and we have
b b ⎡ ⎤
g(x) ⎢⎢⎢ 1 N
g(xi ) ⎥⎥⎥⎥
g(x) dx = p(x) dx ≈ (b − a) ⎢⎣ ⎢ ⎥, (4.3)
a a p(x) N i=1 p(xi ) ⎦
where the numbers xi have been sampled according to p(x). Thus, we have changed
our way of sampling by using the distribution function p(x). This manifests itself in
the summand above. One could state that p(x) helps us to pick our sampling points
according to their importance (e. g., select more points close to a singularity); in fact,
this kind of sampling is called importance sampling.
4.2 Integration Errors
We have outlined that the aforementioned Monte Carlo integration methods provide
approximations of certain integrals and not exact solutions. Naturally, the question
arises what error one has to expect in the course of such an approximation.
4.2 Integration Errors 69
Error in Conventional Methods

Let us first consider the error in conventional methods in which one uses a regular
spacing. As an example, let us focus on the trapezoidal rule: Consider the Taylor series
expansion of an integral over a function f (x) from x0 to x0 + Δx:
x0 +Δx
1 1
f (x) dx = f (x0 )Δx + f (x0 )Δx2 + f (x0 )Δx3 + . . .
x0 2 6

1 1 1
= f (x0 ) + ( f (x0 ) + f (x0 )Δx + f (x0 )Δx + . . . ) + . . . Δx
2
2 2 6
12 3
= f (x0 ) + f (x0 + Δx) Δx + O Δx3 .
2
This approximation, represented by 21 [ f (x0 ) + f (x0 + Δx)]Δx is called the trapezoidal
rule
(see
Figure 4.2 for a geometric interpretation). We see that the error is of order
O Δx3 . Suppose we take Δx to be one third its previous value, then the error will
decrease by a factor of 27! At the same time though, the size of the domain would also
be divided by three. The net factor is thus only 9 and not 27.
We now subdivide our interval [x0 , x1 ] into N subintervals of size Δx = x1N−x0 . The
trapezoidal rule approximation of the whole integral is therefore

Δx
x1 N−1
f (x) dx ≈ f (x0 + jΔx) + f (x0 + ( j + 1)Δx)
x0 2 j=0
Δx
=
[ f (x0 ) + 2 f (x0 + Δx) + 2 f (x0 + 2Δx) + . . .
2
+ 2 f (x0 + (N − 1)Δx) + f (x1 )] .

We thus see that while the error for each step is of order O Δx3 , the cumulative error

is N times this error and thus of order O Δx2 ∝ O N −2 .
The generalization to two dimensions is rather straightforward. Instead of each
interval contributing an error of O Δx3 , we now have a two-dimensional grid and

find for each cell of the grid an error that is proportional to O Δx4 . The cumulative

error is then proportional to NO Δx4 ∝ O Δx2 since N ∝ Δx−2 .
Graphical interpretation of the trapezoidal rule. Figure 4.2

We can do the same for a d-dimensional domain of the integral. The grid is now
segment will be O(Δx ) with the sum over all
d+2
d-dimensional, and the errorfor each
−d
“subintervals” giving us NO Δx d+2
with N ∝ Δx . Thus, the error is again propor-

tional to NO Δx d+2
∝ O Δx and independent of the dimension. We conclude that
2

the error in conventional methods is of order O Δx2 and given that N ∝ Δx−d , we

have Δx ∝ N − d and the error is of order O Δx2 ∝ N − d .
1 2
Monte Carlo Error

Let us first consider the simplified case of a one-dimensional function of one variable,
g : [a, b] → R.
If we consider a set of random points {xi } with i ∈ {1, . . . , N} that are uniformly
chosen from the interval [a, b], the estimate for the integral over g(x) is

b−a
b N
Q= g(x) dx ≈ g(xi ) = (b − a)ḡ , (4.4)
a N i=1
where ḡ is the mean of the considered samples. For M samples, we denote the ensemble
mean by
1
M
g = ḡ j . (4.5)
M j=1
We estimate the variance of ḡ with
1 2
M
σ2ḡ = ḡ j − g , (4.6)
M − 1 j=1
where the denominator is M − 1 as we want to obtain an unbiased estimator of the

variance.1 According to the central limit theorem, the variance of the ensemble average
Q of the integral Q is
(b − a)2 2 σ2
N
2 g
var(Q) = (b − a)2 σ2ḡ = σ g = (b − a) . (4.7)
N2 i=1
N
An unbiased estimator of σg , the variance of g, is given by
1
N
σ2g = (g(xi ) − ḡ)2 . (4.8)
N − 1 i=1
We estimate the error of Q in terms of

σg
δQ ≈ var(Q) = (b − a) √ . (4.9)
N
1 The expected value of an unbiased estimator is equal to the true value of the quantity to be esti-
mated [127].
We can now generalize this argument to multidimensional integrals:

b1 b2 bn
Q= dx1 dx2 . . . dxn g(x1 , . . . , xn ) . (4.10)
a1 a2 an
The previous integration interval [a, b] becomes a hypercube with volume V ⊆ Rn .

Instead of the interval [a, b] appearing in the variance, we now use the hypercube
volume V to obtain
σ2g
var(Q) = V 2 σ2ḡ = V 2 , (4.11)
N
and the error estimate becomes
σg
δQ ≈ var(Q) = V √ . (4.12)
N
As we can see, the proportionality to √1 still remains.
N
Comparison of Errors
We have seen that for the trapezoidal rule, the error of computing integrals converges
as N − d and thus depends on the dimension, while the error in Monte Carlo methods
2
is independent of the dimension. We can determine the dimension at which Monte

Carlo methods become more efficient,
− d2 ! 1
N 0 = √ , (4.13)
N
and we thus find d0 = 4. We conclude that for d > 4, Monte Carlo methods are more
efficient than the trapezoidal rule and should therefore be used whenever such higher-
dimensional integrals are calculated. Note that for higher-order numerical integration
methods (e. g., Simpson’s method), the corresponding value of d0 will be larger.
Figure 4.3 A cube filled with spheres to

4.3 Hard Spheres in a Box illustrate the sphere-packing problem.
Let us now look at an example of higher-dimensional integration. We consider a

sphere-packing problem, in which N hard spheres of radius R are randomly arranged
in a 3D box of volume V (see Figure 4.3). Our goal is two compute the average dis-
tance between these spheres. For more information on sphere packings, we refer the
interested reader to the textbook Sphere Packings, Lattices and Groups (3rd ed.), by
John Conway and Neil Sloane [128]
We describe the centers of the spheres by their position vector xi = (xi , yi , zi ) , 1 ≤
i ≤ N. The distance between two spheres is
*
2 2 2
ri j xi − x j + yi − y j + zi − z j . (4.14)
Note that the spheres are hard (i. e., they cannot overlap). Thus, the minimal distance
between two neighboring spheres is the distance when the spheres are in contact, which
is equal to 2R. This translates into the following condition for ri j :
!
ri j ≥ 2R . (4.15)
So far, this does not seem to have too much to do with integration because we are
simply placing some spheres in a volume V and measuring some distance ri j . Let us
say that we are interested in the average distance between the centers of spheres in the
box. Then we need to evaluate the following integral to obtain an analytical result:

1 2
ri j = ri j d r1 . . . d rN with Z =
3 3
d3 r1 . . . d3 rN . (4.16)
Z V N (N − 1) i< j V
We implicitly assume that the integrals in eq. (4.16) use distances ri j that satisfy the
“no overlap” condition (eq. (4.15)). Let us stay with this formula for a moment and
have a look at the different factors. The first factor, Z −1 , is a normalization factor.
The following factor is of combinatorial origin: We start with N choices (the whole
population) and pick one sphere uniformly at random. There are only (N − 1) choices
left for the second one (as it cannot be the first one again). This gives a total number
of N(N − 1) possible combinations for picking two spheres. As they are indistinguish-
able, there is an additional factor of 12 . Note that we would need to correct with an
additional factor of 21 if we would sum over i j, since in that case we would have
counted each distance twice which was avoided by simply summing over i < j. The
Monte Carlo approach to approximate eq. (4.16) is relatively simple:
Hard Spheres in a Box
• Choose a particle position (i.e., the center of the new sphere).

• Make sure that the new sphere does not overlap with any other sphere (see condition on ri j ). If it does
overlap, reject the position and try again.
• Once all the spheres have been placed, calculate the distances ri j .
• Calculate the average over all ri j .
This procedure will converge and give a value that approximates the integral under
the right conditions. It is important to note that our result will, of course, depend on
the number of spheres positioned: Imagine we had taken a ridiculously small number
of spheres (let us say only two) in a rather large volume and carried out the distance
measurement. There would not be much to average over and on top of that, due to
the relative liberty of positioning the two spheres in the volume, the distance (and with
it the average distance) would fluctuate wildly depending on the setup. If we repeat
this n times and average, the result will thus not converge as nicely as it would with
a relatively large number N of spheres where the space left between spheres would be
small. Placing many spheres in the volume would improve convergence but will slow
down the algorithm as with more spheres comes a higher rate of rejected positions.
For large N, it is even possible that the last few spheres cannot be placed at all!
Exercise: Hard Spheres in a 3D Box

Consider n hard spheres in a three-dimensional cube with edge length L. The radius of the spheres is R
and they are not allowed to overlap. Thus, the positions ri are not independent of each other and the inte-
gral . . . d3 r1 . . . d3 rn used to calculate average quantities (e.g., the average distance) over all possible
configurations is a conditioned integration. It is not easy to solve this integral numerically with conventional
methods. We are now employing Monte Carlo methods for approximating integrals in N dimensions:

1 (k)
M
f (x1 . . . xN ) dx1 . . . dxN = Z f (x1 . . . xN ) with Z = dx1 . . . dxN .
M k=1
(4.17)
The mean distance between the particles averaged over all possible configurations is defined by the condi-
tioned integration

1
rmean = rmean d3 r1 . . . d3 rn with Z = d3 r1 . . . d3 rn , (4.18)
Z
where the mean particle distance for one configuration is given by

2
rmean = ri j with ri j = (xi − x j )2 + (yi − y j )2 + (zi − z j )2 . (4.19)
n(n − 1) i< j
Calculate rmean by using a Monte Carlo method which is relatively easy to implement:
• Choose n particles with random position, such that they are not overlapping:
– choose particle position
– reject it if the excluded volume condition is not satisfied
– repeat until n particles are placed
• For the given configuration k, calculate rmean
(k)
.
• Repeat for M configurations, thus calculating
1 (k)
M
rmean r . (4.20)
M k=1 mean
Note that this equation can be easily derived from eq. (4.18) when applying eq. (4.17).
• Vary the number M of configurations and plot rmean versus the number M to check the convergence
of the integral.
• Vary the number n of particles to check the influence of the dimensionality of the integral on the con-
vergence. For that you should also try very small numbers of particles (e.g., n = 2, leading to a
six-dimensional integration). For this purpose, you may also consider a “one-dimensional box” (instead
of the three-dimensional one) to further reduce the dimension of the integral.
• The total sphere volume n · 4/3πR3 should be small enough compared to the total volume of the box so
that the particles can easily fit into the box.
• Choose different volume fractions ν = (n · 4/3πR3 )/L3 (total sphere volume divided by total volume).
What are the highest volume fractions you can achieve?
4.4 Markov Chains
In most cases, sampling from the phase space of a physical system based on uniform
random numbers is very inefficient because the underlying distribution may exhibit
peaks and regions where it is virtually zero. As an example, we may consider the kinetic
energy of an ideal gas. The distribution of the mean energy will exhibit a sharp peak,
which becomes sharper with increasing system size (see Figure 4.4).
There exist many different methods which avoid unnecessary sampling of regions
where the system is unlikely to be found (importance sampling). A common way to
efficiently choose appropriate samples out of a large pool of possible configurations is
to explore the phase space using a Markov chain. We therefore introduce the virtual
time τ and note that it only represents the steps of a stochastic process and should not
be confused with physical time. We start in a given configuration X and propose a new
configuration Y with a probability T (X → Y). We want the sum of this probability over
all possible new configurations to be equal to unity (normalization). Furthermore, the
probability associated with the transition from a configuration X to a configuration Y
is chosen to be the same as the transition probability of the inverse process, T (Y → X)
Figure 4.4 Example of an energy (“reversibility”). Last but not least, we recall from thermodynamics the ergodicity
distribution hypothesis (see Section 3.1.1), which states that the Markov process should be able
√ whose width is proportional
to 1/ Ld , where d is the system to attain every configuration. In particular, one must be able to reach any configura-
dimension and L is the linear system size. tion X after a finite number of steps. As a consequence, for a thermodynamic quantity
Q, the time average QT is equal to the ensemble average Q. Let us summarize these
properties:
1. ergodicity: T (X → Y) > 0 or in a more generalized sense, any configuration X in the

phase space must be reachable within a finite number of steps,

2. normalization: Y T (X → Y) = 1,
3. reversibility: T (X → Y) = T (Y → X).
In order to implement importance sampling, not every new configuration Y will be

accepted. To illustrate this, let us consider the Ising model: We may propose a new
configuration that will increase the energy of the system, but depending on the tem-
perature it may not be likely that the new configuration will also be accepted. At low
temperatures, it is unlikely that a spin flip will occur, so we will handle this with an
acceptance probability A. That is, we propose a new configuration and use an accep-
tance probability A such that we only sample the important regions in phase space. We
use A(X → Y) to denote the probability of acceptance of a new configuration Y start-
ing from a configuration X. In practice, we are interested in the overall probability of a
configuration actually making it through these two steps; this probability is the prod-
uct of the transition probability T (X → Y) and the acceptance probability A(X → Y)
and is called the probability of a Markov chain:
W(X → Y) = T (X → Y) · A(X → Y) . (4.21)

4.4 Markov Chains 75
If we are interested in the evolution of the probability p(X, τ) that the system is in state
X at time τ, we can derive the evolution equation in the following way:
• a configuration X is reached by coming from Y (this will contribute positively to

p(X, τ));
• a configuration X is left by going to some other configuration Y (this will decrease
the probability p(X, τ)).
The first of these two processes is proportional to the probability for a system to be in
Y, p(Y), while the second one needs to be proportional to the probability for a system
to be in X, p(X). When we combine all of this, we obtain the so-called master equation:
dp(X, τ)
= p(Y)W(Y → X) − p(X)W(X → Y) . (4.22)
dτ YX YX
We now summarize the properties of W(X → Y): Similarly to T (X → Y), the first
property is ergodicity. Any configuration X must be reachable after a finite number of
steps, which can be achieved through the sufficient condition W(X → Y) > 0 ∀X, Y.
If we sum over all possible configurations, we should obtain unity (normalization).
The third property, homogeneity, is new. It tells us that if we sum over all “initial
configurations,” the product of the probability for a system to be in Y multiplied by
its Markov probability for the transition to X is given by the probability for a system
to be in X. More comprehensibly, the probability for a system to be in X is simply a
result of systems coming from other configurations to X.
Thus, the properties of W(X → Y) are as follows:
1. ergodicity: ∀X, Y W(X → Y) > 0,

2. normalization: Y W(X → Y) = 1,

3. homogeneity: Y p(Y)W(Y → X) = p(X).
A Markov chain that satisfies conditions 1–3 converges toward a unique stationary dis-
tribution pst which does not change in time. To efficiently sample the relevant regions
of the phase space, the probability of the Markov chain W(·) has to depend on the
Hamiltonian. To achieve that, we set the stationary distribution pst , defined by
dpst (X, τ)
= 0, (4.23)
dτ
equal to the desired equilibrium distribution of the considered physical system peq :
!
pst = peq . (4.24)
It then follows from the stationary state condition of the Markov chain that

peq (Y)W(Y → X) = peq (X)W(X → Y ) . (4.25)
YX YX
A sufficient condition for this equation to hold is
peq (Y)W(Y → X) = peq (X)W(X → Y ) . (4.26)

We will refer to this condition as condition of detailed balance. As an example, in the

canonical ensemble at fixed temperature T , the equilibrium distribution is given by the
Boltzmann distribution (eq. (3.13))

1 E(X)
peq (X) = exp − , (4.27)
ZT kB T
0 1
with the partition function ZT = X exp − E(X)kB T .
Additional Information: Markov Chain Probability

The attentive reader might object that as W(X → Y) = A(X → Y)T (X → Y), with both

A(X → Y) ≤ 1 and T (X → Y) ≤ 1, W cannot satisfy Y W(X → Y) = 1. We now
explain that the conditions A(X → Y) ≤ 1 and T (X → Y) ≤ 1 are only seemingly in contradiction to

Y W(X → Y) = 1. We therefore note that we are looking at independent events here; the event of
choosing a new configuration is independent of the event of accepting it. The corresponding probabilities
A(X → Y) and T (X → Y) satisfy

A(X → Y) = 1 and T (X → Y) = 1 . (4.28)
Y Y
Furthermore, for

ai = 1 and bi = 1, (4.29)
i i
we find

ai bi = (a1 + a2 + . . . )(b1 + b2 + . . . )
ij
(4.30)
= a1 b1 + a1 (1 − b1 ) + a2 b2 + a2 (1 − b2 ) + . . .
= a1 + a2 + · · · = 1 ,

where we used i bi = 1 in the second step. For two independent events, we therefore obtain

i j ai bi = 1 again. When we apply this procedure to our example involving A(X → Y) and
T (X → Y), we obtain

T (X → Y)A(X → Y) = W(X → Y) = 1 . (4.31)
Y Y
4.5 M(RT)2 Algorithm
One possible choice of the acceptance probability satisfying the detailed balance
condition is
peq (Y)
A (X → Y) = min 1, , (4.32)
peq (X)
4.5 M(RT)2 Algorithm 77
which can be verified by rewriting eq. (4.26) and distinguishing between two cases.
This algorithm was developed in 1953 at Los Alamos National Laboratory and first
implemented by Arianna W. Rosenbluth (see Figure 4.5) on a MANIAC I computer.
It is often referred to as Metropolis or M(RT)2 algorithm. The original algorithm
has been extended by Wilfred K. Hastings in 1970 to account for nonsymmetric
0 E(X) 1 pro-
posal distributions [129]. For the canonical ensemble with peq (X) = ZT exp − kB T , the
1
acceptance probability becomes

ΔE
A (X → Y) = min 1, exp − , (4.33)
kB T
where ΔE = E(Y) − E(X) is the associated change of energy. Equation (4.33) implies
that the Monte Carlo (MC) update is always accepted if the Figure 4.5 Arianna W. Rosenbluth
energy
decreases, and
ΔE (1927–2020) first implemented the
if the energy increases, it is accepted with probability exp − kB T . It is straightfor-
ward to show that the M(RT)2 update rule satisfies the detailed balance condition M(RT)2 algorithm. M(RT)2 is an
(see eq. (4.26)). We observe that the detailed balance condition is clearly satisfied for abbreviation of the last names of the
ΔE = 0 and thus distinguish between the remaining two cases: (i) ΔE < 0 and authors of the original paper [130] that
(ii)
ΔE > 0. For ΔE < 0, we find that A (X → Y) = 1 and A (Y → X) = exp kΔE . The introduced the algorithm. RT is squared
BT
detailed balance condition (see eq. (4.26)) is indeed satisfied since because except for Nicholas C. Metropolis,
the other four authors of the paper
formed two married couples and
peq (Y) A(Y → X) = peq (X) A(X → Y) (4.34)
therefore carried the same family names

ΔE =1
=exp kB T
(Rosenbluth and Teller). The real
contribution of some of the authors (in
and thus
particular of Metropolis and of
peq (Y) ΔE
= exp − . (4.35) A. H. Teller) is subject to
peq (X) kB T controversy [131, 132]. It has been even
Case (ii) with ΔE > 0 is symmetric to case (i). stated by Roy Glauber and Emilio Segré
We use the M(RT)2 algorithm to explore the phase space of the Ising model by that the original algorithm was invented
flipping spins according to the acceptance probability of eq. (4.33). In summary, the by Enrico Fermi [133].
steps of one MC update of the M(RT)2 algorithm when applied to the Ising model are
as follows.
One MC Update of the M(RT)2 Algorithm
• Randomly choose a lattice site i.

• Compute ΔE = E(Y) − E(X) = 2Jσi hi with hi = i, j σ j andE = −J i, j σi σ j .
• Flip the spin if ΔE ≤ 0, otherwise accept it with probability exp − kΔE
BT
.
This Monte Carlo update is performed for many (randomly or regularly) chosen
sites and the time τ of the Markov process is measured in units of MC updates per
site. We should bear in mind that there is only a limited number of possible energy
differences in a lattice. To speed up the simulation, we can therefore create a lookup
t
(a) (b)
Figure 4.6 We show the spontaneous magnetization MS (T ) and the energy per spin E(T ) of the three-dimensional Ising
model for different temperatures T and linear system sizes L.
table that acts as a storage for all possible nearest-neighbor combinations. For example
hi ∈ {0, ±2, ±4} for an Ising model simulation on a square lattice. In Figure 4.6, we
show the spontaneous magnetization MS (T ) and the energy per spin E(T ) of the three-
dimensional Ising model. We performed these simulations with the M(RT)2 algorithm.
Exercise: 3D Ising Model and M(RT)2 Algorithm

Write a program for a Monte Carlo simulation to solve the 3D Ising model with periodic boundary conditions
and implement the M(RT)2 algorithm. Note that only a few energy values are possible. Therefore, it is helpful
to create look-up tables to avoid recomputing the Boltzmann factor.
1. Measure and plot energy E, magnetization M, susceptibility χ and heat capacity C at different tempera-
tures T .
2. Determine the critical temperature T c .
3. Study how your results depend on the system size (start with small systems).
For getting the average energy/magnetization you should measure the current values (e.g., each system
sweep) and average over them (see also remarks below).
• To have an idea of how many single Monte Carlo steps are reasonable, considering system sweeps may be
helpful.
• As a starting configuration you may consider a ground state, the system will then relax to the state
corresponding to a given temperature.
• For equilibration at a given temperature you may have to apply a couple of system sweeps. Take about 100
system sweeps to be relatively safe. You may also plot the relaxation process for some temperatures (e.g.,
M vs. MC-steps/sweeps).
• Using the quantities after each system sweep for getting the average magnetization and energy will at
most temperatures be enough to get more or less uncorrelated configurations, except close to the critical
point (“critical slowing down”).
4.6 Glauber Dynamics (Heat Bath Dynamics) 79
Exercise: Fractal dimension of the percolating cluster (cont.)
• The critical temperature is T c 4.51.

• To speed up your code, you may consider the following points:
– You can store the values of the exponential function (only a few discrete values are needed).
– Use integer variables inside your main loop (e.g., for the values of magnetization M and energy E).
– The current values of magnetization M and energy E can be easily updated after each spin flip, using
the magnetization change due to the spin flip and ΔE, respectively.
– Reuse equilibrated systems from the previous temperature (T ) when going to the next temperature
(T + ΔT ). This helps avoiding throwing away a lot of configurations to reach equilibrium.
4.6 Glauber Dynamics (Heat Bath Dynamics)
The M(RT)2 algorithm is not the only possible choice to satisfy the detailed balance
condition. Roy J. Glauber (see Figure 4.7) proposed the acceptance probability

exp − kΔE
BT
AG (X → Y) = . (4.36)
1 + exp − kΔEB T
In contrast to the M(RT)2 acceptance probability, updates with ΔE = 0 are not always
accepted but only with probability 1/2. We illustrate this behavior in Figure 4.8. To
prove that eq. (4.36) satisfies the condition of detailed balance, we have to show that
peq (Y)AG (Y → X) = peq (X)AG (X → Y) (4.37)

Figure 4.7 Roy J. Glauber (1925–2018)
since T (Y → X) = T (X → Y). Equation (4.37) is equivalent to received the Nobel Prize for his work on
peq (Y) AG (X → Y) optical coherence.
= , (4.38)
peq (X) AG (Y → X)
A comparison of the acceptance probabilities of M(RT)2 and Glauber dynamics (see Eqs. (4.32) and (4.36)). Figure 4.8
which is satisfied since

peq (Y) ΔE
= exp − (4.39)
peq (X) kB T
and
⎡ ⎤−1
AG (X → Y) exp − kΔE ⎢
⎢
⎢ exp ΔE ⎥⎥⎥ ΔE
B ⎢⎢⎢⎣ ⎥⎥⎥⎦ = exp −
T kB T
= . (4.40)
AG (Y → X) 1 + exp − ΔE 1 + exp ΔE kB T
kB T kB T
As in the M(RT)2 algorithm, only the local configuration around one lattice site is

relevant for the update procedure. Furthermore, E = −J i, j σi σ j for the Ising model
and thus the probability to flip a spin σi is

exp −2JσkB T
i hi
AG (X → Y) = , (4.41)
1 + exp −2Jσ kB T
i hi

where hi = i, j σ j is the local field. A possible implementation is
σi (τ + 1) = −σi (τ) · sign(AG (σi ) − z) , (4.42)
where z ∈ (0, 1) denotes a uniformly distributed random number.
Additional Information: Equivalence of Glauber Dynamics and Heat Bath Monte Carlo
For Glauber dynamics, the spin flip and no flip probabilities can be expressed as
⎧ ⎧
⎪
⎪
⎨ pi for σi = −1 ⎪
⎪
⎨1 − pi for σi = −1,
pflip = ⎪⎪ and pno flip = ⎪ ⎪ (4.43)
⎩1 − pi for σi = +1 ⎩ pi for σi = +1,
where
exp (2Jβhi )
pi = . (4.44)
1 + exp (2Jβhi )
This gives
⎧
⎪
⎪
⎨+1 with probability pi ,
σi (τ + 1) = ⎪
⎪ (4.45)
⎩−1 with probability 1 − pi .
This method does not depend on the spin value at time t and is called heat bath Monte Carlo. As we have
shown, it is equivalent to Glauber dynamics.
4.7 Binary Mixtures and Kawasaki Dynamics
We now consider a system that consists of two species A and B (e. g., two different
gas molecules) that are distributed with given concentrations on the sites of a lattice.
These two species could, for instance, represent two sorts of metallic atoms whose
4.8 Creutz Algorithm 81
numbers are conserved. We show an illustration of such a system in Figure 4.9. This
binary mixture can be modeled by defining
• E AA as the energy of an A − A bond,

• E BB as the energy of a B − B bond,
• E AB as the energy of an A − B bond.
If we set E AA = E BB = 0 and E AB = 1, this model corresponds to an Ising model with

constant magnetization. Systems with constant magnetization can be simulated with
Kawasaki dynamics.
Kawasaki dynamics corresponds to an M(RT)2 or Glauber algorithm with constant Figure 4.9 An example of a binary
numbers of spins in each population. The corresponding algorithm was introduced by mixture consisting of two different atoms
Kyozi Kawasaki (see Figure 4.10) and works as follows: A and B.
Kawasaki Dynamics
• Choose an A − B bond.
• Compute ΔE for A − B → B − A.
• Metropolis: If ΔE ≤ 0 flip, else flip with probability

ΔE
p = exp − .
kB T
• Glauber: Flip with probability

ΔE ΔE
p = exp − / 1 + exp − .
kB T kB T
Figure 4.10 Kyozi Kawasaki received the

Boltzmann Medal in 2001 for his
This procedure is very similar to the previously discussed update schemes. The only contributions to the “understanding of
difference is that the magnetization is kept constant. dynamic phenomena in condensed
matter systems.” Photograph courtesy
Helmut R. Brand.
4.8 Creutz Algorithm
Until now, we only considered canonical Monte Carlo algorithms for simulating sys-
tems at constant temperature. In 1983, Michael Creutz (see Figure 4.11) suggested a
method to perform microcanonical Monte Carlo simulations of systems at constant
energy [134]. The algorithm is therefore usually referred to as Creutz algorithm. The
condition of energy conservation is slightly relaxed in this algorithm and energy is not
fully conserved.
The movement in phase space is therefore not strictly constrained to a subspace of
constant energy, but there is a certain additional volume in which the system can freely
move. The condition of constant energy is softened by introducing a so-called demon.
The demon corresponds to a small reservoir of energy ED that can store a certain
maximum energy Emax . The Creutz algorithm is defined by the following steps:
Creutz Algorithm
• Choose a site.
• Compute ΔE for the spin flip.
• Accept the change if Emax ≥ ED − ΔE ≥ 0 and update the demon energy according to ED →
ED + ΔE.
Besides the fact that we can randomly choose a site, this method involves no random
numbers and is thus deterministic and therefore reversible. However, the temperature
of the system is not known.
It is possible to estimate the temperature with the Boltzmann distribution. By
plot-

ting a histogram of the energies ED one observes a distribution P(ED ) ∝ exp − kEBDT .
We show an example of a distribution P(ED ) in Figure 4.12. The fit is not optimal due
Figure 4.11 Michael Creutz works at the to the small number of different values of ED . The method is faster for larger values
Brookhaven National Laboratory. of Emax , because the condition of constant energy is relaxed and the exploration of
Photograph courtesy Brookhaven phase space is less restricted to certain regions, but then the ensemble becomes “less”
National Laboratory. microcanonical.
Exercise: Microcanonical Monte Carlo and Q2R

Simulate a microcanonical Ising system using the Creutz algorithm:
1. Start with an initial spin configuration X of a given energy E and define a demon energy ED such that
Emax ≥ ED ≥ 0.
2. Choose a spin at random and flip it to obtain the configuration Y.
3. Calculate the energy difference ΔE between configurations X and Y.
4. If Emax ≥ ED − ΔE ≥ 0 choose a new spin and repeat the process. If not, revert the spin flip and
choose a new spin.
Determine the corresponding temperature T using
ED
P(ED ) ∝ e− kB T .
Additional tasks:
1. Compute T for different E. Plot energy and magnetization as a function of temperature and compare your
results to the results obtained with the M(RT)2 algorithm.
2. Repeat the above tasks for different system sizes and compare your results.
3. What happens in the case Emax = 0 (Q2R algorithm)? Discuss the issue of ergodicity.
4.10 Application to Interfaces 83
4.9 Boundary Conditions
Computer simulations are always restricted to finite domains. Therefore, one of the
finite-size effects that we have to take into account are boundary effects of the consid-
105
ered domain. We can set the values of the boundary spins to a certain value or just
have no boundary spins at all. Alternatively, we may introduce periodic boundaries.
In some situations, certain boundary conditions also correspond to a real physical 104
situation. For finite lattices, the following boundary conditions may be used:
103
• open boundaries (i. e., no neighbors at the edges of the system), N
• fixed boundary conditions,
102 β = 2.25
• periodic boundaries.
10
Usually, periodic boundary conditions converge faster toward the thermodynamic
limit compared to open ones, while fixed boundary conditions typically induce a
surface field. It is also possible to identify the last element of a row (or column) 1
0 1.0 2.0 3.0
with the first element of the next row (or column), which leads to a so-called heli- ED
cal boundary condition. This is done to store the system in a one-dimensional array,
for instance, to increase the length of the inner loop in a computer program. There Figure 4.12 The demon energy ED is
are more effects related to systems of finite size that we describe in more detail in exponentially distributed. From the
Chapter 5. Boltzmann factor, it is possible to extract
the inverse temperature
β = (kB T )−1 = 2.25. The figure is
4.10 Application to Interfaces taken from Ref. [134].
The Ising model is used not only as a model of ferromagnetism, but has also found
applications in various other areas. One application is the simulation of interfaces
(i.e., boundaries between regions that are occupied by different sorts of matter). Let
us consider two materials A and B with an arbitrary interface (see Figure 4.13). The
surface tension γ is given by the difference in free energy between the mixed system
(A + B) and the pure system A:
γ = fA+B − fA .
We start with binary variables σ = ±1 again, but we apply fixed boundaries, namely,
“+” in the upper half for A and “-” in the lower half for B. We use Kawasaki dynamics Figure 4.13 Interface between two
(see Section 4.7) to simulate the evolution of the interface at temperature T since the materials A and B.
number of particles A and B is constant. The Hamiltonian is an Ising Hamiltonian
without field
N
H({σ}) = −J σi σ j . (4.46)
i, j
Each bond of the interface has a height h measured from the bottom and we can thus
define the width W of the interface according to
)
1
W= (hi − h̄)2 , (4.47)
N i
where h̄ is the average height and N is the number of bonds belonging to the interface.
A special case is the flat surface, where all points at the interface have the same
height, hi = h̄ and thus W = 0. In general, the width W increases very quickly with tem-
perature T , exhibiting a singularity at the critical temperature T c . As we get closer to
T c , the interface becomes more and more diffuse and the system starts to “ignore” the
boundary conditions, placing “+” entries next to a boundary where we had imposed
“−.” The transition is known as roughening transition.
4.10.1 Next-Nearest Neighbors

So far, we have only considered nearest-neighbor interactions. One obvious extension
of this simplified model is to also consider next-nearest-neighbor interactions:

N
H({σ}) = −J σi σ j − K σi σ j , (4.48)
i, j nn i, j nnn
where nnn denotes the next-nearest neighbors.

The new term introduces stiffness to the interface and thus counteracts curvature.
Curved surfaces will have an increased energy due to this new term, which makes the
system relax to a flatter configuration.
4.10.2 Shape of a Drop

We can go even further and also include gravity, which will, for instance, help us repro-
Figure 4.14 Simulation of a duce drops that are attached to a wall. We start with a block of +1 sites attached to a
two-dimensional drop with L = 257, wall of an L × L system filled with −1s. The Hamiltonian is given by
g = 0.001 forming a contact angle θ.

N
N
Black indicates spins with value +1, and H({σ}) = −J σi σ j − K σi σ j − hj σi , (4.49)
white corresponds to spins with a value of i, j, nn i, j, nnn j line j
−1. The figure is taken from Ref. [135].
where gravity g is introduced via a height-dependent magnetic field
( j − 1)(hL − h1 ) hL − h1
h j = h1 + and g = .
L−1 L
To generate interface configurations, we use Kawasaki dynamics (mass conservation)
and do not permit disconnected clusters of +1s (i. e., isolated water droplets in the air).
In Figure 4.14 we show an example of a drop simulation (L = 257, g = 0.001 after
5 × 107 MC updates averaged more than 20 samples). Based on such simulations, we
can measure the contact angle Θ, which is a function of the temperature and goes to
zero when approaching the critical temperature T c .
Phase Transitions 5
5.1 Temporal Correlations
Each time we accept a spin flip in our sampling chain, a new configuration is gen-
erated. Unfortunately, both new and previous configurations are strongly correlated.
As a consequence,
√ we cannot expect that the error of our Monte Carlo method scales
with 1/ N, if we average over all N sampled configurations. We thus have to find a
measure that tells us whether we are already in equilibrium or not, and which enables
us to generate uncorrelated configurations. According to the definition of a Markov
chain, the dependence of a quantity A on virtual time τ is [136]

A(τ) = p (X, τ) A (X) = p (X, τ0 ) A (X(τ)) . (5.1)
X X
In the second step of eq. (5.1), we used that the average is taken over an ensemble of
initial configurations X(τ0 ) that evolve according to eq. (4.22) [136].
If one begins the simulation at a time τ0 ≤ τ, the nonlinear correlation function
A(τ) − A(∞)
A (τ) =
Φnl (5.2)
A(τ0 ) − A(∞)
is a measure to quantify the deviation of A(τ) from the steady state value A(∞) relative
to the deviation of A(τ0 ) from A(∞). Due to its normalization, the nonlinear correla-
tion function ΦnlA (τ) equals 1 at τ0 and 0 for τ → ∞. Strictly speaking, ΦA (τ) is not a
nl
correlation function, but a measure to investigate correlations with respect to the ini-
tial configuration. The nonlinear correlation time τnl
A describes the relaxation toward
equilibrium and is defined as1
∞
τnl
A = Φnl
A (τ) dτ . (5.3)
0
1 If we consider an exponential decay of Φnl

A (τ), we find that this definition yields its characteristic time
since
∞
A dτ = τA .
exp −τ/τnl nl
85
86 Phase Transitions
Close to a critical point, we observe the so-called critical slowing down of the dynamics,
which means that the nonlinear correlation time diverges as a power law
−zA nl
A ∝ |T − T c |
τnl , (5.4)
where znl
A is the nonlinear dynamical critical exponent. This critical slowing down
causes difficulties because eq. (5.4) implies that the time needed to reach equilibrium
becomes infinite at T c .
The linear correlation function of two quantities A and B in equilibrium is
A(τ0 )B(τ) − A B
ΦAB (τ) = , (5.5)
AB − A B
where

A (τ0 ) B (τ) = p (X, τ0 ) A (X (τ0 )) B (X (τ)).
X
As τ goes to infinity, ΦAB (τ) decreases from unity to zero. If A = B, we call eq. (5.5) the
autocorrelation function. For the spin–spin correlations in the Ising model, we obtain
σ(τ0 )σ(τ) − σ(τ0 )2

Φσ (τ) = 4 2 5 . (5.6)
σ (τ0 ) − σ(τ0 )2
The linear correlation time τAB describes the relaxation in equilibrium

∞
τAB = ΦAB (τ) dτ . (5.7)
0
As for the nonlinear correlation time, in the vicinity of T c , we observe critical slowing
down that is described by
τAB ∝ |T − T c |−zA , (5.8)
where zA is the linear dynamic critical exponent. The dynamic exponents for spin
correlations of the Ising model for Glauber and M(RT)2 updating turn out to
be [137–140]
zσ = 2.166(7) (2D) ,
(5.9)
zσ = 2.055(10) (3D) .
There exist conjectured relations between these Ising critical exponents and the critical
dynamic exponents for spin and energy correlations. The relations [141]
zσ − znl
σ = β,
(5.10)
zE − znl
E =1−α
are numerically well established, however, not yet proven.

5.3 Finite-Size Scaling 87
5.2 Decorrelated Configurations
The behavior of the correlation length described in the previous Section 5.1 is valid
for an infinite lattice and diverges at T c (see eq. (3.33)). In a finite system, however, we
cannot observe a quantity that diverges toward infinity. As we schematically illustrate
in Figure 5.1, the correlation length ξ must be less than the system size L at T c (i. e.,
ξ ≈ L at T c ).
We connect the behavior of ξ(T ) close to the critical temperature with its scaling law
ξ(T ) ∝ |T − T c |−ν at T c (see eq. (3.33)) and the corresponding correlation-time scaling
of eq. (5.7):
zAB
τAB ∝ |T − T c |−zAB ∝ L ν . (5.11)
This equation implies that the number of samples that have to be discarded to obtain
uncorrelated configurations is finite, although it increases with the system size L. When
we study large system sizes, the computation may take very long.
To ensure to sample uncorrelated configurations, we should
Figure 5.1 Schematic behavior of the
• first reach equilibrium (discard n0 = cτnl (T ) configurations), correlation length ξ(T ) as a function of
• use only every nthe = cτ(T ) configuration for averaging, and temperature close to T c . The correlation
znl z
• at T c use n0 = cL ν and ne = cL ν , length diverges in an infinite system
(blue solid line) at T c according to
where c ≈ 3 is a “safety factor” to make sure to discard enough samples. A trick eq. (3.33). In a finite system (orange solid
for reducing critical slowing down is to use cluster algorithms, which we introduce in line), however, we observe a roundoff
Chapter 6. since the correlation length cannot
become larger than the system size L.
5.3 Finite-Size Scaling
In the Ising model, spins tend to form clusters and spatial correlations emerge in
addition to the temporal correlations that we discussed in Section 5.1. The spatial cor-
relation function decays exponentially with distance and the correlation length defines
a characteristic length scale (see Section 3.2.3). According to eq. (3.33), the correla-
tion length ξ diverges with an exponent of −ν at T c . Moreover, susceptibility and heat
capacity also exhibit a divergent behavior at T c (see Eqs. (3.28) and (3.29)). The larger
the system size, the more pronounced is the divergence.
In finite systems, we cannot observe a divergence toward infinity but a peak of some
γ
finite height. The height of the susceptibility peak scales with L ν while the width of
the critical region shrinks as L− ν . If we rescale the values for different system sizes
1
accordingly, we obtain a data collapse. That is, we observe that all the values fall onto
a single curve. This can be used to compute critical exponents.
The finite size scaling relation of the susceptibility is [41]
γ 0 11
χ (T, L) = L ν Fχ (T − T c ) L ν , (5.12)
t
(a) (b)
Figure 5.2 System-size dependence of the susceptibility (a) and the corresponding finite-size scaling of eq. (5.12) (b). The inset
in panel (b) shows the two branches of χ (T, L) L−γ for T > T c and T < T c in a log-log plot; the black solid line
is a guide to the eye with slope −γ ≈ −1.237. The data are based on a simulation of the Ising model on a cubic
lattice for different linear system sizes L. The number of samples, which we generated with the Wolff algorithm (see
Section 6.5), is 25 × 103 .
where Fχ is the scaling function of the susceptibility.2 We show an example of a finite-

size scaling data collapse of the susceptibility in Figure 5.2. For the spontaneous
magnetization, the corresponding finite-size scaling relation is
β 0 11
MS (T, L) = L− ν F MS (T − T c ) L ν . (5.13)
Exercise: Finite-Size Scaling

Perform simulations of the three-dimensional Ising model for different system sizes to determine the critical
exponents γ and ν of the magnetic susceptibility χ.
1. Use “decorrelated configurations” to measure the temperature dependence of the susceptibility for
different system sizes.
2. The critical temperature is T c 4.51.
3. Vary γ/ν and 1/ν until you get the best possible data collapse. The best way to judge the quality of the
data collapse is by eye. (Within the linear regions of the scaling function, the slope in a log-log plot should
be −γ. Check if the slope is correct as a verification of your data collapse.)
You may find the following points useful as well:
• You can get an estimate for the ratio γ/ν by plotting the maximal value of χ as a function of the system
size in a log-log plot.
• You may try a more systematic way of determining the exponents:
1. You can get an estimated value γ/ν as a starting value. Try to get the best possible data collapse by
tuning 1/ν only.
2. Within the linear regions of the scaling function, you can measure the slope −γ.
3. Take the new values of γ and ν as new starting value for γ/ν.
5.4 Binder Cumulant 89
5.4 Binder Cumulant
In the previous section, we have outlined that the finite size scaling relation of the
susceptibility (see eq. (5.12)) constitutes a method to determine critical exponents.
However, we still need a way to determine the critical temperature more precisely. To
overcome this hurdle, we make use of the so-called Binder cumulant

M4
U L (T ) = 1 − 4 5L2 , (5.14)
3 M2 L
which is independent of the system size L at T c since
4β 0 1
L− ν F M4 (T − T c ) L ν
1
M4 0 11
4 2 52 = 6 2β = − ,
L
0 17 2
F C (T T c ) L ν (5.15)
M L −ν 1
L F M2 (T − Tc L
) ν
where we used eq. (5.13) in the first step. The Binder cumulant was introduced by
Kurt Binder [142] (see Figure 5.3). At the critical temperature (T = T c ), the scaling Figure 5.3 Kurt Binder (Johannes
function FC , which is nothing but the ratio of two other scaling functions, is a system Gutenberg-Universität Mainz) made
size-independent constant. As shown in the left panel of Figure 5.4, for T > T c , the numerous important contributions to the
magnetization exhibits a Gaussian distribution field of computational statistical physics.
) ⎡ ⎤
Ld ⎢⎢⎢ M 2 Ld ⎥⎥⎥
PL (M) = exp ⎢
⎣ − ⎥⎦ , (5.16)
2πσ2L 2σ2L
with σ2L = kB T χL . Since, for Gaussian distributions, the fourth moment equals three
times the second moment squared,
2
M4 = 3 M2 , (5.17)
L L
it follows that U L (T ) must be zero for T > T c . Below the critical temperature (T < T c ),
there exist two ground states, one with positive and one with negative magnetization;
the corresponding distribution is [143]
) 8 9
1 Ld (M − MS )2 Ld (M + MS )2 Ld
PL (M) = exp − + exp − , (5.18)
2 2πσ2L σL σL
2
as illustrated in the right panel of Figure 5.4.For this distribution, M 4 = M 2 and
L L
therefore U L (T ) = 23 for T < T c . In summary, we obtained
⎧2
⎪
⎪
⎪
⎪
⎪ 3 , for T < T c ,
⎨
U L (T ) = ⎪
⎪const. , for T = T c , (5.19)
⎪
⎪
⎪
⎩0 , for T > T c ,
2 Based on eq. (3.28), we can infer that Fχ [x] ∝ x−γ as L → ∞.

t
(a) (b)
Figure 5.4 Schematical distribution P(M) of the magnetization M above and below the critical temperature T c .
for L → ∞. In Figure 5.5, we show the behavior of the Binder cumulant for the
three-dimensional Ising model and different system sizes. This constitutes a very effi-
cient way to compute the critical temperature based on the very temperature sensitive
Binder cumulant U L . For infinite systems, the cumulant exhibits a jump at T c as
described by eq. (5.19).
All our discussions related to the behavior of different thermodynamic quantities
at the critical temperature assumed a pure power law behavior as, for example, those
described by Eqs. (3.23), (3.28), and (3.29). However, we can observe such a power law
dependence only for temperature values very close to the critical temperature. Farther
away from T c , we cannot observe a clear power law behavior anymore, and corrections
to scaling have to be included:
M (T ) = A0 (T c − T )β + A1 (T c − T )β1 + . . . , (5.20)
−ν −ν1
ξ (T ) = C0 (T c − T ) + C1 (T c − T ) + ... , (5.21)
with β1 > β and ν1 < ν being universal correction to scaling exponents. These correc-
tions can become very important for high-quality data, where the errors are small and
the deviations become visible. The scaling functions must also be generalized as
β 0 11 0 11
M (T, L) = L− ν F M (T − T c ) L ν + L−x F 1M (T − T c ) L ν + . . . , (5.22)
0β β β 1
where x = max ν1 , ν1 , ν − 1 .
These corrections to scaling are responsible for the fact that the curves for Binder
cumulants do not cross at exactly the same point for small systems as shown in the
right panel of Figure 5.5.
5.5 First-Order Transitions
Until now, we only focused on the temperature dependence of thermodynamic quanti-

ties in the context of the second-order phase transition of the Ising model. For T < T c ,
5.5 First-Order Transitions 91
(a) (b)
The temperature dependence of the Binder cumulant for the three-dimensional Ising model for different system sizes. Figure 5.5
Panel (a) shows the crossing point of panel (b) with higher resolution.
For small system sizes and below T c , the magnetization M(τ) switches sign after a certain number of Monte Carlo Figure 5.6
time steps τ if the magnetic field is zero. We simulated a 2D Ising system with 16 × 16 spins.
the Ising model exhibits a jump in the magnetization, which is a first derivative of the
free energy, at H = 0. As outlined in Section 3.1.7, this is a transition of first order.
Because of the bimodal distribution of the magnetization (see Figure 5.4), we can
observe stochastic switching between the two magnetic phases for small system sizes
(see Figure 5.6). The typical switching time is also called “ergodic time” and increases
with system size and decreases with temperature.
Kurt Binder
showed that the magnetization as a function of the field H is described
by tanh αLd if the distribution of the magnetization is given by eq. (5.18) [143].
In particular, he showed that the magnetization and susceptibility are given
by [143]

M(H) = χD L H + ML tanh βHML L ,
d
∂M βML Ld (5.23)
Figure 5.7 The first-order scaling
χL (H) = = χDL + : ; .
∂H cosh2 βHML Ld behavior as described by eq. (5.23) for an
Ising model on the square lattice at
Similar to the scaling of a second-order transition, we can scale the maximum of the kB T/J = 2.1. The figure is taken from
susceptibility (χL (H = 0) ∝ Ld ) and the width of the peak (ΔχL ∝ L−d ). In Figure 5.7, Ref. [143].
we show an illustration of the first-order scaling behavior as described by eq. (5.23).

To summarize, a first-order phase transition is characterized by
1. a bimodal distribution of the order parameter,
2. stochastic switching between the two states in small systems,
3. hysteresis of the order parameter when changing the field,
4. a scaling of the order parameter/response function according to eq. (5.23).
Cluster Algorithms 6
Based on our discussion of correlation effects in Sections 5.1 and 5.2, we have learned
that larger system sizes require longer simulation times close to critical points due
to critical slowing down (see Section 5.2). All algorithms we considered so far are
based on single spin flips only. As a consequence of this update scheme, we have to
wait sufficiently long before adding a new uncorrelated configuration to our averaging
procedure. Otherwise, the considered samples may not be statistically independent.
The aim of cluster algorithms is to reduce the computation time by flipping multiple
spins at the same time. It is essential that the group of spins to flip is chosen with a
sufficiently large acceptance probability. To do this, we will generalize the Ising model
and adapt this generalization to our needs.
6.1 Potts Model
The Potts model [144, 145] was introduced by Renfrey B. Potts as a generalization of
the Ising model with q ≥ 2 states. It is a very versatile model due to its applications in
many fields including sociology, biology and material science. The Hamiltonian of the
Potts model is

H({σ}) = −J δσi σ j − H σi , (6.1)
i, j i
where σi ∈ {1, . . . , q} and δσi σ j is unity when nodes i and j are in the same state
(Kronecker delta). The Potts model exhibits a first order transition at the critical tem-
perature in two dimensions for q > 4, and for q > 2 for dimensions larger than three.1
For q = 2, the Potts model is equivalent to the Ising model. Moreover, there exists a
connection between the Potts model and bond percolation. Kasteleyn (see Figure 6.1)
and Fortuin demonstrated that the two models have related partition functions [147].
A thermodynamic system is characterized by its partition function from which all ther-
modynamic quantities can be derived. Therefore, the partition function completely
describes the thermodynamic properties of a system. If two systems have the same
partition function (up to a multiplicative constant), we consider these two systems as
equivalent. We will derive the relation between the Potts model and bond percolation
in the following section.
1 In three dimensions, the transition from second to first order is at qc = 2.620(5) [146].
93
94 Cluster Algorithms
6.2 The Kasteleyn and Fortuin Theorem
We consider the Potts model not just on a square lattice but on an arbitrary graph
that consists of nodes and bonds ν. Each node can be in q possible states and each
connection leads to an energy cost of unity if two connected nodes are in a different
state and of zero if they are in the same state:
⎧
⎪
⎪
⎨0 , if endpoints are in the same state,
E=J ν with ν = ⎪ ⎪ (6.2)
⎩1 , otherwise .
ν
As we show in Figure 6.2, we introduce contraction and deletion operations of a bond

ν on the graph. As shown in detail in the gray box on the next page, the partition
function of the Potts model can be split up into two partition functions, one for the
Figure 6.1 Pieter W. Kasteleyn
deleted and one for the contracted graph. After applying these operations to every
(1924–1996) worked at the University of
bond, the graph is reduced to a set of separated points corresponding to clusters of
Leiden. Photograph courtesy KNAW.
nodes that are connected and in the same state out of q states. The partition function

of the Potts model Z = X e−βE(X) reduces to

Z= q# of clusters pc (1 − p)d = q# of clusters , (6.3)
b
configurations of
bond percolation
where c and d are, respectively, the numbers of contracted and deleted bonds and
p = 1 − exp(−βJ). In the limit q → 1, one can derive the generating function of
bond percolation. In bond percolation, an edge of a graph is occupied with proba-
bility p and vacant with probability 1 − p. Equation (6.3) constitutes a fundamental
t
Figure 6.2 Contraction and deletion of bond ν1 on a graph.
6.3 Coniglio–Klein Clusters 95
relation between a purely geometrical model (bond percolation) and a magnetic model
described by a Hamiltonian (Potts model) [148]. Interestingly, we can now choose non-
integer values for q. This was meaningless in the original definition of the Potts model,
because q was introduced as the state of each site.
Additional Information: A Proof for the Theorem of Kasteleyn and Fortuin

The partition function is the sum over all possible configurations weighted by the Boltzmann factor and thus
given by
(6.2)
<
Z= e−βE(X) = e−βJ ν ν = e−βJν . (6.4)
X X X ν
We now consider a graph where bond ν1 connects two nodes i and j with states σi and σ j , respectively. If
we delete bond ν1 , the partition function is
<
ZD = e−βJν . (6.5)
X νν1
We can thus rewrite eq. (6.4) as

<
Z= e−βJν1 e−βJν
X νν1
< <
= e−βJν + e−βJ e−βJν ,
X:σi =σ j νν1 X:σi σ j νν1
where the first part is the partition function of the contracted graph ZC and the second part is given by the
identity
< < <
e−βJν = e−βJν − e−βJν = ZD − ZC . (6.6)
X:σi σ j νν1 X νν1 X:σi =σ j νν1
We thus obtain
Z = ZC + e−βJ (ZD − ZC ) = pZC + (1 − p)ZD , (6.7)
where p = 1−e−βJ . To be more precise, we expressed the partition function Z as the contracted and deleted
partition functions at bond ν1 . We apply this procedure to another bond ν2 and find
Z = p2 ZCν1 ,Cν2 + p(1 − p)ZCν1 ,Dν2 + (1 − p)pZDν1 ,Cν2 + (1 − p)2 ZDν1 ,Dν2 . (6.8)
After applying deletion and contraction operations to all bonds, we obtain eq. (6.3).
6.3 Coniglio–Klein Clusters
The first who described the dilution of clusters of equal states with probability p =
1 − e−βJ in the Potts model were Antonio Coniglio and William Klein (see Figure 6.3).
Kurt Binder nicely summarized the historical developments around the theorem of
Kastelyn and Fortuin, Coniglio–Klein Clusters, and cluster algorithms in Ref. [149].
We should bear in mind that the equivalence between bond percolation and the Potts
model only concerns their thermodynamic functions since there is no direct relation
between bond configurations and spin configurations. The fact that the actual spin
values are absent in eq. (6.3) forms the basis for cluster algorithms. The probability of
a given cluster C to be in a certain state σ0 is independent of the state itself:

p(C, σ0 ) = pcC (1 − p)dC q# of clusters pc (1 − p)d . (6.9)
bond percolation
without cluster C
This implies that changing the state of this particular cluster has no effect on the
partition function (and therefore the energy) so that it is possible to accept the flip
with probability one. This can be seen by looking at the detailed balance condition of
the system
p(C, σ1 )W [(C, σ1 ) → (C, σ2 )] = p(C, σ2 )W [(C, σ2 ) → (C, σ1 )] , (6.10)
and using p(C, σ1 ) = p(C, σ2 ). We then obtain for the acceptance probabilities

p(C, σ2 )
A [(C, σ2 ) → (C, σ1 )] = min 1, = 1 for M(RT)2 ,
p(C, σ1 )
(6.11)
p(C, σ2 ) 1
A [(C, σ2 ) → (C, σ1 )] = = for Glauber.
p(C, σ1 ) + p(C, σ2 ) 2
Figure 6.3 Antonio Coniglio (Università Based on these insights, we can focus on cluster algorithms which are much faster than
di Napoli Federico II) and William Klein single-spin flip algorithms and less prone to the problem of critical slowing down.
(Boston University) introduced the
concept of cluster dilution with
probability p = 1 − e−βJ [150]. 6.4 Swendsen–Wang Algorithm
The Swendsen–Wang algorithm [151, 152] is a refined Monte Carlo technique that
uses the advantage of updating whole clusters of spins. It was developed by Robert
Swendsen and Jian-Sheng Wang (see Figures 8.6 and 7.3).
For a certain configuration, we iterate over all bonds connecting spins. Whenever
two bonded sites are in the same state, the two sites belong to the same cluster with
probability p = 1 − e−βJ . Once the clusters are determined, they can be flipped using
any of the updating schemes mentioned before. The basic procedure is as follows:
Swendsen–Wang Algorithm
• Occupy the bonds with probability p = 1 − e−βJ if sites are in the same state.
• Identify the clusters with the Hoshen–Kopelman algorithm (see Section 2.3.4).
• Flip the clusters with probability 1/2 for Ising or always randomly choose a new state for q > 2.
• Repeat the procedure.
6.5 Wolff Algorithm 97
Besides moving faster through phase space, the Swendsen–Wang algorithm also
reduces critical slowing down considerably since the dynamical critical exponents
defined in eq. (5.9) become z = 0.35(1) (2D) and z = 0.75(1) (3D) [151].
6.5 Wolff Algorithm
The Wolff algorithm [153] just considers one cluster instead of all clusters of a
configuration and is based on the following steps:
Wolff Algorithm
• Choose a site randomly.

• If the neighboring sites are in the same state, add them to the cluster with probability p = 1 − e−βJ .
• Repeat this for every site on the boundary of the cluster until all bonds of the cluster have been visited
exactly once.
• Choose a new state for the cluster.
• Repeat the procedure.
Exercise: Cluster Algorithms

Cluster algorithms can be used to reduce the critical slowing down substantially. In this task you can either
pick the Swendsen–Wang or the Wolff algorithm. Both algorithms use the same probability for connecting
sites of the same state:
p = 1 − exp(−βJ) .
The difference between the algorithms is that in Swendsen–Wang simulations all clusters will be determined
and flipped with probability 1/2 (Ising model), or randomly get a new value (Potts model). In the Wolff
algorithm only one cluster is determined by randomly choosing a spin on the lattice. In this case, the cluster
will be flipped (Ising model) or gets a new state that is different from the old one (Potts model).
Task
1. Implement either the Swendsen–Wang or the Wolff algorithm (or both).
2. Check your code by plotting E(T ) and |M|(T ).
3. Show that critical slowing down is substantially reduced by plotting the time dependence of relaxation of
E (or |M|) around/at T c into equilibrium in comparison to the single spin-flip algorithm.
4. Measure the nonlinear correlation time for different system sizes at T c and thus the exponent z and
compare it to the values for the single spin-flip algorithms and/or other cluster algorithms.
Exercise: Cluster Algorithms (cont.)

• Note that the probability p = 1 − exp(−βJ) is derived for the Potts model. This, of course, includes
the Ising model, however, using an energy notation (terms σi σ j in the energy sum take values 0 or 1)
different from the one originally used for the Ising model (terms σi σ j in the energy sum take values ±1).
In the latter case, the probability will be p = 1 − exp(−2βJ).
• For the Swendsen–Wang algorithm you will need a working Hoshen–Kopelman algorithm (see
Section 2.3.4).
• Using these cluster algorithms for temperatures below T c , the magnetization of your system will not be
“frozen” in one ground state as it was for the single spin-flip algorithms.
6.6 Continuous Degrees of Freedom: The n-Vector Model
Before focusing on further simulation methods, we briefly discuss other generalizations

of the Ising model. The Potts model [144, 145] is only one example. In fact, there exists
a large number of models that are modified versions of the Ising model and used to
describe related physical phenomena such as antiferromagnetism, spin glasses, and
metamagnetism.
One of the possible generalizations of the Ising model is the so called n-vector
model [154]. It describes spins as vectors with n components. This model has
applications in modeling magnetism and the Higgs mechanism. Its Hamiltonian

H = −J Si · S j + H Si , (6.12)
Figure 6.4 Harry Eugene (“Gene”)
i, j i
Stanley from Boston University
(Boltzmann Medal 2004) has made resembles
the one of the Potts model in the sense that it favors spin alignment. Here,
many contributions to statistical and Si = S i , S i , . . . , S i is an n-dimensional vector and |Si | = 1. For n = 1, we obtain the
1 2 n
computational physics. Ising model, the case n = 2 corresponds to the XY-model, for n = 3 we find the so-
called Heisenberg model and finally, for n = ∞ one obtains the spherical model which
was solved exactly by Gene Stanley (see Figure 6.4). For n > 1, the degrees of freedom
of the n-vector model are continuous and as a consequence the XY-model and the
Heisenberg model do not exhibit a phase transition from an ordered to a disordered
state in two dimensions at a nonzero critical temperature.2 The proof of this statement
can be found in Ref. [156], and it is a consequence of the Mermin–Wagner theorem
which states that in dimensions d ≤ 2 continuous symmetries cannot be spontaneously
broken at finite temperature in systems with sufficiently short-range interactions. In
three dimensions, however, both the XY and the Heisenberg model exhibit a phase
transition at a nonzero critical temperature T c . The scaling behavior of these models
2 The two-dimensional XY model does, however, exhibit a transition to a low-temperature phase with
power law decaying correlation functions, the Kosterlitz–Thouless transition (Nobel Prize 2016) [155].
6.6 Continuous Degrees of Freedom: The n-Vector Model 99
close to the critical temperature is very similar to the behavior of the Ising model only
with different critical exponents.
For Monte Carlo simulations with vector-valued spins we have to adapt our simu-
lation methods to continuous degrees of freedom. The classical strategy is to flip spins
by modifying the spin locally through adding a small ΔS such that S i = Si + ΔS and
ΔS ⊥ Si . The classical Metropolis algorithm can then be used in the same fashion as
in the Ising model.
In order to use cluster methods, one can project a group of spins onto a plane,
and then reflect the projection of the spins with respect to the plane. In addition,
one has to find a method to identify equal spins. The probability to find equal
spins in a vector-valued model vanishes and one therefore considers them to be in
the same state when the difference between their components is less than a certain
tolerance.
Exercise: Heisenberg Model

The Heisenberg model can be seen as a generalization of the Ising model, where a more realistic description
of classical magnetic dipoles is represented by the Hamiltonian

H = −J Si · Sj ,
i, j
where Si ∈ S2 ⊂ R3 points to the surface of a three-dimensional sphere of radius one.

While for the XY model (spins in S1 ) it is advantageous to use polar coordinates, it is unclear if in the 3D spin
case the added complexity introduced by spherical coordinates justifies the memory savings.
1. Implement a MC simulation of the Heisenberg model with a M(RT)2 algorithm. Remember to normalize
your vectors if you use a Cartesian representation.
• Compute the critical temperature for J = 1. You can use either Binder cumulants, or the magnetic
susceptibility. You should find T c ≈ 1.443.
• Compute the autocorrelation time, either for E or M2 , at T c and find the critical dynamical exponent
given by the relation τ ∝ Lzc .
2. The cluster algorithm can be extended to systems with spins of arbitrary dimensions, with the trick of
considering reflections around a random plane at each MC step. For each step select a random unit vector
r which is chosen to be orthogonal to the reflection plane. Decide if you want to implement Swendsen–
Wang or Wolff algorithm, grow a cluster with bond probability
pi, j = 1 − exp[−2β(Si · r)(Sj · r)], (6.13)
and flip the spins according to
Si = Si − 2r (Si · r). (6.14)
• Repeat the computation of T c and zc .
7 Histogram Methods
For computing the thermal average as defined by eq. (3.15), we need to sample dif-
ferent configurations at different temperatures. One possibility would be to determine
an average at a certain temperature T 0 and extrapolate to another temperature T . For
the canonical ensemble, an extrapolation can be achieved by reweighing histograms
of energies pT0 (E) with the Boltzmann factor exp(E/T − E/T 0 ). Such histogram meth-
ods have first been described in Ref. [157]. We now reformulate the computation of
the thermal average of a quantity Q and of the partition function as a sum over all
possible energies instead of over all possible configurations and find
1
Q (T 0 ) = Q (E) pT0 (E) with ZT0 = pT0 (E) , (7.1)
ZT0 E E
− E
where pT0 (E) = g (E) e kB T0 and g (E) defines the degeneracy of states (i.e., the number
of states with energy E). This takes into account the fact that multiple configura-
tions can have the same energy. Our goal is to compute the quantity Q at another
temperature T according to
1
Q (T ) = Q (E) pT (E) . (7.2)
ZT E
The degeneracy of states contains all the information needed. Using the definition of
g (E) yields

− k ET E E
pT (E) = g (E) e B = pT0 (E) exp − + (7.3)
kB T kB T 0
0 1
and with fT0 ,T (E) = exp − kBET + kBET0 we finally obtain

Q (E) pT0 (E) fT0 ,T (E)
E
Q (T ) = . (7.4)
E pT 0 (E) fT 0 ,T (E)
With eq. (7.4) we found a way to compute a quantity Q at any temperature T based
Figure 7.1 An example of the histogram on a sampling at a certain temperature T 0 . The drawback of this method is that the
and the broad histogram method for the values of Q (E) are sampled around the maximum of pT (E), which converges to a
Ising model on the square lattice with Dirac delta distribution for large systems (see Figure 7.1). This means that the statis-
two different system sizes (L = 32 and tics is very poor if T 0 and T are substantially different, and results become unreliable.
L = 256). The figure is taken from One possible solution is to interpolate between several temperatures (multicanonical
Ref. [158]. method) but this involves computations for many temperatures which is also inefficient
for large systems. A more efficient solution to this problem, the so-called broad
100
7.1 Broad Histogram Method 101
histogram method, has been presented in 1996 by Oliveira et al. [159] (see Figure 7.2).
7.1 Broad Histogram Method
The aim of the broad histogram method is to directly compute the degeneracy of
states over a broad energy range. We show an illustration of this concept in Fig-
ure 7.1. We need to define a Markov process in energy space that is not only exploring
regions of certain energies. Let Nup (E) and Ndown (E) denote the numbers of processes
which, starting from a configuration of energy E, lead to an increase and decrease
in energy, respectively. Furthermore, we have to keep in mind that the degeneracy of
states increases exponentially with energy E, because the number of possible config-
urations increases with energy. To explore all energy regions equally, one imposes a
condition equivalent to the one of detailed balance:
g (E + ΔE) Ndown (E + ΔE) = g (E) Nup (E) . (7.5)
The motion in phase space toward higher energies can then be penalized with a
Metropolis-like dynamics:
Broad Histogram Method
• Choose a new configuration.

• If the new energy is lower, accept the move.
• If the new energy is higher, then accept with probability Ndown (E+ΔE)
Nup (E)
.
We obtain the function g(E) by taking the logarithm of eq. (7.5) and dividing by ΔE
2 3 2 3 0 1
log g (E + ΔE) − log g (E) = log Nup (E) − log [Ndown (E + ΔE)] . (7.6)
In the limit of small energy differences, we can approximate this equation by

2 3
∂ log g (E) 1 Nup (E)
= log , (7.7)
∂E ΔE Ndown (E + ΔE)
which we can numerically integrate over E to obtain g (E). The functions Nup and Ndown Figure 7.2 Paulo Murilo Castro de
can be obtained by checking all possible spin-flips and their corresponding energy Oliveira is a professor emeritus at the
change for each new configuration. They become more and more precise as the simu- Universidade Federal Fluminense in Rio
lation evolves. In addition, we also need to store the values of the quantity Q (E) for de Janeiro.
which we wish to compute the thermal average according to
− k ET
Q (E) g (E) e
E B
Q (T ) = . (7.8)
− k ET
E g (E) e B
Having the degeneracy of states g (E), we can now compute quantities at any
temperature.
102 Histogram Methods
7.2 Flat Histogram Method
A variant of the broad-histogram method that aims at generating a flat energy his-
togram was proposed in 2000 by Jian-Sheng Wang (see Figure 7.3) [160]. It is based
on the following procedure:
Flat Histogram Method
• Choose a new configuration with energy E = E + ΔE.

• Flip the spin with probability

Ndown (E + ΔE)
r(E |E) = min 1, . (7.9)
Nup (E)
• Accumulate statistics for Nup (E).
According to eq. (7.5), the spin-flip probability is

g(E)
r(E |E) = min 1, , (7.10)
g(E + ΔE)
Figure 7.3 Jian-Sheng Wang (National
which corresponds to entropic sampling Monte Carlo [161] and multicanonical Monte
University of Singapore) made significant
Carlo [162, 163]. After the appearance of Ref. [160], another very similar variant was
contributions to the development of
proposed by Wang and Landau [164, 165]:
cluster algorithms and histogram
methods.
Wang–Landau Method
• Start with g (E) = 1 and set f = e.

• Make a Monte Carlo update with p(E) = 1/g(E).
• If the attempt is successful at E: g(E) → f · g(E).
• Obtain a histogram of energies H(E).

• If H(E) is “flat” (i.e., uniform) enough, then f → f .
• Stop when f is smaller than some predefined value (e.g., f ≤ 1 + 10−8 ).
By setting the Monte Carlo update probability to 1/g(E), we obtain smaller transition
probabilities for larger energies due to the larger number of possible configurations
at the same energy. Therefore, the method tends toward energies with fewer configu-
rations. After a successful update, the values of g and E have to be updated. Once a
reasonably flat histogram has been obtained, we can increase the precision by decreas-
ing f . The flatness of the histogram can be measured as the ratio of the minimum to
the maximum value.
7.3 Umbrella Sampling 103
7.3 Umbrella Sampling
The Umbrella sampling technique was developed and proposed in Ref. [166]. The aim
of this method is to overcome the problem of not reaching some configurations for cer-
tain energy landscapes. As an example, in the Ising model below T c , the system could
have difficulties in jumping from a positive to a negative magnetization or vice versa if
the system is very large. The basic idea is to multiply the probability for configuration
X with a function w(X) that is large at the free energy barrier and later on remove this
correction in the averaging step:
− E(X)
w (X) e kB T A/ww
=
p (X) = with A = , (7.11)
− E(X) 1/ww
X w (X) e B
k T
where the average ·w is taken with respect to the distribution =

p (X).
Exercise: Broad Histogram Method
• Start from a certain Ising configuration.

• Propose single spin flips for each spin of the configuration (i.e., go sequentially through the whole sys-
tem) and count the number Nup and Ndown leading to energy increase and energy decrease by ΔE,
respectively. (Do not flip the spins, just check whether energy increases or decreases.)
• Generate a new configuration by choosing one site randomly and flip the spin if energy is decreased. If
energy would be increased, flip the spin with probability Ndown /Nup (usually Nup is larger than Ndown
which would otherwise preferably lead to increasing energy).
• At each step one accumulates the following values for each energy E for later averaging: Nup (E),
Ndown (E), Q(E) (the quantity one wants to measure, could be also more than one quantity).

After the simulation you calculate the averages for each energy (“microcanonical average”): Nup (E) ,
Ndown (E) and Q(E). From these averages you can determine g(E) via solving eq. (7.7). The ther-
modynamic average of Q at arbitrary temperatures can then be calculated according to eq. (7.8).
Task
1. Implement the broad histogram algorithm as described above (alternatively you may also try the flat
histogram method).
2. Plot the number of times each energy was visited (see Figure 7.1).
3. Measure g(E) and plot it in a semilogarithmic plot.
4. You may plot the energy as a function of temperature (see eq. (7.8)) based on Q(E) = E (i.e.,
measuring Q(E) during the simulation is not necessary).
5. Measure the specific heat and susceptibility (see Eqs. (3.26) and (3.27)).
6. Discuss your results.
104 Histogram Methods
Exercise: Broad Histogram Method (cont.)

• In three dimensions, single spin flips lead to three possible values for ΔE. You can store values for one
value of ΔE only (preferably the lowest value) or you store Nup (E) and Ndown (E) for each possible
value ΔE, resulting in three curves for g(E), which should show the same behavior. Alternatively you
can store (Nup (E))1/ΔE and (Ndown (E))1/ΔE .
• As initial configuration, start with a random configuration thermalized at the critical temperature T c for a
few lattice sweeps.
• In order to obtain good correlation free results throw away about 10 lattice sweeps of configuration
between the successive values for averaging.
• To accelerate this procedure, you can thermalize the system using Metropolis algorithm at a temperature
corresponding to the energy. This may be necessary to avoid blocking of states at low temperatures.
Renormalization Group 8
In this chapter, we discuss renormalization group methods and the importance of sym-
metries at criticality to improve our simulations. In particular, we focus on the main
concepts of renormalization theory and present Monte Carlo renormalization group
approaches to determine critical exponents of Ising-like systems. For more details, we
refer the interested reader to Ref. [167].
Usually, the more information there is available about a system, the better certain
quantities can be computed. Close to critical points, changes in the scale of the system
can be used to better extrapolate the values to an infinite system. Furthermore, it is
possible to develop scale-invariant theories to describe the properties of models in the
vicinity of their critical points.
For self-similar patterns (see Section 2.4), the invariance under scale transforma-
tions is obvious. A curve described by the function f (x) is said to be scale-invariant
if f (λx) = λΔ f (x), where λ is a scale factor and Δ is a corresponding exponent. Here,
we want to generalize this intuitive treatment of scale changes to concepts from sta-
tistical physics. One possibility is to look at the free energy density and its invariance
under scale transformations. To renormalize a system means to change its scale by a
factor l such that the renormalized length becomes L̃ = L/l. This can be done either in
position, or in momentum space.
8.1 Real Space Renormalization
If a system is invariant under a certain transformation, we can apply this transforma-

tion infinitely often without changing the observables of the system. At the critical
point, there exist no finite correlation length scales and the properties of the system
are invariant under scale changes. We thus regard criticality as a fixed point under
such renormalization transformations [169]. In order to put the concept of renor-
malization into a mathematical framework, we discuss the examples of free energy
renormalization and decimation of the one-dimensional Ising model. We then gener-
alize the concept in Section 8.5 and present the implementation of renormalization
within Monte Carlo procedures in Section 8.6.
8.2 Renormalization and Free Energy
We first focus on a scale transformation of our system’s characteristic length L →

L̃ = L/l and consider the partition function of an Ising system as defined by the
105
106 Renormalization Group
Hamiltonian given in eq. (3.20). At the critical point, a scale transformation with
L̃ = L/l leaves the partition function

Z= e−βH , (8.1)
{σ}
and the corresponding free energy invariant [121]. Due to the extensivity of the free
energy and in order to keep its density constant, we find

f (, H) = l−d f˜ ˜ , H̃ , (8.2)
where = (T c −T )/T c is the so-called reduced temperature. The physical law describing
the scale invariance at the critical point states that the free energy is a generalized
homogeneous function, which means that there exist exponents yT and yH such that
Figure 8.1 Leo P. Kadanoff (1937–2015)
eq. (8.2) is valid with ˜ = l yT and H̃ = l yH H [123], so we obtain
contributed to the development of
renormalization group theory. He and f˜ ˜ , H̃ = f˜ (lyT , lyH H) . (8.3)
H. J. Maris describe in “Teaching the
The exponents yT and yH are known as thermal and magnetic critical exponents.
Renormalization Group” [168] that
Renormalization also affects the correlation length
“communicating exciting new
developments in physics to ξ ∝ |T − T c |−ν = ||−ν . (8.4)
undergraduate students is of great The renormalized correlation length ξ̃ = ξ/l scales as
importance. There is a natural tendency
for students to believe that they are a ξ̃ ∝ ˜ −ν . (8.5)
long way from the frontiers of science Due to
where discoveries are still being made.” ξ − 1ν
lyT = ˜ ∝ ξ̃− ν =
1 1
Photograph courtesy Thomas A. Witten. ∝ l ν , (8.6)
l
we find yT = 1/ν. For the magnetic critical exponent, one obtains
d + 2 − η d + γ/ν
yH = = . (8.7)
2 2
The critical point is a fixed point of the renormalization transformation since = 0 at
T c and thus does not change for any value of the scaling factor.
8.3 Majority Rule
A straightforward example for a renormalization of spin systems in real space is the

majority rule. Instead of considering all spins in a certain cell separately, we just take
the direction of the net magnetization within this cell as a new spin value:
⎛ ⎞
⎜⎜⎜ ⎟⎟⎟
σ̃i = sign ⎜⎜⎜⎝ ⎜ σi ⎟⎟⎟⎟⎠ . (8.8)
region
We have to be careful when applying this transformation. It is not applicable to cells

with an even number of spins since this may lead to renormalized spin values with
8.4 Decimation of the One-Dimensional Ising Model 107
zero net magnetization. The fact that one numerically deals with systems of finite size
is also something that we have to take into account. We can only renormalize up to a
certain scale before finite size effects become important.
8.4 Decimation of the One-Dimensional Ising Model
Another possible rule is decimation, which eliminates certain spins, generally in a regu-
lar pattern. As a practical example, we consider the one-dimensional Ising model. The
spins only interact with their nearest neighbors and the coupling constant K = J/(kB T )
is the same for all spins. We show an example of such a spin chain in Figure 8.3.
To further analyze this system, we compute its partition function Z by distinguishing
between even and odd spins:
⎡ ⎤
< ⎢⎢⎢ ⎥⎥
σi ⎢⎢⎢ K(σ2i σ2i+1 +σ2i+1 σ2i+2 ) ⎥⎥⎥⎥
Z= eK i
= ⎣ e ⎦
{σ} σ2i =±1 2i σ2i+1 =±1
< (8.9)
= {2cosh [K (σ2i + σ2i+2 )]} , Figure 8.2 An illustration of the majority
σ2i =±1 2i
rule renormalization.
where we used that eK(σ2i +σ2i+2 ) + e−K(σ2i +σ2i+2 ) = 2cosh [K (σ2i + σ2i+2 )] in the last step.
Instead of summing over all configurations explicitly, we first evaluate the contribution
of odd spins and then account for the interactions of even spins. This procedure is
called decimation. In the next step, we rewrite the derived partition function in terms
of a renormalized coupling constant K̃:
<
Z= z (K) eK̃σ2i σ2i+2
σ2i =±1 2i
< (8.10)
eK̃σ2i σ2i+2 ,
N
= [z (K)] 2
σ2i =±1 2i
where the function z(K) is a spin-independent multiplicative part of the parti-

tion function that does not influence the physics, since it just corresponds to an
>
energy shift. The term σ2i =±1 2i eK̃σ2i σ2i+2 in eq. (8.10) describes a partition func-
tion with coupling constant K̃ and N/2 spins. We thus obtain the following recursion
relation:
N
Z(K, N) = [z (K)] 2 Z(K̃, N/2) . (8.11)
Decimation of a one-dimensional Ising chain. Figure 8.3

t
Figure 8.4 An illustration of the fixed point iteration defined by eq. (8.14). The red line corresponds to the function
K̃ = 12 ln [cosh (2K)].
To connect Eqs. (8.9) and (8.10), we compute the relation

z (K) eK̃ s2i s2i+2 = 2cosh [K (s2i + s2i+2 )] explicitly and find
⎧
⎪
⎪
⎨2cosh (2K) if s2i = s2i+2 ,
z (K) e K̃ s2i s2i+2
=⎪⎪ (8.12)
⎩2 otherwise .
We divide and multiply the two expressions of eq. (8.12) with each other and obtain
e2K̃ = cosh (2K) and z2 (K) = 4 cosh (2K) . (8.13)
The renormalized coupling constant K̃ in terms of K is therefore
1
K̃ =
ln [cosh (2K)] . (8.14)
2
We illustrate the fixed point iteration of eq. (8.14) in Figure 8.4. For any chosen initial
condition, the fixed point iteration approaches the stable fixed point K ∗ = 0. We have
thus obtained a rule that describes the change of the coupling constant under renor-
malization. For a given partition function, we now compute the free energy according
to F = −kB T N f (K) = −kB T ln (Z), where f (K) is the dimensionless free energy density.
Taking the logarithm of eq. (8.11) yields
1 1
ln [Z(K, N)] = N f (K) = N ln [z(K)] + N f K̃ . (8.15)
2 2
Based on this equation, we can derive the following recursive relation for the free
energy density using eq. (8.13):
0 1
f K̃ = 2 f (K) − ln 2 cosh (2K)1/2 . (8.16)
There exists one stable fixed at K ∗ = 0 and another unstable one at K ∗ → ∞. At every
fixed point, eq. (8.16) can be rewritten as
0 1
f (K ∗ ) = ln 2 cosh (2K ∗ )1/2 . (8.17)
The case of K ∗ = 0 corresponds to the high-temperature limit where the free energy
approaches the value
F = −NkB T f (K ∗ ) = −NkB T ln(2) . (8.18)
8.5 Generalization 109
In this case, entropy dominates the free energy. For K ∗ → ∞, the system approaches
the low temperature limit and the free energy is
F = −NkB T f (K ∗ ) = −NkB T K = −N J . (8.19)
That is, the free energy is dominated by the internal energy.
8.5 Generalization
For the decimation process that we outlined for the one-dimensional Ising system, we
found that one renormalized coupling constant was sufficient to perform a renormal-
ization iteration because it was possible to find a solution of eq. (8.12) satisfying all
spin values. In general, more coupling constants are necessary to satisfy the equiva-
lent of eq. (8.12) (e. g., in the two-dimensional Ising model). We consider a generalized
Hamiltonian of the form

M <
H(σ) = Kα Oα with Oα = σi+k , (8.20)
α=1 i k∈cα
where cα is the (spin) subset that belongs to operator Õα and M is the total number of
operators we want to consider. An example is the Ising Hamiltonian of eq. (3.20) or a
three-spin coupling

K3 σi σ j σk . (8.21)
i, j,k
We define the renormalized Hamiltonian H̃ obtained from a renormalization trans-

formation P(σ̃, σ) according to [167]

eG+H̃(σ̃) = P(σ̃, σ)eH(σ) with P(σ̃, σ) = 1 . (8.22)
σ σ̃
To satisfy the conservation of the free-energy density, we set

eG+H̃(σ̃) = eH(σ) . (8.23)
σ̃ σ
The renormalized Hamiltonian should have the same structure as eq. (8.20):

M <
H̃(σ̃) = K˜α Õα with Õα = σ̃i+k . (8.24)
α=1 i k∈cα
Inserting Eqs. (8.20) and (8.24) into Eqs. (8.22) and (8.23) and requiring that these
equations be valid for all spin configurations, we obtain a set of equations for the
renormalized coupling constants as function of the unrenormalized ones, giving the
renormalization equation:
K̃α (K1 , . . . , K M ) with α ∈ {1, . . . , M} . (8.25)

t
Figure 8.5 Schematic renormalization flow diagram. The point (K1∗ , K2∗ ) is the fixed point of the flow.
It is important to note that using only M interaction terms instead of an infinite num-
ber is a truncation, leading to systematic errors.
In the presence of a critical point
there exists a nontrivial fixed point Kα∗ = K̃α K1∗ , . . . , K M
∗
. Solving the generally non-
linear system of equations (8.25), we can now construct a flow diagram of the coupling
constant (see Figure 8.5) and obtain values for K̃ for each vector K = (K1 , . . . , K M ).
A convenient procedure to obtain the corresponding critical exponents is the lin-
earization of eq. (8.25) around this fixed point. Thus, we compute the Jacobian
T α,β = ∂∂KK̃αβ and obtain
-
K̃α − Kα∗ = T α,β --K ∗ Kβ − Kβ∗ . (8.26)
β
To analyze the behavior of the system close to criticality, we consider the eigenval-
ues λ1 , . . . , λ M and eigenvectors φ1 , . . . , φ M of the linearized transformation defined by
eq. (8.26). The eigenvectors satisfy φ˜α = λα φα and the fixed point is unstable if there
exist eigenvalues λα > 1. Such eigenvalues are called relevant and the relevant eigen-
value that corresponds to operators with even numbers of spins (temperature sector)
is called λT . We can identify the scaling field ˜ = lyT with the eigenvector of the trans-
formation, and the scaling factor with the eigenvalue λT = lyT . Then, we can compute
the exponent ν according to
1 ln (l)
ν= = . (8.27)
yT ln (λT )
Similarly, one obtains yH from the operators with an odd number of spins like the field
term in eq. (3.20) or the three spin coupling of eq. (8.21). Based on eq. (8.27) and using
yH = (d + 2 − η)/2 (see eq. (8.7)), we obtain
ln (λH )
d + 2 − η = 2yH = 2 (8.28)
ln (l)
and can now compute all the other critical exponents using Eqs. (3.34)–(3.37).
8.6 Monte Carlo Renormalization Group 111
8.6 Monte Carlo Renormalization Group
The implementation of real space renormalization with Monte Carlo techniques was
first proposed by Shang-keng Ma [170] and then made operational by Bob Swendsen
(see Figure 8.6) [171]. Since we are dealing with generalized Hamiltonians with many
interaction terms, we compute the thermal average of operator Oα according to

β Kβ Oβ
{σ} Oα e ∂F
Oα = = , (8.29)
{σ} e
β Kβ Oβ ∂Kα
where F is the free energy and α = 1, . . . , M. Based on the fluctuation-dissipation

theorem (see Section 3.2.2), we can also numerically compute the response functions

∂ Oα
χα,β = = Oα Oβ − Oα Oβ ,
∂Kβ

∂ Õα Figure 8.6 Robert Swendsen (Carnegie
χ̃α,β = = Õα Oβ − Õα Oβ . Mellon University) developed the
∂Kβ
Swendsen–Wang algorithm together
with Jian-Sheng Wang and made
We find with eq. (8.29) the following set of M 2 equations:
important contributions to the Monte
∂ Õ(n)
α ∂K̃γ ∂ Õ(n)
α Carlo Renormalization Group.
(n)
χ̃α,β = = = T γ, β χ(n)
α,γ . (8.30)
∂Kβ γ
∂K β ∂K γ γ
It is thus possible to calculate all M 2 components of the matrix T γ,β from the correla-
tion functions by solving eq. (8.30). At the fixed point K = K ∗ , we can compute the
eigenvalues of the matrix and determine critical exponents via eq. (8.27) and (8.28).
The fixed point can be localized by iterating the renormalization transformation back-
ward. The accuracy of this method depends on the number of iterations that we want
to take into account. There are many error sources in this technique, that originate
from the fact that we are using a combination of several approximations to obtain our
results:
• statistical errors,
• truncation of the Hamiltonian to M operators,
• finite number of renormalization iterations,
• finite size effects,
• no precise knowledge of K ∗ .
Nevertheless, it is one of the most accurate numerical methods to obtain critical
exponents.
9 Learning and Optimizing
There is a growing interest in machine learning and artificial intelligence methods due
to the growing amount of training data, refined algorithms, and the increase in com-
puting power [172, 173]. Many algorithms that had been around for decades have been
developed further and new methods emerged during the past few years.
The general idea behind machine learning is the development of computational
methods to perform a certain task with a performance that can be improved with
more training data. In more abstract terms, we could think of the learning problem
as a function that depends on multiple parameters and maps a given input to an out-
put. We distinguish between supervised and unsupervised learning tasks. In supervised
learning, we are given input–output pairs and adjust the parameters of the function
such that we obtain good performance in mapping given inputs to desired outputs.
In the case of unsupervised learning, we aim at extracting the underlying probability
distribution of the sample data.
Many machine-learning techniques lack a solid theoretical basis and are therefore
Figure 9.1 Geoffrey Hinton contributed not yet very well understood. Attempts have been made to characterize certain learn-
significantly to the development of ing algorithms with methods from statistical physics such as renormalization group
artificial neural networks. He is currently approaches [174]. In this chapter, we focus on one particular example of unsuper-
professor for computer science at the vised learning, so-called Boltzmann machine learning, whose origins lie in statistical
University of Toronto. physics [175]. This allows us to transfer some of our acquired knowledge on computa-
tional sampling techniques in statistical physics to the field of machine learning. After
providing an overview of the necessary theoretical concepts, we will discuss the appli-
cation of Boltzmann machines as generative neural networks, which can be used to
learn distributions and generate corresponding samples. Boltzmann machine learning
is closely linked to the name Geoffrey Hinton (see Figure 9.1) who made important
contributions to this field [176–179].
9.1 Hopfield Network
We begin our excursion to Boltzmann machines with a network consisting of N

McCulloch–Pitts neurons [180] that are fully connected (i. e., every single neuron is
connected to all other neurons). A McCulloch–Pitts neuron (henceforth neuron) rep-
resents a node in a neural network and is nothing but a function of a bias term b and N
inputs {xi }i∈{1,...,N} that are weighted by {wi }i∈{1,...,N} to compute and output y. We show
a single neuron in Figure 9.2 and an example of a Hopfield network with 8 neurons in
Figure 9.3.
112
9.1 Hopfield Network 113
In a Hopfield network (named after John Hopfield [181], see Figure 9.4), we use
discrete inputs xi ∈ {−1, 1} and weights that satisfy wi j = w ji and wii = 0. The activation
of neuron i is described by

ai = wi j x j + bi , (9.1)
j
where bi is the bias term of neuron i. Similarly to the Ising model, the associated energy
is
1
E=− wi j xi x j − bi xi . (9.2)
2 i, j i
The dynamics of a Hopfield network evolves in discrete time steps Δt according to Figure 9.2 An illustration of a single
⎧
⎪
⎪
⎨1 , if ai (t) ≥ 0 , neuron with output y, inputs
xi (t + Δt) = ⎪
⎪ (9.3) {xi }i∈{1,...,N} , weights {wi }i∈{1,...,N} ,
⎩−1 , otherwise ,
and bias b.
where ai (t) is the activation of neuron i at time t (see eq. (9.1)).
A closer look at the update in eq. (9.3) and the energy as defined by eq. (9.2) shows
that the energy in a Hopfield network is never increasing for asynchronous updates: If

ai = j wi j x j + bi ≥ 0, we set xi (ai ) = 1 and if ai < 0 we set xi (ai ) = −1. In both cases, we
obtain a negative energy contribution in eq. (9.2). The energy difference ΔEi between
neuron i in states −1 and +1 is
⎛ ⎞
⎜⎜⎜ ⎟⎟⎟
ΔEi = E(xi = −1) − E(xi = 1) = 2 ⎜⎝⎜bi + ⎜ wi j x j ⎟⎟⎠⎟ . (9.4)
j
We thus showed that the activation defined by eq. (9.1) equals one half of the energy
difference ΔEi . We can absorb the bias bi in the weights wi j by associating an extra
active unit to every node in the network. In this way, the energy difference is
Figure 9.3 An example of a Hopfield

ΔEi = 2 wi j x j . (9.5) network with four neurons. Each neuron
j is connected to all other neurons through
colored edges. Black lines indicate inputs
One possible application of Hopfield networks is to store certain information in
and outputs.
local minima by adapting the weights. For example, we could be interested in stor-
ing n binary (black and white) patterns in a Hopfield network. These n patterns are
described by the set {pνi = ±1|1 ≤ i ≤ N} with ν ∈ {1, . . . , n}. To learn the weights that
describe our binary patterns, we can employ the Hebbian learning rule [182].
Hebbian Learning
• Set the weights to

1 μ μ
wi j = w ji = p p . (9.6)
n μ i j
For Hopfield networks with a large number of neurons (thermodynamic limit), the
number of random patterns that can be stored is approximately 14 percent of the
number of neurons [183]. After employing the Hebbian learning rule, we may con-
sider certain initial configurations of xi and let them evolve according to eq. (9.3).
A Hopfield network is said to correctly represent a pattern {pνi = ±1|1 ≤ i ≤ N} if
xi (t + Δt) = xi (t) = pνi (t) for all i (i. e., if pνi is a fixed point of the system). Depending on
the number of stored patterns and the chosen initial configuration, the Hopfield net-
work may approach a local minimum that does not correspond to the desired pattern.
To avoid getting trapped in local minima, it is possible to employ a stochastic update
rule instead of the deterministic one of eq. (9.3) as will be discussed in the next section.
Figure 9.4 John Hopfield is a professor

emeritus at Princeton University.
9.2 Boltzmann Machine Learning
In Hopfield networks one can take an initial configuration that is sufficiently close to
the desired local energy minimum and retrieve the corresponding pattern. For some
applications, finding a local minimum based on the deterministic update rule defined
by eq. (9.3) might not be sufficient. In general constraint satisfaction tasks, such as
learning a certain distribution, our goal is that the underlying system is able to escape
from local minima to find a global minimum given a certain input [176]. We therefore
consider Boltzmann machines that are similar to Hopfield networks but use the update
probability
1
σi = σ(ΔEi /T ) = σ(2ai /T ) = (9.7)
1 + exp(−ΔEi /T )
to set neuron i to unity (i. e., activate it) independently of its state [176]. Here,
: ;
σ(x) = 1/ 1 + exp(−x) denotes the sigmoid function (see Figure 9.5). Such a prob-
abilistic update rule is similar to the Monte Carlo methods that we utilized to study
Ising-like systems in Chapter 4. As defined in eq. (9.4), ΔEi is the energy difference
between an inactive neuron i and an active one. The parameter T acts as temperature
equivalent.1
A closer look at Eqs. (9.2) and (9.7) tells us that we are simulating a Hamiltonian
system with Glauber dynamics (see Section 4.6). Due to the satisfied detailed balance
condition, we reach thermal equilibrium and find for the probabilities of the system to
be in state X or Y 2

peq (Y) E(Y) − E(X)
= exp − . (9.8)
peq (X) T
Independent of the initial configuration, this stochastic update procedure always leads
to a thermal equilibrium configuration that is fully determined by its energy.
1 For T → 0, we recover deterministic dynamics as described by eq. (9.3).

2 Here we set kB = 1.
The sigmoid function σi (eq. (9.7)) for different temperatures T . Figure 9.5
Boltzmann Machines
For Boltzmann machines (BMs), we distinguish between two types of nodes: visible
units and hidden units (see Figure 9.6) that are represented by the nonempty set V and
the possibly empty set H, respectively. While training a BM, visible units are set by
the environment (e. g., an empirical distribution), whereas hidden units may be used to
account for constraints involving more than two states in the input data.
Let P ({ν}) be the probability distribution of all configurations of visible units {ν}
in a freely running network. The term freely running just describes that no external
inputs are clamped/connected to the visible units. We obtain the distribution P ({ν})
by summing (i. e., marginalizing) over the corresponding joint probability distribution:

P ({ν}) = P ({ν}, {h}) , (9.9)
{h}
where we sum over all possible configurations of hidden units {h}. The goal is to come Figure 9.6 Boltzmann machines are
up with a method such that P ({ν}) approaches the unknown environment (i. e., data) composed of visible units (blue) and
distribution P ({ν}). We measure the difference between P ({ν}) and P ({ν}) in terms of hidden units (red). In the shown example,
6 {v7 i }i∈{1,...,5}
the Kullback-Leibler (KL) divergence (or relative entropy) there are five visible units
and three hidden units h j .
P ({ν}) j∈{1,...,3}
G(P, P ) = P ({ν}) ln . (9.10) The network underlying a Boltzmann
{ν}
P ({ν})
machine is complete.
To minimize G, we perform (as outlined in the gray box on pages 117–118) a gradient
descent according to [176]
∂G 1
=− pi j − p i j , (9.11)
∂wi j T
where pi j is the probability that two units i and j are active if the environment is deter-
mining the states of the visible units and p i j is the corresponding probability in a
freely running network without a coupling to the environment. Both probabilities are
measured at thermal equilibrium. The weights wi j of the network are then updated
according to [176]

w(n+1)
ij = w(n) (n)
i j − Δwi j with Δwi j = pi j − p i j , (9.12)
where n denotes the training epoch index and is the learning rate. Note that, for
the sake of brevity, we did not include the index n in the definition of Δwi j . To reach
thermal equilibrium, we need to activate all visible and hidden units according to the
probability defined in eq. (9.7). In summary, to train a Boltzmann machine we have to
do the following:
Boltzmann Machine Learning
1. Clamp the input data (environment distribution) to the visible units.

2. Activate all hidden units according to eq. (9.7) until the system reaches thermal equilibrium.
3. Update all weights according to eq. (9.12) and return to step 2 or stop if the weight updates are sufficiently
small.
After training a machine, we can unclamp the visible units from the environment
and generate samples to assess the quality of the samples that our machine generates.
We therefore consider different initial configurations and activate neurons according
to eq. (9.7) until we reach thermal equilibrium. If the learning of our BM was suc-
cessful, the distribution of states of the unclamped visible units should follow the
environment distribution.
Restricted Boltzmann Machines

Boltzmann machines are impractical for general learning tasks, because for large
system sizes the underlying complete network structure requires an enormous com-
putational effort to reach thermal equilibrium.
So-called restricted Boltzmann machines (RBMs) are not taking into account con-
Figure 9.7 Restricted Boltzmann nections within the set of hidden and visible units (see Figure 9.7), and turned out to
machines are composed of a visible layer be more suitable for learning tasks. The resulting network structure is called bipartite.
(blue) and a hidden layer (red). In the For RBMs, the updates of visible and hidden units are done in an alternating way.
shown example, the respective layers Due to the missing intralayer connections, we can first update all visible units in par-
consist of six visible units {vi }i∈{1,...,6} allel. In particular, a visible unit i that we denote by vi is activated with the conditional
and four hidden units {hi } j∈{1,...,4} . The probability
network structure underlying a restricted ⎛ ⎞
⎜⎜⎜⎜ ⎟⎟⎟
Boltzmann machine is bipartite and p(vi = 1|h) = σ ⎜⎜⎝ wi j h j + bi ⎟⎟⎟⎠ , (9.13)
undirected. j
where bi is the bias of visible unit vi and h a given configuration of hidden units. In
the second step, we activate all hidden units in parallel according to the conditional
probability
⎛ ⎞
⎜⎜⎜ ⎟⎟
p(h j = 1|v) = σ ⎜⎝ wi j vi + c j ⎟⎟⎟⎠ ,
⎜ (9.14)
i
where h j denotes hidden unit j, c j its corresponding bias, and v a given configuration
of visible units. This sampling technique is known as block Gibbs sampling.
Snapshots of 32 × 32 Ising configurations for T ∈ {1.5, 2.5, 4}. The configurations are based on M(RT)2 and RBM Figure 9.9
samples, respectively.
Training an RBM is similar to training a BM. The major difference is that we have
to account for the bipartite network structure in eq. (9.12) and thus obtain for the
weight update

Δwi j = νi h j data − νi h j model . (9.15)
Instead of sampling the configurations for computing νi h j data and νi h j model at ther-
mal equilibrium, we could also just consider a few relaxation steps. This method is
called contrastive divergence [184] and defined by the following update rule:

i j = νi h j data − νi h j model ,
ΔwCD 0 k
(9.16) Figure 9.8 Temperature dependence of
the magnetization of the Ising model on a
where the superscript index k indicates the number of updates. square lattice with 32 × 32 lattice sites.
Restricted Boltzmann machines have been applied to study phase transitions [186, Blue triangles correspond to the mean
187] and represent wave functions [188, 189] and other physical features [190]. In magnetization of 20 × 104 M(RT)2
Figure 9.8, we show the temperature dependence of the magnetization of RBM- samples. Orange squares indicate
generated Ising samples. Three snapshots of M(RT)2 and corresponding RBM- samples that we generated with RBMs
generated Ising configurations with 32 × 32 spins are shown in Figure 9.9. The that were trained on the shown M(RT)2
RBM has been trained with 20 × 104 realizations of Ising configurations at different samples. Each RBM was trained for a
temperatures [185]. distinct temperature
T ∈ {1.5, 2, 2.5, 2.75, 3, 4} [185].
Additional Information: Gradient Descent in RBMs

To derive eq. (9.11) by minimizing the KL divergence G(P, P ), we proceed according to Ref. [176]. We first
note that the environment (or clamped) distribution P ({ν}) is independent of wi j and thus
∂G P ({ν}) ∂P ({ν})
=− . (9.17)
∂wi j {ν}
P ({ν}) ∂wi j
Next, we have to calculate the gradient ∂P ({ν}) /∂wi j . For a freely running network, the probability
distribution of the visible units at equilibrium is
Additional Information: Gradient Descent in RBMs (cont.)
−E({ν,h})/T
{h} e
P ({ν}) = P ({ν}, {h}) = −E({ν,h})/T
, (9.18)
{h} {ν,h} e
with the energy

1
E ({ν, h}) = − wi j xi{ν,h} x{ν,h}
j , (9.19)
2 i, j
where xi{ν,h} is the state of neuron i for a system in state {ν, h}. Using
∂e−E({ν,h})/T 1
= xi{ν,h} x{ν,h}
j e
−E({ν,h})/T
(9.20)
∂wi j T
leads to

∂P ({ν})
1
T
xi{ν,h} x{ν,h}
{h} j e
−E({ν,h})/T
{h} e−E({ν,h})/T T1 {ν,h} xi{ν,h} x{ν,h}
j e
−E({ν,h})/T
= − 2
∂wi j {ν,h} e
−E({ν,h})/T
−E({ν,h})/T
{ν,h} e
⎡ ⎤
1 ⎢⎢⎢⎢ {ν,h} {ν,h} ⎥⎥⎥
= ⎢⎢⎣ xi x j P ({ν}, {h}) − P ({ν}) xi x j P ({ν}, {h})⎥⎥⎥⎦ .
{ν,h} {ν,h}
T {h} {ν,h}
(9.21)
We now plug this result into eq. (9.17) and find
⎡ ⎤
∂G P ({ν}) 1 ⎢⎢⎢ ⎥⎥⎥
=− ⎢⎢
⎢ x {ν,h}
x {ν,h}
P
({ν}, {h}) − P
({ν}) x {ν,h}
x {ν,h}
P
({ν}, {h}) ⎥⎥⎥
∂wi j P ({ν}) T ⎣ i j i j ⎦
{ν} {h} {ν,h}
⎡ ⎤
1 ⎢⎢⎢ {ν,h} {ν,h} ⎥⎥⎥
= − ⎢⎢⎢⎣ xi x j P ({ν}, {h}) − xi{ν,h} x{ν,h}
j P ({ν}, {h})⎥
⎥⎥⎦ , (9.22)
T {ν,h} {ν,h}
P({ν})
where we used that {ν} P ({ν}) = 1 and P ({ν})
P ({ν}, {h}) = P ({ν}, {h}) . This equation follows
from the definition of joint probability distributions
P ({ν}, {h}) = P ({h}|{ν}) P ({ν}) and P ({ν}, {h}) = P ({h}|{ν}) P ({ν}) (9.23)
and P ({h}|{ν}) = P ({h}|{ν}). The distributions P ({h}|{ν}) and P ({h}|{ν}) are identical, because in
an equilibrated system the probability of observing a certain hidden state h given a visible one ν is indepen-
dent of the origin of ν. That is, for the conditional probability it is irrelevant if ν is given by the environment
(clamped) or by a freely running machine. With the definitions

pi j xi{ν,h} x{ν,h}
j P ({ν}, {h}) and p i j xi{ν,h} x{ν,h}
j P ({ν}, {h}) , (9.24)
{ν,h} {ν,h}
we finally obtain eq. (9.11).

Exercise: Learning the Ising Model with RBMs

In this exercise, we are going to use an RBM to generate two-dimensional Ising configurations with L = 32
at a certain temperature T . Therefore, we choose the number of visible nodes to be |V| = 32 × 32.
Before samples can be drawn, the machine has to be trained. By training we mean updating the weights
and biases according to some training data. First, we have to generate about 5000 two-dimensional Ising
configurations on a square lattice of linear size L = 32 using the M(RT)2 algorithm. We then train the
machine via contrastive divergence. The update rule for the weights is given by

wi j → wi j − vi h j 0data − vi h j kmodel ,
where is the learning rate. For completeness, the update rules for the biases (visible and hidden) are
given by

bi → bi − vi 0data − vi kmodel ,

c j → c j − h j 0data − h j kmodel .
The expectation values are understood to be averages over the whole set of training data. Note that this proce-
dure is in general very slow, because the averages are always computed over the whole training data. Instead
of performing this kind of optimization for the weights and biases, it is beneficial to divide the whole set of
training data into subsets and compute the averages only over these subsets. The quantity (vi h j )data is com-
puted by taking a vector v from the training data and determining the corresponding vector h as described
above. For the quantity (vi h j )kmodel one has to take a vector from the training data, compute the corresponding
vector h, compute the new v, and perform k more back-and-forth operations.
Use your training data to find the optimal weights and biases for your RBM.
Once the machine is trained, it can be used to generate new samples. We do this in the following way:
1. Set the nodes in the visible layer to random values (either 0 or 1).
2. Let the machine evolve. That is, go several times back and forth between the visible and hidden layer.
3. Read out the nodes in the visible layer and store this RBM-based Ising sample.
Use the RBM to obtain new Ising configurations for a certain temperature and store these samples in a sep-
arate file. You can examine the quality of the RBM samples by computing their magnetization, energy, and
correlation function.
9.3 Simulated Annealing
In Sections 9.1 and 9.2, we discussed how Hopfield networks and Boltzmann machines
can be used to learn distributions and generate corresponding samples. The under-
lying optimization problem (i. e., approximating a distribution based on a given set
of samples) was solved with tailored learning rules. For the approximate solution of
general optimization problems, we may use simulated annealing, or SA for short,
which is a stochastic optimization technique that was introduced by Scott Kirk-

patrick (see Figure 9.10) and collaborators in 1983 [191]. Consider, for instance, a
given set S of possible solutions and a cost function F : S → R. We seek its global
minimum:
s∗ ∈ S ∗ {s ∈ S : F(s) ≤ F(t) ∀t ∈ S } .
Finding the solution will become increasingly difficult for large S , particularly when,
for instance, |S | = n!. The name of this optimization technique stems from annealing
in metallurgy (a technique that involves heating and controlled cooling of a material
to increase the size of its crystals and reduce their defects).
In each step of the SA algorithm, we replace the current solution by a random
“nearby” solution that is chosen and accepted with a given probability which depends
Figure 9.10 Scott Kirkpatrick is a
on the difference between the corresponding values of the cost function and a global
computer scientist and professor at the
parameter T (temperature) which is gradually decreased during the process. The cur-
Hebrew University, Jerusalem.
rent solution changes almost randomly for high temperatures but comes closer to
the minimum as we let T approach zero. We permit nonetheless “uphill” movements
in order to avoid that the algorithm gets stuck in a local minimum (this will all be
explained in the following example).
Traveling Salesman
An optimization task that can be tackled with simulated annealing is the so-called
traveling salesman problem. To illustrate this problem, we consider n cities σi and
denote the traveling cost from city σi to city σ j by c(σi , σ j ).
We now look for the cheapest trip through all these cities in order to minimize our
costs. The set S of solutions is
S = {permutations of {1, . . . , n}} (9.25)
and the cost function (for a given order, characterized by the permutation π) is

n−1
F(π) = c(σπi , σπi+1 ) for π∈S. (9.26)
i=1
A quick word about permutations: Let us say we have eight cities; then one possible
permutation (the trivial one) is π1 = {1, 2, 3, 4, 5, 6, 7, 8}, and a second one may be π2 =
{1, 8, 3, 4, 5, 6, 7, 2} (numbers describe the order of the cities on the trip). For the second
permutation (see Figure 9.11), the first three summands in F(π2 ) will be c(σ1 , σ8 ) +
c(σ8 , σ3 ) + c(σ3 , σ4 ). Finding the best trajectory is a NP-complex problem. That is,
the time that is necessary to solve the problem grows faster than any polynomial of
degree n.
We can now apply the simulated annealing algorithm to find the order of cities that
minimizes the cost function F(σ), which is here the total length of the path. We start
from a certain initial configuration of cities σ, generate a permutation σ , and apply
the following procedure:
Two possible permutations of cities in the traveling salesman problem. Figure 9.11
Simulated Annealing
• If F(σ ) ≤ F(σ): replace σ by σ .

• If F(σ ) > F(σ): replace σ by σ with probability exp(−ΔF/T ) (where ΔF = F(σ ) −
F(σ) > 0).
The T appearing in the probability is an artificial parameter of SA (temperature).

In the course of the algorithm we slowly let T go to zero in order to find the global
minimum.
Exercise: Traveling Salesman

Apply the simulated annealing algorithm to find paths that minimize the cost function of the traveling
salesman problem.
Consider a configuration of n cities (e.g., n = 10) and assign corresponding costs c(σi , σ j ) between city
σi and σ j . You can use uniformly distributed random numbers as cost values.
After initializing the sets of cities and costs, start from an initial permutation of cities π1 and apply the
simulated annealing algorithm to find permutations of cities that minimize the total cost.
• It might be helpful to try out different ways (slower/faster) of decreasing the temperature T in your
simulated annealing implementation.
• You can plot the time until you reach a certain total cost value against T to determine temperatures that
are useful for your specific optimization task.
10 Parallelization
After having discussed different algorithms for the simulation of equilibrium systems,
we now discuss parallelization techniques that can help making the necessary compu-
tations faster. Computation times of parallel codes can be orders magnitudes faster
compared to purely serial simulations! The easiest way to parallelize and obtain bet-
ter statistics is farming (i. e., executing the same program on different computers or
processors with different initial conditions). This technique is useful for Ising Monte
Carlo and related simulations for which one has to repeat the exact same sequence
of operations as often as possible. Farming is therefore just a method to improve
statistics by generating more samples than one could obtain from a single machine.
More sophisticated parallelization techniques are presented in the subsequent
sections.
10.1 Multispin Coding
The idea behind multispin coding [192] is based on the idea to efficiently use all
available bits (e. g., 64 bits if you are using a 64-bit system) for computations. Mul-
tispin coding is useful when one deals with only a few integer variables. For the
Ising model, we are simulating interactions of spins σi ∈ {±1}. Storing such spin
values in 64 bit integers is not only a waste of memory but also leads to a waste
of computation time since most of the bits are not carrying any information. On a
cubic lattice, the local field hi ∈ {0, ±2, ±4, ±6} exhibits only seven distinct values.
In binary representation, three bits are sufficient to store up to eight different num-
bers. Therefore, three bits would be enough to represent all possible local field and
energy values of any site σi while 61 bits remain unused. Spin configurations can be
also stored more efficiently using the sequences (000) and (001) which correspond
to spin down and spin up, respectively. Applying the XOR function (⊕) to two spin
sequences in the aforementioned representation yields (000) whenever two spins are
parallel and (001) for antiparallel spins. For the truth table of the XOR function, see
Table 10.1.
We can now store 21 spin values σi , . . . , σi+20 in a single 64-bit integer word:
N = (δ1 , δ2 , δ3 , δ4 , . . . , δ62 , δ63 , δ64 ) , (10.1)

σi+20 σi
where δ(·) ∈ {0, 1}. Thus, 63 bits of each integer are used and only one bit remains
unused. Similarly, we also store the six neighbors of each spin in N in integer words
122
10.1 Multispin Coding 123
Table 10.1 XOR truth table
a b a⊕b
0 0 0
0 1 1
1 0 1
1 1 0
K1 , . . . , K6 . No spins in N shall be neighbors to avoid memory conflicts and one has to

be very careful with the organization of storing spins. The flipping probabilities for all
21 spins are determined by the energy
E = N ⊕ K1 + N ⊕ K2 + N ⊕ K3 + N ⊕ K4 + N ⊕ K5 + N ⊕ K6 . (10.2)
Here ⊕ is the bitwise XOR, which realizes the XOR function for each bit in parallel,
executing 64 logical operations simultaneously. To determine the flipping probabilities,
each value has to be extracted and compared with a random number. We apply the
mask 7 = (0, . . . , 0, 1, 1, 1) and the bitwise AND (&) function to eq. (10.1) and illustrate
how to extract information of one lattice site:
7&N j = (0, . . . , 0, 1, 1, 1) ∧ (δ1 , . . . , δ62 , δ63 , δ64 ) = (0, . . . , 0, δ̃62 , δ̃63 , δ̃64 ) . (10.3)
Here δ̃(·) are the last three bit values of the integer word defined by eq. (10.1). With the
bitwise shift operator to the right or to the left we can also access the remaining bits
of eq. (10.1). A possible implementation of Monte Carlo updates of the Ising model
using the multispin coding algorithm would be the following:
cw =0;
f o r ( i =1; i <=21; i ++)
{
z=r a n f ( ) ;
i f ( z<P (E&7)) cw=(cw | 1 ) ;
cw=r o r ( cw , 3 ) ;
E=r o r ( E , 3 ) ;
}
cw=r o r ( cw , 1 ) ;
N=(N^cw ) ;
In the above pseudocode, we use the abbreviation “cw” to denote the “changer word,”
which is 1 if the spin is flipped and 0 if the spin is not flipped. Furthermore, we use
“&,” “|,” and “∧” to denote bitwise AND, OR, and XOR operations, respectively.
The circular-right-shift operation is abbreviated by “ror.” With multispin coding,
124 Parallelization
we can update 21 sites simultaneously and reduce the memory requirement by a

factor 21.
According to Ref. [193], “multi spin coding makes a program more complicated and
error-prone but may save a lot of memory and computer time for large systems. The
more complicated the interaction is, the less useful is multispin coding, and for continuous
degrees of freedom it does not seem to work at all.”
10.2 Vectorization
Vectorization is a technique which allows to perform multiple operations at the same

time like on a car assembly line. It only works in the innermost loops of a program.
As an example, we consider the following loop:
i_max = 1 0 0 0 0 ;
f o r ( i =1; i <=i_max ; i ++)
{
A( i ) = B( i ) ∗ (C( i ) + D( i ) ) ;
}
With the help of vectorizing compilers, this loop can be made more effi-
cient for simulations by executing multiple operations simultaneously and the
longer the loop the better. In the case of more complicated routines, this may
have to be implemented by hand. The following situations are problematic for
vectorization:
• multiple short loops,

• conditional branchings like if-statements,
• indirect addressing.
For optimal performance, the instructions must be repetitive without interruptions or

exceptions such as “if” statements. However, there are ways to handle some of such
cases. An example would be replacing
i f ( P ( i )> z )
{
s = −s ;
}
by
s = s ∗ s i g n ( z−P ( i ) ) ;
10.3 Domain Decomposition 125
Domain decomposition procedure. Figure 10.1
Moreover, one has to make sure that the inner loop is the longest and loops that
cannot be vectorized should be split up. It is also recommended to make use of
vectorized random number generators. Today most compilers automatically try to
vectorize programs in order to fill the pipelines of the processors.
10.3 Domain Decomposition
If one needs to simulate very large systems with local interactions, one can use domain
decomposition methods on several processors (see Figure 10.1). Instead of using
one CPU/core to simulate a certain dynamics (e. g., spin interactions) on the whole
domain, one can decompose the simulation domain into subdomains and parallelize
the computations. This method has the advantage that a large proportion of inter-
actions can be treated in parallel. Communication between processing units is only
necessary to appropriately describe interactions between boundary regions. Monte
Carlo simulations of systems with local interactions on regular lattices are well suited
for parallelization with domain decomposition.
Depending on the given domain, certain decompositions may have an advan-
tage over others. In general, one has to find a good trade-off between the speed-
up gain resulting from a larger number of subdomains and the corresponding
communication between CPUs that may slow down the simulations substantially.
It is also possible to dynamically change the domain sizes and interface posi-
tions depending on the load that each processor is just handling (dynamic load
sharing).
To parallelize a routine one has to use programming languages created specifi-
cally for such tasks or embed special libraries in which the parallelization has been
implemented in such a way that it can be summoned in standard programming lan-
guages such as C++. For programs that are intended to run on a graphics processing
unit (GPU), one can use CUDA (Compute Unified Device Architecture), a parallel Figure 10.2 Gene M. Amdahl
computing platform and programming model created by NVIDIA and implemented (1922–2015) contributed to the
in their GPUs. Moreover, MPI (Message Passing Interface) is a standardized and development of mainframe computers at
portable message-passing system. This means that instructions can be passed between IBM and later founded the company
processors within programs written in languages such as Java or C++. Amdahl Corporation.
126 Parallelization
The bottleneck of domain decomposition methods is the communication between

processors. Processors are not isolated units that completely work on their own,
and the more one slices the system into domains, the more communication between
processors is needed. This is generally not efficient and usually slows down the paral-
lelization. In addition, some parts of a program are not parallelizable. Amdahl’s law
(after Gene Amdahl, see Figure 10.2) established a relation between the theoretical
speed-up and the fraction of parallelizable code. For its derivation, we assume that a
fraction p of a computer program can be parallelized. The runtime of this fraction p
will be distributed among n processors, whereas the runtime of the non-parallelizable
part of the program stays the same. The total runtime is thus
p
T (n) = (1 − p)T 0 + T 0 , (10.4)
n
where T 0 is the runtime before parallelization. The resulting speedup is simply
Figure 10.3 Amdahl’s law (see
T0 1
eq. (10.5)) for different fractions of S (n) = = p . (10.5)
parallelizable code p and numbers of T (n) (1 − p) + n
processors n. We show some characteristic speed-up curves in Figure 10.3.
Nonequilibrium Systems 11
11.1 Directed Percolation and Gillespie Algorithms
Systems that approach thermal equilibrium obey reversibility. That is, their underlying
probability flows cancel out each other. The Ising model is one important example of
an equilibrium model in statistical physics. In Section 6.2, we outlined that the Ising
model is closely related to isotropic percolation (see Section 2.1), which consequently
can be also considered an equilibrium model. In addition to isotropic percolation,
one can define directed percolation, where sites can be only occupied (made “active”)
along a certain direction. The dynamics thus has an irreversible character. For this
reason, directed percolation is a nonequilibrium process with universal properties
different from those of isotropic percolation [121, 194]. Both processes are very fun-
damental in the sense that they are simple to define, and include features that are
observed in a variety of contexts. For instance, models that describe certain hadronic
interactions [195] and the spreading of opinions [196–200] or diseases [201–206] have
been mapped to nonequilibrium (directed percolation) processes [19, 121]. Isotropic
percolation (see Section 2.1) is an equilibrium process in which a site or bond is occu-
pied with a certain probability p and thus no time is needed to make a configuration
– all sites can be occupied in parallel. In the case of directed percolation, an addi-
tional constraint is defined, namely, at each time step occupied sites or bonds occupy
their neighbors with probability p only along a given direction [207]. Time is there-
fore a relevant parameter in directed-percolation simulations. Both models, directed
and isotropic percolation, have in common that there exists a critical probability pc at
which a percolating phase occurs. We illustrate this behavior in Figure 11.1.
In the vicinity of pc , the percolation order parameter P(p), the density of active sites
(directed percolation), takes the form P(p) ∝ (p − pc )β . In directed percolation and
related nonequilibrium systems, we define an absorbing phase (p ≤ pc ) in which the
“spreading” stops and a fluctuating phase (p > pc ) that is characterized by P(p) > 0.
Although the free energy is not defined for nonequilibrium processes, we can distin-
guish between first-order and continuous phase transitions based on the behavior of
the order parameter (see Section 3.1.7).
Another formulation of directed percolation is given by the contact process (CP).
On a d-dimensional lattice with active (si (t) = 1) and inactive (si (t) = 0) sites, we
sequentially update the system according to the following dynamics: Let the number

of active neighbors neighbors be ni (t) = i, j s j (t). A new value si (t + dt) ∈ {0, 1} is
obtained according to the transition rates.
127
128 Nonequilibrium Systems
t
(a) (b)
Figure 11.1 The order parameter of the directed-percolation second-order phase transition (a). The gray lines illustrate the effect
of a field-like term. Three clusters of directed bond percolation on a tilted square lattice (i. e., the square lattice is
tilted by 45 degrees, such that all bonds are equivalent) with different occupation probabilities p (b). At around
pc = 0.6447, a second-order phase transition occurs [121].
tFigure 11.2 Update rules of the contact process on a two-dimensional lattice (spatial rotation invariance of the reactions is
assumed). Red (green) nodes represent active (inactive) states.
w[0 → 1, ni ] = (λni )/(2d) and w[1 → 0, ni ] = 1 , (11.1)

where λ is the spreading rate.
We show an illustration of the CP updates in Figure 11.2. If λ ≤ λc , the “epidemic”
will die out. However, for λ > λc , we observe an epidemic (or fluctuating) phase. That
is, we observe a mean density of active sites

n(t) = si (t) , (11.2)
i
which is larger than zero for λ > λc . Starting with every site active at λc , the mean
density of active sites scales as n(t) ∝ t−δ (see Figure 11.3). In Table 11.1, we summarize
the critical exponents β and δ in different dimensions, and the critical values λc that are
observed in CP simulations on d-dimensional lattices. Note that the contact process is
also often referred to as (asynchronous) susceptible-infected-susceptible (SIS) model.
In the SIS model (see Figure 11.4), susceptible nodes become infected and infected
nodes recover as for the contact process with rates as defined by eq. (11.1). The
Table 11.1 The critical infection rate λc and the critical exponents
β and δ of the contact process on a square lattice for different
dimensions d
(1+1)D (2+1)D (3+1)D MF
λc 3.2978(2) 1.64872(3) 1.31683(2) –

β 0.2767(3) 0.580(4) 0.818(4) 1
δ 0.15944(2) 0.4510(4) 0.7398(10) 1
Note: At and above the critical dimension dc = 4, the
critical exponents are described by mean-field (MF)
theory [121]. The values of λc are taken from Ref. [208],
and the values of β and δ are taken from Ref. [209].
Note that λc is dimension dependent, even in the MF
case [208].
We show the relaxation of the density n(t). For λ = 1.6488 ≈ λc , the density decays as n(t) ∝ t−δ with Figure 11.3
δ = 0.4510(4) [209]. The upper guide-to-the-eye (black solid line) has slope ∝ t−0.38 , and the one below has slope
∝ t−0.45 .
SIS model is used to model epidemic diseases without (long-lasting) immunization.

Examples include bacterial infections, common cold, and influenza.
The contact process has the same critical exponents as directed percolation. This Figure 11.4 Schematic of the
implies that they both belong to the same universality class. According to a conjec- suscpetible-infected-suscpetible (SIS)
ture by Grassberger [210] and Janssen [211], the directed percolation universality class model. Susceptible individuals become
requires [121]: infected at rate β and recover at rate γ.
The use of β and γ as transition rate
1. a continuous phase transition from a fluctuating phase to a unique absorbing state, symbols is common in the
2. the transition having a one-component order parameter, epidemic-modeling literature. Note that
3. local process dynamics, the transition rate β should not be
4. no additional attributes, such as conservation laws or special symmetries. confused with a critical exponent. In
terms of a contact process description and
Nonequilibrium systems are, in general, not described by a Hamiltonian and we can-
according to eq. (11.1), we associate β
not use the same algorithms that we employed for simulating Ising-like systems or
with λni /(2d ) and set γ = 1.
other equilibrium models. Time is now a physical parameter and not a mere virtual
property of a Markov chain.
A standard algorithm for simulating nonequilibrium dynamics is kinetic Monte

Carlo that is also known under the name Gillespie algorithm (after Daniel
T. Gillespie [212, 213]).
As an example, we apply the kinetic Monte Carlo algorithm to the contact process
on a square lattice. According to eq. (11.1), recovery and infection define N = 2 pro-
cesses with corresponding rates R1 = 1 and R2 = λni /4. At time t, the total recovery

rate is given by Q1 (t) = i si (t). On a square lattice only nearest-neighbor interactions

are considered and the total rate of the second process (infection) is Q2 (t) = i λni /4
(see Figure 11.2).
The following steps summarize a general kinetic Monte Carlo update [201]:1
Kinetic Monte Carlo (Gillespie) Algorithm
1. For all N processes, identify the corresponding rates (per node) {R1 , R2 , . . . , RN }.
2. Determine the overall rates (all nodes) {Q1 , Q2 , . . . , QN } by summing up all the individual rate
contributions {R1 , R2 , . . . , RN } in the whole system.
N
3. Let η ∈ [0, 1) be a uniformly distributed random number and Q = i=1 Qi . The process with rate R j
j−1 j
occurs if i=1 Qi ≤ ηQ < i=1 Qi .
4. Let ∈ [0, 1) be a uniformly distributed random number. Update the time according to t → t + Δt,
where Δt = −Q−1 ln() and go back to step 2.
Additional Information: Kinetic Monte Carlo (Gillespie) Time Step and Process Selection
The kinetic Monte Carlo (Gillespie) algorithm uses a specific time update schema t → t + Δt, where
Δt = −Q−1 ln(1 − ) and ∈ [0, 1). The attentive reader may realize that this update rule is related
to the discussion of transforming a uniform distribution to an exponential distribution (see Section 1.6). To
better understand the origin of the specific form of the time step, we now present a derivation that is based
on Refs. [212, 213].
Our goal is to transform a given uniformly distributed random number to a random number that is distributed
according to a distribution which determines the time evolution of the underlying model. We use P(τ, i) dτ
to denote the probability that the next process with rate Qi will occur in the time interval (t +τ, t +τ+dτ).
We rewrite the reaction probability density function P(τ, i) according to
P(τ, i) dτ = P0 (τ)Qi dτ , (11.3)
where P0 (τ) is the probability that no reaction occurred in (t, t + τ), and Qi dτ is the probability that a Qi
process will occur in (t + τ, t + τ + dτ). To determine the functional form of P0 (τ), we use
1 To guarantee efficient simulations, one usually stores the site identifiers/numbers that belong to a certain
reaction in corresponding lists. After each update, one should only update the list entries of the sites that
were involved in a certain reaction and not recompute all lists.
Additional Information: Kinetic Monte Carlo (Gillespie) Time Step and Process Selection (cont.)
P0 (τ + dτ ) = P0 (τ ) (1 − Qdτ ) , (11.4)
where (1 − Q dτ ) is the probability that no process occurs in dτ . For sufficiently small dτ , the solution
of eq. (11.4) is
P0 (τ) = exp(−Qτ) . (11.5)
Inserting eq. (11.5) into eq. (11.3) yields
P(τ, i) = Qi exp(−Qτ) . (11.6)
We now use P(τ, i) = P(i|τ)P(τ), where P(τ) is the probability that a process occurs in (t + τ, t +
τ + dτ) and P(i|τ) is the corresponding conditional probability that the process with rate Qi is selected.

Moreover, we obtain P(τ) = i P(τ, i) = Q exp(−Qτ) and consequently P(i|τ) = Qi /Q. Finally,
we apply the transformation method (see Section 1.6) to P(τ) and find
Δt
= P(τ) dτ = 1 − exp(−QΔt) , (11.7)
0
where is a uniformly distributed random number between 0 and 1. Inverting eq. (11.7) yields the time step
Δt = −Q−1 ln(1 − ) = −Q−1 ln() since ∈ [0, 1).
The outlined algorithm is based on the assumption that the underlying dynamics
is governed by independent Poisson processes. Thus, the corresponding interevent
times are exponentially distributed. However, the interevent time distributions of
many natural processes are nonexponential. Examples of such processes include earth-
quakes [214], firing of neurons [215], and social dynamics [216, 217]. Deviations
from a Poisson description occur, for instance, because of correlation and memory
(non-Markovian) effects [218].
To be able to simulate processes with nonexponential interevent time distributions,
Marian Boguñá (see Figure 11.5) and collaborators introduced the non-Markovian Figure 11.5 Marian Boguñá works at the
Gillespie algorithm (nMGA) [219]: Universitat de Barcelona.
Non-Markovian Gillespie Algorithm
1. Let ψi (τ) be the probability density function of interevent times τ of process i ∈ {1, 2, . . . , N}) and
Ψi (τ) the corresponding survival probability.
2. Initialize the elapsed times {t j } ( j ∈ {1, 2, . . . , N}) for all processes (e.g., t j = 0).
3. Determine the time until the next event, Δt, according to Φ(Δt|{t j }) = , where ∈ [0, 1) is a
> Ψ (t +Δt)
uniformly distributed random number and Φ(Δt|{t j }) = nj=1 Ψj jj(t j ) .

4. Let η ∼ [0, 1) be a uniformly distributed random number and Q(Δt|{t j }) = Nj=1 Q j (t j + Δt) with
j−1
Q (t + Δt) = ψ j (t j + Δt)/Ψ j (t j + Δt). Process i occurs if i=1 Qi (ti + Δt) ≤ ηQ(Δt|{t j }) <
jj j
i=1 Qi (ti + Δt).
Non-Markovian Gillespie Algorithm (cont.)
5. Update the times {t j } according to

t j → t j + Δt ∀ j i and ti = 0 . (11.8)
6. Go back to step 2.
We describe the details of this algorithm in the gray box below. Potential drawbacks
of the nMGA are that it (i) is exact only for a sufficiently large number of processes
N and (ii) requires one to update the rates Q j (t j ) of all processes after the occurrence
of an event. Algorithms that overcome these problems are the shortest-path kinetic
Monte Carlo algorithm [220] and the Laplace Gillespie algorithm [221].
Additional Information: Non-Markovian Gillespie Algorithm

Let ψi (τ) be the probability density function of interevent times τ of process i ∈ {1, 2, . . . , N}). Further-
more, let ti be the time elapsed since the last occurrence of process i. The waiting time τ until the next event
is distributed according to
ψi (τ + ti )
ψi (τ|ti ) = , (11.9)
Ψi (ti )
∞
where Ψi (ti ) = t ψi (τ ) dτ is the survival probability of process i. That is, Ψi (ti ) is the probability that
i
the interevent time is larger than ti . We have to bear in mind that process i coexists with N − 1 other pro-
cesses. The probability density that process i and not another process generates the next event after time Δt is
<
φ(Δt, i|{t j }) = ψi (Δt|ti ) Ψ j (Δt|t j ) , (11.10)
ji
where ∞
Ψ j (t j + Δt)
Ψ j (Δt|t j ) = ψ j (τ|t j ) dτ = (11.11)
Δt Ψ j (t j )
is the conditional survival probability of process j (i.e., the probability that process j occurs with an
interevent time larger than Δt, given that process j occurred at time t j before). We thus obtain
ψi (τ + ti )
φ(Δt, i|{t j }) = Φ(Δt|{t j }) , (11.12)
Ψi (ti )
where
<
n
Ψ j (t j + Δt)
Φ(Δt|{t j }) = (11.13)
j=1
Ψ j (t j )
is the probability that no event occurs for time Δt. We therefore have to generate , a uniformly distributed
random number between 0 and 1, and solve = Φ(Δt|{t j }) to determine the time until the next event,
Δt. Now we approximate Φ(Δt|{t j }) for the limit of a large number of processes N and thus small time
steps Δt according to
Additional Information: Non-Markovian Gillespie Algorithm (cont.)

⎡ N ⎤
⎢⎢⎢ Ψ j (t j + Δt) ⎥⎥⎥
Φ(Δt|{t j }) = exp ⎢⎢⎣⎢ ln ⎥⎥⎥
⎦
j=1
Ψ j (t j )
⎡ N ⎛ ⎞⎤
⎢⎢⎢ ⎜⎜⎜ Ψ j (t j ) − ψ j (t j )Δt + O Δt2 ⎟⎟⎟⎥⎥⎥
= exp ⎢⎢⎢⎣ ln ⎜⎜⎜⎝ ⎟⎟⎟⎥⎥⎥
⎠⎦ (11.14)
j=1
Ψ j (t j )
⎡ ⎛ N ⎞⎤
⎢⎢⎢ ⎜⎜⎜ ⎟⎟⎟⎥⎥⎥ 0 1
⎢
≈ exp ⎢⎣⎢−Δt ⎜⎝⎜ ⎜ Q j (t j )⎟⎟⎠⎟⎥⎥⎦⎥ = exp −Δt Q(t j ) ,
j=1
where we used that Q j (t j ) = ψ j (t j )/Ψ j (t j ). In contrast to the standard Gillespie algorithm, the rates
Q j (t j ) depend on the times t j elapsed since the last event. Finally, we solve Φ(Δt|{t j }) = to obtain
Δt = −Q(t j )−1 ln(). In the Markovian case where Q j (t j ) = Q j , we recover the standard Gillespie
algorithm with Δt = −Q−1 ln().
Exercise: Kinetic Monte Carlo Simulation of the Contact Process

Write a program to simulate the contact process on a 2D square lattice with the kinetic Monte Carlo method.
First, consider a lattice with N = 128 × 128 sites and study the evolution of n(t) at λ = λc ≈ 1.6488.
Can you observe that n(t) ∝ t−δ with δ ≈ 0.45? Your results will improve, if you consider larger lattices
(e.g., N = 1024 × 1024 sites).
Next, we focus on the stationary behavior for different values of λ. Wait until the system reaches its stationary
state nst (λ). For λ ≤ λc ≈ 1.6488 and large system sizes, we observe nst = 0 for λ < λc and nst > 0
for λ > λc . After reaching a stationary state, measure nst (λ). We have to measure nst (λ) a certain number
of times to compute the average density nst (λ). Compute nst (λ) for different values of λ, plot nst (λ)
against λ, and determine the exponent β.
Consider the following points:
1. Define an initial seed (e.g., a circular infected region or a small fraction of randomly infected nodes).
2. Let the system evolve for sufficiently long times. After a certain number of updates, the system will reach
a stationary state or fluctuate around a stationary state. You can monitor the equilibration process by
keeping track of the variance within a certain time interval.
11.2 Cellular Automata
Cellular automata (CA) were introduced by Stanislaw Ulam and John von Neumann.
A cellular automaton is a model that consists of a grid of cells and each one of the cells
can be in a finite number of states (e. g., ±1). Cellular automata use the configuration
of cells at time t to compute the subsequent configuration at time t + 1 based on
deterministic rules. To summarize:
• Cellular automata are defined on a grid.

• The variables of each grid cell have discrete values (e. g., Boolean).
• At each finite time step, we deterministically compute the configuration of cells at
the next step t + 1 based on the configuration of the last time step t.
If we consider a binary variable σi at a certain lattice site i, the next time step t + 1 is
determined by
σi (t + 1) = fi (σ1 (t), . . . , σk (t)) , (11.15)
where k is the number of inputs that are assumed to be fixed in time for every site.
In this example, the binary function fi determines the value of σi (t + 1) based on
k
the values of σ1 (t), . . . , σk (t). We see that there are thus 22 possible rules! We dis-
tinguish between sequential and parallel updates. For sequential updates, we update
one variable σi per time step, whereas for parallel updates all variables are updated
at once.
In the following sections, we will discuss only a few properties of cellular
automata and note that much more information on this topic can be found in Refs.
[222, 223].
11.2.1 Classification of Cellular Automata

We consider k = 3 (3 inputs) as an example. There are 23 = 8 possible binary entries
(n = 111, 110, 101, 100, 011, 010, 001, 000) and the rule needs to define an output for
each of these entries. Let us consider the following rule:
entries: 111 110 101 100 011 010 001 000
f (n): 0 1 1 0 0 1 0 1
Furthermore, we define
2
k
−1
c= 2n f (n),
n=0
which for the mentioned rule yields
c = 0 · 27 + 1 · 26 + 1 · 25 + 0 · 24 + 0 · 23 + 1 · 22 + 0 · 21 + 1 · 20 = 64 + 32 + 4 + 1 = 101 .
We can thus identify a rule by its c number. We consider a few rules to see how this
works:
Evolution for two different CA rules. The horizontal describes a one-dimensional chain of cells, time goes from top to Figure 11.6
bottom, black is 1, and white is 0. The figure is taken from Ref. [222].
Rule number c Entries n: 111 110 101 100 011 010 001 000
4 f4 (n): 0 0 0 0 0 1 0 0
8 f8 (n): 0 0 0 0 1 0 0 0
20 f20 (n): 0 0 0 1 0 1 0 0
28 f28 (n): 0 0 0 1 1 1 0 0
90 f90 (n): 0 1 0 1 1 0 1 0
11.2.2 Time Evolution

We can study the time evolution of a given rule. The evolution patterns differ signifi-
cantly from rule to rule. In fact, in some cases special patterns start to appear while in
other cases every site ends up in the same state. Two examples starting with a random
initial configuration are illustrated in Figure 11.6.
11.2.3 Classes of Automata

Stephen Wolfram (see Figure 11.7) divided the rules into four groups according to
their dynamical characteristics [224]:
• Class 1: Initial patterns evolve quickly into a stable, homogeneous state, any trace
of the initial patterns disappears.
• Class 2: Almost all initial patterns evolve quickly into stable striped or oscillat-
ing structures. Some traces of initial patterns remain. Local changes to the initial Figure 11.7 Stephen Wolfram
pattern tend to remain local. investigated the patterns formed by
• Class 3: Nearly all initial patterns evolve in a pseudo-random or chaotic manner. cellular automata and developed the
Any stable structures that appear are quickly destroyed by the surrounding noise. scientific computing software
Local changes to the initial pattern tend to spread indefinitely. Mathematica.
t
Figure 11.8 Examples of persistent structures of a class 4 (rule 20 for k = 5) CA. The figure is taken from Ref. [222].
• Class 4: Nearly all initial patterns evolve into structures that interact in complex
and interesting ways. These structures collide and sometimes annihilate each other
slowly disappearing and eventually exhibiting class 2 behavior; the time necessary
to reach that point is very large and diverges with system size.
We show some persistent structures of a class 4 automaton in Figure 11.8.
11.2.4 The Game of Life
One example of a class 4 cellular automaton is the Game of Life, which was invented
by Conway (see Figure 11.9) in 1970. Let us consider a square lattice, and let n be
the number of nearest and next-nearest neighbors that are in state “1.” We use the
following rule for a certain lattice site:
• if n < 2 : set the site to 0,

Figure 11.9 John Horton Conway
• if n = 2 : stay as before,
(1937–2020) was a British
• if n = 3 : set the site to 1,
mathematician.
• if n > 3 : set the site to 0.
This rule is known under the name Game of Life. Cellular automata that only depend
on the number n of 1s in a given neighborhood like the Game of Life are called total-
istic. During its temporal evolution, the Game of Life keeps producing “gliders” from
a “gun” (see Figure 11.10).
t
Figure 11.10 Glider gun of ones (black) within zeros (white). Steps 0 and 16 (in a period of 30 time steps) of the Game of
Life [225].
Exercise: Game of Life

Simulate Conway’s Game of Life (see Section 11.2.4) on a square lattice with periodic boundary conditions.
Implement different initial configurations and study their evolution. Can you observe the glider behavior that
we show in Figure 11.10?
Tasks
1. How many types of different gliders do you observe in your simulations for different initial conditions?
2. Study the density of 1s for different system sizes as a function of time.
11.2.5 Q2R
For Emax → 0, the Creutz algorithm (see Section 4.8) for microcanonical simulations of
the Ising model on the square lattice becomes a cellular automaton called Q2R [226].
The update rules on a square lattice for spins σi j ∈ {0, 1} are
σi j (τ + 1) = f (xi j ) ⊕ σi j (τ) , (11.16)
where ⊕ is the binary XOR operation and
⎧
⎪
⎪
⎨1 if x = 2
xi j = σi−1 j + σi+1 j + σi j−1 + σi j+1 and f (x) = ⎪
⎪ . (11.17)
⎩0 if x 2
For the outlined Q2R dynamics, spins are flipped if and only if the change in energy
is zero. This can be implemented in a very efficient way using multi spin coding (see
Section 10.1). The update rule of eq. (11.16) can be expressed in a very elegant way
using logical functions:
σ(τ + 1) = σ(τ) ⊕ {[(σ1 ⊕ σ2 ) ∧ (σ3 ⊕ σ4 )] ∨ [(σ1 ⊕ σ3 ) ∧ (σ2 ⊕ σ4 )]} , (11.18)
where ∧ and ∨ are the logical AND and logical OR operators, respectively. This logical
expression can be computed in roughly 12 cycles that last typically around 10 ns for 64
sites in parallel for bitwise logical functions. The method is extremely fast, determinis-
tic, and reversible. It represents the fastest possible simulation for an Ising model [227].
The problem is that it is not ergodic and that it strongly depends on the initial spin
configuration.
11.3 Irreversible Growth
For many applications, we cannot assume that the underlying system is in thermal
equilibrium. Many natural phenomena are essentially irreversible. Examples include
• deposition and aggregation,

• fluid invasion,
• electric breakdown,
• biological morphogenesis,
• fracture and fragmentation.
In the following sections, we will get to know some methods that help us understand
irreversible growth. More details can be found in Ref. [228].
11.3.1 Self-Affine Scaling
Let us consider a growing interface, like the front of an invading fluid or the surface of
a tumor. We can analyze the dependence of the interface width W (see eq. (4.47)) on the
variables L (lattice length) and t (time). The following relation is called Family–Vicsek
scaling [230]
t
W(L, t) = Lξ f z , (11.19)
L
where ξ is the roughening exponent and z the dynamic exponent. This scaling relation
was introduced by Fereydoon Family and Tamás Vicsek (see Figure 11.11). We can
now consider two limits: (i) t → ∞ and (ii) L → ∞. To simplify the expression, we
substitute the argument of f by u = t/Lz and find
Figure 11.11 Fereydoon Family (Emory • t → ∞ : W ∝ Lξ =⇒ f (u → ∞) = const,

University) (top) and Tamás Vicsek • L → ∞ : W ∝ tβ =⇒ f (u → 0) ∝ uβ ,
(Eötvös Loránd University) (bottom)
where we have introduced the so-called growth exponent β = ξz (see Figure 11.12). We
introduced the concept of dynamic
can numerically verify these laws by observing a data collapse.
scaling.
11.3.2 Random Deposition

The simplest model for irreversible growth models is random deposition. It is based
on only one rule:
Random Deposition
• Pick a column uniformly at random and add a particle of unit height to that column (see Figure 11.13).
The height is directly proportional to time (i. e., number of depositions), h ∝ t and the
1
width W is proportional to t 2 . The roughening exponent ξ and the growth exponent β
are both 12 , so the Family–Vicsek scaling law (see eq. (11.19)) is
√ t
W(L, t) = Lf . (11.20)
L
We show a typical random-deposition pattern in Figure 11.13. One immediately
notices the sharp spikes which are completely independent from one another and that
the height varies wildly between columns.
11.3.3 Random Deposition with Surface Diffusion
For many applications, the random deposition model is oversimplified. If we would

like to get something more realistic/less spiky, we need to give the particles a bit more Figure 11.12 Interface width W versus
freedom. This is done by permitting the particles to move a short distance to find a time t for ballistic deposition and four
more stable configuration. different linear lattice sizes L. The figure
The following procedure accounts for such short-distance movements: is taken from Ref. [229].
Random Deposition with Surface Diffusion
• Pick a random column i.

• Compare the heights h(i) and h(i + 1).
• The particle is added to the lower column. For equal heights, the new particle is added to column i or i + 1
with equal probability.
We illustrate this procedure (along with a typical growth pattern) in Figure 11.14.
√
Using this growth model, the average height now increases with t while the width W
1
increases with t 4 . The roughening exponent ξ is 12 and the growth exponent β is 14 , so
the Family–Vicsek scaling law is
√ t
W(L, t) = Lf . (11.21)
L2
In Figure 11.14, we observe that the surface is much “smoother” and not as jagged
anymore as without diffusion.
11.3.4 Restricted Solid on Solid Model

Figure 11.13 Growth by adding one
One can make the surfaces even smoother by requiring that neighboring sites may not particle and result after adding many
have a height difference larger than one unit. The procedure then needs to reflect this particles using the random deposition
restriction with an additional condition (which can lead to the rejection of a particle algorithm. The colors represent the times
addition, if it does not meet the condition). The recipe incorporating this idea defines at which certain particles were added.
the restricted solid on solid model (RSOS):
Restricted Solid on Solid Model
• Pick a random column i.

• Add a particle if |h(i + 1) − h(i − 1)| ≤ 1.
• If the particle could not be placed, pick a different column.
We illustrate the procedure, again along with a typical growth pattern, in

Figure 11.15. Because of the rather stringent requirement introduced in the RSOS,
the height now varies less than for the previous models and the interfaces become
smoother. In Figure 11.12, we see how for the RSOS model the width increases with
time for two different system sizes. Again, we may ask for the roughening exponent ξ
which is 12 and the growth exponent β which is 13 . The Family–Vicsek scaling law for
RSOS is thus

√ t
W(L, t) = L f 3
.
L2
11.3.5 Eden Model

Figure 11.14 Growth by adding one The Eden model was introduced 1961 by Murray Eden [231] to describe tumor growth
particle and result after adding many and epidemic spreading. In this model, each neighbor of an occupied site has the same
particles using random deposition with probability of being occupied. The growth algorithm is as follows:
surface diffusion. The color represents the
time at which the particle was attached.
Eden Model
• Make a list of all empty nearest neighbors (red in Figure 11.16).

• Pick one empty nearest neighbor at random from the list. Add it to the cluster (green in
Figure 11.16).
• Add the empty neighbors of this site to the list, if they are not already there.
There may be more than one growth site in a column, so there can be overhangs and
there can also be lakes of empty sites inside the occupied zone. We illustrate the Eden
model together with a typical realization in Figure 11.16. The roughening exponent ξ
is 12 again while the growth exponent β is 13 . The Family–Vicsek scaling is the same as
for RSOS.
We can grow a cluster from a point-like seed according to the Eden model (a
so-called Eden cluster) and use it as a simple model for simulating tumor growth.
We show a typical growth pattern in Figure 11.17 and see that it is essentially just a
big blob with rough edges.
11.3.6 Ballistic Deposition

In yet another model, we drop particles from above onto the surface; they stick on
the surface either when they touch a particle below or a neighbor on one side. The
algorithm is as follows:
Ballistic Deposition
• Pick a column i.
• Let a particle fall vertically until it touches a neighbor (either below or on either side).
We show the growth sites (red) of ballistic deposition and a typical realization
in Figure 11.18. Considering only the external surface (not the internal holes),
the roughening and growth exponents are again ξ = 12 and β = 13 , respectively.
The size of the largest hole in fact grows logarithmically with time. The model is
often generalized to particles impacting with a certain angle instead of falling down
vertically.
11.3.7 Continuum Description

We have seen a handful of models to grow irreversible patterns according to a given
set of rules. Let us now formulate an equation for the surface height h(x, t) based on
some symmetry arguments:
1. Invariance under translation in time t → t + δt, where δt is a constant.

2. Translation invariance along the growth direction: h → h + δh, where δh is a
constant.
3. Translation invariance along the substrate: x → x + δx, where δx is a Figure 11.15 Growth rule and typical
constant. simulation of the RSOS algorithm. The
4. Isotropic growth dynamics (e.g., x → −x). color represents the time at which the
particle was attached.
5. Conserved relaxation dynamics: The deterministic part can be written as the
gradient of a current.
Of course, not all growth algorithms are going to respect every single one of these sym-
metries (nor would we want them to!). Depending on the growth process we envision,
a given set of symmetries will need to be respected.
An equation for a growth process respecting all five symmetries is the Edwards–
Wilkinson (EW) equation [232]
∂h(x, t)
= νΔh + η(x, t) , (11.22)
∂t
where ν is the surface tension and η is Gaussian white noise with average
η(x, t) = 0 (11.23)
and second moment
η(x, t)η(x , t ) = 2Dδ(x − x )δ(t − t ) . (11.24)
The resulting time evolution of h(t) follows Family–Vicsek scaling with exponents
ξ = 12 and β = 14 . Random deposition with surface diffusion belongs to this universality
class.
When we include the next leading order in the gradient (which is nonlinear), we
drop the fifth requirement (conserved relaxation dynamics) and obtain the Kardar–
Parisi–Zhang (KPZ) equation [233]
Figure 11.16 Growth sites (red) and a ∂h λ

= νΔh + (∇h)2 + η(x, t) . (11.25)
typical realization of the Eden algorithm. ∂t 2
The color represents the time at which
the particle was attached. This equation was first proposed by M. Kardar, G. Parisi, and Y.-C. Zhang (in 1986).
Its scaling exponents are ξ = 12 and β = 13 . Thus, the RSOS model, the Eden model, and
ballistic deposition belong all to the universality class of KPZ.
The EW and KPZ equations both describe the evolution of the height h. We see
that one phenomenon that is common to both equations is noise, parametrized as
η(x, t), which means that these are stochastic equations that require specific numerical
methods [234]. Note that only the KPZ equation takes into account differences in
height along the cluster (via the term ∝ (∇h)2 ), which describes lateral growth.
11.3.8 Diffusion-Limited Aggregation
Figure 11.17 Eden cluster on a square Another irreversible growth model is diffusion-limited aggregation (DLA), which
lattice grown from a single seed. The has been introduced by Witten and Sander [235] (see Figure 11.19). It describes the
figure was taken from Ref. [228]. aggregation of diffusing particles to a cluster. The algorithm is as follows:
Diffusion-Limited Aggregation
1. Consider a particle far away from the surface of the cluster.

2. Let the particle diffuse (i.e., perform a random walk). If the particle moves too far away, go back
to step 1.
3. Attach the particle to the growth cluster if it touches it. Then go back to step 1.
Exercise: Diffusion-Limited Aggregation

In diffusion-limited aggregation (DLA), we start with a seed point. In each of the following steps a particle is
released from the outside and performs a random walk until it reaches the perimeter of the structure, where it
is attached to the structure (or seed surface). This can be implemented using continuous space or on a lattice
(to simulate snow flakes, a triangular lattice is used).
For simplicity, generate DLA clusters on a two-dimensional lattice (square lattice, or/and triangular lattice). As
a seed, you may use a site or a surface (in 2D a line). For a seed site, it is practical to define a start radius, from
where the random walker starts, and a maximal radius. When the walker reaches the maximal radius, the
random walk is stopped and a new walker starts at the start radius (for a seed surface it can be implemented
similarly).
• Color sites according to their aggregation time. Which sites of the perimeter are occupied? What does it
mean for the probability of occupying perimeter sites (e.g., as compared to the Eden model where each
perimeter site is occupied with equal probability)?
• Implement a method of “noise reduction” such that a perimeter site has to be visited m times before it
is occupied (i.e., store number of visits at each perimeter site). What is the qualitative difference when
varying m?
• Determine the fractal dimension with the methods you know already (sandbox, box counting, and Figure 11.18 Ballistic deposition growth
ensemble method). sites (red) and typical simulation. The
• Generate a DLA cluster in continuous space (i.e., no lattice). color represents the time at which the
particle was attached.
• On the lattice, a list of all perimeter sites can be helpful to determine if a site belongs to the perimeter
(this would also allow you to easily change your implementation to the Eden model). Another (simpler)
possibility is to check if the neighboring sites are occupied.
• For efficiency, the start radius and the maximal radius can be adjusted to the actual radius of the cluster
at each time step (e.g., start radius two sites larger than current cluster radius, maximal radius twice the
current cluster radius).
We show an example of a DLA growth process in Figure 11.20. DLA clusters are
fractal with fractal dimensions of d ≈ 1.71 in two and df ≈ 2.53 in three dimensions.
Applications of DLA include electro- and mineral deposition, viscous fingering, and
the formation of ice dendrites.
11.3.9 Dielectric Breakdown Model (DBM) Figure 11.19 Thomas Witten is a

professor at the University of Chicago
A very interesting generalization of DLA is a model originally designed to describe (top), and Leonard M. Sander is a
dielectric breakdown [236]. professor emeritus at the University of
Michigan (bottom).
We start out from the Laplace equation

φ = 0. (11.26)
In the center of the system, we place the seed for the growing cluster on which we
impose the boundary condition φ = 0, while on the external boundary of the system,
we impose φ = 1. After finding the solution, one adds a site on the surface of the
cluster with probability
p ∝ (∇φ)η , (11.27)
where the gradient is taken perpendicular to the surface and η is a parameter of the
model. If one replaces the electric field φ by pressure, the above problem describes
viscous fingering, if one identifies φ with temperature, one describes crystal growth in
an undercooled melt, and if φ is chosen to be the probability to find a random walker,
we model diffusion, like in Section 11.3.8. We recover two previous models for special
values of η:
• For η = 1, we recover DLA.

• For η = 0, growth becomes independent of φ and DBM produces Eden clusters.
We show DBM configurations for different values of η in Figure 11.21. For large η, the
clusters become more anisotropic resembling electric discharges in a random medium,
which gives the model its name.
Figure 11.20 Growth rule and a typical
realization of diffusion-limited
aggregation. The color represents the
time at which the particle was attached.
t
Figure 11.21 Clusters obtained by the dielectric breakdown model for different values of η. The field φ is shown in gray.
11.3.10 Random Fuse Model and Fracture

The random fuse model was introduced by Lucilla de Arcangelis et al. [237] (see
Figure 11.22) to describe brittle fracture of disordered solids. This model is based on
a network of fuses that are defined as electrical resistors which follow Ohm’s law until
they reach a randomly chosen breaking current at which they fail (see Figure 11.23).
If one applies an electrical potential drop U between two sides of the system that one
slowly increases, one by one, the fuses will break each time the current flowing through
them reaches their threshold. When all conducting paths are disrupted, the system
fractures in two pieces and no current can flow anymore. In order to simulate this
model, one must calculate the currents passing through each resistor after each break-
ing. This is done by imposing Kirchhoff’s law at each point of the grid, which produces
a system of coupled linear equations. These can be solved with LU decomposition
since the matrix is sparse.
I
U Ifuse
Ic
Figure 11.22 Lucilla de Arcangelis is a

professor at the University of Campania
“Luigi Vanvitelli” in Naples, Italy.
0 Uc Ufuse
I
(a) (b)
(a) Resistor network with a potential difference U between the top and bottom parts. Note that some resistors are Figure 11.23
broken (indicated by broken bonds). (b) If the current exceeds a threshold Ic , the fuse breaks, and current can no
longer flow through it.
PART II
M O L ECULAR DYNAMICS
Basic Molecular Dynamics 12
12.1 Introduction
In Part I, we introduced different stochastic methods to describe the properties of

a large number of interacting units, for example particles, spins, and other entities.
Instead of a statistical description, we now focus on modeling microscopic particle
interactions by solving the corresponding equations of motion. This technique is also
known as molecular dynamics (MD). There are several excellent books dedicated to
this method [238]. We start with some straightforward and intuitive methods such as
the so-called Verlet and leapfrog schemes for solving Newton’s equations of motion.
In what follows, we discuss Lagrange multipliers and long-range potential methods
as tools to simulate composed and long-range interacting particle systems. For hard-
sphere interactions and long-lasting contacts, we give an overview of event-driven MD
and contact dynamics. We conclude this second part of our book with a discussion of
different quantum-mechanical (ab initio) MD approaches.
In classical MD simulations, we have to solve Newton’s equations of motion. The
mathematical framework of classical mechanics dates back to the work of Isaac New-
ton in the seventeenth century [239]. But only with the rise of computers in the second
half of the twentieth century, it became possible to perform first MD simulations.
Many of the techniques contained in this chapter form the basis of modern commer-
cial softwares that are frequently applied to many engineering and industrial problems.
Some examples of fields where MD simulations are relevant include
• simulation of atoms and molecules,

• gravitational interactions in astrophysics,
Additional Information: History of Numerical Methods

Numerical methods are almost as old as mathematics itself. Simple calculations such as linear interpolation
and square root approximations were developed a few thousand years ago (see, e.g., the Babylonian clay
tablet). More refined methods in differential analysis started to become of great interest with the rise of phys-
ical sciences in the seventeenth century. As an example, Euler’s method was described by Leonhard Euler in
1768 but was certainly known for a long time before. Runge–Kutta methods, which are often used today
for very precise calculations (e.g., in celestial mechanics), were developed around 1900. For Hamiltonian
problems, there exist different tailored numerical methods [240].
149
150 Basic Molecular Dynamics
• fluid dynamics (including traffic and pedestrians [241]),

• biopolymers,
• granular materials,
• dislocations, voids, quasi-particles,
• electrons (Car–Parrinello [242]),
• virtual reality simulations.
One of the pioneers of this field is Berni Alder (see Figure 12.1). He was one of the
first who developed MD methods to computationally study particle interactions [243].
12.2 Equations of Motion
To model interacting particle systems, we use generalized coordinates

qi = q1i , . . . , qdi and pi = p1i , . . . , pdi (12.1)
in a system where each particle i has d degrees of freedom. We then describe the system
consisting of N particles by
Q = (q1 , . . . , qN ) and P = (p1 , . . . , pN ) , (12.2)
and the Hamiltonian

H (P, Q) = K (P) + V (Q) (12.3)
( pk )2
with K (P) = i,k 2mi i being the kinetic energy, mi the mass of the ith particle, and
V(Q) the potential energy. The sum over k ∈ {1, . . . , d} accounts for the d degrees
Figure 12.1 Berni Alder (1925–2020) of freedom. The potential (e. g., an attractive or repulsive electromagnetic potential)
was a professor at the University of determines particle interactions, and therefore their dynamics. The potential energy
California, Davis. He won the Boltzmann can be decomposed in the following way:
Medal in 2001. The Berni Alder Prize is
the most prestigious European prize for V (Q) = v1 (qi ) + v2 qi , q j + v3 qi , q j , qk + . . . . (12.4)
computational physics. i i j>i i j>i k> j>i
The used summation ( j > i, k > j > i) avoids to sum over any pair twice. The first term
v1 (·) describes interactions of particles with the environment (e. g., gravity and bound-
ary conditions for particle–wall interactions). The second term v2 (·) accounts for
pairwise interactions, and v3 (·) for three-body interactions between particles. Exam-
ples of potentials involving three particles are the Stillinger–Weber potential [244]
and the Axilrod–Teller potential [245]. Usually, three or more body interactions are,
however, neglected and their effect is considered in an effective two-body interaction
described by
- -
veff q , q = vattr (r) + vrep (r) with r = --q − q -- ,
2 i j i (12.5)
j
12.2 Equations of Motion 151
where vattr (r) and vrep (r) represent attractive and repulsive parts of the potential,
respectively. The simplest potential is the hard-sphere interaction potential
⎧
⎪
⎪
⎨∞ , if r < σ ,
v (r) = ⎪
rep
⎪ (12.6)
⎩0 , if r ≥ σ .
An illustration of the hard-sphere potential is shown in Figure 12.2. This potential

is not well suited for numerical simulations due to the fact that the force f = −∇vrep (r)
becomes infinite at r = σ. A softer potential should be used in numerical simulations.
One possibility is to use a spring-like repulsion potential
⎧
⎪
⎪
⎨ 2 (R − r) , if r < R ,
k 2 Figure 12.2 An example of a hard-sphere
v (r) = ⎪
rep
⎪ (12.7) potential. In the blue shaded region the
⎩0 , if r > R ,
potential is infinite.
where k is the elastic spring constant, and R = R1 + R2 is the sum of the radii of
the individual particles. Such a soft-sphere interaction may lead to unrealistic sim-
ulation outcomes (overlapping particles). Another important potential for modeling
molecular interactions is the Lennard–Jones potential
12 6
σ σ
vLJ (r) = 4 − , (12.8)
r r
where is the depth of the potential well and σ is the interaction range. It is a
mathematically simple model that approximates the spherically symmetric interaction
between a pair of neutral atoms or molecules. The first term accounts for Pauli repul-
sion at short ranges and the second term describes attractive van der Waals forces. An
example of a Lennard–Jones potential is shown in Figure 12.3. It is possible to define
a cutoff for certain short-range potentials to reduce the computational effort. This is,
however, not applicable to long-range potentials that are therefore more challenging
to simulate.
Once the interaction potential has been defined, we can easily derive the equations
of motion using the Hamilton equations Figure 12.3 The Lennard–Jones
potential approaches zero for large
∂H ∂H
q̇ik = and ṗik = − , (12.9) distances r and reaches a minimum at
∂pik ∂qik r = 21/6 σ.
where k ∈ {1, . . . , d} and i ∈ {1, . . . , N}. Let us identify qi with the position vector xi and
q̇i = ẋi with the velocity vector vi . Using ẋi = vi = pi /mi and ṗi = −∇i V (Q) = fi , we
obtain Newton’s equations of motion

mi ẍi = ṗi = fi = fi j , (12.10)
j
where fi j is the force exerted on particle i by particle j. Simulating Hamiltonian dynam-

ics implies that our molecular system exhibits certain conservation laws. First, the
total energy is conserved if our considered Hamiltonian H (P, Q) is not explicitly time
dependent (i.e., if ∂t H = 0). Second, if the system is translational invariant in a certain
direction, the corresponding momentum is conserved. And third, if the simulated sys-
tem is rotational invariant about a certain axis, the corresponding angular momentum
component is conserved. For example, a cubic box with periodic boundary conditions
leads to a conserved total momentum

P= pi . (12.11)
i
For a spherically symmetric box, the total angular momentum

L= xi ∧ pi (12.12)
i
about the center of symmetry is conserved.
12.3 Contact Time
We now have to solve the equations of motion defined by eq. (12.10) with the help of
appropriate numerical integration methods. We are going to numerically compute the
motion of particles via an explicit forward integration and it is thus very important
to use a sufficiently small time step Δt. If our chosen time step is too large, we may
encounter unrealistic overlaps and particles may pass through the interaction range of
other particles without interacting with them. We therefore need a measure that esti-
mates the required time step and the corresponding integration error. For potentials
of finite range, such a measure is the so-called contact time that corresponds to the
time span during which a contact lasts. Since we consider an interaction force which
only depends on distance, we first analyze our particle dynamics in a one-dimensional
setting. The interaction of the particle with the potential is illustrated in Figure 12.4.
Using energy conservation
t
Figure 12.4 Derivation of the contact time. The particle with a kinetic energy K approaches another particle at distance rmin and
comes to a halt at a distance rmax , when its potential energy equals K. Then it returns and leaves the collision region
with kinetic energy K.
12.4 Verlet Method 153
1 2
E= mṙ + V (r) = const. , (12.13)
2
we obtain the velocity
12
dr 2
= (E −V (r)) (12.14)
dt m
and the contact time
1 tc rmax rmax − 12
2 dt 2
tc = 2 dt = 2 dr = 2 (E − V (r)) dr , (12.15)
0 rmin dr rmin m
where rmin and rmax are the range of the potential and the turning point of a collid-
ing particle, respectively. The contact time sets the scale for the appropriate time step
of the explicit forward integration of the equations of motion in an MD simulation.
We expect accurate results only if the time step is at least ten times smaller than the
smallest contact time. The time integration of the equations of motion is then possible
using an explicit forward integration method such as
• Euler’s method,
• Runge–Kutta methods,
• predictor-corrector methods,
• Verlet method,
• leapfrog method.
In the subsequent sections, we only focus on the last two methods which have been
developed specifically for solving Newton’s equations.
12.4 Verlet Method
The Verlet integration method was proposed by Loup Verlet (see Figure 12.5) to solve
Newton’s equations of motion [246]. The procedure is simple and related to forward
Euler integration. We begin with a Taylor expansion of the position xi (t ± Δt) of the
ith particle for sufficiently small time steps Δt:
1
xi (t + Δt) = xi (t) + Δt vi (t) + Δt2 v̇i + O Δt3 ,
2 (12.16)
1 2
xi (t − Δt) = xi (t) − Δt vi (t) + Δt v̇i − O Δt3 .
2
Adding these two expressions yields
Verlet Integration

xi (t + Δt) = 2xi (t) − xi (t − Δt) + Δt2 ẍi (t) + O Δt4 . (12.17)
Newton’s second law enables us to express the acceleration ai (t) = v̇i (t) = ẍi (t) as
1
ẍi (t) = fi j (t) with fi j (t) = −∇V ri j (t) . (12.18)
mi j
The particle trajectories are then computed by plugging these results into eq. (12.17).
Ideally, we should use a time step of approximately Δt ≈ tc /20. Some general remarks
about the Verlet method:
• Two time steps need to be stored (t and t − Δt).

• Velocities can be computed with vi (t) = xi (t+Δt)−x i (t−Δt)
.
2Δt
Figure 12.5 Loup Verlet (1931–2019) • The local numerical error is of order O Δt (i.e., it is globally a third order
4
was a research director at CNRS in Orsay. algorithm).

Photograph courtesy Jean-Jaques Banide • The numbers which are added are of order O Δt0 and O Δt2 , posing a challenge
and the Les Houches School of Physics to numerical accuracy.
(reprinted with kind permission of • We can include higher-order terms (very inefficient).
Christiane DeWitt).
• The method is time reversible, which allows us to estimate the accumulation of
numerical errors by reversing the process and comparing it to the initial configu-
ration.
12.5 Leapfrog Method
The leapfrog method was introduced by Hockney in 1970 [247]. For the derivation
of the leapfrog method, we start with the velocities of the ith particle at intermediate
steps

1 1
vi t + Δt = vi (t) + Δtv̇i (t) + O Δt2 ,
2 2
(12.19)
1 1
vi t − Δt = vi (t) − Δtv̇i (t) + O Δt . 2
2 2
Taking the difference of these two equations yields

1 1
vi t + Δt = vi t − Δt + Δt ẍi (t) + O Δt3 , (12.20)
2 2
where ẍi (t) is given in eq. (12.18). Finally, the position is updated according to

1
xi (t + Δt) = xi (t) + Δt vi t + Δt + O Δt4 . (12.21)
2
The leapfrog method has the same order of accuracy as the Verlet method. However,
both methods differ in the way in which the variables are integrated. To summarize,
the leapfrog method is based on the following update scheme:
12.5 Leapfrog Method 155
Leapfrog Integration
1
ẍi (t) = fi j (t) ,
mi j

1 1
vi t + Δt = vi t − Δt + Δt ẍi (t) , (12.22)
2 2

1
xi (t + Δt) = xi (t) + Δt vi t + Δt .
2
Additional Information: Comparison of Euler and Leapfrog Methods

If we compare the leapfrog algorithm with a forward Euler scheme
1
ẍi (t) = fi j (t) ,
mi j
(12.23)
vi (t + Δt) = vi (t) + Δt ẍi (t) ,
xi (t + Δt) = xi (t) + Δt vi (t) ,
we see that the leapfrog method uses velocity and acceleration at different time steps to compute the new
particle positions.
Verlet and leapfrog schemes are microcanonical (i. e., energy conserving). Therefore,
for sufficiently small time steps Δt, energy is conserved on average during a simulation
and fluctuations are due to roundoff errors. Large fluctuations in energy are usually
a hint for the time step being too large or an erroneous code. It is also possible to
estimate the accuracy of the method by analyzing the underlying energy fluctuations.
As in the case of the Verlet integration scheme, the leapfrog method is completely
time reversible and we can assess the error by looking at the change in the initial con-
figuration after having simulated forward and then backward in time. The difference
between the two configurations can be taken as a measure for the error of the method.
Another approach is to take two initial configurations that differ only slightly in the
velocity or the position of a particle, and to observe the difference in the time evolution
of the two systems. Initially, both systems should behave similarly, whereas the differ-
ence increases exponentially with time. For instance, let ΔE be the energy difference
between both systems. The slope of log ΔE as function of time is an indicator of the
precision of the method. The slope is the so-called Lyapunov exponent, and describes
the divergence of the simulated trajectory from the true one. To reduce this sensitive
dependence on initial conditions, other integration methods (e. g., splitting methods)
can be used to simulate the long-term evolution of such Hamiltonian systems.
Exercise: Simulating Interacting Particles in a Box

We now simulate a system of N interacting particles in a box using the Verlet scheme.
The programming code of this exercise will be the basis for further improvements and modifications in the
following exercises.
Consider a system of N point particles of mass m (in a box of length L with periodic boundary conditions)
that interact via the Lennard–Jones potential
⎡ 12 6 ⎤
⎢⎢ 1 1 ⎥⎥⎥⎥
V(r) = 4 ⎢⎢⎣⎢ − ⎥.
r r ⎦
The total Hamiltonian of the system is given by
1 2 1 -- -
H= m vi + V(- ri − r j --) .
2 i
2 i j
Task 1: Simulate the system by using the Verlet scheme. Use sufficiently small time steps and a cutoff length
rc = 2.5 for the Lennard–Jones potential.
Task 2: Investigate the influence of the choice of the time step size.
Task 3: Measure the energy of the system as well as its fluctuations. Is the energy conserved?
Visualization of the results: For visualizing the particles you have different options. You can simply use a graph-
ical tool (e.g., gnuplot, Matlab, matplotlib) which reads in the coordinates for a given time step. Another
possibility would be to use either vmd (www.ks.uiuc.edu/Research/vmd/) or paraview (www
.paraview.org/) to visualize the dynamics and create animations.
Optimizing Molecular Dynamics 13
Usually, we are interested in performing MD simulations for a large number of parti-

cles. This makes it necessary tocompute
the forces between any pair of all N particles –
an operation of complexity O N 2 . However, there exist different ways of optimizing
our MD simulations. For example, let us consider a potential that is a function of even
powers of the particle distance r. That is, our potential has the form v(r) ∝ r−2n with
n ≥ 1. If we now determine the force between two particles i and j
fi j = −∇ri−2n 2(n−1)
j ∝ ri j ri j , (13.1)
we realize that it is possible to omit the computation of the square root in

?
@
A d
2
ri j = xik − xkj . (13.2)
k=1
We can simply evaluate

⎡? ⎤2(n−1) ⎡
⎢⎢⎢ @A d
⎥⎥ ⎢⎢⎢d
⎤(n−1)
2 ⎥⎥⎥
⎢⎢ 2⎥
⎢ k ⎥ ⎥
ri2(n−1) = ⎢⎢⎢ x − x j ⎥⎥⎥⎥
k
= ⎢⎢⎣⎢ xi − x j ⎥⎥⎥⎦
k k
. (13.3)
j
⎣ k=1 i ⎦ k=1
If the potential has a complicated form which makes it computationally expensive

to evaluate, we may discretize the potential and store its values in a lookup table.
For short range potentials, we define a cutoff rc and divide the interval (0, rc2 ) into K
pieces with subinterval lengths
j 2
lj = r , (13.4)
K c

where j ∈ {1, . . . , K}. The force values stored in a lookup table are f j = f l j , and
the corresponding index j is given by
⎢⎢⎢ ⎥⎥
⎢⎢⎢
d 2 ⎥⎥⎥⎥
j = ⎢⎢⎣S ⎢ xi − x j ⎥⎥⎥⎦ + 1,
k k
(13.5)
k=1
where · denotes the floor function1 and S = K/rc2 . Interpolation methods are useful
to obtain intermediate values. The definition of a cutoff introduces a discontinuity
1 The floor function x of a real number x corresponds to the greatest integer that is less than or equal to x.
157
158 Optimizing Molecular Dynamics
of the potential at rc and thus an infinite force at rc . To avoid this problem, we can
introduce a modified potential ṽ(r) according to
⎧ -
⎪
⎪ ∂v -
⎨v(r) − v(rc ) − ∂r -r=rc (r − rc ) if r ≤ rc ,
ṽ (r) = ⎪
⎪ (13.6)
⎩0 if r > rc ,
where v(rc ) is the value of the original potential at rc . Subtracting v(rc ) removes the
jump which represents a physically irrelevant shift in the energy scale. The subtrac-
tion of the first derivative makes the potential differentiable at rc but slightly modifies
the physics. For the Lennard–Jones potential, a common cutoff value is rc = 2.5σ.
Potentials that decay slower than r−d are called long-range potentials and for them it
is not possible to introduce any cutoff, because the forces at large distances are not
negligible.
13.1 Verlet Tables
To reduce the amount of computations, we should ignore particle–particle interactions

whenever their force contributions are negligible. Therefore, we only consider particles
in a certain range rl > rc . An illustration of this concept is shown in Figure 13.1. For
every particle, we store the coordinates of the neighboring particles in a list which
is referred to as Verlet table. As the particles move, the table has to be updated
after
r l − rc
n= (13.7)
Δtvmax
time steps with vmax being the largest velocity occurring
in the system. Updating the
whole list is still an operation of complexity O N 2 .
The Verlet table LIST is a one-dimensional vector in which all the particles in the
Figure 13.1 An illustration of the Verlet neighborhood of particle i are sequentially stored for all particles i, starting with i = 1
table method. Only particles with their and ending with i = N. A second vector POINT[i] contains the index of the first particle
center being within a distance of rl > rc in the neighborhood of particle i in LIST. Therefore, the particles in the neighborhood
from particle i are considered. of particle i are: LIST[POINT[i]], . . . ,LIST[POINT[i + 1]-1]. The force is calculated by
going through this list and finding all neighbors of particle i.
13.2 Linked-Cell Method
A more efficient way to identify interactions between neighboring particles in our MD

simulations is the linked-cell method [248, 249] (see Figure 13.2). In this method, the
domain is discretized using a regular grid with a grid spacing of M > rc where rc is the
range of the potential (or the cutoff range). Each particle is located within a certain
13.2 Linked-Cell Method 159
grid cell. Due to our choice of M, we only have to account for particle interactions in
certain cell neighborhoods. Particles located in more distant cells do not contribute to
the force computation. In d dimensions there are 3d neighboring cells of interest. On
average, we thus have to compute the interactions of N3d N/N M particles where N M is
the number of grid cells. To keep track of the locations of all particles, we use linked
lists which were introduced by Donald E. Knuth (see Figure 13.3) [9]. In a vector (or
list) FIRST of length N M , we store the index of a particle located in cell j in FIRST[j]. If
cell j is empty, then FIRST[j]=0. The indices of the remaining particles that are located
in the same cell are stored in a second vector LIST of length N. If particle i is the last
one in a cell, then LIST[i]=0.
In the following code listing, we show an example of how to extract all particles that
are located in cell i = 2.
Figure 13.2 An illustration of the
linked-cell method. A grid with grid
i =2;
spacing M ( r2c < M < rc ) is placed on top
M[ 1 ] = FIRST [ i ] ;
of the system. Only interactions between
j =1;
particles in a certain cell neighborhood
w h i l e ( LIST [M[ j ] ] ! = 0 )
have to be considered.
{
M[ j +1]=LIST [M[ j ] ] ;
j = j +1;
}
When a particle changes its cell, FIRST and LIST can be updated locally. The algo-
rithm is thus of order O (N). In addition, this method is well suited for parallelization
(domain decomposition).
Exercise: Linked-Cell Method

In this exercise, we are going to make use of the linked-cell method in our MD simulations. Note that the
length of the particle lists may alter from one time step to another. If you use C++, you have the choice of
storing lists of pointers to a particle, or lists of particle indices.
Task
1. Implement the linked-cell method in your existing MD program for particles with Lennard–Jones poten-
tials (see previous exercise).
2. How many particles can you simulate?
3. Try different optimizations:
• Work with r2 rather than r (avoiding calculating square roots). Figure 13.3 Donald E. Knuth is a
• Use a look-up table for the force as a function of discrete values of r2 and interpolate for other professor emeritus at Stanford University.
values. He is the inventor of the typesetting
system “Tex” and author of the famous
textbook The Art of Computer
Programming [9].
160 Optimizing Molecular Dynamics
Exercise: Linked-Cell Method (cont.)

• Since the particles are in a cubic box, divide the system into smaller cubic regions of minimum size rc ,
where rc = 2.5σ is the interaction range (cutoff of the Lennard–Jones potential).
• Note that for small systems (a few hundred particles) checking all possible particle pairs can be faster than
using the linked-cell method. For large systems, the saved computation time can be enormous.
Dynamics of Composed Particles 14
Until now, we just considered single-particle interactions with potentials that depend
on the distance between particles. In nature, however, there exist many examples of
systems in which the interactions also depend on size and shape of the considered
particles. Examples of such systems include sand grains and molecules such as the
water molecule that we show in Figure 14.1. To describe the time evolution of such
composed systems, we have to consider their shapes and constituents.
14.1 Lagrange Multipliers
One possibility to model composed particle systems is to introduce additional con-

straint forces [250]. Such constraint forces are used to establish rigid bonds or angles
between individual particles. The idea is to rewrite the equation of motion for each
particle i according to
mi ẍi = fi + gi , (14.1)

interaction forces constraint forces
where the first term accounts for interactions between particles and the second one for
constraint forces. We impose the constraints forces to account for the geometric struc-
ture of molecules (e. g., certain distances d12 and d23 between atoms). We therefore Figure 14.1 A water molecule as a
define a potential such that the constraint forces gi are proportional to the difference composed particle system consisting of
between the actual and the desired distance of the particles. As an example, we con- two hydrogen atoms and one oxygen
sider a water molecule that consists of three particles (see Figure 14.1). We assume atom.
that distance and angles between atoms are constant as it is the case in most practical
applications. The differences between actual and desired distances
χ12 = r12
2
− d12
2
and χ23 = r23
2
− d23
2
(14.2)
are zero if the particles have the desired distance, with ri j = !ri j ! and ri j = xi − x j . We
define the force
λ12 λ23
gk = ∇xk χ12 + ∇xk χ23 , (14.3)
2 2
for k ∈ {1, 2, 3}, and λ12 and λ23 are the yet undetermined Lagrange multipliers.
We determine these multipliers by imposing the constraints. Inserting eq. (14.2) into
eq. (14.3) yields
g1 = λ12 r12 , g2 = λ23 r23 − λ12 r12 , g3 = −λ23 r23 . (14.4)

161
162 Dynamics of Composed Particles
These equations describe nothing but linear springs with yet to be determined spring
constants λ(·) . To obtain the values of the Lagrange multipliers λ(·) , the Verlet algo-
rithm is executed in two steps. We first compute the Verlet update without constraint
to obtain
fi
x̃i (t + Δt) = 2xi − xi (t − Δt) + Δt2 . (14.5)
mi
Next, we correct the value using the constraints according to
gi
xi (t + Δt) = x̃i (t + Δt) + Δt2 . (14.6)
mi
Combining Eqs. (14.6) and (14.4), the updated positions are
λ12
x1 (t + Δt) = x̃1 (t + Δt) + Δt2 r12 (t) , (14.7)
m1
λ23 λ12
x2 (t + Δt) = x̃2 (t + Δt) + Δt2 r23 (t) − Δt2 r12 (t) , (14.8)
m2 m2
λ23
x3 (t + Δt) = x̃3 (t + Δt) − Δt2 r23 (t) . (14.9)
m3
With these expressions, we now obtain λ12 and λ23 by inserting Eqs. (14.7)–(14.9) into
the constraint conditions:
|x1 (t + Δt) − x2 (t + Δt)|2 = d12
2
,
(14.10)
|x2 (t + Δt) − x3 (t + Δt)|2 = d23
2
,
which gives
-- --2
-- 1 1 2 λ23 --
-- 12r̃ (t + Δt) + Δt 2
λ 12 + r12 (t) − Δt r23 (t) -- = d12 ,
2
m1 m2 m2
-- --2 (14.11)
-- 1 1 2 λ12 -
--r̃23 (t + Δt) + Δt λ23
2
+ r23 (t) − Δt r12 (t)-- = d23 2
,
m2 m3 m2 -
where r̃i j = x̃i − x̃ j . These coupled quadratic equations are solved for λ12 and λ23
to compute the next position xi (t + Δt). Depending on the precision needed, we may
ignore the higher-order terms of Δt and just solve two coupled quadratic equations.
Exercise: Composed Particle Dynamics

Consider a system of N diatomic molecules (in a box of length L with periodic boundary conditions) that
interact with each other. For simplicity, we assume that all atoms are identical. Atoms of different molecules
interact via the Lennard–Jones potential
⎡ 12 6 ⎤
⎢⎢ 1 1 ⎥⎥⎥⎥
V(r) = 4 ⎢⎢⎣⎢ − ⎥.
r r ⎦
Exercise: Composed Particle Dynamics (cont.)

To ensure that the two particles in a molecule keep their distance r, we have to add the constraint
forces
g1 = λr and g2 = −λr
to the equation(s) of motion.
Task 1: Derive an analytical expression for the Lagrange multiplier λ.
Task 2: Simulate the system using the Verlet scheme.
• The distance between the two atoms in a molecule should be small enough such that other molecules do
not cross their bond.
• It might be easier to consider the 2D case and check if the behavior of the molecules is correct before you
go to the 3D case.
Task 3: Extend your program to simulate three-component molecules.
14.2 Rigid Bodies
Systems whose n constituents of mass mi are located at fixed positions xi with

respect to each other are referred to as rigid bodies. The motion of such objects is
described by translations of the center of mass and rotations around it. The center of
mass is
1
n n
xcm = xi mi with M= mi . (14.12)
M i=1 i=1
The equation of motion of the center of mass and the corresponding torque are
given by

n
n
M ẍcm = fi = fcm and T = di ∧ fi , (14.13)
i=1 i=1
where di = xi − xcm . In two dimensions, the rotation axis always points in the direction
of the normal vector on the plane. Therefore, there only exist three degrees of freedom:
two translational and one rotational. In three dimensions, there are six degrees of
freedom: three translational and three rotational.
We first discuss the treatment of rigid bodies in two dimensions and then generalize
it to the three-dimensional case.
14.2.1 Two Dimensions
In two dimensions, the moment of inertia and torque are

I= r ρ (r) dA and T =
2
r ft (r) dA , (14.14)
A A
where ρ(r) is the mass density, A the area, and ft (r) the tangential force (see
Figure 14.2). In general, the mass density may be constant or depend on the actual
position. The equation of motion is
I ω̇ = T , (14.15)
where ω is the angular velocity. We now apply the Verlet algorithm to the position x
Figure 14.2 An example of a rigid body and rotation angle φ and compute the corresponding time evolutions according to
in two dimensions. The black dot shows T (t)
the center of mass (CM), and ft (r) φ (t + Δt) = 2φ (t) − φ (t − Δt) + Δt2 ,
I
represents the tangential force (14.16)
x (t + Δt) = 2x (t) − x (t − Δt) + Δt2 M −1 f j (t) ,
component at position r. j∈A
where the total torque T is the sum over all the torques acting on the rigid body.
That is,
0 1
T (t) = f jy (t) d xj (t) − f jx (t) dyj (t) . (14.17)
j∈A
14.2.2 Three Dimensions

In three dimensions, the moment of inertia is a tensor that needs to be diagonal-
ized in order to solve the equation of motion. Only in the body-fixed coordinate
system a systematic diagonalization of this tensor can be achieved. To describe the
motion of rigid bodies in three dimensions, we consider a lab-fixed and a body-fixed
coordinate system with coordinates x and y, respectively. We assume that the point
y = 0 corresponds to x = 0 in the lab-fixed system. The transformation between both
systems is
x = R(t)y , (14.18)
where R(t) ∈ SO(3) denotes a rotation matrix. Furthermore, we define Ω = R Ṙ and
1 T
find with RT R = 1 that

RT Ṙ + ṘT R = Ω + ΩT = 0 . (14.19)
This equation implies that Ω is skew-symmetric and thus of the form
⎛ ⎞
⎜⎜⎜ 0 −ω ω ⎟⎟⎟
⎜⎜⎜⎜ ⎟⎟⎟⎟
3 2
Ω = ⎜⎜⎜⎜ ω3 0 −ω1 ⎟⎟⎟⎟ and Ωy = ω ∧ y , (14.20)

⎜⎜⎝ ⎟⎟⎠
−ω2 ω1 0
1 SO(3) is the so-called three-dimensional rotation group, or special orthogonal group. All rotation matrices
R ∈ SO(3) satisfy RT R = RRT = 1.
where ω = (ω1 , ω2 , ω3 ) is the angular velocity in the body-fixed frame and ∧ denotes
the cross product. Therefore, the angular momentum is

n
n
L= mi xi ∧ ẋi = mi Ryi ∧ Ṙyi . (14.21)
i=1 i=1
Combining Eqs. (14.20) and (14.21) yields

n
n
2 3
L=R mi yi ∧ (ω ∧ yi ) = R mi ω (yi · yi ) − yi (ω · yi ) . (14.22)
i=1 i=1
The components of the inertia tensor are

n 0 1
I jk = mi (yi · yi ) δ jk − yij yki (14.23)
i=1
and thus

3
L = RS with S j = I jk ωk . (14.24)
k=1
Considering a coordinate system whose axes are parallel to the principal axes of the
moment of inertia of the body, the inertia tensor takes the form
⎛ ⎞
⎜⎜⎜I1 0 0 ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟
I = ⎜⎜⎜⎜⎜ 0 I2 0 ⎟⎟⎟⎟⎟ and S j = I j ω j . (14.25)
⎜⎜⎝ ⎟⎟⎠
0 0 I3
We use eq. (14.24) to obtain the equations of motion
L̇ = ṘS + RṠ = =
T, (14.26)
where T= = RT represents the torque in the lab-fixed coordinate system. Multiplying
the prior equation with RT , we find Euler’s equations in the principal axes coordinate
system

T1 I2 − I 3
ω̇1 = + ω 2 ω3 ,
I1 I1

T2 I3 − I 1
ω̇2 = + ω 3 ω1 , (14.27)
I2 I2

T3 I1 − I 2
ω̇3 = + ω 1 ω2 .
I3 I3
We then integrate the angular velocities according to

T 1 (t) I2 − I3
ω1 (t + Δt) = ω1 (t) + Δt + Δt ω 2 ω3 ,
I1 I1

T 2 (t) I3 − I1
ω2 (t + Δt) = ω2 (t) + Δt + Δt ω 3 ω1 , (14.28)
I2 I2

T 3 (t) I1 − I2
ω3 (t + Δt) = ω3 (t) + Δt + Δt ω 1 ω2 .
I3 I3
From these expressions, we obtain the angular velocity in the laboratory frame
= (t + Δt) = Rω (t + Δt) .
ω (14.29)
Since the particles are moving all the time, the rotation matrix is not constant. We
have to find an efficient way to determine and update R at every step in our simulation.
In the following section, we therefore discuss three-dimensional rotations introducing
Euler angles and quaternions.
14.2.3 Euler Angles
One possible parameterization of the rotation matrix R is given by the Euler angles
Figure 14.3 To transform the lab-fixed (φ, θ, ψ). These angles correspond to three rotations that transform the lab-fixed system
x-system into the body-fixed y-system, into a body-fixed system (see Figure 14.3). First, we rotate the lab-fixed system at an
we first rotate the lab-fixed system angle φ around the x3 -axis and refer to the rotated x1 -axis as K-axis. In the second
around the x 3 -axis at an angle φ. Next, step, we rotate the system around the K-axis at an angle θ and finally apply a rotation
we rotate the system around the K-axis at around the y3 -axis at an angle ψ. Mathematically, this means
an angle θ and finally apply a rotation
around the y3 -axis at an angle ψ. R = R(φ, θ, ψ)
⎛ ⎞⎛ ⎞⎛ ⎞
⎜⎜⎜cos φ − sin φ 0⎟⎟⎟⎟ ⎜⎜⎜1 0 0 ⎟⎟⎟ ⎜⎜⎜cos ψ − sin ψ 0⎟⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ (14.30)
= ⎜⎜⎜⎜⎜ sin φ cos φ 0⎟⎟⎟⎟ ⎜⎜⎜⎜⎜0 cos θ
⎟
− sin θ⎟⎟⎟⎟ ⎜⎜⎜⎜⎜ sin ψ cos ψ 0⎟⎟⎟⎟ .
⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠
0 0 1 0 sin θ cos θ 0 0 1
As a consequence of the occurrence of products of multiple trigonometric functions,

this parameterization is not well suited for efficient computations. We have to keep in
mind that this operation has to be performed for every particle and every time step,
leading to a high computational effort. For the computation of angular velocities, also
derivatives of eq. (14.30) would have to be considered. Specifically, we would have to
compute
sin φ cos θ cos φ cos θ
φ̇ = −ω̃ x + ω̃y + ω̃z ,
sin θ sin θ
θ̇ = ω̃ x cos θ + ω̃y sin φ , (14.31)
sin φ cos φ
ψ̇ = ω̃ x − ω̃y ,
sin θ sin θ
which, for some angles, θ (e. g., θ = 0) gives diverging terms.
14.2.4 Quaternions
Figure 14.4 Denis J. Evans is a professor
at the Australian National University in Denis J. Evans (see Figure 14.4) came up with a trick to optimize the computation of
Canberra. rotational velocities [251, 252].
This trick is based on quaternions, which are a generalization of complex numbers

with four basis vectors that span a four-dimensional space. The sum of the squares of
these basis vectors is equal to unity. We use the quaternion parameters
θ φ + ψ
q0 = cos cos , (14.32)
2 2
θ φ − ψ
q1 = sin cos , (14.33)
2 2
θ φ − ψ
q2 = sin sin , (14.34)
2 2
θ φ + ψ
q3 = cos sin , (14.35)
2 2

with 0 < qi < 1 and i q2i = 1 for i ∈ {1, . . . , 4}, to represent the Euler angles
(φ, θ, ψ). Since the Euclidean norm of q = (q1 , q2 , q3 , q4 )T equals unity, there exist only
three independent parameters. The rotation matrix, as defined in eq. (14.30), has a
quaternion representation
⎛ ⎞
⎜⎜⎜q2 + q2 − q2 − q2 2 (q1 q2 + q0 q3 ) 2 (q1 q3 − q0 q2 ) ⎟⎟⎟⎟
⎜⎜⎜ 0 1 2 3
⎟⎟
R = ⎜⎜⎜⎜⎜ 2 (q1 q2 − q0 q3 ) q20 − q21 + q22 − q23 2 (q2 q3 + q0 q1 ) ⎟⎟⎟⎟⎟ . (14.36)
⎜⎜⎝ 2⎠
⎟⎟
2 (q1 q3 + q0 q2 ) 2 (q2 q3 + q0 q1 ) q0 − q1 − q2 + q3
2 2 2
This offers a more efficient way to compute rotations without the necessity of com-
puting products of trigonometric functions. The angular velocities can be computed
according to
⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜⎜⎜q̇0 ⎟⎟⎟ ⎜⎜⎜q0 −q1 −q2 −q3 ⎟⎟⎟ ⎜⎜⎜ 0 ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟ ⎜⎜⎜ ⎟⎟⎟
⎜⎜⎜⎜ 1 ⎟⎟⎟⎟ = 1 ⎜⎜⎜⎜ 1
q̇ q q0 −q3 q2 ⎟ ⎟⎟⎟ ⎜⎜⎜ω x ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ 2 ⎜⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ . (14.37)
⎜⎜⎜q̇2 ⎟⎟⎟ ⎜⎜⎜q2 q3 q0 −q1 ⎟⎟⎟⎟⎟ ⎜⎜⎜⎜⎜ωy ⎟⎟⎟⎟⎟
⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠
q̇3 q3 −q2 q1 q0 ωz
Since quaternion and Eulerian angle representations are connected by a diffeomor-

phism, there is always the possibility of calculating the values of the Euler angles if
needed:
⎡ ⎤
⎢⎢⎢ 2 (q0 q1 + q2 q3 ) ⎥⎥⎥
φ = arctan ⎢⎢⎣⎢ ⎥⎥⎦⎥ , (14.38) Figure 14.5 Rotating spheres in a sphere
1 − 2 q21 + q22
assembly as an example of rigid body
2 3
θ = arcsin 2 (q0 q2 − q1 q3 ) , (14.39) dynamics [253].
⎡ ⎤
⎢⎢⎢ 2 (q0 q3 + q1 q2 ) ⎥⎥⎥
ψ = arctan ⎢⎢⎣⎢ ⎥⎥⎦⎥ . (14.40)
1 − 2 q22 + q23
There is no need of calculating the Euler angles except if one wants to render the
rotating object in space graphically. With the help of quaternions, we can now simulate
rigid body dynamics in the following way:
Rigid Body Simulation with Quaternions
• = and the rotation matrix R (see eq. (14.36)).

Determine the torque in the lab frame T
• Compute the torque in the body frame, T = RT T, =
• Obtain ω (t + Δt) according to eq. (14.28),
• Update the rotation matrix as defined in eq. (14.36) by computing q(t +Δt) according to eq. (14.37).
In Figure 14.5, we show an example of a system of rotating particles whose

dynamics has been simulated using quaternions.
Long-Range Potentials 15
In Chapter 13, we discussed different optimization techniques to speed up our MD

simulations. If the potential decays sufficiently fast, it is possible to define a cutoff
to neglect vanishingly small force contributions at large distances. However, if the
potential of a system in d dimensions decays as V(r) ∝ r−d or even slower, it is not
possible to define a cutoff anymore. The reason is that there is a nonnegligible energy
contribution at any distance. Examples of such potentials occur in electrostatic, gravi-
tational, and dipolar interactions. To simulate such systems, we discuss the following
methods:
• Ewald summation,
• particle-mesh methods (PM, PPPM & APPPM),
• reaction field method.
15.1 Ewald Summation
Paul Peter Ewald (see Figure 15.1) developed a method to compute long-range inter-
actions in periodic systems. It has been originally developed to study ionic crystals and
our goal is to apply this method to MD simulations.
Until now, we used periodic boundary conditions for systems of finite size. This was
only possible, because the interaction correlations were decaying with increasing dis-
tance. For long-range potentials, the particles are no longer uncorrelated since they
are able to interact even at large distances. What we can do is periodically repeat-
ing our system by attaching its own image at its borders. This leads to a larger
system size and allows us to model farther-reaching interactions. We illustrate this
procedure in Figure 15.2. We treat the aforementioned repeated systems as really
existing, and compute interactions between particles in our “box” of interest with Figure 15.1 Paul Peter Ewald
all remaining particles in the system. Every copy of the system is characterized by (1888–1985) was a professor and rector
a vector that ranges from the origin of the central system of interest to the origin at the University of Stuttgart, Germany.
of the copy. The resulting potential is thus the sum over particle interactions that He was an important crystallographer
repeat over several boxes. For gravitational or electrostatic V(r) ∝ r−1 potentials, and vehemently opposed the Nazis. For
we obtain further reading, see Ref. [254].
1 - -−1
V= zi z j --ri j + n-- , (15.1)
2 n i, j
169
170 Long-Range Potentials
t
Figure 15.2 An example of an Ewald summation procedure. From left to right, images of the system are periodically attached at
the boundaries.
t
Figure 15.3 The Ewald sum is decomposed into one sum in real and another sum in reciprocal space. The charges are screened
with Gaussian charge distributions. Other distributions are also possible.
where zi and z j represent the corresponding charges or masses. For the sake of brevity,
we are using units, in which 4π0 or 4πG are unity in eq. (15.1). The sum over

n = n x L, ny L, nz L (15.2)
denotes the summation over all copies of the system of linear size L and integers n x ,
ny , and nz . The prime indicates that we are excluding the potential-energy contribution
for n = 0 when i = j. The first sum comprises the central box at n = 0 and the following
terms are those for which |n| = L. That is, n = (±L, 0, 0), n = (0, ±L, 0), and n = (0, 0, ±L).
The Ewald sum is only conditionally convergent (i. e., the convergence depends on the
order of the summation) and the convergence is very slow. Since Ewald never used a
computer, he intended to sum over an infinite number of cells. In practice, we have to
truncate the sum at some point and try to estimate the error. Because of the numerical
difficulties of the algorithm, it is only used for systems consisting of a few particles.
Several approaches are possible to improve the convergence [255, 256].
Additional Information: Screening with Gaussian Charge Distributions

One can reduce the long-range effect of point-like charges by screening them with Gaussian charge distribu-
tions of opposite sign
zi κ3
ρi (r) = 3 exp −κ2 r2 (15.3)
π2
with some appropriate inverse width κ. After having introduced the screening charges, we have to cancel
them out again using a charge distribution like the one of eq. (15.3) but of opposite sign. This is done in
Fourier space. The basic idea behind this summation is illustrated in Figure 15.3. The result is one sum in real
and another sum in reciprocal space which converges faster [257]
⎡ - -
1 ⎢⎢⎢⎢ erfc κ --ri j + n--
V= ⎢⎢ zi z j -- -
2 i, j ⎣ n -ri j + n--
2 ⎤
1 4π2 ⎥⎥
zi z j 2 exp − 2 cos kri j ⎥⎥⎦⎥
k
+ (15.4)
πL3 k0 k 4κ
-- --2
κ 2 2π -- --
− √ zi + 3 -- zi ri --
π i 3L - i -
with the complementary error function
∞
2
erfc (x) = √ exp −t2 dt.
π x
These formulas are rather clumsy and only applicable to small systems and we therefore focus on other
approaches in the next sections.
15.2 Particle-Mesh Method
Another possibility of simulating long-range particle interactions is the so-called

particle-mesh (PM) method. This method was invented by Eastwood and Hockney,
and is very well suited for reasonably homogeneous distributions of particles [249,
258]. It is based on the following steps.
The mesh concept is similar to the linked-cell method (see Figure 13.2). Criteria for
a particle-mesh scheme are as follows:
• Errors should vanish for large distances between particles.
• Momentum should always be conserved (Fi j = −F ji ).
• Charges or masses on the mesh and the interpolated forces should vary smoothly.
This method is not very efficient and time consuming for
• inhomogeneous distributions of charges or masses,
• strong correlations, like bound states in molecules,
Particle-Mesh Method
• Put a fine mesh on top of the system so that only a few particles are in each cell.
• Distribute the charges onto the mesh points.
• Calculate the electrostatic potential by solving the Poisson equation on the mesh using FFT.
• Calculate the force on each particle by numerically differentiating the potential and interpolating back from
the mesh to the particle position.
• deviation from the point-like object approximation,

• complex geometries of the system.
In those cases, we may use the particle-particle-particle-mesh (P3 M) or adaptive
particle-particle-particle-mesh (AP3 M) algorithms. These methods are presented later
in the subsequent sections. The PM method is well suited for charged gases, because
such systems are characterized by overall homogeneous particle densities.
15.2.1 Implementation
Once we implemented the grid for our d-dimensional system, there exists more than
one possibility to assign charge (or mass) values to vertices. Two common options are
as follows:
• nearest grid point: assign the charge to nearest grid point.
• cloud-in-cell: assign charge proportionally to the 2d nearest grid points.
In two dimensions, a cloud in cell implementation of a charge q located at (x, y) is given
by the charge distribution
q
ρi j = (xi+1 − x)(yi+1 − y) , (15.5)
2Δl2
q
ρi+1 j = (x − xi )(yi+1 − y) , (15.6)
2Δl2
q
ρi j+1 = (xi+1 − x)(y − yi ) , (15.7)
2Δl2
q
ρi+1 j+1 = (x − xi )(y − yi ) , (15.8)
2Δl2
where Δl is the grid spacing. The potential at a certain point r can be computed by the
convolution of ρ (r) with the Green’s function:

: ; : ;
φ (r) = ρ r G r, r dr . (15.9)
For electrostatic and gravitational interactions, one has to solve the Poisson equa-
tion for which the corresponding Green’s function is G (r, r ) ∝ |r − r |−1 . In Fourier
space, eq. (15.9) corresponds to a multiplication with the Fourier transform G B (k) ∝ k−2
B
such that φ̂ (k) = ρ̂ (k) G (k) (convolution theorem). After applying the inverse Fourier
Additional Information: Fourier Transform

∞
Let f (t) be a square-integrable function (i.e., −∞ | f (t)| dt < ∞), then the Fourier transform is
∞
fˆ(ω) = f (t)e−iωt dt, (15.10)
−∞
and the inverse Fourier transform is

∞
1
f (t) = fˆ(ω)eiωt dω . (15.11)
2π −∞
transform, one obtains φ. If one performs the Fourier transform using fast Fourier
: ;
transform (FFT) [259, 260], the method has computational complexity O N log N ,
where N is the number of mesh points. See the gray box on page 174 and Ref. [261] for
more details on the Fourier transform.
Knowing the potential, we compute the force

F ri j = −∇φ ri j (15.12)
at every vertex of the mesh and then interpolate the forces on the corners of the mesh
to find the force that is exerted by the field on a particle inside the cell. Note that
the PM algorithm is not very well suited for inhomogeneous distributions of parti-
cles, strong correlations like bound states, and complex geometries. In these cases,
the P3 M and AP3 M methods, tree codes, and multiple expansions constitute possi-
ble extensions of the method. One important application is the very heterogeneous
distribution of stars and galaxies in astrophysics. For further reading, we recommend
Ref. [262].
15.2.2 P3 M (Particle-Particle-Particle-Mesh)
For particles with close neighbors (e. g., binary stars), we decompose the force acting
on it into two terms:
F = F s + Fl , (15.17)
where F s accounts for short-range interactions and Fl describes interactions over large
distances. The short-range interactions are simulated by solving Newton’s equation of
motion whereas the PM method is applied to the long-range forces. This approach
has the drawback that the long-range potential field has to be computed for each close
particle and we still have not solved the problem of strong heterogeneity.
15.2.3 AP3 M (Adaptive Particle-Particle-Particle-Mesh)

For simulations of the mass distribution in the Universe, most of the grid points
will be empty, making the previous algorithms very inefficient. A possibility to
Additional Information: Discrete and Fast Fourier Transform

Computing the Fourier transform of discrete data points on a computer is done in terms of the discrete Fourier
transform (DFT). To define the DFT of a sequence of N data points vk , we first construct an appropriate basis
using the Nth roots of unity

2πi
ωN exp − . (15.13)
N
Then, the DFT v̂ of a vector v ∈ CN is defined as
1 kj
N−1
v̂ j = √ ωN vk . (15.14)
N k=0
The computation of v̂ j requires O(N 2 ) operations. Fast Fourier transform methods such as the algorithm of
Cooley and Tukey [259] achieve performances of O(N log N) by computing the DFT in terms of smaller even
and odd-indexed parts of vk :
⎛N/2−1 ⎞
1 ⎜⎜⎜⎜ 2k j ⎟⎟⎟
N/2−1
v̂ j = √ ⎜⎜⎝ ωN v2k + ω(2k+1) j
v ⎟⎟⎟ . (15.15)
N k=0
N 2k+1 ⎠
k=0
Defining an even-indexed term

1 2k j
N/2−1
ej = √ ωN v2k
N k=0
and an odd-indexed term
1
N/2−1
o j = √ ωNj ω2k j
N v2k+1 ,
N k=0
we can rewrite eq. (15.15) in terms of

v̂ j = e j + ωNj o j . (15.16)
Similarly, we can write v̂ j+N/2 = e j − ωNj o j . We can further decompose e j and o j in even- and odd-
indexed parts and so forth. This recursive way of computing v̂ j is the basis of the Cooley and Tukey FFT
algorithm.
deal with strong heterogeneities is to adapt the mesh to the particle density (i.e.,
using an adaptive mesh with higher resolution where the particle density is higher).
One implementation is to refine a cell as long as the number of particles in the cell
exceeds a certain threshold. Tree codes can be used to locate the cells that emerge in
the refinement process [262]. The advances in the simulation of many-body systems
often depend on the development of such bookkeeping techniques. Current simula-
tion frameworks are able to simulate the interaction of up to 1012 particles [263].
For further reading, we recommend the textbook Computer Simulations Using
Particles [249].
15.3 Reaction Field Method 175
Exercise: Simulating Gravitational Interactions with the Particle-Mesh Method

Start from a set of massive point-like particles randomly distributed in space. Next, generate a grid with N =
L × L grid cells. You can, for instance, start with N = 128 × 128 grid cells.
For each grid cell, aggregate the total mass density according to eq. (15.8), solve the Poisson equation for
Newtonian gravity
∇2 φ = 4πGρ , (15.18)
where φ is the gravitational potential, G is the gravitational constant, and ρ is the local mass density. Use the
gravitational potential φ to compute the resulting gravitational forces according to eq. (15.12). Note that one
possible discretization of eq. (15.12) is (Δx = Δy = 1)
⎛ ⎞
⎜⎜⎜φ(ri+1 j ) − φ(ri j )⎟⎟⎟
F ri j = − ⎜⎜⎜⎝ ⎜ ⎟⎟⎟ . (15.19)
⎟
φ(ri j+1 ) − φ(ri j )⎠
Figure 15.4 Lars Onsager (1903–1976)
After determining the forces, advance the particle positions in time and determine the new particle densities
was a professor for theoretical chemistry
in each grid cell.
at Yale University. He received the Nobel
Prize in Chemistry in 1968.
15.3 Reaction Field Method
For more complex interactions such as composed molecules or non-point-like par-

ticles, a good solution is to ignore the complexity of distant particles and only take
into account their mean effect while explicitly calculating the interaction with close
particles. The concept has its root in the work of Onsager (see Figure 15.4) on the
dielectric constant but it was introduced as an explicit computational technique in
the 1970s [264–266]. This method is mostly used for the simulation of dipole–dipole
interactions.
To outline this method, we consider a sphere Ni of radius rc around particle i
(see Figure 15.5). The region outside the sphere is described by an effective dielec-
tric medium and the dipole moments μ j of the Ni particles j within the sphere induce
an electric field (reaction field) in the external dielectric medium
Figure 15.5 In the reaction-field method,
2( s − 1) 1 particle i is surrounded by a spherical
Ei = μ , (15.20)
2 s + 1 rc3 j∈N j region of radius rc . The interactions
i
between all particles within the sphere
where s is the effective permittivity of the dielectric continuum outside the sphere. Ni (blue disks) are treated as dipole
The resulting total force exerted on particle i is then interactions with dipole moments μ j

Fi = Fi j + μi ∧ Ei . (15.21) ( j ∈ Ni ). The region outside the sphere
j∈Ni is described by an effective dielectric
constant s .
As the particles are moving, the number of particles inside the cavity changes. This
causes force discontinuities because of the sudden addition or removal of a particle.
To avoid this effect, one can weight interactions between particles i and j in eq. (15.21)
according to
⎧
⎪
⎪
⎪ 1, for r j < rt ,
⎪
⎪
⎨ rc −r j
g(r j ) = ⎪
⎪ rc −rt , for rt ≤ r j ≤ rc ,
(15.22)
⎪
⎪
⎪
⎩0 , for rc < r j ,
where rt = 0.9rc .
Canonical Ensemble 16
After having discussed various techniques to simulate the interaction of particles, we

now focus again on their statistical description. Often we are not interested in the
properties of single particles but rather in macroscopic quantities such as temper-
ature, pressure, and density. Up until now, we have not introduced any method to
control such macroscopic properties in our MD simulations. We only imposed energy
conservation and were thus simulating the microcanonical ensemble. Experiments
are, however, usually conducted at constant temperature and not at constant energy
because systems are usually able to exchange energy with their environment. In this
chapter, we will therefore discuss some important methods such as the Nosé–Hoover
thermostat and the Parrinello–Rahman barostat to control temperature and pressure
in our MD simulations.
To be able to control temperature, we must first couple our system to a heat bath.
There are various options to do this [267]:
• rescaling of velocities,
• introducing constraints (Hoover),
• Nosé–Hoover thermostat,
• stochastic method (Anderson).
However, before focusing on a more detailed discussion of these methods, we will

precisely define the concept of temperature used in the subsequent sections. We start
from the equipartition theorem [268]
. / . /
∂H ∂H
qμ = pμ = δμν kB T, (16.1)
∂qν ∂pν
for a Hamiltonian H with the generalized coordinates p and q. We consider a classical

system whose Hamiltonian is given by
N
p2i
H= + V (x1 , . . . , xN ) , (16.2)
i=1
2mi
where pi = mi xi . We define the instantaneous temperature [257]
2 N
p2i
T = , (16.3)
3kB (N − 1) i=1 2mi
177
178 Canonical Ensemble
which, after inserting eq. (16.2) into eq. (16.1), is, when averaged, equal
to T .
16.1 Velocity Rescaling
We now discuss a straightforward yet physically incorrect method to simulate a canon-

ical ensemble of particles. Equation (16.3) shows how the temperature depends on
momentum (i. e., on velocity). Intuitively, we should be able to adjust the instanta-
neous temperature of the system by rescaling the particle velocity. The algorithm is as
follows:
Velocity Rescaling
Rescale the velocities of all particles according to
vi → αvi . (16.4)
The measured temperature is proportional to the squared velocities, so
T → α2 T . (16.5)
Therefore, we have to set
*
T
α= , (16.6)
T
to simulate at a certain desired temperature T .
This method is very easy to implement. However, the problem is that the result-
ing velocity distribution deviates from the Maxwell–Boltzmann distribution that
one expects in the canonical ensemble. A modification of this method makes
use of an additional parameter τT , which describes the coupling to a heat
bath.
Berendsen Thermostat
For the so-called Berendsen thermostat, the velocity-scaling factor is [269]
*
Δt T
α= 1+ −1 . (16.7)
τT T
Still, we do not recover the canonical velocity distribution. Velocity rescaling should
only be applied to initialize an MD simulation at a given temperature.
16.2 Constraint Method
Another possibility to keep the temperature of the system constant is to add a friction
term to the equations of motion:
ṗi = fi − ξpi . (16.8)
Various definitions of the friction coefficient ξ are possible. The original proposal
of Hoover (see Figure 16.1) is based on the following constant temperature condi-
tion [271, 272]:
⎛ N ⎞
d ⎜⎜⎜⎜ 2 ⎟⎟⎟⎟
N
Ṫ ∝ ⎜⎝ pi ⎟⎠ ∝ ṗi pi = 0 . (16.9)
dt i=1 i=1
Combining Eqs. (16.8) and (16.9) yields Figure 16.1 William G. Hoover is a
N computational physicist who worked at
fi pi
ξ = Ni=1 . (16.10) the University of California’s Livermore
i=1 |pi |
2
Laboratory. The photograph is taken from
This method makes it necessary to already start at the desired temperature. Instead Ref. [270].
of eq. (16.10), another possibility is to determine the friction coefficient according to
([269, 273], see Figure 16.2)
T
ξ =γ 1− , (16.11)
T
or [273]
f kB
ξ̇ = (T − T ) . (16.12)
Q
The parameters γ and Q determine how fast the temperature attains the desired one,
and f is the number of degrees of freedom. Still, all these methods have the drawback Figure 16.2 Herman Berendsen
that they either do not recover the Maxwell–Boltzmann velocity distribution or not (1934–2019) was a professor of physical
satisfy the H-theorem. In addition, the approaches of Eqs. (16.11) and (16.12) are not chemistry at the University of Groningen.
time reversible. He was one of the pioneers of MD
simulations of molecules [250].
Photograph courtesy the University of
16.3 Nosé–Hoover Thermostat Groningen.
In order to overcome the problem of the wrong velocity distribution, we are now going
to discuss the Nosé–Hoover thermostat as the correct method to simulate heat bath
particle dynamics. Shuichi Nosé (see Figure 16.3) introduced a new degree of freedom
s that describes the heat bath [273–275]. The potential and kinetic energies associated
with s are
V (s) = (3N + 1) kB T ln s ,
1 (16.13)
K (s) = Q ṡ2 ,
2
where the parameter Q is called thermal inertia. The new degree of freedom s is used
to apply a Sundman time-transformation and a change of variables [276]:
dt = s dt and p i = s pi . (16.14)
The resulting rescaled velocities are

dxi dxi dt vi
v i =
=
= . (16.15)
dt dt dt s
The Hamiltonian is thus
N
p i 2 ps 2
H= + + V (x1 , . . . , xN ) + V (s) , (16.16)
i=1
2mi s2 2Q
Figure 16.3 Shuichi Nosé (1951–2005)
was a professor of theoretical physics at where we have set p s = Q ṡ. Applying the Hamilton equations (12.9), the velocities are
the Keio University, Tokyo. dxi p i
= ∇ p H = ,
dt i
mi s2
(16.17)
ds ∂H ps
= = .
dt ∂p s Q
With p i = mi s2 ẋi (in virtual time t ) we find
dp i
fi = = −∇xi H = −∇xi V (x1 , . . . , xN ) = 2mi s ṡẋi + mi s2 ẍi , (16.18)
dt
and
⎡ N ⎤
dp s ∂H 1 ⎢⎢⎢⎢ pi 2 ⎥⎥⎥
⎥⎥⎦ .
= − = ⎢
⎣ − (3N + 1) k B T (16.19)
dt ∂s s i=1 mi s2
Based on these equations, we find for the equations of motion in rescaled time t (see
eq. (16.14))
mi s2 ẍi = fi − 2mi ṡsẋi with i ∈ {1, . . . , N} , (16.20)
and

N
1
Qs̈ = mi sẋ2i − (3N + 1) kB T . (16.21)
i=1
s
Note that Eqs. (16.20) and (16.21) are coupled. This reflects the fact that the two
systems are not isolated and exchange energy in form of heat. In order to obtain the
equations of motion in real time, we have to remind ourselves that dt = dt /s and
p i = spi . Thus, we find for the velocities
dxi dxi p pi
=s = i = ,
dt dt mi s mi
(16.22)
ds ds ps
=s =s ,
dt dt Q
and for the forces,

dpi d p i dp 1 ds 1 ds
=s = i −
pi = fi − pi ,
dt dt s dt s dt s dt
(16.23)
dp s p2i
N
dp s
=s = − (3N + 1) kB T .
dt dt i=1
mi
Nosé–Hoover Equations of Motion
We use ξ = d ln(s)
dt
= ṡs to denote a friction term and rewrite the equations of motion (16.20) and (16.21)
in real time:
fi
ẍi = − ξẋi , (16.24)
mi
and
1
N
1 1
Qξ̇ = mi ẋ2i − (3N + 1) kB T . (16.25)
2 2 i=1 2
The first term in eq. (16.25) represents the measured kinetic energy and the second
term corresponds to the desired kinetic energy. The thermal inertia Q describes the
coupling to the heat bath. The smaller its value, the stronger the system reacts to
temperature fluctuations. For Q → ∞, we recover microcanonical MD, and for too
small values of Q we find spurious temperature oscillations. We show such oscillations
in Figure 16.4. A reasonable value of Q is characterized by temperature fluctuations
that are normally distributed:
*
2
ΔT = T, (16.26)
Nd
where d is the system’s dimension and N is the number of particles.
Temperature fluctuations of the Nosé–Hoover thermostat set at T = 1 for 64 particles and different values of the Figure 16.4
thermal inertia Q. Particle interactions are described by a Lennard–Jones potential.
Additional Information: Nosé–Hoover Particles Follow a Canonical Velocity Distribution

We now show that the Nosé–Hoover thermostat recovers the canonical partition function. Therefore, we start
from microcanonical MD of the closed system of particles and heat bath, and the corresponding partition
function
Z= δ (H − E) ds dp s d3N x d3N p , (16.27)
where integration with respect to x and p has to be taken over a three-dimensional space for N particles.
With H = H1 + (3N + 1) kB T ln (s) in real time, we find

2 3
Z= δ (H1 − E) + (3N + 1) kB T ln (s) s3N ds dp s d3N x d3N p
(16.28)
H1 −E
− (3N+1)k s3N+1
= δ s−e B T 3N
ds dp s d x d p,3N
(3N + 1) kB T
2 3
where we used the identity δ f (s) = δ (s − s0 ) / f (s) with f (s0 ) = 0 in the second step. Integrat-
ing eq. (16.28) over s yields

1 H1 −E
Z= e− kB T dp s d3N x d3N p
(3N + 1) kB T
p2s (16.29)
H
− k 0T 3N 1 E− 2Q
= e B d xd p 3N
e kB T dp s ,
(3N + 1) kB T
p2
with H1 = H0 + 2Qs . The first part of the last expression is the canonical partition function and the last
part is a constant factor. In 1985, Hoover proved that the Nosé–Hoover thermostat is the only method with a
single friction parameter that gives the correct canonical distribution. Hoover also showed that this thermostat
satisfies the Liouville equation. That is, the density of states is conserved in phase space.
Exercise: Canonical MD Simulations

We will now simulate a system in contact with a thermal reservoir at fixed temperature T . There are several
techniques to do this. We will use the Nosé–Hoover thermostat, which is the best in the sense that it repro-
duces the correct energy distribution corresponding to a canonical system. This technique introduces a new
degree of freedom s into the system with its own potential. Here, we rewrite the final equations of motions
for a system having these extra degrees of freedom:
ẋi = vi ,
mi v̇i = fi − ξvi ,
⎛ N ⎞
1 ⎜⎜⎜⎜ pi 2 ⎟⎟
ξ̇ = ⎜⎝ − G kB T ⎟⎟⎟⎠ ,
Q i=1 mi
16.4 Stochastic Method 183
Exercise: Canonical MD Simulations (cont.)

where G = 3N + 1 is the number of degrees of freedom, T is the (desired) temperature of the system,
kB is the Boltzmann factor, and Q is the inertia corresponding to the extra degree of freedom s. The friction
coefficient ξ is defined as
ṡ
ξ= .
s
Note that the coordinate s of the extra degree of freedom is not needed in the simulation. One can discretize
the above equations easily with the Verlet method.
Extend your MD code to simulate a canonical system with a given temperature.
1. Start with an initial configuration with instant temperature T0 , and set the desired temperature T to a
different value and observe the behavior of T in time.
2. When the system reaches the equilibrium state, compute the total energy of the system and compute its
distribution. Do you get the correct distribution?
3. Compute the heat capacity of the system as a function of temperature using the fluctuation-dissipation
theorem (see eq. (3.27)).

• Try to adjust the value of Q so that the system reaches the equilibrium fast and still has reasonably small
fluctuations (in T and/or the total energy). (For large Q, the convergence to equilibrium is slow and the
fluctuations are small while for small Q it is the other way around.)
16.4 Stochastic Method
Another method to perform canonical MD simulations has been proposed by

Andersen [277] (see Figure 16.5). The so-called stochastic method is a combination
of MD and a Monte Carlo algorithm.
Stochastic Method
At every nth time step, randomly select one particle and give it a new momentum p, which is distributed
according to the Maxwell–Boltzmann distribution
1 p2
P (p) = 3
e− kB T , (16.30)
(πkB T ) 2
where T is the desired temperature.
At every nth time step, a particle is selected uniformly at random and given a new
momentum according to eq. (16.30). If n is too small, we recover pure Monte Carlo
and loose the ability to simulate particle interactions in real physical time. If n is too
large, the coupling to the heat bath is too weak, equilibration is slow, and we will
essentially simulate a microcanonical system.
16.5 Constant Pressure
In addition to the simulation of MD systems with constant temperature, another

important situation is that of constant pressure. To formulate constant pressure
MD, we again consider the equipartition theorem and a Hamiltonian as defined by
Eqs. (16.1) and (16.2), and find
. N /
1 2 3
xi · ∇xi V (x) = NkB T . (16.31)
3 i=1
We now distinguish between particle–particle and particle–wall forces f part and f ext ,
respectively. This leads to
. N / . N /
1 2 3 1
xi · ∇xi V (x) = − xi · fipart + fiext
Figure 16.5 Hans C. Andersen is a 3 i=1 3 i=1
. N / . N / (16.32)
1 1
professor emeritus of chemistry at
part
Stanford University. =− xi · fi − xi · fi
ext
.
3 i=1 3 i=1
We define
1
N
w= xi · fipart (16.33)
3 i=1
as the virial. Using

. N /
1 1 1
xi · fi
ext
=− pxdA = − p (∇ · x) dV = −pV , (16.34)
3 i=1 3 Γ 3 V
we define the instantaneous pressure P via
PV = NkB T + w . (16.35)
Similar to the thermal inertia Q of the Nosé–Hoover thermostat, we introduce a

parameter W which adjusts the pressure of the system. The ratio V between the new
and old volumes now becomes the new degree of freedom and instead of time, we now
rescale the spatial coordinate x according to
1
x → V3x. (16.36)
16.6 Parrinello–Rahman Barostat 185
Similar to rescaling the time derivative in Section 16.3, here we have to rescale the spa-
tial derivative. The outlined rescaling is only valid in isotropic systems where a change
in length is the same in every direction. The rescaled Hamiltonian is then given by

N
1 1
H= mi ẋ2i + W V̇2 + V (x1 , . . . , xN ) + pV , (16.37)
i=1
2 2
where the new variable V is the relative volume change controlled by a piston of mass
W (see Figure 16.6) that also defines the canonical momentum
pV = W V̇ . (16.38)
We again derive the equations of motion from the Hamiltonian. The velocities are
dxi pi
= ∇ pi H = 1
,
dt mi V 3
(16.39)
dV ∂H pV
= = ,
dt ∂pV W
and the corresponding forces are
dpi 1
= −∇xi H = −∇xi V V 3 xi = fi , Figure 16.6 A weight of mass W exerts a
dt pressure p on the system with volume V.
1 0
N 1 1 (16.40)
dpV ∂H
=− =− xi · ∇xi V V 3 xi − p .
dt ∂V 3V i=1
Finally, the equations of motion of the so-called Berendsen barostat are

fi V̇
ẍi = − ẋi ,
mi 3V
(16.41)
1 1
N N
W V̈ = mi ẋ2i + fi xi − p .
3V i=1 3V i=1
16.6 Parrinello–Rahman Barostat
For the Berendsen barostat, an isotropic rescaling of space and an isotropic medium
are assumed. However, this assumption may not be satisfied for all systems since, for
instance, the volume of certain crystals cannot be rescaled equally in every direction.
The corresponding generalization of the Berendsen barostat is an anisotropic scaling
of a box, spanned by three vectors, a, b, and c, of volume
V = a · (b ∧ c) = det (H) with H = {a, b, c} . (16.42)
We describe the position of a particle i in the box by
ri = Hsi = xi a + yi b + zi c with 0 < xi , yi , zi < 1 . (16.43)

Therefore, the distance between two particles i and j is

ri2j = siTj G si j with G = H T H . (16.44)
According to Parrinello (see Figure 22.3) and Rahman (see Figure 16.7) [331], the
Hamiltonian that we defined previously in eq. (16.37) can be rewritten in terms of the
rescaled coordinates (16.44):
1 1
H= mi ṡiT Gṡi + V ri j + WTr Ḣ T Ḣ + pV . (16.45)
2 i ij
2
The corresponding equations of motion are

mi s̈i = H −1 fi − miG−1 Ġṡi ,
T (16.46)
Figure 16.7 Aneesur Rahman W Ḧ = pV H −1 .
(1927–1987) is known for his The Parrinello–Rahman barostat is important for simulating crystals with applica-
contributions to computational physics tions in solid-state physics and material science. Unlike the Hoover thermostat, here
and molecular dynamics. He was a the new degree of freedom is not a simple scalar but a matrix H that accounts for the
professor at the University of Minnesota. anisotropy of the system.
The Aneesur Rahman Prize for
Computational Physics awarded by the
American Physical Society is the most
prestigious prize for computational
physics.
Inelastic Collisions in Molecular Dynamics 17
In many cases, loss of energy is not just a minor unwanted effect of some otherwise
well-behaved and converging mathematical method. From landslides to turbulent flow
in aircraft engines, dissipation is often necessary for the simulations to be stable and
important in practice. Often macroscopic systems are not closed and energy leaks out
in the form of wear, shape deformation, acoustic emission, and other degrees of free-
dom that one cannot or does not want to include explicitly in the calculation. To
mathematically describe inelastic collisions, we outline some concepts such as damped
oscillators and plastic deformations.
17.1 Restitution Coefficient
The kinetic energy of colliding billiard balls is not constant because of friction, plas-
tic deformation, and sound emission. As an example, we could think of a basketball
dropped from a certain height (see Figure 17.1). Due to energy dissipation, the ball
will not reach the same height as before after bouncing back from the ground. In our
MD simulations, we account for energy dissipation effects in an effective manner by
introducing the restitution coefficient
after 2
E after v
r= before
= , (17.1) Figure 17.1 A basketball dropped from a
E vbefore
certain height will not reach its initial
where E after and E before are the energies before and after the collision event, and vafter height again because of dissipation
and vbefore are the corresponding velocities. Elastic collisions correspond to r = 1, effects.
whereas perfect plasticity is described by r = 0. We can also distinguish between
normal and tangential velocity components vn and vt , and define the corresponding
coefficients
vafter
en = n
, (17.2)
vbefore
n
vafter
et = t
. (17.3)
vbefore
t
For a bouncing ball, the restitution coefficient accounts for effects such as wear, plastic
deformations, and sound emission. These coefficients strongly depend on the material,
angle of impact, and weakly on other factors like particle shape and relative impact
velocity, and can be determined experimentally. In MD simulations, we incorporate
187
188 Inelastic Collisions in Molecular Dynamics
inelasticity with the help of a damped

-- -- oscillator. For the corresponding equations of
motion, we use r = Ri + R j − -xi − x j - and meff = m1 +m2 to denote the radial distance
m1 m2
between two particles and the effective mass, respectively. The equation of motion is
then
meff r̈ = −kr − γṙ , (17.4)
where k is the spring constant and γ is the viscous damping coefficient. The solution
of this differential equation is given by
vbefore
r (t) = sin (ωt) exp (−δt) , (17.5)
ω

where ω = ω20 − δ2 , ω0 = mkeff , and δ = 2mγeff are the oscillator frequency, the ground
frequency of the undamped oscillator, and the damping coefficient, respec-
tively. Damping the oscillator reduces its frequency and generally increases the
period T .
We regard the collision between two particles as a half-cycle of a damped oscilla-
tor, for which some energy is dissipated due to damping. The collision lasts half the
period of an oscillation. The kinetic energy of the two particles decreases when they
come together, and when the particles bounce back their kinetic energy increases. We
therefore compute the restitution coefficient according to
⎛ ⎞
ṙ (t + T ) ⎜⎜⎜ γπ ⎟⎟⎟
en = = exp (−δT ) = exp ⎜⎜⎝− ⎟⎟⎠ . (17.6)
ṙ (t) 4meff k − γ2
This relation implies that a restitution coefficient uniquely determines a viscous
damping coefficient according to
)
meff k
γ = 2 ln (en ) . (17.7)
ln (en )2 + π2
An important assumption that we made is that the restitution coefficient is con-

stant throughout the complete interaction. This assumption is violated for a Hertzian
contact between spheres in three dimensions for which the contact area depends on
their distance. The dependence between energy loss and interaction overlap becomes
nonlinear and our approach has then to be corrected since the restitution coefficient
is not constant anymore [278].
17.2 Plastic Deformation
In many engineering and industrial applications, collisions between two deformable

objects are characterized by an irreversible change of their shapes, a so-called
17.2 Plastic Deformation 189
plastic deformation. We show an example of a plastic deformation in Figure 17.2. One

approach to model plasticity has been introduced by Walton and Braun [279, 280].
The stiffnesses before and after the collision are modeled by springs with two different
spring constants, k1 and k2 , so that after the collision the particle does not come back
to its original position, leaving a remnant deformation as we illustrate in Figure 17.3.
The interaction is dissipative and the force is
⎧
⎪
⎪
⎨k1 δ , loading ,
f =⎪
⎪ (17.8)
⎩k2 (δ − δ0 ) , unloading .
First, the objects approach each other according to Hooke’s law and overlap up to
δmax so that k1 = fmax /δmax . During this initial contact, the objects are deformed and
therefore repulsion occurs with a different spring constant k2 = fmax / (δmax − δ0 ) such
that the repulsive forces vanish before the overlap does.
The collision time is not symmetric since loading lasts longer than unloading. We Figure 17.2 Two particles undergoing
therefore describe the contact time by elastic and plastic interactions.
* *
π mi j mi j
tc = + . (17.9)
2 k1 k2
The dissipated energy is equal to the area enclosed by the two curves (i.e.,
the blue area in Figure 17.3). From this observation, we obtain the restitution
coefficient
E after k1
r= before
= . (17.10)
E k2
Force f as a function of deformation δ for a plastic collision. The diagram shows a plastic deformation according to Figure 17.3
Walton and Braun [279].
190 Inelastic Collisions in Molecular Dynamics
17.3 Coulomb Friction and Discrete Element Method
Coulomb friction occurs when two solids are sheared against each other, that is, when
there is a difference in tangential velocity at their contact point. The energy dissipation
due to Coulomb friction is proportional to the normal force. Two friction coefficients
are used to take such dissipation effects into account. The static friction coefficient
describes friction effects of nonmoving objects and the dynamic coefficient accounts
for friction of moving objects. The classic example illustrating Coulomb friction is an
object located on an inclined plane (see Figure 17.4). At a certain inclination angle,
the static friction angle, the object begins to slide down the plane. The tangent of this
angle gives the static friction coefficient. The object only stops moving when the angle
Figure 17.4 Solid block on inclined plane is decreased to the dynamic friction angle, the tangent of which defines the dynamic
illustrating the forces acting in Coulomb friction coefficient [281]. Note that the friction force is independent of the contact area.
friction. Numerically, the problem is very difficult to handle because the friction coefficient
μ is described by a discontinuous function of the relative tangential velocity vt . For
nonzero tangential velocities vt > 0, the dynamic friction coefficient μd is different
from the static one μ s when vt = 0.
One approach of implementing Coulomb friction is to distinguish between a small
shear velocity vt below which static friction is implemented and above which dynamic
friction is used [282]. For two colliding particles, the tangential momentum transfer is
then given by

Δpt = μ s meff (et + 1) vbefore
i − vbefore
j t for vt 1 ,
(17.11)
Δpt = μd meff (et + 1) vbefore
i − vbefore
j t for vt " 1 ,
where t is the tangential unit vector.

For multiple particles interacting via Coulomb friction such as sand or stones,
a method proposed by Cundall and Strack is established nowadays [283]. This
forward integration technique is widely used in engineering and industrial appli-
cations and is known as discrete element method (DEM). The company Itasca
founded by Cundall himself is a leading provider of DEM software. In DEM
simulations, we define two tangential forces, one for static and one for dynamic
friction:
2 3
ft = − min γvt , μd fn sign (vt ) , (17.12)
2 3
ft = − min |k s ξ| , μ s fn sign (vt ) . (17.13)
If the tangential velocities are smaller than a certain value, then a spring with spring
constant k s is attached between both sides of the contact. If |k s ξ| > μ s fn , the spring is
removed and the particles are free to move with dynamic friction coefficient. Equations
(17.12) and (17.13) only describe the tangential force component, and for a complete
17.3 Coulomb Friction and Discrete Element Method 191
DEM program, one has to add the normal forces using eq. (17.7). Discrete-element-
method simulations are used very much in mining and geotechnical and chemical
engineering.
Exercise: Sliding on an Inclined Chute

In this exercise, we study the train model on an inclined chute with inclination angle θ. The train model consists
of a chain of blocks of mass m. All blocks are connected through linear springs with spring constant k. The first
block is pulled with constant velocity v. We account for Coulomb friction between each block and the chute
by using static and dynamic friction coefficients μ s and μd . Use eq. (17.11) to add Coulomb friction in your
simulations.
Tasks
• Simulate the motion of the train model for different inclination angles θ.
• Vary the spring constant k, and compare the motion of the train for large and small spring constants. The
observed stop and go motion is called stick-slip.
18 Event-Driven Molecular Dynamics
Standard molecular dynamics (MD) methods, such as Verlet or leapfrog meth-

ods, are based on explicit forward integration schemes with a finite time step Δt
to numerically solve certain MD problems (see Sections 12.4 and 12.5). These
approaches are, however, not suited to handle hard-core potentials because of infi-
nite repelling forces that occur if particles collide. Instead, event-driven methods are
the right choice to simulate rigid particle collisions. As opposed to batch program-
ming, the flow of an event-driven program is not determined by loops but by events.
Therefore, event-driven MD simulations are characterized by branching points and
conditional logic and are intrinsically sequential. In the next sections, we discuss
different examples of event-driven dynamics and their application to simulate rigid
particles.
18.1 Event-Driven Procedure
As a first example of the application of event-driven methods to MD, we consider

a work by Alder [243] (see Figure 12.1). In this work, Alder studied rigid bodies
of finite volume such as billiard balls. A hard-core potential is assumed to model
elastic collisions between rigid particles not taking friction into account. Our stan-
dard MD integration schemes fail to describe such systems and we therefore focus
on an event-driven formulation of particle interactions. We regard particle colli-
sions as instantaneous events between which particles are not interacting. In this
method, only the exchange of the particles’ momenta is taken into account and no
forces are computed. Furthermore, only binary collisions are considered and simul-
taneous interactions between three or more particles are neglected. Between two
collision events, particles follow ballistic trajectories. To perform an event-driven
MD simulation, we need to determine the time tc at which the next collision will
occur to then obtain the velocities of the two particles after the collision based on
the particle velocities before the collision. A look-up table can be used to perform
these computations efficiently. To determine tc , we have to identify the next collision
event.
Let us consider a two-dimensional collision problem involving two disks i and j
with radii Ri and R j , respectively. We show such a collision in Figure 18.1. The collision
angle θ is the angle between the connecting vector ri j = ri − r j and the relative velocity
vi j = vi − v j . For the moment, we are not taking into account the influence of friction
and thus neglect the exchange of angular momentum. We compute the times ti j at
192
18.1 Event-Driven Procedure 193
Two particles i and j collide. Particle i has radius Ri and angular velocity ωi , and particle j has radius R j and Figure 18.1
angular velocity ω j . The relative velocity of both particles is vi j = vi − v j , and the collision angle θ is the angle
between ri j = ri − r j and vi j . We use n = rij /|ri j | to denote the normal vector that points from the center of
particle i to the center of particle j. The corresponding tangential unit vector is t.
which the next collision between particles i and j will occur. At time ti j , the distance
between the two particles is
-- -- - -
--ri j ti j -- = --Ri + R j -- . (18.1)
Given a relative velocity vi j at time t0 , we obtain ri j (ti j ) = ri j (t0 ) + vi j ti j . The corre-

sponding contact time ti j for particles i and j can be found via
0 1 0 12 2
v2i j ti2j + 2 ri j (t0 ) vi j ti j + ri j (t0 ) − Ri + R j = 0 . (18.2)
If solutions ti j of eq. (18.2) are not real (i. e., ti j R), the corresponding trajectories of
particles i and j never cross. The next collision occurs at time

tc = min ti j , (18.3)
ij
where the minimum is taken over all real-valued ti j . Thus, in the time interval [t0 , tc ],
positions and angular orientations of all particles evolve according to
ri (t0 + tc ) = ri (t0 ) + vi (t0 ) tc and φi (t0 + tc ) = φi (t0 ) + ωi (t0 ) tc . (18.4)
It is also possible to add gravitational and electrostatic force terms. The main bottle-
neck is the computation of contact times ti j because this operation scales quadratically
with the number of particles. For each particle i, we have to identify the particle j that
first collides with particle i to then compute ti j . After performing this procedure for
all particles, we have to find the minimum in the set of all collision times ti j according
to eq. (18.3). For high-density particle systems, very distant particles will not collide
and it is thus not necessary to determine corresponding collision times ti j . Still, the
distances between all particles need to be determined and thus the algorithm may still
be very inefficient. Instead of looking at distances between particles, it is also possible
to divide the system into sectors and treat those separately. The crossing of particles
through the boundaries of a sector is then considered as a “collision” event at which
lists are recalculated but no velocities change. Another possibility to speed up this
algorithm is the so-called Lubachevsky method [284].
18.2 Lubachevsky Method
Unfortunately, the loop to calculate tc scales quadratically with the number of particles
when simply checking all pairs of particle interactions. Checking and updating all
particles is, however, very inefficient because (i) most collisions would only occur after
O (N) steps and (ii) most particles are not participating in a finally chosen collision
but their positions are always updated. This can be improved by only considering and
updating those particles that are part of a collision event.
An algorithm that has been developed by Boris Lubachevsky [284] reduces the
computational complexity for this improved algorithm is based on lists of events
and binary stacks. In addition to the particle position and velocity, Lubachevsky’s
method stores the last and next events for each particle. In this way, we are keep-
ing track of the time of the event and the partner particle involved in the event. In
practice, this can be implemented by updating at each collision six arrays (event time,
new partner, and for each particle position and velocity) of dimension N (number of
particles in the system). Alternatively, we can create a list of pointers pointing to
a data structure for each particle consisting of six variables. Storing the last event
is needed as particles are only updated after being involved in an event. For each
particle i, the time t(i) is the minimal time of all possible collisions involving this
particle:

t(i) = min ti j . (18.5)
j
Considering the collision of particle i with N − 1, others can be improved by dividing

the computational domain in sectors such that only neighboring sectors have to be
considered in this step. Sector boundaries have to be treated similar to obstacles such
that a collision event happens during which no velocities change when particles cross
sector boundaries. For each particle i, this step would then be of order O (1) instead of
O (N). The next collision occurs at time

tc = min t(i) . (18.6)
i
To summarize, the Lubachevsky method is as follows:
Lubachevsky Trick
We store t(i) in increasing order in a stack:
• The vector PART[m] points to particle i which is at position m in the stack. (Sometimes also a vector
POS [i] is used to store position m of particle i in the stack.)
18.3 Collision with Perfect Slip 195
Lubachevsky Trick (cont.)
• Establish an implicit ordering of the collision times t(i) , where m = 1 points to the shortest time.
• PART[1] is the particle with minimal collision time: tc = t(PART[1]) .
• After the collision events, all six entries (event times, new partners, positions, and velocities) have to be
updated for both particles. Additionally, the vector PART[m] has to be reordered.
: ;
Reordering the times t(i) after each event is of order O log N when using, for exam-
ple, binary tournament trees for sorting [284]. The advantages of this method are that
it is neither necessary to calculate the collision times for all particle pairs at every step
nor to update the positions of particles that do not collide. Only position and velocity
of the particles involved in the collision event are updated.
18.3 Collision with Perfect Slip
After having identified the next collision event, we have to update the positions, veloc-
ities, and angular orientations of the colliding particles. With eq. (18.4), we only
described the time evolution until the collision occurs. In this section, we now focus on
the particle dynamics after a collision event. In a first approximation, we are assuming
perfect slip and neglect any tangential exchange of momentum. Only linear momen-
tum and no angular momentum is exchanged. Conservation of momentum implies
that the velocities after the collision are
Δp
vi after = vi before + , (18.7)
mi
Δp
v j after = v j before − , (18.8)
mj
and energy conservation yields
1 before 2 1 before 2 1 after 2 1 after 2
mi vi + mj vj = mi vi + mj vj . (18.9)
2 2 2 2
Thus, the exchanged momentum is
0 1
Δp = −2meff vbefore
i − vbefore
j · n n, (18.10)
mm
where meff = mi +m i j
j
is the effective mass and n = ri j /|ri j | is the collision normal vector.
If mi = m j , the velocity updates are
vi after = vbefore
i − vnij · n , (18.11)
v j after = vbefore
j + vnij · n , (18.12)

with vnij = vibefore − v jbefore · n. The values can be stored once in a look-up table such
that there is no need of calculating the new velocities at every collision.
18.4 Collision with Rotation
We now consider two spherical particles i and j of the same radius R and mass m.
Due to friction, the tangential component of angular momentum is exchanged if
particles collide with nonzero tangential velocity. To account for such an angular
momentum exchange in our MD simulations, we begin with the equation of motion
for rotation [280]
dωi
I = r ∧ fi , (18.13)
dt
where ∧ denotes the cross product and ωi , I, and fi are angular velocity, moment of
inertia, and the forces exerted on particle i, respectively. For more details on rigid
body dynamics, see Section 14.2. For the two colliding spherical particles of radius R,
moment of inertia I, and mass m, angular momentum conservation yields
: ; : ;
I ω i − ωi = −mRn ∧ v i − vi ,
(18.14)
I ω j − ω j = mRn ∧ v j − v j ,
with the primed velocities representing velocities after the collision. Using conserva-
tion of momentum
v i + v j = vi + v j , (18.15)
the angular velocities after the collision are
Rm : ;
ω i − ωi = ω j − ω j = − n ∧ v i − vi . (18.16)
I
The relative velocity between particles i and j is

ui j = vi − v j − R ωi + ω j ∧ n , (18.17)
where n is the unit vector pointing from particle i to particle j. We decompose the
relative velocity into its normal and tangential components unij and uti j , respectively.
For the angular momentum exchange, we have to consider the relative velocity of the
particle surfaces at the contact point. The normal and tangential velocities are given by

unij = ui j n n ,

uti j = ui j − unij = −n ∧ n ∧ ui j (18.18)

= vi − v j − R ωi + ω j ∧ n ,
t t
and general slips are described by

uti j = et uti j , (18.19)
where the tangential restitution coefficient et accounts for different slip types. The per-
fect slip collision is recovered for et = 1 which implies that no rotation energy is
transferred from one particle to the other. No slip corresponds to et = 0. Energy
conservation only holds if et = 1 and energy is dissipated if et < 1.
18.5 Inelastic Collisions 197
If we compute the difference between the relative tangential velocities before and
after the slip event, we find

(1 − et ) uti j = uti j − uti j
0
1 (18.20)
= − vti − vti − vtj + vtj − R ω i − ωi + ω j − ω j ∧ n .
Combining this equation with Eqs. (18.15) and (18.16), we obtain an expression
without angular velocities

(1 − et ) uti j = uti j − uti j
0 1 (18.21)
= − 2 vti − vti + 2q vti − vti
and thus
(1 − et ) t mR2
vti = vti − u with q= . (18.22)
2 (1 + q) i j I
Analogously, we find for the remaining quantities
(1 − et ) t
vtj = vtj + u ,
2 (1 + q) i j
(1 − et )
ω i = ωi − : ; ut ∧ n , (18.23)
2R 1 + q−1 i j
(1 − et )
ω j = ω j − : ; ut ∧ n .
2R 1 + q−1 i j
Finally, the updated velocities are
(1 − et ) t
vi = vi − unij − u ,
2 (1 + q) i j
(18.24)
(1 − et ) t
v j = v j + unij + u .
2 (1 + q) i j
18.5 Inelastic Collisions
The relative velocity of two colliding particles at their contact point is

0 1
unij = ui j n n = vi − v j n n . (18.25)
For an inelastic collision, dissipation effects due to various physical mechanisms lead
to reduced normal velocities
unij = −en unij , (18.26)
where en is the normal restitution coefficient defined in eq. (17.2). For en = 1, there is
no dissipation, whereas dissipation occurs for en < 1. The velocities after the collision
are given by
(1 + en ) n
vi = vi − ui j ,
2 (18.27)
(1 + en ) n
v j = v j + ui j .
2
In the normal direction, we include the normal restitution coefficient in eq. (18.10)
and describe the momentum exchange according to
0 1
Δpn = −meff (1 + en ) vi − v j n n . (18.28)
mR2
With q = I , the generalized set of Eqs. (18.22) and (18.23) is given by
(1 + en ) n (1 − et ) t
v i = vi − ui j − u ,
2 2 (1 + q) i j
(1 + en ) n (1 − et ) t
v j = v j + ui j + u ,
2 2 (1 + q) i j
(18.29)
(1 − et )
ω i = ωi − : ; ut ∧ n ,
2R 1 + q−1 i j
(1 − et )
ω j = ω j − : ; ut ∧ n ,
2R 1 + q−1 i j
describing inelastic collisions of rotating particles [278].
Because dissipation occurs in macroscopic particle systems, it is important to
incorporate energy dissipation effects in simulations by accounting for normal and
tangential restitution coefficients [280].
18.6 Inelastic Collapse
A prominent consequence of dissipation in hard-sphere event-driven MD simula-

tions is the so-called inelastic collapse, which is characterized by particles that collide
infinitely often in finite time. This effect is often confused with the clustering insta-
bility [285, 286] that can occur in regions of high particle densities, where the large
number of inelastic collisions leads to large dissipation effects and allows particles to
form clusters of locally even higher densities. Such clustering effects are common in
the formation of mass concentrations out of intergalactic dust (for more details, see
Figure 18.5). Without clustering, stars, galaxies, and other celestial objects cannot be
formed. Note, however, that clustering is a hydrodynamic instability that occurs due to
dissipation if both the density and the system are large enough [287], notably without
long-range forces like gravity [288].
When simulating inelastic collisions of hard, rigid particles, we may encounter
so-called finite-time singularities as described by McNamara (see Figure 18.2) and
Young [287, 289]. Such singularities are likely to occur in event-driven simulations of
high-density particle systems because of the hard-sphere interaction potential and the
associated vanishing interaction duration. Instantaneous interactions in event-driven
Figure 18.2 Sean McNamara is a
simulations are neither resembling particle collisions that we can observe in nature nor
professor at the University of Rennes 1.
those that we simulated using soft-sphere, MD/DEM-type frameworks.
Unlike the clustering instability, the inelastic collapse can occur in systems that
only involve a small number of particles. One could, for instance, think of a parti-
cle bouncing between two other particles that approach each other along a straight
line (see Figure 18.3). To understand the emergence of finite-time singularities with
The orange particle at the center is bouncing between the blue particles, which approach each other. Figure 18.3
an even simpler model, we consider a ball bouncing vertically on a hard surface (see
Figure 17.1) [280]. Every time the ball hits the surface, its kinetic energy is lowered
according to eq. (17.1). As a consequence, the ball will not reach the initial height
anymore and the time between two contacts with the surface decreases after each col-
lision. After a finite time, the ball comes to a rest, but the simulation takes an infinite
time to run. In an event-driven simulation, the ball only stops its motion after infinitely
many collisions, which means that the simulation never ends. A related problem is the
famous Zenon paradox.1 The total time needed for the bouncing ball to come to rest
is the sum over an infinite number of times between two successive surface contacts ti .
Since the height is proportional to the potential energy, it scales with the restitution
coefficient at every surface contact. Consequently, after the ith surface contact, the
damping of the height is proportional to ri . The total time is thus

∞
ttot = ti
)
i=1
2hinitial √ i
∞
=2 r (18.30)
g
) i=1

2hinitial 1
=2 √ −1 ,
g 1− r
where hinitial is the initial height and g the gravitational acceleration. Mathematically,
the bouncing ball jumps infinitely often during the finite time given by eq. (18.30).
In event-driven particle simulations, interactions between particles are modeled as
instantaneous events. This is the origin of the inelastic collapse, a wrong assumption
that lies at the heart of event-driven rigid-particle models. However, real collisions have
a certain duration. In order to handle the inelastic collapse numerically, Luding (see
Figure 18.4) and McNamara introduced a restitution coefficient that depends on the
time elapsed since the last event [286]. In their so-called TC model, a collision takes a
finite time tcontact . If the time since the last collision of both interacting particles t(i) is
less than tcontact , the restitution coefficient is set to unity and otherwise to r. We thus
obtain ⎧
⎪
⎪
⎨r , for t > tcontact or t > tcontact ,
(i) ( j)
r =⎪
(i, j)
⎪ (18.31)
⎩1 , otherwise .
Figure 18.4 Stefan Luding is a professor
With this new definition of the restitution coefficient, the collision type changes from at the University of Twente.
inelastic to elastic if too many collisions occur during tcontact . Depending on the
1 The Zenon paradox is also known as the paradox of Achilles and the turtle.
t
Figure 18.5 Snapshot of an event-driven simulation with N = 99,856 particles in a 2D system with periodic boundary conditions.
The gas cools due to dissipation so that 10 percent of the relative velocity is lost per collision, i. e., the restitution
coefficient is r = 0.9. Red, green, and blue correspond to large, medium, and small particle densities, respectively.
Figure courtesy Stefan Luding [290].
material properties of the colliding particles, it is also possible to use more com-
plex functional dependencies of the restitution coefficient. For very dense and viscous
materials, setting the restitution coefficient to zero instead of unity may be a better
choice. In this case, the particles form clusters and stick together.
In Figure 18.5, we show an example of a simulation involving 99,856 particles that
collide inelastically. We observe the formation of clusters in certain regions of the
simulation snapshot shown. Such simulations are important to model the instabil-
ity of intergalactic dust before gravity becomes relevant. For certain systems, it may
be important to account for additional long-range repulsive interactions between par-
ticles (see Chapter 15), which counteract clustering, while attractive interactions do
enhance clustering [288].
We now understand the main differences between hard-sphere event-driven simu-
lations and soft-potential approaches. Binary and instantaneous collisions determine
the dynamics in an event-driven simulation, whereas multiple interactions might occur
at the same time in the other case.
Exercise: Event-Driven MD
So far, in our MD simulations, we were using particles interacting with a smooth potential (Lenard-Jones
potential). Event-driven simulations are able to model particles interacting by a hard-core potential consid-
ering binary collisions (i.e., instantaneous contacts). Between binary collisions, particle trajectories can be
calculated analytically. In systems without external forces (e.g., gravity) the particles follow straight paths
between collisions. The collision time between pairs can be calculated by solving the quadratic equation
-- -
-r (t ) + v t -- = R + R ,
ij 0 ij ij i j
where ri j (t0 ) is the distance vector between particles at time t0 , vi j the velocity difference, ti j the possible
collision time, and Ri j are the particle radii. In this method one has to calculate the collision time between all
particle pairs getting the “global” collision time tc = mini j (ti j ).
1. Implement the event-driven dynamics either in one, two, and three dimensions. You can use your old code
as basis or start implementing a completely new code (e.g., for the one-dimensional case).
2. Several simplifications can be applied:
• For a one-dimensional system, the calculation of the collision time is easier (there is no need to solve
a quadratic equation and only neighboring particles interact). You can also set the radii to zero in this
case.
• In two- and three-dimensional simulations, consider collisions with perfect slip. Then, you do not need
angular velocities.
• Consider particles with the same mass m and radius R.
3. The aim of this exercise is to solve a few simple one-dimensional problems (you can also try higher-
dimensional setups, if you like):
• Chain of N beads with restitution coefficient en = 1 (N = 5 resembles a Newton’s cradle). How do you
implement the synchronous motion of multiple particles? You can also vary the restitution coefficient
in this system, if you like.
• Chain of N beads with restitution coefficient en < 1 hitting a resting wall (en = 1): Measure the

effective restitution coefficient eeff = E f /Ei of multiple synchronously moving particles, where Ei
and E f are the initial and final kinetic energies of the chain, respectively. Vary the number N of the
beads (for fixed en ). Above which value of N does the effective restitution coefficient practically vanish
(i.e., synchronously moving particles do not rebounce any more)?
4. Please consider the following points:
• With all the above simplifications, in the one-dimensional case, the velocities are simply exchanged for
perfect restitution (en = 1). For en < 1, the relation is also quite simple.
• To speed up your code, you can store the events for each particle in a priority queue. While checking,
all particle pairs require O(N 2 ) time, generating events only for the colliding particles and inserting
them in the queue has a runtime of order O(N log N) instead. Methods with insertion time of O(1)
(and O(N) in total) have been proposed in Ref. [291]. Only interacting particles need to be advanced,
but bear in mind that the collision can invalidate previously predicted events. You can either remove
them manually from the queue in time O(N) or, more efficiently, keep a counter of each particle’s
collisions and identify an event as invalid only when resolving it.
19 Nonspherical Particles
Up until now, we focused on the simulation of spherical particles. However, it would

be also desirable to simulate interactions between arbitrarily shaped particles. Such
methods are relevant to describe realistic granular systems consisting of particles of
various shapes. For some cases such as ellipsoidal particles, it is possible to obtain
analytical solutions of the interaction dynamics. For the more general treatment of
arbitrarily shaped particles, only numerical approximations are feasible.
19.1 Ellipsoidal Particles
To simulate the interaction between ellipsoidal particles, we consider two ellipses

whose parameterization is given by
2 3 2 3
(x − xa ) cos(θa ) + (y − ya ) sin(θa ) 2 (x − xa ) sin(θa ) − (y − ya ) cos(θa ) 2
+ = 1 , (19.1)
a21 a22
2 3 2 3
(x − xb ) cos(θb ) + (y − yb ) sin(θb ) 2 (x − xb ) sin(θb ) − (y − yb ) cos(θb ) 2
+ = 1 , (19.2)
b21 b22
where θa and θb are the rotation angles, a1 and b1 are the major axis, and a2 and
b2 are the minor axis of ellipses A and B, respectively. We show the two ellipses in
Figure 19.1. As proposed by Perram and Wertheim [292], we take the overlap of the
two ellipses as being proportional to the interaction force. To calculate the overlap, we
transform the ellipses into circles using an appropriate metric. Let u1 , u2 and v1 , v2 be
orthonormal vectors along the principal axes of the ellipses A and B, respectively. The
rotated ellipses are defined by the matrices

A= a−2
k uk ⊗ uk
T
and B = b−2
k vk ⊗ vk .
T
(19.3)
k k
Ellipse A can be described by the functional
G A (r) = (r − ra )T A (r − ra ) , (19.4)
which is smaller than 1 inside the ellipse, larger then 1 outside and exactly 1 on the
ellipse. That is,
⎧
⎪
⎪
⎪< 1 , if r is inside the ellipse ,
⎪
⎪
⎨
G A (r) = ⎪
⎪1, if r is on the ellipse , (19.5)
⎪
⎪
⎪
⎩> 1 , if r is outside the ellipse .
202
19.1 Ellipsoidal Particles 203
Two overlapping ellipses. Figure 19.1
Analogously, we use G B (r) to denote the corresponding functional for ellipse B. Next,
we define another functional
G (r, λ) = λG A (r) + (1 − λ) G B (r) , (19.6)
which combines the two ellipses with a parameter λ ∈ [0, 1] that interpolates between
the two centers of the ellipses. With the help of G (r, λ), we compute the contact point.
To do this, we have to minimize the functional according to
∇r G (r, λ)|r=rm = 0 . (19.7)
This yields
rm (λ) = [λA + (1 − λ) B]−1 [λAra + (1 − λ) Brb ] . (19.8)
If we start from λ = 0 and arrive at λ = 1, the quantity rm (λ) describes a path from the
center of the first ellipse to the center of the second one. We rewrite the path in terms
of rab = rb − ra and find
0 1−1
rm (λ) = ra + (1 − λ) A−1 (1 − λ) A−1 + λB−1 rab ,
0 1−1 (19.9)
rm (λ) = rb − λB−1 (1 − λ) A−1 + λB−1 rab .
If the value of the functional along this path between the centers is always smaller than
unity, we know that we never left the interior of the ellipses, and hence they overlap.
We define the overlap function
S (λ) = G (rm (λ) , λ) , (19.10)
and by inserting eq. (19.9) we obtain

0 1−1
S (λ) = λ (1 − λ) rTab (1 − λ) A−1 + λB−1 rab . (19.11)
t
Figure 19.2 Overlap vector and contact point for two ellipses.
This corresponds to the value of the functional (19.6) on the path that connects the
two centers. We are interested to know if the path ever exhibits values larger than unity.
If λmax is the value of λ ∈ [0, 1] at which S has its maximum, we will be able to tell if
the ellipses are overlapping or not:
⎧
⎪
⎪
⎪ < 1 , if the ellipses overlap ,
⎪
⎪
⎨
S (λmax ) = ⎪
⎪ 1, if the ellipses touch , (19.12)
⎪
⎪
⎪
⎩> 1 , if the ellipses are separated .
Figure 19.3 Examples of superellipsoids Based on this result, we compute the contact point for stiff ellipses for which over-
with parameters as defined by laps in the simulations are not allowed. We set S (λmax ) to unity and find the distance
eq. (19.16). The figure is taken from between the centers of the ellipses for which they just touch:
Ref. [293]. C 0 1−1 D− 12
r∗ = λmax (1 − λmax ) eTr (1 − λmax ) A−1 + λmax B−1 er , (19.13)
where er = rab /|rab |. We define the overlap vector as

r∗ er − rab
ξ= . (19.14)
2
The contact points of the overlapping and the rigid case are related via (see
Figure 19.2)
r pc = r∗pc + ξ . (19.15)
The algorithm described above can be easily expanded to three-dimensional ellip-

soids. Moreover, there exist generalized versions of ellipsoids, so-called superellipsoids,
which are mathematically defined by [293]
Figure 19.4 An example of a z m
superellipsoid packing with 1,125 xm + ym + = 1, (19.16)
particles. The shape of the superellipsoids α
is determined by eq. (19.16) with m = 4 where the parameters m and α control the shape (see Figure 19.3). The macroscopic
and α = 0.3. The figure is taken from properties of superellipsoids exhibit interesting features such as high packing densities
Ref. [293]. [293, 294]. We show an example of a superellipsoid packing in Figure 19.4. The
Possible overlaps of two triangles. Figure 19.5
contact of superellipsoids can be treated by generalizing the outlined algorithm for

ellipsoids.
We saw that even for the simplest generalization of spheres, it was already necessary
to develop complex analytical methods. This substantially increases the computational
complexity of corresponding simulation algorithms. Ellipsoids and superellipsoids can
describe only a limited number of the shapes encountered in nature and, therefore, it
is necessary to develop approaches to handle objects of even more complex shapes.
19.2 Polygons
A particular class of macroscopic particles are those described by polygons in two

dimensions or polyhedra in three dimensions. Examples of such objects include dice
and crystalline grains like sugar. Again, a measure for the repulsive forces between
“particles” is the overlap area. The advantage of simple polygons is that it is possible
to compute the overlap area dividing them into triangles. However, between triangles,
there can be various different types of contacts and their identification is cumbersome
as we show in Figure 19.5.
Additional complexities arise when the overlap area (blue) does not include all parts
that were deformed (see lower panels of Figure 19.5). Furthermore, rotations of par-
ticles require an additional treatment, as the torque strongly depends on the shape of
the particles. Due to the mentioned reasons, simulating objects that are composed of
polygons is rather complex.
Figure 19.6 Fernando Alonso-Marroquin
is currently a senior lecturer at the
19.3 Spheropolygons University of Sydney.
Another class of methods that is more efficient than the mere division into poly-
gons relies on spheropolygons [295]. This technique was invented by Fernando
Alonso-Marroquin (see Figure 19.6), who uses the two-dimensional Minkowski cow
(see Figure 19.7) as an example of an arbitrarily shaped object in his simulations. This
particular shape is obtained by sweeping a disk around a polygon that resembles the
shape of a cow.
The important feature of the spheropolygon method is that it is possible to simu-
late arbitrary shapes. Once the boundary of the shape is locally described by a disk,
we only have to compute the contact between all pairs of disks that define the edges
and vertices of the shape, which is easier than considering polygons (see Section 19.2).
There are of course a number of constraints when approximating the shapes in such a
way. For example, too large disks would not appropriately describe the original shape
Figure 19.7 The Minkowski cow is an of a certain object and may even lead to wrong simulation results. Imagine that the
example of an irregular shape composed diameters of the disks were larger than the distance between the hooves or the dis-
of spheropolygons. The figure is taken tance between the tail and the rest of the cow of Figure 19.7. Substantial features
from Ref. [295]. of the shape would be lost. A generalization of this technique to three dimensions is
straightforward.
Many other techniques have been developed to describe arbitrarily shaped objects.
Effective implementations have been developed by engineers, physicists, and mathe-
maticians. An important example is the field of mathematical morphology developed
by Jean Serra in the 1960s [296]. The techniques of Marroquin are related to
the so-called dilation techniques, but there are also other methods like the erosion
technique.
Exercise: Triangle Jumping on a Plane

We simulate the interaction of a pair of particles under the influence of gravitational acceleration g using
spheropolygons.
The spheropolygons fall under the influence of gravity from an initial height with zero initial velocity and
hit a “sphero-segment” at the bottom. We use the following shapes: (i) a circular particle, (ii) a convex
spheropolygon square, and (iii) a nonconvex spheropolygon boomerang.
The contact force F between one pair of edge-vertex contact can be written as:
F = kn δn en . (19.17)
Exercise: Triangle Jumping on a Plane (cont.)

where kn is the normal stiffness, δn is the length of the overlap of the contact layer, and en is the normal unit
vector pointing in the direction of the contact force. The corresponding equations of motion of particle i are

ma = Fc + mg . (19.18)
c
where m is the mass of particle i and a is its acceleration. The index c accounts for potentially multiple
contact pairs between particles i and j.
Task 1: How many contact points define the interaction between particles i and j for each case? Where is the
location of the contact, and how is Fc defined (edge-to-vertex or vertex-to-edge) after a contact?
Task 2: Implement the numerical simulation of the falling spheropolygon. Investigate the influence of the
radius of particle i. How does decreasing the radius affect the numerical stability of the simulation?
Task 3: Measure the potential and kinetic energies of particle i and check whether the energy is conserved.
20 Contact Dynamics
We started our discussion of MD simulations by introducing integration methods to

directly solve Newton’s equations of motion for a certain number of interacting par-
ticles. Such direct integration methods are able to simulate the dynamics of particles
which interact via soft potentials. To overcome the problem of infinite forces in the
case of hard-sphere potentials, we discussed event-driven MD techniques in Chap-
ter 18. The advantage of event-driven MD is that it is applicable to rigid bodies, but it
is not capable to compute forces since we are only considering binary collision events
and accounting for the corresponding momentum and angular momentum transfer.
Event-driven methods also have difficulties to describe long-lasting contacts. However,
very often one actually wants to compute forces and model contact phenomena that
may occur in granulates, rocks, and machine parts. An example of a system that can
be modeled with contact dynamics is a woodpecker toy (see Figure 20.1).
Simulating such and related systems requires one to resolve the observed contact
interactions. This is done in the context of contact dynamics. The basic idea is that
particle interactions are fully determined by constraints. To avoid overlapping par-
ticles, we impose constraint forces at their contacts. Per Lötstedt first described the
Figure 20.1 A woodpecker toy is a idea of the method and Jean Jacques Moreau (see Figure 20.2) provided a concrete
benchmark example of contact dynamics formulation and wrote the first working program. Contact dynamics is an applica-
simulations [297]. An animation can be tion of nonsmooth mechanics where the time evolution of the particles’ positions
found at https://github.com/gabyx/ and momenta are not assumed to be described by smooth functions anymore. An
Woodpecker. important problem in this field is the ambiguous boundary condition problem or Sig-
norini problem which has been originally formulated by Alberto Signorini. It consists
of determining the shape of a rigid body. Perfectly rigid bodies with perfect volume
exclusion are described by the Signorini graph. If particles are not in contact, there are
also no forces between them. Forces only occur if the distance between both particles
is zero. As we show in Figure 20.3, this causes force discontinuities which are diffi-
cult to handle numerically. Such force discontinuities may also occur due to friction
between particles. Let us consider two particles that are in contact. If their relative
velocity is zero, many values of the friction forces are possible. We show the resulting
discontinuous behavior in the Coulomb graph in Figure 20.3.
20.1 One-Dimensional Contact
To familiarize ourselves with the concepts of contact dynamics, we first focus on

the example of a one-dimensional contact as shown in Figure 20.4. In this example,
208
we neglect friction and only consider normal forces at the contact point of the two
particles. Both particles have mass m and velocities v1 and v2 , respectively. We have
to make sure that these two particles do not overlap. Therefore, we impose constraint
forces Ri (t) in such a way that they compensate all other forces which would lead to
overlaps. These constraint forces should be defined in such a way that they have no
influence on the particle dynamics before and after the contact. The time evolution of
the particles’ positions and velocities is described by an implicit Euler scheme which
is given by
1
vi (t + Δt) = vi (t) + Δt Fi (t + Δt) ,
mi (20.1)
ri (t + Δt) = ri (t) + Δtvi (t + Δt) . Figure 20.2 Jean Jacques Moreau
(1923–2014) was working in Montpelier
The forces Fi (t) = Fext
i (t) + Ri (t) consist of an external force Fi (t) and a contact
ext
and mainly focused on elliptic equations.
force Ri (t). After his retirement, he developed
So far, we only considered forces that act on the center of mass. However, contact contact dynamics [298]. Photograph
forces act locally on the contact point and not on the center of mass. We therefore courtesy Franck (Farhang) Radjai.
introduce a matrix H which transforms local contact forces into particle forces, and
the corresponding transpose H T transforms particle velocities into relative velocities.
We proceed with example of Figure 20.4 and find
⎡ ⎤ ⎡ ⎤
⎢⎢v ⎥⎥ ⎢⎢v1 ⎥⎥⎥
⎢⎢⎢ 1 ⎥⎥⎥ T ⎢
vn ≡ v2 − v1 = −1 1 ⎢⎢⎣ ⎥⎥⎦ = H ⎢⎢⎢⎢⎣ ⎥⎥⎥⎥⎦
loc
(20.2)
v2 v2
Signorini (left) and Coulomb (right) graphs. Figure 20.3
A one-dimensional contact of two particles that are a distance d apart. Both particles have mass m and velocities v1 Figure 20.4
and v2 , respectively. Additional external forces are F1ext and F2ext .
210 Contact Dynamics
and
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢⎢⎢R1 ⎥⎥⎥ ⎢⎢⎢−Rloc ⎥⎥⎥ ⎢⎢⎢−1⎥⎥⎥
⎢⎢⎢ ⎥⎥⎥ = ⎢⎢⎢ n ⎥⎥⎥ = ⎢⎢⎢ ⎥⎥⎥ Rloc = HRloc , (20.3)
⎢⎣ ⎥⎦ ⎢⎣ loc ⎥⎦ ⎢⎣ ⎥⎦ n n
R2 Rn 1
where the superscript “loc” refers to variables in the space of contacts. Note that we
used momentum conservation in the first step of eq. (20.3). The equations of motion
for both particles are
⎡ ⎤ ⎡⎡ ⎤ ⎡ ⎤⎤
d ⎢⎢⎢⎢⎢v1 ⎥⎥⎥⎥⎥ 1 ⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢R1 ⎥⎥⎥⎥⎥ ⎢⎢⎢⎢⎢F1ext ⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥
⎢ ⎥ = ⎢⎢ ⎥ + ⎢ ⎥⎥ . (20.4)
dt ⎢⎣v2 ⎥⎦ m ⎢⎣⎢⎣R2 ⎥⎦ ⎢⎣F ext ⎥⎦⎥⎦
2
We combine this equation with the transformation of eq. (20.3) to obtain the equation
of motion of the contact
⎡⎡ ⎤ ⎡ ⎤⎤
1 ⎢⎢⎢⎢−1⎥⎥
dvloc ⎢⎢⎢⎢⎢⎢ ⎥⎥⎥ loc ⎢⎢⎢⎢⎢F1ext ⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥
n
= −1 1 ⎢⎢ ⎥ R + ⎢⎢⎣ ⎥⎥⎥⎥
dt m ⎢⎣⎢⎣ 1 ⎥⎦ n F2ext ⎦⎦ (20.5)
1 loc 1 ext
= R + F − F1 , ext
meff n m 2

where meff = m/2 is the effective mass and m1 F2ext − F1ext is the acceleration without
contact forces. We integrate the prior equation with an implicit Euler method (see
eq. (20.1)) and find
n (t + Δt) − vn (t)
vloc loc
1 loc 1 ext
= Rn (t + Δt) + F2 − F1ext . (20.6)
Δt meff m
The unknown quantities in this equation are vloc n (t + Δt) and Rn (t + Δt). First, one can
loc
loc, free
compute the relative velocity vn (t + Δt) and the local contact force Rloc
n (t + Δt) in
the absence of any constraints:
1 ext
vloc,
n
free
(t + Δt) = vloc
n (t) + Δt F2 − F1ext (20.7)
m
and
loc, free
n (t + Δt) − vn
vloc (t + Δt)
n (t + Δt) = meff
Rloc . (20.8)
Δt
To find a solution, we make use of the Signorini constraint. For particles that are not in
contact, the contact force vanishes and the relative velocity is different from zero. This
case corresponds to the open solution. In the case of a persisting contact, the contact
force is different from zero and the relative velocity vanishes. We distinguish between
the possible cases:
• particles are not in contact,

• particles are closing contact,
• particles are in persisting contact,

• particles are opening contact.
It is not necessary to account for the influence of contact forces if the particles are not
in contact. However, for particles that are approaching each other, we have to take into
account the possibility of overlaps at every time step. In the case of a potential overlap,
constraint forces have to be applied such that the overlap vanishes. No overlap (e. g.,
no contact or closing contact) at time t + Δt corresponds to
n (t + Δt)Δt ≥ 0,
d(t + Δt) = d(t) + vloc (20.9)
n (t+Δt) = 0. A vanishing or negative distance d describes persisting and forming

and Rloc
contacts during the time step (i. e., the gap d closes). Thus, we impose
d(t)
n (t + Δt) = −
vloc . (20.10)
Δt
The contact force is then given by
−d(t)/Δt − vloc,
n
free
(t + Δt)
n (t + Δt) = meff
Rloc . (20.11)
Δt
The Signorini condition can be reformulated in terms of velocities to find the inter-
section of the Signorini graph with eq. (20.8) (see Figure 20.5). In addition to the
normal contact forces, we may also consider tangential contact forces to model
friction. For simplicity, we may assume that the static and dynamic friction coef-
ficients are equal (μ s = μd ). Similar to the solution of eq. (20.8), the solution for
the tangential contact force can be obtained with the help of the Coulomb graph of
Figure 20.3.
Intersection of eq. (20.8) with the Signorini graph (shifted by −d/Δt). The slope of the black solid line is meff /Δt, Figure 20.5
n = vn n = 0 (see
loc, free
and the intersection point between this line and the Signorini graph occurs at vloc and Rloc
eq. (20.8)). This solution (i. e., no constraint forces) corresponds to no contact or an opening contact. For a
n = −d/Δt (see eq. (20.10)), but Rn
loc
persisting or forming contact (black dashed line), the local velocity is vloc
is finite.
20.2 Generalization to N Particles
To describe the interactions of N particles utilizing contact dynamics, we use the

generalized coordinates
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
⎜⎜⎜ v1 ⎟⎟⎟ ⎜⎜⎜ R1 ⎟⎟⎟ ⎜⎜⎜Fext ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ 1 ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟
⎜⎜⎜⎜ ω1 ⎟⎟⎟⎟ ⎜⎜⎜ T1 ⎟⎟⎟ ⎜⎜⎜ 0 ⎟⎟⎟⎟⎟
⎜⎜ . ⎟⎟ ⎜
⎜⎜ . ⎟⎟ ⎟ ⎜
⎜⎜ . ⎟⎟⎟
q̇ = ⎜⎜⎜⎜ .. ⎟⎟⎟⎟ , R = ⎜⎜⎜⎜ .. ⎟⎟⎟⎟ , and F ext = ⎜⎜⎜⎜ .. ⎟⎟⎟⎟ , (20.12)
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ext ⎟⎟⎟⎟⎟
⎜⎜⎜ vN ⎟⎟⎟ ⎜⎜⎜RN ⎟⎟⎟ ⎜⎜⎜FN ⎟⎟⎟
⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟
ωN TN 0 ⎠
where ωi are the angular velocities and Ti are the torques. The number of compo-
nents depends on the dimension of the simulated system. In two dimensions, we have
3N components (two translational and one rotational per particle), whereas in three
dimensions there are 6N components (three translational and three rotational per
particle). Let c be the number of contacts. In two dimensions, there exist 2c contact
variables (one normal and one tangential at each contact), while in three dimensions
the number of contact variables is 3c (one normal and two tangential for each contact).
The contact velocities and forces are
⎡ ⎤ ⎡ ⎤
⎢⎢⎢vloc ⎥⎥⎥ ⎢⎢⎢Rloc ⎥⎥⎥
⎢⎢⎢⎢ ⎥⎥⎥⎥
1 ⎢⎢⎢⎢ 1 ⎥⎥
. . ⎥⎥
u = ⎢⎢⎢⎢⎢ .. ⎥⎥⎥⎥⎥ and Rloc = ⎢⎢⎢⎢⎢ .. ⎥⎥⎥⎥⎥ . (20.13)
⎢⎢⎢ ⎥⎥⎥ ⎢⎢⎢ ⎥
⎣ loc ⎦ ⎣ loc ⎥⎥⎦
vc Rc
The corresponding transformation from contact into particle variables is given by
u = H T q̇ and R = HRloc . (20.14)
The matrix H is not constant as in the case of the one-dimensional interaction of two
particles. At every time step, the particles involved in a contact change; some contacts
open and some new contacts form. The dimension of the matrix H is 2c × 3N in two
dimensions and 3c × 6N in three dimensions.
In two dimensions, we use the diagonal mass
⎛ ⎞ ⎛ ⎞
⎜⎜⎜ξ1 · · · 0 ⎟⎟⎟ ⎜⎜⎜mi 0 0 ⎟⎟⎟
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎜ ⎟⎟⎟
⎜. .⎟
M = ⎜⎜⎜⎜⎜ .. . . . .. ⎟⎟⎟⎟⎟ with ξi = ⎜⎜⎜⎜⎜ 0 mi 0 ⎟⎟⎟⎟⎟ . (20.15)
⎜⎜⎜ ⎟⎟⎟ ⎜⎜⎝ ⎟⎟⎠
⎝ ⎠
0 · · · ξN 0 0 Ii
to obtain
M q̈ (t) = R (t) + Fext . (20.16)
Based on the relation

u̇ = H T q̈, (20.17)
20.2 Generalization to N Particles 213
we find

u̇ = H T M −1 R (t) + M −1 Fext
(20.18)
= H T M −1 R (t) + H T M −1 Fext .
Using contact variables yields
u̇ = H T M −1 HRloc + H T M −1 Fext . (20.19)
We define the effective inverse mass matrix
−1
Meff = H T M −1 H, (20.20)
and solve the resulting system of equations with an implicit Euler scheme (as for the
one-dimensional contact):
u (t + Δt) − ufree (t + Δt)
Rloc (t + Δt) = Meff . (20.21)
Δt
The dimensions of the vectors change over time because the number of contacts is
−1
not constant. To solve eq. (20.21), we perform a direct inversion of Meff = H T M −1 H
and to determine u(t + Δt), we have to account for all constraints by checking if the
contacts close. Friction, rolling friction, and cohesion can be simulated by modifying
the constraint forces.
Up until now, we have only focused on contact dynamics simulations of spheri-
cal particles. As described in Section 19.3, nonspherical objects can be composed of
spherical objects (e. g., by introducing certain constraints between particles). For gen-
eral nonspherical rigid objects of finite volume (e. g., polyhedra), one can generalize
the above procedure systematically by determining the conditions for touching and
slipless motion for each contact using the technique described in Chapter 19 [299].
Since in general the number of particle variables is larger than the number of contact
variables, the solution for particle positions and velocities is not unique. Obtaining the
physically correct solution requires excluding unphysical cases by imposing adequate
conditions, which adds some subtlety to the programming. For further reading, we
recommend the textbook The Physics of Granular Media [300].
Exercise: Contact Dynamics
1. Implement the contact dynamics for spherical particles either in 2D or 3D. For more information, see
Ref. [300].
2. Consider the following simplifications:
• Use a frictionless contact because then the transformation matrices are very simple even for the 3D case.
Then you do not need angular velocities.
• For the frictionless contact, there is only a normal force component and one constraint (Signorini). That
is, the force calculation is the same as described in Section 20.1 for a 1D contact.
Exercise: Contact Dynamics (cont.)
• Consider particles with the same mass m and radius R (possible extension: vary R).
• For dealing with contact networks (more than one contact per particle), when calculating the force at
one contact, treat the other contact forces acting on the two particles involved as external forces.
• Start with a verification of your force calculation within a system including only two particles, or one
particle (under the influence of gravitation) and a wall, or similar systems.
• Then apply the method to several particles, where the forces for each time step have to be calculated
iteratively. The simplest implementation is to use a fixed number of iterations for each time step.
• A simple 3D example would be settling of particles under gravity, supported by a wall in vertical direc-
tion and periodic boundaries in horizontal directions (alternatively you can also use walls as horizontal
boundaries instead).
• You could first use a 2D setup within your 3D code to test it. Note that for settling of frictionless disks with
the same radius one gets strong crystallization effects, which are avoided in 3D (or using disks with slightly
different radii in 2D).
• Using this iterative scheme for the force calculation, the exact solution will usually not be reached, leading
to small particle overlaps. The average and maximal overlap in the system can be used as an simulation
error measure.
• To avoid an artificially large repulsive force by these overlaps, one typically treats overlapping contacts as
being directly in contact (in practice: set distance d to zero if d < 0), so that the contact force only acts to
avoid further overlap (not compensating already existing overlap).
• When using a fixed number of iterations NI the system macroscopically behaves like a system of elas-
tic particles. For a given time step Δt, the effective elasticity scales like κ ∝ NI /(Δt)2 . That is, you
can either use a large number of iterations or a small time step to reduce “quasi-elasticity.” As within
the force iteration, the information is spreading diffusively, equalizing the diffusion length of this pro-
cess. The maximal system extension Lmax (measured in number of particles with diameter dparticle )
leads to a lower bound for the number of iterations needed to really model perfectly rigid particles:
NI > (Lmax /dparticle )2 [301].
• It is useful to store the total force Ri acting on a particle i due to all its contacts in a contact network. This is
particularly helpful when calculating the contact force at a contact considering all other contact forces like
external forces. Before updating the force at one contact (involving particles i and j), one simply subtracts
the current contact’s contribution from Ri and R j (for both particles i and j) involved in the contact,
defining R i and R j . Then calculate the force at the contact by setting Fext
i + Ri and F j + R j as
ext
external forces, and add the new contact’s contribution. Then you can continue in the same way with the
next contact.
Discrete Fluid Models 21
There are two possibilities to model fluid flow. First, in a continuum approach the fluid
is described by its equations of motion, namely, the Navier–Stokes equations. Second,
it is possible to use particle-based models of fluids. These two approaches are equiva-
lent, since they are both based on local momentum conservation. In fact, there exist
many different discrete particle methods today, including
• lattice gas automata (LGA),

• lattice Boltzmann method (LBM),
• stochastic rotation dynamics (SRD),
• direct simulation Monte Carlo (DSMC),
• dissipative particle dynamics (DPD),
• smooth particle hydrodynamics (SPH).
21.1 Lattice Gas Automata
Lattice gas automata (LGA) use cellular automata methods to simulate fluid flows.
When particles meet, they follow the collision rules shown in Figure 21.1. With a
Chapman–Enskog expansion, we can retrieve the Navier–Stokes equations in the
continuum limit of LGA [302].
Unfortunately, the anisotropy of a certain lattice has strong influence on LGA.
In two dimensions, the only triangular lattice is Galilean invariant and there exists
no regular three-dimensional Galilean-invariant lattice. Only by making a projection
from a four-dimensional lattice, one can achieve Galilean-invariant three-dimensional
simulations. For these reasons, LGA has been of little practical use.
21.2 Lattice Boltzmann Method
Based on the Chapman–Enskog theory, it is possible to derive the Navier–Stokes equa-

tions from the Boltzmann equation [303]. This connection between fluid dynamics
and Boltzmann transport theory allows us to simulate the motion of fluids by solv-
ing the corresponding Boltzmann equation on a lattice with a discrete set of velocities
vi , where i labels a certain bond. In the lattice Boltzmann method (LBM), we use a
215
216 Discrete Fluid Models
tFigure 21.1 Rules for collisions in a lattice gas automaton. The rules are analogously applied in the other lattice directions. Only in
the shown cases, momentum is exchanged – in all other cases, particles do not interact. The upper rule is probabilistic
in the sense that the black and the gray outcomes occur each with probability 1/2.
velocity distribution function f (x, vi , t) to describe the density of particles with velocity
vi at position x and time t. The mass and momentum densities are

ρ(x, t) = f (x, vi , t) and ρu(x, t) = vi f (x, vi , t) , (21.1)
i i
respectively. The velocity distribution can be updated following Bhatnagar, Gross, and
Krook (BGK) [304] as
1 0 eq 1
f (x + Δtvi , vi , t + Δt) = f (x, vi , t) + fi − f (x, vi , t) Δt , (21.2)
τ
where fieq is the equilibrium distribution and τ a characteristic relaxation time scale.
An expansion of the Maxwell–Boltzmann distribution in Hermite polynomials is used
as the equilibrium distribution

3 9 3
fieq = wi ρ 1 + 2 uvi + 4 (uvi ) − 2 uu ,
2
(21.3)
cs 2c s 2c s
where c2s = (1/3)Δx2 /Δt2 is the lattice speed and wi are weights that depend on the
Figure 21.2 Directions of the lattice lattice and on the direction of the bond. The weights are obtained through Gaussian
Boltzmann weights on a square lattice quadrature (see the gray box on page 217). One possible choice of the weights in two
with next-nearest neighbors (D2Q9). dimensions is (D2Q9, see Figure 21.2)
21.2 Lattice Boltzmann Method 217
⎧
⎪
⎪
⎪ 4/9 i = 0,
⎪
⎪
⎨
wi = ⎪
⎪ 1/9 i = 1, 2, 3, 4 , (21.4)
⎪
⎪
⎪
⎩1/36 i = 5, 6, 7, 8 .
The first step in a lattice Boltzmann simulation is the so-called collision (or
relaxation) step:
1 0 eq 1
f ∗ (x, vi , t) = f (x, vi , t) + fi − f (x, vi , t) Δt . (21.5)
τ
The second step is propagation (or streaming):
f (x + Δtvi , vi , t + Δt) = f ∗ (x, vi , t) . (21.6)
Implementing the LBM on a computer is straightforward.
Lattice Boltzmann Method
1. Compute ρ(x, t) and u(x, t) according to eq. (21.1).

2. Determine the equilibrium distribution fieq (see eq. (21.3)).
3. Perform collision (or relaxation) step (see eq. (21.5)).
4. Perform propagation (or relaxation) step (see eq. (21.6)).
5. Advance time from t to t + Δt and go back to step 1.
In the D2Q9 model, the characteristic relaxation time scale τ and kinematic
viscosity ν are connected via
2τ − 1 2
ν= c s Δt . (21.7)
6
The LBM is very well suited to simulate fluids in complex geometries like porous
media and due to the underlying regular lattice it is particularly well suited for Figure 21.3 Sauro Succi is working at the
parallelization. For a detailed treatment of lattice Boltzmann approaches, we rec- Italian Institute of Technology.
ommend The Lattice Boltzmann Equation: For Fluid Dynamics and Beyond [303] (see
Figure 21.3).
Additional Information: Derivation of Lattice Boltzmann Weights

For the derivation of the weights wi , we start with the Gaussian quadrature theorem. For a polynomial g(x)
of at most degree 2n + 1 and weight function w(x), we have
b n
g(x)w(x) dx = wi g(xi ), (21.8)
a i=0
with
<n
b
x − xk
wi = w(x) dx , i = 0, . . . , n, (21.9)
a x − xk
ki i
Additional Information: Derivation of Lattice Boltzmann Weights (cont.)

if there exists a polynomial p(x) of degree n + 1 such that
b
xk p(x)w(x) dx = 0 , ∀k = 0, . . . , n, (21.10)
a
and xi are the zeros of p(x). The equilibrium distribution f eq (v) is given by the Maxwell–Boltzmann velocity
distribution
ρ − (v−u)
2
f eq (v) = e 2c2s
. (21.11)
m(2πc2s )d/2
A Taylor expansion for small ||u||/c s yields

ρ − v2
2
vu (vu)2 u2
f eq (v) ≈ e 2c s
1 + + − . (21.12)
m(2πc2s )d/2 c2s 2c4s 2c2s
Identifying the expansion in Hermite polynomials in the square brackets of eq. (21.12) with p(x) and the
prefactor with w(x), one can use the above theorem of Gaussian quadrature to determine the weights wi
and thus approximate eq. (21.11).
Exercise: Flow around an Obstacle

Use the lattice Boltzmann method to simulate the fluid flow through a two-dimensional channel around a
rectangular obstacle in a two-dimensional box. Discretize the computational domain using a square lattice
and apply no-slip boundary conditions at the obstacle and walls. The bottom and top parts of the box are
solid walls. To create a fluid flow with a certain entry velocity, set the fluid velocity at the left boundary of the
system to a constant value. The resulting flow pattern is known under the name “von Kárman vortex street”
Figure 21.4 Example of a two
(see Figure 21.4).
dimensional vortex street simulation in
which a fluid flows around a circular
Implement the D2Q9 method and simulate vortex streets for different Reynolds numbers
obstacle (white) [305].
uD
Re = ,
ν
where D is the width of the channel and u the flow velocity. You can use eq. (21.7) to determine
a characteristic relaxation time τ that corresponds to a certain viscosity ν and thus fix the Reynolds
number.
21.3 Stochastic Rotation Dynamics
Stochastic rotation dynamics (SRD) is a particle-based fluid modeling approach [306].

This method has been developed by Malevanets and Kapral (see Figure 21.5). It is
also known as multiparticle collision dynamics (MPC). In this method, space is being
21.4 Direct Simulation Monte Carlo 219
Stochastic Rotation Dynamics

The particle positions and velocities are updated according to
x i = xi + Δtvi , (21.13)
v i = u + Ω (vi − u) , (21.14)
where u = v is the mean velocity of particles in the respective cell and Ω is the rotation matrix. It is
given by
⎛ ⎞
⎜⎜⎜ cos (α) ± sin (α) 0⎟⎟⎟
⎜⎜⎜⎜ ⎟⎟⎟⎟
Ω = ⎜⎜⎜⎜∓ sin (α) cos (α) 0⎟⎟⎟⎟ , (21.15)
⎜⎜⎜ ⎟⎟⎟
⎝ ⎠
0 0 1
where α is a randomly chosen angle. Figure 21.5 Raymond Kapral is a
professor of chemistry at the University of
Toronto.
discretized into cells and fluids are composed of N point-like particles with mass m,
positions xi , and velocities vi .
The momentum exchange between particles is modeled by rotations of local particle
velocities. In this model, a certain stochastic motion is intrinsic, so that one can effec-
tively model small Péclet numbers. Since eq. (21.14) conserves the momentum after
coarse graining, SRD recovers hydrodynamics correctly.
21.4 Direct Simulation Monte Carlo
Direct simulation Monte Carlo (DSMC), as introduced by Graeme Bird, is a particle-

based simulation technique, which is appropriate to model rarified gases (i.e., particle
systems at large Knudsen numbers) [307, 308]
λ
Kn = , (21.16)
L
where λ is the mean free path and L is a characteristic system length scale. This method
is very popular in aerospace modeling because the atmosphere has very low density at
high altitudes and the corresponding Knudsen numbers are large.
In this method, we divide the system into cells and generate particles in each cell
according to a desired density. To impose fluid velocities and temperatures, we assign
a velocity to each particle which is distributed according to the Maxwell–Boltzmann
distribution. In DSMC the number of point-like simulation particles (simulators) is
much smaller than the number of physical molecules. Each simulator represents a
certain number of physical molecules. For traditional DSMC, about 20 particles per
collision cell is the rule of thumb.
Direct Simulation Monte Carlo

Collisions are modeled by sorting particles into cells. We iterate over all cells and
1. randomly select collision partners within a cell,
2. implement a collision that conserves momentum.
We note that collision pairs with a large relative velocity are more likely to col-
lide. The temperature T W may be implemented via collisions with a thermal wall. This
wall resets the momentum of a particle by choosing it from a Maxwell–Boltzmann
distribution such as
m − mv x
2
Pvx (v x ) = ± v x e 2kB TW ,
kB T W
* m m(v −u )2
− y W
Pvy vy = e 2kB TW , (21.17)
2πkB T W
*
m mv 2
− z
Pvz (vz ) = e 2kB TW .
2πkB T W
Here, uW is the velocity of the thermal wall in y-direction. For more information on
DSMC, we refer the reader to Bird’s textbook [308].
21.5 Dissipative Particle Dynamics
Another particle-based fluid simulation technique is dissipative particle dynamics

(DPD) [309]. In this method, particle interactions are described by

Fi = fiCj + fiRj + fiDj , (21.18)
i j
where fiCj denotes the conservative forces, fiRj a random force, and fiDj the dissipative
forces. In DPD, momentum is conserved. The dissipative forces are proportional to
the particle velocities and the weights of the random and dissipative forces must be
chosen such that thermal equilibrium is reached. More details on DPD and asso-
ciated remaining challenges can be found in Ref. [310]. This method is mostly used
Figure 21.6 Julia M. Yeomans is a in soft condensed matter physics to study polymers and colloids. Important contri-
professor of physics at the University of butions in this field were made by Julia Yeomans (see Figure 21.6) [311], who also
Oxford and works on the physics of soft contributed to the development of other discrete fluid models, including LBM and
matter and biological systems. SRD [312].
21.6 Smoothed Particle Hydrodynamics 221
21.6 Smoothed Particle Hydrodynamics
Another important technique in the field of computational fluid dynamics is smoothed

particle hydrodynamics (SPH) [313]. This method uses smooth kernel functions W to
represent properties of particles in a weighted sense [314, 315]. Instead of localized
positions and velocities, the particle characteristics are smoothened over a smoothing
length h. The value of an arbitrary quantity A is then given by
- - : ; m j - -
A (r) = W --r − r -- , h A r dr ≈ W --r − r j -- , h A j . (21.19)
Ω j
ρj
In this method, no spatial discretization is necessary and even free surfaces and fluid
drops can be simulated easily with SPH. This makes this method broadly applicable
in many different fields where fluids exhibit interfaces.
Examples of kernel functions include Gaussians and spline functions like [316]
⎧
⎪
⎪
⎪ 1 − 32 q2 + 34 q3 , if 0 ≤ q ≤ 1 ,
⎪
⎪
σ ⎨1
W (r, h) = d ⎪ 4 (2 − q) ,
3
if 1 < q ≤ 2 , (21.20)
h ⎪
⎪
⎪
⎪
⎩0 , otherwise ,
where d is the system dimension, σ = [ 32 , 7π , π ], q = hr , and r = |ra − rb |. Kernels may

10 1
be changed without much effort for different purposes [317].

22 Ab Initio Simulations
Simulating interactions of (classical) objects such as billiard balls, sand grains, and
stars is based on classical mechanics (i. e., solving Newton’s equations of motion). To
a certain extent, it is also possible to consider classical approaches for the simulation
of atoms and molecules if one can describe interactions well by a known potential
or if the bond length can be assumed to be constant given certain environmental
conditions (see Section 14.1). However, for an ab initio (i. e., from first principles) sim-
ulation of molecular and atomic interactions, we have to take quantum-mechanical
effects into account. In the subsequent sections, we therefore briefly discuss the main
quantum-mechanical methods which are necessary for ab initio MD simulations. In
particular, we introduce some basic concepts of quantum mechanics including the
Born–Oppenheimer and Kohn–Sham approximations. We conclude this chapter by
discussing the Car–Parrinello method.
22.1 Introduction
In quantum mechanics, the wave function ψ(r, t) at position r is represented by a

vector |ψ that belongs to the Hilbert space L2 R3 , with scalar product

ψ|φ = ψ∗ (r) φ (r) dr . (22.1)
R3
In general, a quantum state is a vector |ψ of a Hilbert space with scalar product

ψ|φ. All states satisfy the normalization condition

ψ|ψ = ψ∗ (r) ψ (r) dr = 1. (22.2)
R3
Additional Information: Square-Integrable Functions

The Hilbert space L2 R3 is the space of square-integrable functions. That is, elements |ψ ∈ L2 R3
satisfy

ψ|ψ = ψ∗ (r) ψ (r) dr < ∞ .
R3
222
and are uniquely defined up to phase eiα with α ∈ [0, 2π). That is,
|ψ ∝ eiα |ψ . (22.3)
According to the statistical interpretation of the wave function (Born rule), the
probability density |ψ (r)|2 denotes the probability of finding a particle described by
ψ at position r. Observables are represented by self-adjoint operators A with the
expectation value ψ|A|ψ.
Given a Hamilton operator
p2
+ V (r) with p = −i∇ ,
H= (22.4)
2m
the time evolution of the function |ψt is described by the time-dependent Schrödinger
equation
d |ψt
i = H |ψt . (22.5)
dt
To determine the time evolution of ψt (r), we use separation of variables with ψt (r) =
ψ (r) φ (t) and find
φ (t) = e−iEt/ , (22.6)
where E is the expectation value of the Hamilton operator. The time-independent
Schrödinger equation of a single particle is thus
2 2
∇ |ψ + V (r) |ψ = E |ψ .
− (22.7)
2m
For simulating quantum-mechanical particle interactions, we have to consider the
many-body wave function of N particles
ψ (r1 , s1 , r2 , s2 , . . . , rN , sN ) , (22.8)
where si is the spin of particle i. Many-body operators A are defined by their action on
all individual particles. For the following discussion of many-body wave functions, we
use ψ (1, 2, . . . , N) to abbreviate the wave function of eq. (22.8) and A (1, 2, . . . , N) as
an abbreviation of a corresponding many-body operator. We now define the exchange
operator Pi j (1 ≤ i, j ≤ N) according to
Pi j ψ (1, 2, . . . , i, . . . , j, . . . , N) = ψ (1, 2, . . . , j, . . . , i, . . . , N) ,
(22.9)
Pi j A (1, 2, . . . , i, . . . , j, . . . , N) = A (1, 2, . . . , j, . . . , i, . . . , N) .
The Hamiltonian commutes with the exchange operation because it is invariant under
particle exchange. That is,
0 1
H, Pi j = 0 . (22.10)
−1 2
We note that Pi j = Pi j and Pi j = 1. The two possible eigenvalues are thus +1 and
−1, which are both realized in nature with the corresponding wave functions
⎧
⎪
⎪
⎨+ψ (1, 2, . . . , j, . . . , i, . . . , N) , (bosons) ,
ψ (1, 2, . . . , i, . . . , j, . . . , N) = ⎪
⎪ (22.11)
⎩−ψ (1, 2, . . . , j, . . . , i, . . . , N) , (fermions) .
Additional Information: Commutator

The commutator of two operators A and B is defined as [A, B] = AB − BA. If [A, B] = 0, the two
operators A and B commute since AB = BA.
The wave functions of bosons are completely symmetric under particle exchange,
whereas fermions exhibit completely antisymmetric wave functions. For fermions, this
implies that two particles cannot be in the same state since
ψ (1, 2, . . . , i, . . . , i, . . . , N) = −ψ (1, 2, . . . , i, . . . , i, . . . , N ) = 0 . (22.12)
This is the so-called Pauli exclusion principle. Note that a quantum-mechanical
description of N particles requires one to construct many-body wave functions by con-
sidering all possible particle permutations since all particles (unlike classical ones) are
indistinguishable.
For N noninteracting particles, the corresponding Hamilton operator is

N
p2i
H= Hi with Hi = + V (ri ) . (22.13)
i=1
2m
To construct many-body wave functions, we use ψν to denote the eigenstates of the
single-particle Hamiltonian such that
Hi ψν (ri , si ) = ν ψν (ri , si ) , (22.14)
where ν is some quantum number to label the states. For bosons, the many-body wave
function is
1, 2, . . . , N|ψB = ψB (1, 2, . . . , N)

= P ψν1 (r1 , s1 ) ψν2 (r2 , s2 ) · · · ψνN (rN , sN ) , (22.15)
P∈S N
and for fermions we find

1, 2, . . . , N|ψF = ψF (1, 2, . . . , N)

= sign (P) P ψν1 (r1 , s1 ) ψν2 (r2 , s2 ) · · · ψνN (rN , sN ) . (22.16)
P∈S N
Here, P is the exchange operator for the state νi and sign (P) is the sign of the per-
mutation P which is +1 and −1, if P is composed of an even and odd number of
transpositions, respectively. For N particles, the set of possible exchange operators is
described by the symmetric group S N . For fermions, it is possible to rewrite eq. (22.16)
in terms of the so-called Slater determinant
--⎛ ⎞-
--⎜⎜⎜ ψν1 (1) · · · ψνN (1) ⎟⎟⎟---
--⎜⎜⎜⎜ ⎟⎟-
. .. .. ⎟⎟⎟⎟--
ψF (1, . . . , N) = --⎜⎜⎜⎜⎜ .. . . ⎟⎟⎟⎟-- . (22.17)
--⎜⎜ ⎟-
--⎝⎜ ⎟
⎠-
ψν1 (N) · · · ψνN (N) -
22.3 Born–Oppenheimer Approximation 225
We still have to normalize the many-body wave functions by dividing them by

ψB |ψB = N!nν1 !nν2 ! · · · nνN ! (bosons) (22.18)
and
ψF |ψF = N! (fermions) . (22.19)
For bosons, a certain index νi may appear multiple times in eq. (22.15). Therefore, we
have to account for the additional normalization factors nνi which denote the number
of particles in the stationary single particle state νi . For fermions, the occupation num-
bers nνi are either zero or one. The calculation of eq. (22.17) is computationally very
time consuming.
22.2 Implementation of Wave Functions
Single particle wave functions are often expanded in an orthonormal basis

ψj = c jk χk , (22.20)
k
where {χk } is a set of orthonormal basis functions and c jk are the corresponding expan-
sion coefficients for ψ j in the considered basis. In solid state physics, a plane wave basis
χk = exp (ikx) is often used (plane wave functions are not localized and can therefore
describe valence electrons in metals).
We typically use as many plane waves as necessary until to obtain the energy with a
desired precision. For 16 water molecules, a cutoff of about 70 Ry (1 Ry = 2.179 . . . ×
10−18 J [318]) is needed. This corresponds typically to about 15,000 plain waves per
electron.
Using localized basis sets can help to decrease the number of basis functions for
covalent bondings. Wave functions are then dependent on the ion positions. One
possible choice are Gaussian-type orbitals (GTOs)

χl (r) = cl rl exp −αr2 . (22.21)
For localized basis sets, so-called Pulay forces appear [319]. These forces are due to
numerical artifacts originating from a finite basis. They may be up to an order of
magnitude bigger than the physical forces and have to be corrected for.
22.3 Born–Oppenheimer Approximation
For the simulation of atomic and molecular interactions, we have to describe the
quantum-mechanical evolution equations of all involved particles. Instead of a full
quantum-mechanical description for both nuclei and electrons, we make use of the fact
that their masses differ by three orders of magnitude. In this way, we determine the
electron distribution for fixed ion positions. This is the so-called Born–Oppenheimer
approximation.
The Born–Oppenheimer approximation implies that the time step for the motion
of ions has to be sufficiently small, such that electrons do not skip any energy level
transition when ions move. Mathematically, the energies i of energy level i have to
satisfy
|i (r + Δr) − i (r)| |i+1 (r) − i (r)| , (22.22)
where r is the ion position before the update, Δr = uΔt, u is the ion velocity, and Δt is
the corresponding time step.
22.4 Hohenberg–Kohn Theorems
If the ground state is nondegenerate (i. e., there exists only one quantum state with the
energy of the ground state), all information on this ground state is contained in its
density distribution
N
n (r) = |ψi (r)|2 . (22.23)
i=1
It is possible to obtain all observables based on n (r). The knowledge of the wave func-
tions is not needed. To formalize this approach, we consider N electrons described
by a wave function Ψ with normalization Ψ|Ψ = N. These electrons orbit around a
nucleus with potential
Ze2
V (r) = − , (22.24)
|r − ri |
Figure 22.1 Pierre C. Hohenberg i
(1934–2017) worked at different where Z is the nuclear charge number and e the elementary charge. The corresponding
institutions including Bell Laboratories, Hamiltonian operator is given by
TU München, Leiden University, and
p2 1 e2
others. He spoke 12 languages. H= i
+ -- -- + V (ri ) . (22.25)
Photograph courtesy Michael Marsland, i
2m 2 i ji -ri − r j - i

Yale University. F V
We abbreviate the Hamiltonian by H = F + V, where F represents the kinetic and

electron–electron interaction energy, and V is the potential energy. The ground state
|Ψ0 is thus completely determined by N and V (r). We define the ground state particle
density as

n0 (r) = Ψ0 |ρ |Ψ0 = N |Ψ0 (r, r2 , . . . , rN )|2 dr2 dr3 · · · drN ,
(r) (22.26)

where ρ (r) = i δ (r − ri ) is the particle density operator.
22.4 Hohenberg–Kohn Theorems 227
The theorems of Hohenberg and Kohn [320] (see Figures 22.1 and 22.2) state that
(i) the external potential V is uniquely determined by the ground-state electron den-
sity, and that (ii) the ground state energy can be obtained by minimizing an energy
functional. Because of the Hohenberg–Kohn theorems, it is not necessary anymore to
consider wave functions for nondegenerate ground states. Instead, we can use a density
functional approach.
Additional Information: Proof of Hohenberg–Kohn Theorems

Claim (i) can be proven by contradiction. Thus, we assume that there exist two external potentials V (r) and
V (r) which give rise to the same particle density n0 (r). The corresponding Hamiltonians are H = F +V
and H = F + V and the corresponding ground-state energies are
E0 = Ψ0 |H|Ψ0 and E0 = Ψ 0 |H |Ψ 0 . (22.27)
With these definitions, we obtain
E0 < Ψ 0 |H|Ψ 0 = Ψ 0 |H |Ψ 0 + Ψ 0 | (H − H ) |Ψ 0

2 3 (22.28)
= E0 + n0 (r) V (r) − V (r) dr

and
E0 < Ψ0 |H |Ψ0 = Ψ0 |H|Ψ0 + Ψ0 | (H − H) |Ψ0

2 3 (22.29)
= E0 + n0 (r) V (r) − V (r) dr .
Adding Eqs. (22.28) and (22.29) leads to the contradiction

E0 + E0 < E0 + E0 . (22.30)
Thus, the potential V is (up to a constant) uniquely determined by n0 (r). We are now able to recast the
problem of solving the Schrödinger equation for our N particle system in variational form by defining the
energy functional

EV [n] = n (r) V (r) dr + F [n] , (22.31)
where F [n] = Ψ|F|Ψ. The particle density n (r) determines the potential V (r) and the ground-state
Ψ. For another potential Ṽ (r), we find
Ψ|H̃|Ψ = Ψ|F|Ψ + Ψ|Ṽ|Ψ

(22.32)
= F [n] + n (r) Ṽ (r) dr = EṼ [n] ≥ E0
according to the variational principle. The functional EṼ [n] equals E0 if Ψ is the ground-state
for Ṽ (r).
22.5 Kohn–Sham Approximation
Based on the Hohenberg–Kohn theorems, we are able to describe a nondegenerate

ground state with density functionals. This is the basis of density functional theory
(DFT). Earlier works by Thomas and Fermi were too inaccurate for most applica-
tions. Further developments by Walter Kohn and Lu Jeu Sham (see Figure 22.2)
made it possible to approximate the electron–electron interactions in terms of non-
interacting single particles moving in a potential which only depends on the density
distribution [321, 322]. Additional correction terms can be used to account for many-
particle effects. In the Kohn–Sham approximation, electrons are treated as single
particles interacting with the effective Kohn–Sham potential VKS . The corresponding
wave functions ψi (r) with i ∈ {1, . . . , N} and energies i obey
2

− ∇2 + VKS (r) ψi (r) = i ψi (r) . (22.33)
2m
The Kohn–Sham potential is given by

n (r )
VKS (r) = V (r) + e 2
dr + Vxc (r) , (22.34)
|r − r |
where V (r) is an external potential (e. g., a nucleus potential), Vxc (r) = δExc [n] /δn (r)
is a potential term accounting for exchange and correlation effects, and Exc (r) is the
exchange-correlation energy. To use the Kohn–Sham approximation in simulations,
we have to start with an initial guess for n (r) and then solve eq. (22.33) to obtain the
wave functions ψi (r). This allows us to determine the particle density (see eq. (22.23))
and the kinetic energy
Figure 22.2 Walter Kohn (1923–2016)
N
2
(top) and Lu Jeu Sham (bottom). Kohn
T s [n] = − ψ∗i (r) − ∇2 ψi (r) dr . (22.35)
was awarded the Nobel Prize in i=1
2m
Chemistry in 1998.
The orbital energies i have to sum up to

N
i = T s [n] + n (r) VKS (r) dr . (22.36)
i=1
Unlike many-particle orbitals, the eigenvalues of the single particle orbitals have no
physical meaning. The sum of them is not the total energy. The only justification of
the single particle wave functions is that they yield the correct density.
Until here, density functional theory is describing a many-body quantum system
in an exact way. However, for the computation of the exchange-correlation term
Exc [n] = Ex [n] + Ec [n] approximations have to be applied. Under the assumption of
a homogeneous electron gas, the exchange energy in the local-density approximation
(LDA) is
1/3
3 3
Ex [n] = −
LDA
n (r)4/3 dr . (22.37)
4 π
22.6 Hellmann–Feynman Theorem 229
For the correlation energy

EcLDA [n] = c [n (r)] dr , (22.38)
no analytical expression is known for c [n (r)]. LDA has been used to compute band
structures and the total energy of systems in solid state physics. In quantum chemistry,
it has been much less popular since the simulation of chemical bonds requires higher
accuracy. It is possible to improve the LDA by adding a dependence on the gradient of
the density. This approach is called general-gradient approximation (GGA). The phys-
ical intuition for this correction is that quantum-mechanical effects are very strong
when there is a change in the slope of the wave function and, in particular, when two
identical fermions come closer. The GGA approximation of the exchange energy is

ExGGA
[n] = x [n (r) , |∇n (r)|] dr . (22.39)
One possibility of a GGA approximation of the exchange energy is [323]

(∇n)2
E GGA
x [n] = E LDA
x [n] − β dr, (22.40)
n4/3
where β is an additional free parameter. Very popular is also another empirical

approach suggested by Axel Becke [324]

x2
E Becke [n] = E LDA [n] − β n4/3 dr , (22.41)
1 + 6β x sinh−1 (x)
x x
where x = |∇n| /n4/3 and β = 0.0042. The corresponding paper is among the most
cited ones in computational physics. The general-gradient approximation gives much
better results than LDA. They can be used to calculate all kinds of chemical bonds like
covalent, ionic, metallic, and hydrogen bridges. However, DFT approaches that utilize
GGA fail to describe van der Waals interactions properly. A possible improvement
has been suggested by Grimme [325], which leads to an additional energy term

N−1 N
C6i j 1
Edisp = −s6 fdmp Ri j with fdmp Ri j = , (22.42)
−d(Ri j /Rr −1)
R
i=1 j=i+1 i j 1+e
where s6 , C6i j , Rr , and d are model parameters and Ri j is the distance between two ions.
22.6 Hellmann–Feynman Theorem
With the previous methods, we are able to describe the motion of electrons in the
potential of the (slower) moving nuclei. After a sufficiently small time step, we also
have to update the position of the ions. Given a Hamiltonian operator H of the system,
the forces acting on the ions are (Hellmann–Feynman theorem) [326, 327]
. - /
d2 Rα -- ∂H ---
mα 2 = − ψ -- -ψ , (22.43)
dt ∂Rα -
where mα and Rα are the mass and position of ion α. Combining DFT approaches
with the Hellmann–Feynman theorem yields an iterative method for solving the
motion of molecules and other quantum-mechanical systems according to the
following steps:
Ab Initio MD
• Solve the electron dynamics in the ion potential.

• Solve the motion of the ions using classical MD techniques in combination with the Hellmann–Feynman
theorem.
The problem of this method is that the first step is computationally very expensive.
For this reason, the contribution of Car and Parrinello has been very important.
22.7 Car–Parrinello Method
Roberto Car and Michele Parrinello (see Figure 22.3) reformulated the ab initio MD
problem in terms of the Hamiltonian [328, 329]
0 6 7 6 71 MI μ
HCP {RI } , ṘI , {ψi } , ψ̇i = Ṙ2I + ψ̇i |ψ̇i
I
2 i
2 (22.44)
+ EKS ({RI } , {ψi }) + Eion ({RI }) ,
where MI , RI , and ψi are the masses and positions of the nuclei, and the orbital wave
functions, respectively. The parameter μ is a fictitious mass which is used to control
the energy transfer from nuclei to electrons. The remaining terms are the Kohn–Sham
energy EKS and the ion–ion interaction energy Eion . The Kohn–Sham energy is the
sum of the kinetic energy of eq. (22.35) and the potential energy of eq. (22.34). The
corresponding equations of motion are [330]
Car–Parrinello Equations of Motion

MI R̈I = −∇RI EKS + Λi j ∇RI ψi |ψ j ,
ij
δEKS δEion (22.45)
μψ̈i (r, t) = − − + Λi j ψ j (r, t) .
δψ∗i (r, t) δψ∗i (r, t) j
22.7 Car–Parrinello Method 231
The additional Lagrange multiplier Λi j is used to ensure the orthonormality constraint

ψi |ψ j = ψ∗i (r, t) ψ j (r, t) dr = δi j , (22.46)
where δi j is the Kronecker delta.

Similar to the Nosé–Hoover approach, ions and electrons are treated as two dif-
ferent systems coupled by the Hamiltonian as defined in eq. (22.44). The coupling
is realized with a fictitious mass μ which should not be confused with the electron
mass. The fictitious mass is a tunable parameter of the method describing how stiff
the electron motion is coupled to the nuclei. If μ → 0, the electronic response is very
rapid and the electrons remain to a sufficiently high degree in the ground state. But
numerically, we want to avoid too large accelerations, so electron dynamics is made
artificially slower than in reality. Since the electron configuration is now following an
equation of motion, the integration step has to be small enough to resolve the elec-
tronic motion. Since μ is usually smaller than the mass of the ions, time steps used in
the Car–Parrinello method are usually smaller than the time steps of the direct Born–
Oppenheimer approach. Often time steps of the order of a tenth of a femtosecond
are used.
Solving the Car–Parrinello equations requires much less computational effort per
time step than the previously discussed ab initio methods. In addition, the energy fluc-
tuations are smaller since we always consider the same bonds. However, it is difficult
to simulate light ions.
Figure 22.3 Roberto Car is a professor of
chemistry at Princeton University (top)
and Michele Parrinello was a professor at
ETH Zurich (bottom).
References
[1] D. Landau and K. Binder, A Guide to Monte Carlo Simulations in Statistical

Physics. Cambridge, UK: Cambridge University Press, 2009.
[2] D. H. Lehmer, “Mathematical methods in large-scale computing units,” Annu.
Comput. Lab. Harvard Univ., vol. 26, pp. 141–146, 1951.
[3] W. Payne, J. R. Rabung, and T. Bogyo, “Coding the Lehmer pseudo-random
number generator,” Communications of the ACM, vol. 12, no. 2, pp. 85–86, 1969.
[4] R. D. Carmichael, “Note on a new number theory function,” Bulletin of the
American Mathematical Society, vol. 16, no. 5, pp. 232–238, 1910.
[5] G. Woltman et al., “GIMPS, the Great Internet Mersenne Prime Search,” 2007,
www.mersenne.org/.
[6] S. K. Park and K. W. Miller, “Random number generators: Good ones are hard
to find,” Communications of the ACM, vol. 31, no. 10, pp. 1192–1201, 1988.
[7] G. Marsaglia, “Random number generators,” Journal of Modern Applied Statis-
tical Methods, vol. 2, no. 1, p. 2, 2003.
[8] G. Marsaglia “Random numbers fall mainly in the planes,” Proceedings of the
National Academy of Sciences of the United States of America, vol. 61, no. 1,
p. 25, 1968.
[9] D. E. Knuth, The Art of Computer Programming. Vol. 3. London, UK: Pearson
Education, 1997.
[10] R. C. Tausworthe, “Random numbers generated by linear recurrence modulo
two,” Mathematics of Computation, vol. 19, no. 90, pp. 201–209, 1965.
[11] S. Kirkpatrick and E. P. Stoll, “A very fast shift-register sequence random num-
ber generator,” Journal of Computational Physics, vol. 40, no. 2, pp. 517–526,
1981.
[12] J. R. Heringa, H. Blöte, and A. Compagner, “New primitive trinomials
of Mersenne-exponent degrees for random-number generation,” International
Journal of Modern Physics C, vol. 3, no. 03, pp. 561–564, 1992.
[13] R. P. Brent and P. Zimmermann, “Random number generators with period
divisible by a Mersenne prime,” in International Conference on Computational
Science and Its Applications. Berlin, Germany: Springer, 2003, pp. 1–10.
[14] M. Matsumoto and T. Nishimura, “Mersenne twister: a 623-dimensionally
equidistributed uniform pseudo-random number generator,” ACM Transactions
on Modeling and Computer Simulation (TOMACS), vol. 8, no. 1, pp. 3–30, 1998.
[15] G. Marsaglia, “The Marsaglia random number CD-ROM including the
Diehard Battery of Tests of Randomness.” [Online]. Available: https://web
.archive.org/web/20160125103112/http://stat.fsu.edu/pub/diehard/
232
References 233
[16] G. E. P. Box and M. E. Muller, “A note on the generation of random normal

deviates,” The Annals of Mathematical Statistics, vol. 29, pp. 610–611, 1958.
[17] S. R. Broadbent and J. M. Hammersley, “Percolation processes: I. crystals and
mazes,” in Mathematical Proceedings of the Cambridge Philosophical Society,
vol. 53, no. 3. Cambridge, UK: Cambridge University Press, 1957, pp. 629–641.
[18] D. Stauffer and A. Aharony, Introduction to Percolation Theory. Oxfordshire,
UK: Taylor & Francis, 2018.
[19] P. Grassberger, “On the critical behavior of the general epidemic process and
dynamical percolation,” Mathematical Biosciences, vol. 63, no. 2, pp. 157–172,
1983.
[20] C. Moore and M. E. Newman, “Epidemics and percolation in small-world
networks,” Physical Review E, vol. 61, no. 5, p. 5678, 2000.
[21] D. Sornette, Why Stock Markets Crash: Critical Events in Complex Financial
Systems. Vol. 49. Princeton, NJ: Princeton University Press, 2017.
[22] S. Galam, “Sociophysics: a review of Galam models,” International Journal of
Modern Physics C, vol. 19, no. 03, pp. 409–440, 2008.
[23] A. S. Perelson, “Immune network theory,” Immunological Reviews, vol. 110,
no. 1, pp. 5–36, 1989.
[24] B. Gauthier-Manuel, R. Meyer, and P. Pieranski, “The sphere rheometer:
I. Quasistatic measurements,” Journal of Physics E: Scientific Instruments,
vol. 17, no. 12, p. 1177, 1984.
[25] D. Stauffer, A. Coniglio, and M. Adam, “Gelation and critical phenomena,” in
Polymer Networks. Berlin, Germany: Springer, 1982, pp. 103–158.
[26] H. Herrmann, D. Hong, and H. Stanley, “Backbone and elastic backbone
of percolation clusters obtained by the new method of ‘burning’,” Journal of
Physics A: Mathematical and General, vol. 17, no. 5, p. L261, 1984.
[27] V. W. D. Stauffer, F. W. Hehl, and J. Zabolitzky, Computer Simulation and
Computer Algebra. Berlin, Germany: Springer, 1988.
[28] G. Grimmett, “What is percolation?,” in Percolation. Berlin, Germany: Springer,
1999, pp. 1–31.
[29] J. L. Jacobsen, “High-precision percolation thresholds and potts-model critical
manifolds from graph polynomials,” Journal of Physics A: Mathematical and
Theoretical, vol. 47, no. 13, Article 135001, 2014.
[30] J. W. Essam, “Percolation theory,” Reports on Progress in Physics, vol. 43, no. 7,
p. 833, 1980.
[31] M. F. Sykes and J. W. Essam, “Exact critical percolation probabilities for site
and bond problems in two dimensions,” Journal of Mathematical Physics, vol. 5,
no. 8, pp. 1117–1127, 1964.
[32] C. D. Lorenz and R. M. Ziff, “Universality of the excess number of clusters and
the crossing probability function in three-dimensional percolation,” Journal of
Physics A: Mathematical and General, vol. 31, no. 40, p. 8147, 1998.
[33] C.D. Lorenz and R.M. Ziff, “Precise determination of the bond percolation
thresholds and finite-size scaling corrections for the sc, fcc, and bcc lattices,”
Physical Review E, vol. 57, no. 1, p. 230, 1998.
234 References
[34] X. Xu, J. Wang, J.-P. Lv, and Y. Deng, “Simultaneous analysis of three-
dimensional percolation models,” Frontiers of Physics, vol. 9, no. 1, pp. 113–119,
2014.
[35] S. Mertens and C. Moore, “Percolation thresholds and fisher exponents in
hypercubic lattices,” Physical Review E, vol. 98, no. 2, Article 022120, 2018.
[36] J. Hoshen and R. Kopelman, “Percolation and cluster distribution. I. Clus-
ter multiple labeling technique and critical concentration algorithm,” Physical
Review B, vol. 14, no. 8, p. 3438, 1976.
[37] J. Gracey, “Four loop renormalization of φ 3 theory in six dimensions,” Physical
Review D, vol. 92, no. 2, Article 025012, 2015.
[38] D. Stauffer, “Scaling theory of percolation clusters,” Physics Reports, vol. 54,
no. 1, pp. 1–74, 1979.
[39] H. Hu, H. W. Blöte, R. M. Ziff, and Y. Deng, “Short-range correlations in
percolation at criticality,” Physical Review E, vol. 90, no. 4, Article 042106, 2014.
[40] H. Herrmann, D. Landau, and D. Stauffer, “New universality class for kinetic
gelation,” Physical Review Letters, vol. 49, no. 6, p. 412, 1982.
[41] A. E. Ferdinand and M. E. Fisher, “Bounded and inhomogeneous Ising mod-
els. I. Specific-heat anomaly of a finite lattice,” Physical Review, vol. 185, no. 2,
p. 832, 1969.
[42] Y. Deng and H. W. Blöte, “Monte Carlo study of the site-percolation model in
two and three dimensions,” Physical Review E, vol. 72, no. 1, Article 016126,
2005.
[43] G. Paul, R. M. Ziff, and H. E. Stanley, “Percolation threshold, fisher exponent,
and shortest path exponent for four and five dimensions,” Physical Review E,
vol. 64, no. 2, Article 026115, 2001.
[44] Z. Zhou, J. Yang, Y. Deng, and R. M. Ziff, “Shortest-path fractal dimension
for percolation in two and three dimensions,” Physical Review E, vol. 86, no. 6,
Article 061101, 2012.
[45] J. Chalupa, P. L. Leath, and G. R. Reich, “Bootstrap percolation on a Bethe
lattice,” Journal of Physics C: Solid State Physics, vol. 12, no. 1, p. L31, 1979.
[46] J. Adler, “Bootstrap percolation,” Physica A: Statistical Mechanics and its
Applications, vol. 171, no. 3, pp. 453–470, 1991.
[47] J. Adler and U. Lev, “Bootstrap percolation: visualizations and applications,”
Brazilian Journal of Physics, vol. 33, no. 3, pp. 641–644, 2003.
[48] B. B. Mandelbrot, Fractals: Form, Chance, and Dimension. Vol. 706. San
Francisco, CA: W. H. Freeman, 1977.
[49] D. Weitz and M. Oliveria, “Fractal structures formed by kinetic aggregation of
aqueous gold colloids,” Physical Review Letters, vol. 52, no. 16, p. 1433, 1984.
[50] D. Weitz, M. Lin, and C. Sandroff, “Colloidal aggregation revisited: new
insights based on fractal structure and surface-enhanced raman scattering,”
Surface Science, vol. 158, no. 1–3, pp. 147–164, 1985.
[51] H. Hentschel and I. Procaccia, “The infinite number of generalized dimensions
of fractals and strange attractors,” Physica D: Nonlinear Phenomena, vol. 8,
no. 3, pp. 435–444, 1983.
References 235
[52] T. C. Halsey, M. H. Jensen, L. P. Kadanoff, I. Procaccia, and B. I. Shraiman,

“Fractal measures and their singularities: The characterization of strange sets,”
Physical Review A, vol. 33, no. 2, p. 1141, 1986.
[53] A. Chhabra and R. V. Jensen, “Direct determination of the f (α) singularity
spectrum,” Physical Review Letters, vol. 62, no. 12, p. 1327, 1989.
[54] S. Xu, L. Böttcher, and T. Chou, “Diversity in biology: definitions, quantifica-
tion and models,” Physical Biology, vol. 17, no. 3, Article 031001, 2020.
[55] A. Kapitulnik, Y. Gefen, and A. Aharony, “On the fractal dimension and corre-
lations in percolation theory,” Journal of Statistical Physics, vol. 36, no. 5–6, pp.
807–814, 1984.
[56] S. Panyukov, “Correlation functions of an infinite cluster in percolation theory,”
Zhurnal Eksperimental’noi i Teoreticheskoi Fiziki, vol. 93, pp. 1454–1460, 1987.
[57] E. Sander, L. M. Sander, and R. M. Ziff, “Fractals and fractal correlations,”
Computers in Physics, vol. 8, no. 4, pp. 420–425, 1994.
[58] A. Einstein, “Über die von der molekularkinetischen Theorie der Wärme
geforderte Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen,”
Annalen der Physik, vol. 322, no. 8, pp. 549–560, 1905.
[59] M. Kac, “Random walk and the theory of Brownian motion,” The American
Mathematical Monthly, vol. 54, no. 7P1, pp. 369–391, 1947.
[60] B. B. Mandelbrot, “A multifractal walk down wall street,” Scientific American,
vol. 280, no. 2, pp. 70–73, 1999.
[61] B. G. Malkiel, A Random Walk down Wall Street: Including a Life-Cycle Guide
to Personal Investing. New York, NY: W. W. Norton, 1999.
[62] F. Bartumeus, M. G. E. da Luz, G. M. Viswanathan, and J. Catalan, “Animal
search strategies: a quantitative random-walk analysis,” Ecology, vol. 86, no. 11,
pp. 3078–3087, 2005.
[63] M. Rubinstein and R. H. Colby, Polymer Physics. Vol. 23. Oxford, UK: Oxford
University Press, 2003.
[64] P. D. Gujrati and A. I. Leonov, Modeling and Simulation in Polymers. Hoboken,
NJ: John Wiley, 2010.
[65] L. Page, S. Brin, R. Motwani, and T. Winograd, “The PageRank citation
ranking: bringing order to the web.” Stanford InfoLab, Tech. Rep., 1999.
[66] S. Wald and L. Böttcher, “From classical to quantum walks with stochastic
resetting on networks,” arXiv preprint arXiv:2008.00510, 2020.
[67] G. H. Weiss, Aspects and Applications of the Random Walk. Amsterdam,
Netherlands: North-Holland, 1994.
[68] D. Ben-Avraham and S. Havlin, Diffusion and Reactions in Fractals and Disor-
dered Systems. Cambridge, UK: Cambridge University Press, 2000.
[69] M. L. Boas, Mathematical Methods in the Physical Sciences. Hoboken, NJ: John
Wiley, 2006.
[70] G. F. Lawler, O. Schramm, and W. Werner, “The dimension of the planar
Brownian frontier is 4/3,” Mathematical Research Letters, vol. 8, no. 1, pp.
13–24, 2001.
[71] A. Klenke, Wahrscheinlichkeitstheorie. Vol. 1. Berlin, Germany: Springer, 2006.
236 References
[72] A. Dvoretzky and P. Erdös, “Some problems on random walk in space,” in

Proceedings of the Second Berkeley Symposium on Mathematical Statistics and
Probability, 1951, pp. 353–367.
[73] E. Almaas, R. Kulkarni, and D. Stroud, “Scaling properties of random walks on
small-world networks,” Physical Review E, vol. 68, no. 5, Article 056105, 2003.
[74] M. Shlesinger, B. West, and J. Klafter, “Lévy dynamics of enhanced diffusion:
application to turbulence,” Physical Review Letters, vol. 58, no. 11, p. 1100, 1987.
[75] P. J. Flory and M. Volkenstein, “Statistical mechanics of chain molecules,”
Biopolymers: Original Research on Biomolecules, vol. 8, no. 5, pp. 699–700, 1969.
[76] B. Nienhuis, “Exact critical point and critical exponents of o(n) models in two
dimensions,” Physical Review Letters, vol. 49, no. 15, p. 1062, 1982.
[77] N. Eizenberg and J. Klafter, “Critical exponents of self-avoiding walks in three
dimensions,” Physical Review B, vol. 53, no. 9, p. 5078, 1996.
[78] N. Madras and G. Slade, The Self-Avoiding Walk. Berlin, Germany: Springer
Science & Business Media, 2013.
[79] J. Millan and P. Prieto, eds., Some Selected Topics from Critical Phenomena,
1985, critical Phenomena Phase Transitions Supersymmetry (LASP-85), Cali,
Colombia.
[80] A. Guttmann, “On two-dimensional self-avoiding random walks,” Journal of
Physics A: Mathematical and General, vol. 17, no. 2, p. 455, 1984.
[81] I. Majid, N. Jan, A. Coniglio, and H. E. Stanley, “Kinetic growth walk: a new
model for linear polymers,” Physical Review Letters, vol. 52, no. 15, p. 1257,
1984.
[82] F. T. Wall and F. Mandel, “Macromolecular dimensions obtained by an efficient
Monte Carlo method without sample attrition,” Journal of Chemical Physics,
vol. 63, no. 11, pp. 4592–4595, 1975.
[83] P. H. Verdier and W. Stockmayer, “Monte Carlo calculations on the dynamics of
polymers in dilute solution,” Journal of Chemical Physics, vol. 36, no. 1, pp. 227–
235, 1962.
[84] N. Madras and A. D. Sokal, “The pivot algorithm: a highly efficient Monte
Carlo method for the self-avoiding walk,” Journal of Statistical Physics, vol. 50,
no. 1–2, pp. 109–186, 1988.
[85] S. N. Dorogovtsev and J. F. Mendes, Evolution of Networks: From Biological
Nets to the Internet and WWW. Oxford, UK: Oxford University Press, 2013.
[86] A.-L. Barabási, Network Science. Cambridge, UK: Cambridge University Press,
2016.
[87] M. Newman, Networks. Oxford, UK: Oxford University Press, 2018.
[88] J. S. Andrade Jr., H. J. Herrmann, R. F. Andrade, and L. R. Da Silva,
“Apollonian networks: simultaneously scale-free, small world, euclidean, space
filling, and with matching graphs.” Physical Review Letters, vol. 94, no. 1, Article
018702, 2005.
[89] M. Bastian, S. Heymann, and M. Jacomy, “Gephi: an open source software for
exploring and manipulating networks,” in Third international AAAI Conference
on Weblogs and Social Media, 2009.
References 237
[90] R. Albert and A.-L. Barabási, “Statistical mechanics of complex networks,”

Reviews of Modern Physics, vol. 74, no. 1, p. 47, 2002.
[91] A.-L. Barabási and R. Albert, “Emergence of scaling in random networks,”
Science, vol. 286, no. 5439, pp. 509–512, 1999.
[92] S. N. Dorogovtsev and J. F. Mendes, “Evolution of networks,” Advances in
Physics, vol. 51, no. 4, pp. 1079–1187, 2002.
[93] A.-L. Barabâsi, H. Jeong, Z. Néda, E. Ravasz, A. Schubert, and T. Vicsek, “Evo-
lution of the social network of scientific collaborations,” Physica A: Statistical
Mechanics and Its Applications, vol. 311, no. 3–4, pp. 590–614, 2002.
[94] S. Nauer, L. Böttcher, and M. A. Porter, “Random-graph models and charac-
terization of granular networks,” Journal of Complex Networks, vol. 8, no. 5,
cnz037, 2019.
[95] L. Böttcher, “Extending buffon’s needle problem to ‘random-line graphs’,” arXiv
preprint arXiv:1911.10679, 2019.
[96] M. E. Newman, S. H. Strogatz, and D. J. Watts, “Random graphs with arbitrary
degree distributions and their applications,” Physical Review E, vol. 64, no. 2,
Article 026118, 2001.
[97] D. J. Watts and S. H. Strogatz, “Collective dynamics of ‘small-world’ networks,”
Nature, vol. 393, no. 6684, p. 440, 1998.
[98] A. Barrat and M. Weigt, “On the properties of small-world network models,”
European Physics Journal B, vol. 13, no. 3, pp. 547–560, 2000.
[99] J. Saramäki, M. Kivelä, J.-P. Onnela, K. Kaski, and J. Kertesz, “Generalizations
of the clustering coefficient to weighted complex networks,” Physical Review E,
vol. 75, no. 2, Article 027105, 2007.
[100] R. Cohen and S. Havlin, “Scale-free networks are ultrasmall,” Physical Review
Letters, vol. 90, no. 5, Article 058701, 2003.
[101] A. Smart, P. Umbanhowar, and J. Ottino, “Effects of self-organization on
transport in granular matter: a network-based approach,” Europhysics Letters,
vol. 79, no. 2, p. 24002, 2007.
[102] A. Smart and J. M. Ottino, “Granular matter and networks: three related
examples,” Soft Matter, vol. 4, no. 11, pp. 2125–2131, 2008.
[103] L. Papadopoulos, M. A. Porter, K. E. Daniels, and D. S. Bassett, “Network
analysis of particles and grains,” Journal of Complex Networks, vol. 6, no. 4,
pp. 485–565, 2018.
[104] J. Bang-Jensen and G. Z. Gutin, Digraphs: Theory, Algorithms and Applications.
Berlin, Germany: Springer Science & Business Media, 2008.
[105] W. Ebeling and I. M. Sokolov, Statistical Thermodynamics and Stochastic
Theory of Nonequilibrium Systems. Vol. 8. Singapore: World Scientific, 2005.
[106] J. W. Gibbs, “On the fundamental formula of statistical mechanics, with appli-
cations to astronomy and thermodynamics,” in Proceedings of the American
Association for the Advancement of Science, 1884, pp. 57–58.
[107] R. V. Solé, Phase Transitions. Princeton, NJ: Princeton University Press, 2011.
[108] G. Jaeger, “The ehrenfest classification of phase transitions: introduction and
evolution,” Archive for History of Exact Sciences, vol. 53, no. 1, pp. 51–81, 1998.
238 References
[109] H. Stanley, Introduction to Phase Transitions and Critical Phenomena. Oxford,

UK: Clarendon, 1971.
[110] T. Ising, R. Folk, R. Kenna, B. Berche, and Y. Holovatch, “The fate of Ernst
Ising and the fate of his model,” Journal of Physical Studies, vol. 21, p. 3002,
2017.
[111] L. Onsager, “Crystal statistics. I. A two-dimensional model with an order-
disorder transition,” Physical Review, vol. 65, no. 3–4, p. 117, 1944.
[112] A. Böttcher and B. Silbermann, Introduction to Large Truncated Toeplitz
Matrices. Berlin, Germany: Springer Science & Business Media, 2012.
[113] C. N. Yang, “The spontaneous magnetization of a two-dimensional ising
model,” Physical Review, vol. 85, no. 5, p. 808, 1952.
[114] W. P. Wolf, “The Ising model and real magnetic materials,” Brazilian Journal of
Physics, vol. 30, no. 4, pp. 794–810, 2000.
[115] A. Pelissetto and E. Vicari, “Critical phenomena and renormalization-group
theory,” Physics Reports, vol. 368, no. 6, pp. 549–727, 2002.
[116] S. El-Showk, M. F. Paulos, D. Poland, S. Rychkov, D. Simmons-Duffin, and
A. Vichi, “Solving the 3D Ising model with the conformal bootstrap,” Physical
Review D, vol. 86, no. 2, Article 025022, 2012.
[117] A. M. Ferrenberg, J. Xu, and D. P. Landau, “Pushing the limits of Monte Carlo
simulations for the three-dimensional Ising model,” Physical Review E, vol. 97,
no. 4, Article 043301, 2018.
[118] F. Kos, D. Poland, D. Simmons-Duffin, and A. Vichi, “Precision islands in the
Ising and O(N) models,” Journal of High Energy Physics, vol. 2016, no. 8, p. 36,
2016.
[119] Z. Komargodski and D. Simmons-Duffin, “The random-bond Ising model in
2.01 and 3 dimensions,” Journal of Physics A: Mathematical and Theoretical,
vol. 50, no. 15, Article 154001, 2017.
[120] H. A. Kramers and G. H. Wannier, “Statistics of the two-dimensional ferromag-
net. Part I,” Physical Review, vol. 60, no. 3, p. 252, 1941.
[121] M. Henkel, H. Hinrichsen, and S. Lübeck, Non-Equilibrium Phase Transi-
tions, Vol. 1. Absorbing Phase Transitions. Berlin, Germany: Springer Science
& Business Media, 2008.
[122] B. M. McCoy and T. T. Wu, The Two-Dimensional Ising Model. North
Chelmsford, MA: Courier Corporation, 2014.
[123] H. Stanley, “Scaling, universality, and renormalization: Three pillars of modern
critical phenomena,” 1999. [Online]. Available: http://link.springer.com/chapter/
10.1007%2F978-1-4612-1512-7_39
[124] L. Kadanoff, “Critical behaviour, universality and scaling,” in M. S. Green, edi-
tor, Proceedings of the 1970 Varenna Summer School on Critical Phenomena,
1971.
[125] J. Sengers, J. L. Sengers, and C. Croxton, Progress in Liquid Physics. Chichester,
UK: Wiley, 1978.
[126] W. Krauth, Statistical Mechanics: Algorithms and Computations. Vol. 13.
Oxford, UK: Oxford University Press, 2006.
References 239
[127] W. H. Holtzman, “The unbiased estimate of the population variance and stan-
dard deviation,” American Journal of Psychology, vol. 63, no. 4, pp. 615–617,
1950.
[128] J. H. Conway and N. J. A. Sloane, Sphere Packings, Lattices and Groups, 3rd ed.,
vol. 290 Berlin, Germany: Springer Science & Business Media, 1998.
[129] W. Hastings, “Monte Carlo sampling methods using Markov chains and
their applications,” BioMetrika, 1969. [Online]. Available: http://biomet
.oxfordjournals.org/content/57/1/97
[130] N. Metropolis, A. W. Rosenbluth, M. Rosenbluth, A. Teller, and E. Teller,
“Equation of state calculations by fast computing machines,” Journal of Chem-
ical Physics, 1953. [Online]. Available: http://en.wikipedia.org/wiki/Equation_
of_State_Calculations_by_Fast_Computing_Machines
[131] K. Barth, Oral History Transcript – Dr. Marshall Rosenbluth, 2003. [Online].
Available: www.aip.org/history/ohilist/28636_1.html
[132] J. Gubernatis, “Marshall Rosenbluth and the Metropolis algorithm,” AIP,
2005. [Online]. Available: http://scitation.aip.org/content/aip/journal/pop/12/5/
10.1063/1.1887186
[133] E. Segré, Enrico Fermi, Physicist. Chicago, IL: University of Chicago Press,
1970.
[134] M. Creutz, “Microcanonical monte carlo simulation,” Physical Review Letters,
vol. 50, no. 19, p. 1411, 1983.
[135] S. Manna, H. Herrmann, and D. Landau, “A stochastic method to determine
the shape of a drop on a wall,” Journal of Statistical Physics, vol. 66, no. 3–4,
pp. 1155–1163, 1992.
[136] K. Binder and D. Heermann, Monte Carlo Simulation in Statistical Physics.
Berlin, Germany: Springer, 1997.
[137] P. Grassberger, “Damage spreading and critical exponents for ‘model A’ Ising
dynamics,” Physica A, vol. 214, no. 4, pp. 547–559, 1995.
[138] M. Nightingale and H. Blöte, “Dynamic exponent of the two-dimensional Ising
model and Monte Carlo computation of the subdominant eigenvalue of the
stochastic matrix,” Physical Review Letters, vol. 76, no. 24, p. 4548, 1996.
[139] F.-G. Wang and C.-K. Hu, “Universality in dynamic critical phenomena,”
Physical Review E, vol. 56, no. 2, p. 2310, 1997.
[140] N. Ito, “Nonequilibrium relaxation method – an alternative simulation strat-
egy,” Pramana, vol. 64, no. 6, pp. 871–880, 2005.
[141] Z. Rácz, “Nonlinear relaxation near the critical point: molecular-field and
scaling theory,” Physical Review B, vol. 13, no. 1, p. 263, 1976.
[142] K. Binder, “Finite size scaling analysis of Ising model block distribution func-
tions,” Zeitschrift für Physik B Condensed Matter, vol. 43, no. 2, pp. 119–140,
1981.
[143] K. Binder and D. Landau, “Finite-size scaling at first-order phase transitions,”
Physical Review B, vol. 30, no. 3, p. 1477, 1984.
240 References
[144] R. B. Potts, “Some generalized order-disorder transformations,” in Mathe-

matical Proceedings of the Cambridge Philosophical Society, vol. 48, no. 1.
Cambridge, UK: Cambridge University Press, 1952, pp. 106–109.
[145] F.-Y. Wu, “The Potts model,” Reviews of Modern Physics, vol. 54, no. 1, p. 235,
1982.
[146] F. Gliozzi, “Simulation of Potts models with real q and no critical slowing
down,” Physical Review E, vol. 66, no. 1, Article 016115, 2002.
[147] P. Kasteleyn and C. Fortuin, “Phase transitions in lattice systems with random
local properties,” Journal of the Physical Society of Japan, vol. 26, p. 11, 1969.
[148] F. Wu, “Percolation and the Potts model,” Journal of Statistical Physics, vol. 18,
no. 2, pp. 115–123, 1978.
[149] K. Binder, “Coniglio–Klein droplets: a fruitful concept to understand phase
transitions geometrically,” Journal of Physics A, vol. 50, no. 6, Article 061001,
2017.
[150] A. Coniglio and W. Klein, “Clusters and Ising critical droplets: a renor-
malisation group approach,” Journal of Physics A, vol. 13, no. 8, p. 2775,
1980.
[151] R. H. Swendsen and J.-S. Wang, “Nonuniversal critical dynamics in Monte
Carlo simulations,” Physical Review Letters, vol. 58, no. 2, p. 86, 1987.
[152] J.-S. Wang and R. H. Swendsen, “Cluster Monte Carlo algorithms,” Physica A,
vol. 167, no. 3, pp. 565–579, 1990.
[153] U. Wolff, “Collective Monte Carlo updating for spin systems,” Physical Review
Letters, vol. 62, no. 4, p. 361, 1989.
[154] H. E. Stanley, “Dependence of critical properties on dimensionality of spins,”
Physical Review Letters, vol. 20, no. 12, p. 589, 1968.
[155] J. M. Kosterlitz and D. J. Thouless, “Ordering, metastability and phase tran-
sitions in two-dimensional systems,” Journal of Physics C: Solid State Physics,
vol. 6, no. 7, p. 1181, 1973.
[156] N. Mermin and H. Wagner, “Absence of ferromagnetism or antiferromagnetism
in one- or two-dimensional isotropic Heisenberg models,” Physical Review
Letters, no. 17, p. 1133, 1966.
[157] Z. Salzburg, J. Jacobson, W. Fickett, and W. Wood, “Excitation of non-radial
stellar oscillations by gravitational waves: a first model,” Chemical Physics,
no. 30, p. 65, 1959.
[158] T. Penna and H. Herrmann, “Broad histogram Monte Carlo,” European
Physical Journal B–Condensed Matter and Complex Systems, vol. 1, no. 2,
pp. 205–208, 1998.
[159] P. de Oliveira, T. Penna, and H. Herrmann, “Broad histogram method,”
Brazilian Journal of Physics, no. 26, p. 677, 1996.
[160] J.-S. Wang, “Flat histogram Monte Carlo method,” Physica A: Statistical
Mechanics and Its Applications, vol. 281, no. 1–4, pp. 147–150, 2000.
[161] J. Lee, “New Monte Carlo algorithm: entropic sampling,” Physical Review
Letters, vol. 71, no. 2, p. 211, 1993.
References 241
[162] B. A. Berg and T. Neuhaus, “Multicanonical ensemble: a new approach to sim-

ulate first-order phase transitions,” Physical Review Letters, vol. 68, no. 1, p. 9,
1992.
[163] B. A. Berg, “The multicanonical ensemble: a new approach to computer
simulations,” International Journal of Modern Physics C, vol. 3, no. FSU-HEP-
92-08-17, pp. 1083–1098, 1992.
[164] F. Wang and D. Landau, “Efficient, multiple-range random walk algorithm to
calculate the density of states,” Physical Review Letters, vol. 86, no. 10, p. 2050,
2001.
[165] C. Zhou, T. C. Schulthess, S. Torbrügge, and D. Landau, “Wang-landau algo-
rithm for continuous models and joint density of states,” Physical Review
Letters, vol. 96, no. 12, Article 120201, 2006.
[166] G. Torrie and J. Valleau, “Nonphysical sampling distributions in Monte Carlo
free-energy estimation: umbrella sampling,” Journal of Computational Physics,
vol. 23, pp. 187–199, 1977. [Online]. Available: www.sciencedirect.com/science/
article/pii/0021999177901218
[167] T. Niemeijer and J. Van Leeuwen, Phase Transitions and Critical Phenomena,
vol. 6, London: Academic, pp. 425–505, 1976.
[168] H. J. Maris and L. P. Kadanoff, “Teaching the renormalization group,” Ameri-
can Journal of Physics, vol. 46, no. 6, pp. 652–657, 1978.
[169] K. G. Wilson, “The renormalization group: critical phenomena and the Kondo
problem,” Reviews of Modern Physics, vol. 47, pp. 773–840, 1975. [Online].
Available: http://link.aps.org/doi/10.1103/RevModPhys.47.773
[170] S. Ma, “Monte Carlo renormalization group,” Physical Review Letters, vol. 37,
p. 461, 1976.
[171] R. Swendsen, “Monte Carlo renormalization group,” Physical Review Letters,
vol. 42, p. 859, 1979.
[172] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no.
7553, p. 436, 2015.
[173] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Boston, MA: MIT
Press, 2016.
[174] P. Mehta and D. J. Schwab, “An exact mapping between the variational renor-
malization group and deep learning,” arXiv preprint arXiv:1410.3831, 2014.
[175] G. E. Hinton, “Deterministic Boltzmann learning performs steepest descent in
weight-space,” Neural Computation, vol. 1, no. 1, pp. 143–150, 1989.
[176] D. H. Ackley, G. E. Hinton, and T. J. Sejnowski, “A learning algorithm for
Boltzmann machines,” Cognitive Science, vol. 9, no. 1, pp. 147–169, 1985.
[177] G. E. Hinton and R. R. Salakhutdinov, “Reducing the dimensionality of data
with neural networks,” Science, vol. 313, no. 5786, pp. 504–507, 2006.
[178] V. Nair and G. E. Hinton, “Rectified linear units improve restricted Boltzmann
machines,” in Proceedings of the 27th InterNational Conference on Machine
Learning (ICML-10), 2010, pp. 807–814.
242 References
[179] G. E. Hinton, “A practical guide to training restricted Boltzmann machines,” in

Neural Networks: Tricks of the Trade. Berlin, Germany: Springer, 2012, pp. 599–
619.
[180] W. S. McCulloch and W. Pitts, “A logical calculus of the ideas immanent in
nervous activity,” Bulletin of Mathematical Biophysics, vol. 5, no. 4, pp. 115–133,
1943.
[181] J. J. Hopfield, “Neural networks and physical systems with emergent collective
computational abilities,” Proceedings of the National Academy of Sciences of the
United States of America, vol. 79, no. 8, pp. 2554–2558, 1982.
[182] D. O. Hebb, The Organization of Behavior: A Neuropsychological Theory.
Oxfordshire, UK: Psychology Press, 2005.
[183] J. Hertz, A. Krogh, R. G. Palmer, and H. Horner, “Introduction to the theory
of neural computation,” Physics Today, vol. 44, p. 70, 1991.
[184] G. E. Hinton, “Training products of experts by minimizing contrastive diver-
gence,” Neural Computation, vol. 14, no. 8, pp. 1771–1800, 2002.
[185] F. D’Angelo and L. Böttcher, “Learning the Ising model with generative neural
networks,” Physical Review Research, vol. 2, no. 2, Article 023266, 2020.
[186] D. Kim and D.-H. Kim, “Smallest neural network to learn the Ising criticality,”
Physical Review E, vol. 98, no. 2, Article 022138, 2018.
[187] S. Efthymiou, M. J. Beach, and R. G. Melko, “Super-resolving the ising model
with convolutional neural networks,” Physical Review B, vol. 99, no. 7, Article
075113, 2019.
[188] G. Carleo and M. Troyer, “Solving the quantum many-body problem with
artificial neural networks,” Science, vol. 355, no. 6325, pp. 602–606, 2017.
[189] G. Torlai, G. Mazzola, J. Carrasquilla, M. Troyer, R. Melko, and G. Carleo,
“Neural-network quantum state tomography,” Nature Physics, vol. 14, no. 5,
p. 447, 2018.
[190] P. Mehta, M. Bukov, C.-H. Wang, A. G. Day, C. Richardson, C. K. Fisher, and
D. J. Schwab, “A high-bias, low-variance introduction to machine learning for
physicists,” Physics Reports, vol. 810, pp. 1–124, 2019.
[191] S. Kirkpatrick, C. D. Gelatt, and M. P. Vecchi, “Optimization by simulated
annealing,” Science, vol. 220, no. 4598, pp. 671–680, 1983.
[192] L. Jacobs and C. Rebi, “Multi-spin coding: A very efficient technique for Monte
Carlo simulations of spin systems,” Journal of Computational Physics, vol. 41,
no. 1, pp. 203–210, 1981.
[193] A. Baumgärtner, K. Binder, J.-P. Hansen, M. Kalos, K. Kehr, D. Landau,
D. Levesque, H. Müller-Krumbhaar, C. Rebbi, Y. Saito et al., Applications of the
Monte Carlo Method in Statistical Physics. Vol. 36. Berlin, Germany: Springer
Science & Business Media, 2013.
[194] J. Marro and R. Dickman, Nonequilibrium Phase Transitions in Lattice Models.
Cambridge, UK: Cambridge University Press, 2005.
[195] J. L. Cardy and R. Sugar, “Directed percolation and reggeon field theory,”
Journal of Physics A, vol. 13, p. L423, 1980.
References 243
[196] P. Grassberger, “Directed percolation in 2+ 1 dimensions,” Journal of Physics A,

vol. 22, no. 17, p. 3673, 1989.
[197] L. Böttcher, M. Luković, J. Nagler, S. Havlin, and H. J. Herrmann, “Failure and
recovery in dynamical networks,” Scientific Reports, vol. 7, p. 41729, 2017.
[198] L. Böttcher, J. Nagler, and H. J. Herrmann, “Critical behaviors in contagion
dynamics,” Physical Review Letters, vol. 118, no. 8, Article 088301, 2017.
[199] L. Böttcher, H. J. Herrmann, and H. Gersbach, “Clout, activists and budget:
the road to presidency,” PloS ONE, vol. 13, no. 3, Article e0193199, 2018.
[200] P. Richter, M. Henkel, and L. Böttcher, “Aging and relaxation in bistable
contagion dynamics,” arXiv preprint arXiv:2006.05937, 2020.
[201] M. J. Keeling and P. Rohani, Modeling Infectious Diseases in Humans and
Animals. Princeton, NJ: Princeton University Press, 2011.
[202] R. Pastor-Satorras, C. Castellano, P. Van Mieghem, and A. Vespignani, “Epi-
demic processes in complex networks,” Reviews of Modern Physics, vol. 87, no. 3,
p. 925, 2015.
[203] L. Böttcher, O. Woolley-Meza, N. A. Araújo, H. J. Herrmann, and D. Hel-
bing, “Disease-induced resource constraints can trigger explosive epidemics,”
Scientific Reports, vol. 5, p. 16571, 2015.
[204] L. Böttcher, J. Andrade, and H. J. Herrmann, “Targeted recovery as an effective
strategy against epidemic spreading,” Scientific Reports, vol. 7, no. 1, p. 14356,
2017.
[205] L. Böttcher, O. Woolley-Meza, E. Goles, D. Helbing, and H. J. Herrmann,
“Connectivity disruption sparks explosive epidemic spreading,” Physical Review
E, vol. 93, no. 4, Article 042315, 2016.
[206] L. Böttcher, H. J. Herrmann, and M. Henkel, “Dynamical universality of the
contact process,” Journal of Physics A: Mathematical and Theoretical, vol. 51,
no. 12, Article 125003, 2018.
[207] S. Broadbent and J. Hammersley, “Percolation processes I. Crystals and mazes,”
Mathematical Proceedings of the Cambridge Philosophical Society, vol. 53, no.
3, pp. 629–641, 1957.
[208] M. M. S. Sabag and M. J. de Oliveira, “Conserved contact process in one to five
dimensions,” Physical Review E, vol. 66, no. 3, 036115, 2002.
[209] J. Wang, Z. Zhou, Q. Liu, T. M. Garoni, and Y. Deng, “High-precision Monte
Carlo study of directed percolation in (d+ 1) dimensions,” Physical Review E,
vol. 88, no. 4, Article 042102, 2013.
[210] P. Grassberger, “On phase transitions in Schlögl’s second model,” Zeitschrift für
Physik B, vol. 47, no. 4, pp. 365–374, 1982.
[211] H.-K. Janssen, “On the nonequilibrium phase transition in reaction-diffusion
systems with an absorbing stationary state,” Zeitschrift für Physik B, vol. 42,
no. 2, pp. 151–154, 1981.
[212] D. T. Gillespie, “A general method for numerically simulating the stochastic time
evolution of coupled chemical reactions,” Journal of Computational Physics,
vol. 22, no. 4, pp. 403–434, 1976.
244 References
[213] “Exact stochastic simulation of coupled chemical reactions,” Journal of Physical

Chemistry, vol. 81, no. 25, pp. 2340–2361, 1977.
[214] P. Bak, K. Christensen, L. Danon, and T. Scanlon, “Unified scaling law for
earthquakes,” Physical Review Letters, vol. 88, no. 17, Article 178501, 2002.
[215] R. Baddeley, L. F. Abbott, M. C. Booth, F. Sengpiel, T. Freeman, E. A. Wake-
man, and E. T. Rolls, “Responses of neurons in primary and inferior temporal
visual cortices to natural scenes,” Proceedings of the Royal Society of London,
Series B, vol. 264, no. 1389, pp. 1775–1783, 1997.
[216] A.-L. Barabasi, “The origin of bursts and heavy tails in human dynamics,”
Nature, vol. 435, no. 7039, p. 207, 2005.
[217] L. Böttcher, O. Woolley-Meza, and D. Brockmann, “Temporal dynamics of
online petitions,” PloS ONE, vol. 12, no. 5, Article e0178062, 2017.
[218] K.-I. Goh and A.-L. Barabási, “Burstiness and memory in complex systems,”
Europhysics Letters, vol. 81, no. 4, p. 48002, 2008.
[219] M. Boguná, L. F. Lafuerza, R. Toral, and M. Á. Serrano, “Simulating non-
Markovian stochastic processes,” Physical Review E, vol. 90, no. 4, Article
042108, 2014.
[220] L. Böttcher and N. Antulov-Fantulin, “Unifying continuous, discrete, and
hybrid susceptible-infected-recovered processes on networks,” Physical Review
Research, vol. 2, no. 3, Article 033121, 2020.
[221] N. Masuda and L. E. Rocha, “A Gillespie algorithm for non-Markovian
stochastic processes,” SIAM Review, vol. 60, no. 1, pp. 95–115, 2018.
[222] S. Wolfram, “Universality and complexity in cellular automata,” Physica D:
Nonlinear Phenomena, vol. 10, no. 1–2, pp. 1–35, 1984.
[223] H. J. Herrmann, “Cellular automata,” in Nonlinear Phenomena in Complex
Systems, Amsterdam: Elsevier, 1989, p. 151.
[224] A. Ilachinski, Cellular Automata: A Discrete Universe. Singaore: World Scien-
tific, 2001.
[225] “Game of life glider gun.” [Online]. Available: https://commons.wikimedia.org/
wiki/File:Gospers_glider_gun.gif
[226] G. Y. Vichniac, “Simulating physics with cellular automata,” Physica D, vol. 10,
no. 1–2, pp. 96–116, 1984.
[227] J. G. Zabolitzky and H. J. Herrmann, “Multitasking case study on the cray-2:
the q2r cellular automaton,” Journal of Computational Physics, vol. 76, no. 2,
pp. 426–447, 1988.
[228] H. J. Herrmann, “Geometrical cluster growth models and kinetic gelation,”
Physics Reports, vol. 136, no. 3, pp. 153–224, 1986.
[229] F. Family, “Dynamic scaling and phase transitions in interface growth,” Physica
A: Statistical Mechanics and Its Applications, vol. 168, no. 1, pp. 561–580, 1990.
[230] F. Family and T. Vicsek, “Scaling of the active zone in the Eden process on
percolation networks and the ballistic deposition model,” Journal of Physics A:
Mathematical and General, vol. 18, no. 2, p. L75, 1985.
References 245
[231] M. Eden, “A two-dimensional growth process,” in Proceedings of the fourth

Berkeley symposium on mathematical statistics and probability, vol. 4, Berkeley:
University of California Press, 1961, pp. 223–239.
[232] S. F. Edwards and D. Wilkinson, “The surface statistics of a granular aggregate,”
Proceedings of the Royal Society of London, Series A, Mathematical and Physical
Sciences, vol. 381, no. 1780, pp. 17–31, 1982.
[233] M. Kardar, G. Parisi, and Y.-C. Zhang, “Dynamic scaling of growing inter-
faces,” Physical Review Letters, vol. 56, no. 9, p. 889, 1986.
[234] D. J. Higham, “An algorithmic introduction to numerical simulation of stochas-
tic differential equations,” SIAM Review, vol. 43, no. 3, pp. 525–546, 2001.
[235] T. A. Witten and L. M. Sander, “Diffusion-limited aggregation,” Physical
Review B, vol. 27, no. 9, p. 5686, 1983.
[236] L. Niemeyer, L. Pietronero, and H. Wiesmann, “Fractal dimension of dielectric
breakdown,” Physical Review Letters, vol. 52, no. 12, p. 1033, 1984.
[237] L. de Arcangelis, S. Redner, and H. Herrmann, “A random fuse model for
breaking processes,” Journal de Physique Lettres, vol. 46, no. 13, pp. 585–590,
1985.
[238] D. C. Rapaport and D. C. R. Rapaport, The Art of Molecular Dynamics
Simulation. Cambridge, UK: Cambridge University Press, 2004.
[239] I. Newton, Philosophiae naturalis principia mathematica. Vol. 1. G. Brookman,
1833.
[240] J.-M. Sanz-Serna and M.-P. Calvo, Numerical Hamiltonian Problems. Mineola,
NY: Courier Dover Publications, 2018.
[241] R. Mani, L. Böttcher, H. J. Herrmann, and D. Helbing, “Extreme power law in
a driven many-particle system without threshold dynamics,” Physical Review E,
vol. 90, no. 4, Article 042201, 2014.
[242] R. Car and M. Parrinello, “Unified approach for molecular dynamics and
density-functional theory,” Physical Review Letters, vol. 55, no. 22, p. 2471,
1985.
[243] B. Alder and T. Wainwright, “Phase transition for a hard sphere system,”
Journal of Chemical Physics, vol. 27, pp. 1208–1209, 1957.
[244] R. Vink, G. Barkema, W. Van der Weg, and N. Mousseau, “Fitting the
Stillinger–Weber potential to amorphous silicon,” Journal of Non-crystalline
Solids, vol. 282, no. 2–3, pp. 248–255, 2001.
[245] B. Axilrod and E. Teller, “Interaction of the Van der Waals type between three
atoms,” Journal of Chemical Physics, vol. 11, no. 6, pp. 299–300, 1943.
[246] L. Verlet, “Computer experiments on classical fluids,” Physical Review, vol. 159,
pp. 98–103, 1967.
[247] R. W. Hockney, “The potential calculation and some applications,” Methods in
Computation Physics, vol. 9, p. 136, 1970.
[248] B. Quentrec and C. Brot, “New method for searching for neighbors in molec-
ular dynamics computations,” Journal of Computational Physics, vol. 13, no. 3,
pp. 430–432, 1973.
246 References
[249] R. Hockney and J. Eastwood, Computer Simulations Using Particles. Boca

Raton, FL: CRC Press, 1988.
[250] J. Ryckaert, G. Ciccotti, and H. Berendsen, “Numerical integration of the Carte-
sian equations of motion of a system with constraints: molecular dynamics of
n-alkanes,” Journal of Computational Physics, vol. 239, pp. 327–341, 1977.
[251] D. Evans, “On the representatation of orientation space,” Molecular Physics,
vol. 34, pp. 317–325, 1977.
[252] “Singularity free algorithm for molecular dynamics simulation of rigid poly-
atomics,” Molecular Physics, vol. 34, pp. 327–331, 1977.
[253] D. V. Stäger, N. Araújo, and H. J. Herrmann, “Prediction and control of slip-free
rotation states in sphere assemblies,” Physical Review Letters, vol. 116, no. 25,
Article 254301, 2016.
[254] C. Weiner, Interview with Paul Peter Ewald, 1968. [Online]. Available: www.aip
.org/history/ohilist/4596_1.html
[255] M. Sangster and M.Dixon, “Interionic potentials in alkali halides and their use
in simulations of the molten salts,” Advances in Physics, vol. 25, pp. 247–342,
1976.
[256] D. Wolf, P. Keblinski, S. Phillpot, and J. Eggebrecht, “Exact method for
the simulation of Coulombic systems by spherically truncated, pairwise r- 1
summation,” Journal of Physical Chemistry, vol. 110, no. 17, pp. 8254–8282,
1999.
[257] M. Allen and D. Tildesley, Computer Simulation of Liquids. Oxford, UK:
Clarendon Press, 1987.
[258] J. W. Eastwood, R. W. Hockney, and D. N. Lawrence, “P3M3DP – The
three-dimensional periodic particle-particle/particle-mesh program,” Computer
Physics Communications, vol. 19, no. 2, pp. 215–261 (1980).
[259] J. W. Cooley and J. W. Tukey, “An algorithm for the machine calculation
of complex Fourier series,” Mathematics of Computation, vol. 19, no. 90,
pp. 297–301, 1965.
[260] J. W. Cooley, “The re-discovery of the fast fourier transform algorithm,”
Microchimica Acta, vol. 93, no. 1–6, pp. 33–45, 1987.
[261] I. N. Sneddon, Fourier Transforms. North Chelmsford, MA: Courier Corpora-
tion, 1995.
[262] S. Pfalzner and P. Gibbon, Many-Body Tree Methods in Physics. Cambridge,
UK: Cambridge University Press, 2005.
[263] D. Potter, J. Stadel, and R. Teyssier, “Pkdgrav3: beyond trillion particle cos-
mological simulations for the next era of galaxy surveys,” Computational
Astrophysics and Cosmology, vol. 4, no. 1, p. 2, 2017.
[264] L. Onsager, “Electric moments of molecules in liquids,” Journal of the American
Chemical Society, no. 58, pp. 1486–1493, 1936. [Online]. Available: http://pubs
.acs.org/doi/abs/10.1021/ja01299a050
[265] H. Yoshida, “Construction of higher order symplectic integrators,” Physics
Letters A, vol. 150, 1990. [Online]. Available: http://cacs.usc.edu/education/
phys516/yoshida-symplectic-pla00.pdf
References 247
[266] R. Watts, “Monte Carlo studies of liquid water,” Molecular Physics, vol. 26, pp.
1069–1083, 1974. [Online]. Available: www.tandfonline.com/doi/abs/10.1080/
00268977400102381#.U_Wxd9bPHVM
[267] E. Braun, S. M. Moosavi, and B. Smit, “Anomalous effects of velocity rescaling
algorithms: the flying ice cube effect revisited,” Journal of Chemical Theory and
Computation, vol. 14, no. 10, pp. 5262–5272, 2018.
[268] K. Huang, Introduction to Statistical Physics. London, UK: Chapman and
Hall/CRC, 2009.
[269] H. J. Berendsen, J. v. Postma, W. F. van Gunsteren, A. DiNola, and J. R. Haak,
“Molecular dynamics with coupling to an external bath,” Journal of Chemical
Physics, vol. 81, no. 8, pp. 3684–3690, 1984.
[270] W. G. Hoover, “William G. Hoover’s personal website.” [Online]. Available:
http://williamhoover.info/
[271] W. G. Hoover, A. J. Ladd, and B. Moran, “High-strain-rate plastic flow stud-
ied via nonequilibrium molecular dynamics,” Physical Review Letters, vol. 48,
no. 26, p. 1818, 1982.
[272] D. J. Evans, W. G. Hoover, B. H. Failor, B. Moran, and A. J. Ladd, “Nonequi-
librium molecular dynamics via Gauss’s principle of least constraint,” Physical
Review A, vol. 28, no. 2, p. 1016, 1983.
[273] W. Hoover, “Canonical dynamics: equilibrium phase-space distributions,”
Physical Review, vol. 31, 1985. [Online]. Available: http://journals.aps.org/pra/
abstract/10.1103/PhysRevA.31.1695
[274] S. Nosé, “A molecular dynamics method for simulations in the canonical ensem-
ble,” Molecular Physics, vol. 52, 1984. [Online]. Available: www.tandfonline
.com/doi/abs/10.1080/00268978400101201#.U3Ic-3LI9yA
[275] S. Nosé, “A unified formulation of the constant temperature molecular dynamics
methods,” Journal of Chemical Physics, vol. 81, 1984. [Online]. Available: http://
scitation.aip.org/content/aip/journal/jcp/81/1/10.1063/1.447334
[276] S. D. Bond, B. J. Leimkuhler, and B. B. Laird, “The Nosé–Poincaré method for
constant temperature molecular dynamics,” Journal of Computational Physics,
vol. 151, no. 1, pp. 114–134, 1999.
[277] H. Andersen, “Molecular dynamics simulations at constant pressure and/or
temperature,” Journal of Chemical Physics, vol. 72, 1980. [Online]. Available:
http://scitation.aip.org/content/aip/journal/jcp/72/4/10.1063/1.439486
[278] T. Pöschel and T. Schwager, Computational Granular Dynamics: Models and
Algorithms. Berlin, Germany: Springer Science & Business Media, 2005.
[279] O. Walton and R. Braun, “Viscosity, granular-temperature and stress calcula-
tions for shearing assemblies of inelastic frictional disks,” Journal of Rheology,
no. 30, pp. 949–980, 1986. [Online]. Available: http://arxiv.org/abs/cond-mat/
9810009
[280] S. Luding, “Collisions and contacts between two particles,” in Physics of Dry
Granular Media. Berlin, Germany: Springer, 1998, pp. 285–304.
[281] E. Popova and V. L. Popov, “The research works of coulomb and amontons and
generalized laws of friction,” Friction, vol. 3, no. 2, pp. 183–190, 2015.
248 References
[282] S. Luding, “Particle–particle interactions and models (discrete element

method),” 2010. [Online]. Available: www2.msm.ctw.utwente.nl/sluding/
TEACHING/ParticleTechnology/Luding_PForcesModels.pdf
[283] P. Cundall and O. Strack, “A discrete numerical model for granular assem-
blies,” 1979. [Online]. Available: www.icevirtuallibrary.com/content/article/10
.1680/geot.1979.29.1.47
[284] B. D. Lubachevsky, “How to simulate billiards and similar systems,” Journal of
Computational Physics, vol. 94, no. 2, pp. 255–283, 1991.
[285] S. Luding, “Clustering instabilities, arching, and anomalous interaction proba-
bilities as examples for cooperative phenomena in dry granular media,” T.A.S.K.
Quarterly, Scientific Bulletin of Academic Computer Centre of the Technical
University of Gdansk, vol. 2, pp. 417–443, 1998.
[286] S. Luding and S. McNamara, “How to handle inelastic collapse of a dissipative
hard-sphere gas with the tc model,” Granular Matter, 1998. [Online]. Available:
http://arxiv.org/abs/cond-mat/9810009
[287] S. McNamara and W. Young, “Inelastic collapse in two dimensions,” Physical
Review E, vol. 50, no. 1, p. R28, 1994.
[288] S. Gonzalez, A. R. Thornton, and S. Luding, “Free cooling phase-diagram
of hard-spheres with short-and long-range interactions,” European Physical
Journal Special Topics, vol. 223, no. 11, pp. 2205–2225, 2014.
[289] S. McNamara and W. Young, “Inelastic collapse and dumping in a one-
dimensional granular medium,” Physical Review, 1991. [Online]. Available:
www-pord.ucsd.edu/~wryoung/reprintPDFs/InelasticCollapse.pdf
[290] S. Luding, “Freely cooling granular gas.” [Online]. Available: www2.msm.ctw
.utwente.nl/sluding/pictures/cooling.html
[291] G. Paul, “A complexity O(1) priority queue for event driven molecular dynamics
simulations,” Journal of Computational Physics, vol. 221, no. 2, pp. 615–625,
2007.
[292] J. W. Perram and M. S. Wertheim, “Statistical mechanics of hard ellipsoids:
overlap algorithm and the contact function,” Journal of Computational Physics,
no. 58, pp. 409–416, 1985.
[293] G. W. Delaney and P. W. Cleary, “The packing properties of superellipsoids,”
Europhysics Letters, vol. 89, no. 3, p. 34002, 2010.
[294] A. Donev, I. Cisse, D. Sachs, E. Variano, F. Stillinger, R. Connelly, S. Torquato,
and P. M. Chaikin, “Improving the density of jammed disordered packings
using ellipsoids,” Science, vol. 303, no. 5660, pp. 990–993, 2004. [Online].
Available: www.sciencemag.org/content/303/5660/990.full
[295] F. Alonso-Marroquin, “Spheropolygons: a new method to simulate conservative
and dissipative interactions between 2d complex-shaped rigid bodies,” Euro-
physics Letters, vol. 83, p. 14001, 2008. [Online]. Available: http://iopscience.iop
.org/0295-5075/83/1/14001
[296] G. Matheron and J. Serra, The Birth of Mathematical Morphology, 2000.
[Online]. Available: http://cmm.ensmp.fr/~serra/pdf/birth_of_mm.pdf
References 249
[297] R. I. Leine, C. Glocker, and D. H. Van Campen, “Nonlinear dynamics of the

woodpecker toy,” in Proceedings of DETC 2001 ASME Symposium on Nonlinear
Dynamics and Control In Engineering Systems, 2001.
[298] O. Maisonneuve, Commemoration of Jean-Jacques Moreau, 2014. [Online].
Available: www.isimmforum.tu-darmstadt.de/fileadmin/home/users/186/ISIMM
/JJMoreau-ISIMM-forum.pdf
[299] F. Radjai and V. Richefeu, “Contact dynamics as a nonsmooth discrete element
method,” Mechanics of Materials, vol. 41, no. 6, pp. 715–728, 2009.
[300] L. Brendel, T. Unger, and D. E. Wolf, “Contact dynamics for beginners,” in The
Physics of Granular Media, Weinheim: Wiley, 2004, pp. 325–343.
[301] T. Unger, L. Brendel, D. E. Wolf, and J. Kertész, “Elastic behavior in contact
dynamics of rigid particles,” Physical Review E, vol. 65, no. 6, Article 061305,
2002.
[302] D. H. Rothman and S. Zaleski, Lattice-gas cellular automata: simple models of
complex hydrodynamics. Vol. 5. Cambridge, UK: Cambridge University Press,
2004.
[303] S. Succi, The lattice Boltzmann equation: for fluid dynamics and beyond. Oxford,
UK: Oxford University Press, 2001.
[304] P. L. Bhatnagar, E. P. Gross, and M. Krook, “A model for collision processes
in gases. I. Small amplitude processes in charged and neutral one-component
systems,” Physical Review, vol. 94, no. 3, p. 511, 1954.
[305] J. Rabault, U. Reglade, N. Cerardi, M. Kuchta, and A. Jensen, “Deep rein-
forcement learning achieves flow control of the 2d karman vortex street,” arXiv
preprint arXiv:1808.10754, 2018.
[306] A. Malevanets and R. Kapral, “Mesoscopic model for solvent dynamics,”
Journal of Chemical Physics, vol. 110, no. 17, pp. 8605–8613, 1999.
[307] G. Bird, “Approach to translational equilibrium in a rigid sphere gas,” Physical
Fluids, vol. 6, p. 1518, 1963.
[308] G. A. Bird, Molecular gas dynamics and the direct simulation of gas flows. Vol. 5.
Oxford, UK: Clarendon Press, 1994.
[309] P. Hoogerbrugge and J. Koelman, “Simulating microscopic hydrodynamic phe-
nomena with dissipative particle dynamics,” Europhysics Letters, 1992. [Online].
Available: http://iopscience.iop.org/0295-5075/19/3/001
[310] P. Espanol and P. B. Warren, “Perspective: dissipative particle dynamics,”
Journal of Chemical Physics, vol. 146, no. 15, Article 150901, 2017.
[311] C. Marsh and J. Yeomans, “Dissipative particle dynamics: the equilibrium for
finite time steps,” Europhysics Letters, vol. 37, no. 8, p. 511, 1997.
[312] J. Yeomans, “Mesoscale simulations: Lattice Boltzmann and particle algo-
rithms,” Physica A: Statistical Mechanics and its Applications, vol. 369, no. 1,
pp. 159–184, 2006.
[313] J. J. Monaghan, “Smoothed particle hydrodynamics,” Reports on Progress in
Physics, vol. 68, no. 8, p. 1703, 2005.
250 References
[314] R. Gingold and J. Monaghan, “Smoothed particle hydrodynamics – theory

and application to non-spherical stars,” Monthly Notices of the Royal Astro-
nomical Society, vol. 181, pp. 375–389, 1977. [Online]. Available: http://adsabs
.harvard.edu/abs/1977MNRAS.181..375G
[315] L. Lucy, “Smoothed particle hydrodynamics – theory and application to non-
spherical stars,” Astronomy Journal, vol. 82, pp. 1013–1024, 1977. [Online].
Available: http://adsabs.harvard.edu/abs/1977AJ.....82.1013L
[316] J. J. Monaghan, “Smoothed particle hydrodynamics,” Annual Review of
Astronomy and Astrophysics, vol. 30, no. 1, pp. 543–574, 1992.
[317] T. Harada, S. Koshizuka, and Y. Kawaguchi, “Smoothed particle hydrodynam-
ics on gpus,” 2007. [Online]. Available: http://inf.ufrgs.br/cgi2007/cd_cgi/papers/
harada.pdf
[318] NIST, “Rydberg constant times hc in J.” [Online]. Available: https://physics
.nist.gov/cgi-bin/cuu/Value?rydhcj
[319] P. Pulay, “Ab initio calculation of force constants and equilibrium geometries in
polyatomic molecules: I. theory,” Molecular Physics, vol. 17, no. 2, pp. 197–204,
1969.
[320] P. Hohenberg and W. Kohn, “Inhomogeneous electron gas,” Physical Review,
vol. 136, no. 3B, p. B864, 1964.
[321] W. Kohn and L. J. Sham, “Self-consistent equations including exchange and
correlation effects,” Physical Review, vol. 140, no. 4A, Article A1133, 1965.
[322] C. Lee, W. Yang, and R. G. Parr, “Development of the colle-salvetti correlation-
energy formula into a functional of the electron density,” Physical Review B,
vol. 37, no. 2, p. 785, 1988.
[323] F. Herman, J. P. Van Dyke, and I. B. Ortenburger, “Improved statistical
exchange approximation for inhomogeneous many-electron systems,” Physical
Review Letters, vol. 22, no. 16, p. 807, 1969.
[324] A. Becke, “Density-functional exchange-energy approximation with correct
asymptotic behavior,” Physical Review, no. 38, p. 3098, 1988. [Online]. Available:
http://journals.aps.org/pra/abstract/10.1103/PhysRevA.38.3098
[325] S. Grimme, “Semiempirical GGA-type density functional constructed with a
long-range dispersion correction,” Journal of Computational Chemistry, vol. 27,
pp. 1787–1799, 2006. [Online]. Available: http://onlinelibrary.wiley.com/doi/10
.1002/jcc.20495/abstract
[326] H. Hellmann, Hans Hellmann: Einführung in Die Quantenchemie: Mit Biografis-
chen Notizen von Hans Hellmann Jr. Berlin, Germany: Springer, 2015.
[327] R. P. Feynman, “Forces in molecules,” Physical Review, vol. 56, no. 4, p. 340,
1939.
[328] R. Car and M. Parrinello, “Unified Approach for Molecular Dynamics and
Density-Functional Theory,” Physical Review Letters, no. 55, p. 2471, 1985.
[Online]. Available: http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.55
.2471
References 251
[329] M. Tuckerman, E. J. Bohm, L. V. Kalé, and G. Martyna, Car–Parrinello

Method. Boston, MA: Springer US, 2011, pp. 220–227. [Online]. Available:
https://doi.org/10.1007/978-0-387-09766-4_200
[330] T. D. Kühne, “Second generation Car–Parrinello molecular dynamics,” Wiley
Interdisciplinary Reviews: Computational Molecular Science, vol. 4, no. 4, pp.
391–406, 2014.
[331] Parrinello, M. and Rahman, A. (1980). Crystal structure and pair potentials: A
molecular-dynamics study. Physical Review Letters, 45(14), 1196.

Hans J. Herrmann, Lucas Böttcher - Computational Statistical

Uploaded by

Copyright:

Available Formats

Hans J. Herrmann, Lucas Böttcher - Computational Statistical

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hans J. Herrmann, Lucas Böttcher - Computational Statistical

Uploaded by

Copyright:

Available Formats

Index

Game of Life, 136, see cellular automata interface, 83

Lucilla de Arcangelis, 144, 145 non-smooth mechanics, 208

reaction-field method, 175 spreading rate, 128, see contact process

Lucas Böttcher is Assistant Professor of Computational Social Science at Frankfurt

Hans J. Herrmann is Directeur de Recherche at CNRS in Paris, visiting Professor at the

Cambridge University Press is part of the University of Cambridge.

Part I Stochastic Methods 1

4 Monte Carlo Methods 67

4.9 Boundary Conditions 83

7 Histogram Methods 100

8 Renormalization Group 105

9 Learning and Optimizing 112

11 Nonequilibrium Systems 127

Part II Molecular Dynamics 147

12 Basic Molecular Dynamics 149

13 Optimizing Molecular Dynamics 157

14 Dynamics of Composed Particles 161

15 Long-Range Potentials 169

16 Canonical Ensemble 177

17 Inelastic Collisions in Molecular Dynamics 187

18 Event-Driven Molecular Dynamics 192

19 Nonspherical Particles 202

20 Contact Dynamics 208

21 Discrete Fluid Models 215

22 Ab Initio Simulations 222

Computational physics is the study and implementation of numerical algorithms to

STOCH ASTIC METHODS

• To simulate random events in experimental data (e. g., radioactive decay)

1.1 Deﬁnition of Random Numbers

1.2 Congruential RNG (Multiplicative)

c p−1 mod p = 1 . (1.4)

A Mersenne number is defined as Mn = 2n − 1 with n ∈ N. If the number M is prime,

The number p = 2147483648 is a Mersenne prime and corresponds to the maximal

Additional Information: Marsaglia’s Theorem

Additional Information: Marsaglia’s Theorem (cont.)

Exercise: Congruential RNG

Exercise: Congruential RNG (cont.)

1.3 Lagged Fibonacci RNG (Additive)

A slightly more complicated RNG is the lagged Fibonacci algorithm proposed by

Two-Element Lagged Fibonacci RNG

1.4 Available Libraries

double rnd = u ( mt_rng ) ;

c o u t << rnd << e n d l ;

1.5 How Good is an RNG?

1. Square test (see Section 1.2 for details)

si si+d − s2i , (1.9)

for different values of d.

Additional Information: Marsaglia’s “Diehard” Tests

Additional Information: Marsaglia’s “Diehard” Tests (cont.)

1.6 Nonuniform Distributions

Thus far, we have only considered uniform distributions of pseudo-random numbers.

• We can apply transformation methods if the target distribution is known analytically,

These methods are explained in the following sections.

1.6.1 Transformation Method

The corresponding cumulative distribution function (CDF) is

where z is a uniformly distributed random variable and y a random variable distributed

Of course, these conditions can be overcome to a certain extent by precomputing/tabu-

The Exponential Distribution

By applying the area equality of eq. (1.13), we find

V (C) ∝ rid , (2.25)

The q-mass varies with according to