Department of Applied Analysis and Complex Dynamical Systems

Toward the Analysis and Synthesis of Varied Behavior of Complex Systems

Complex systems refer to systems that exhibit, as a whole, a variety of behavior and functions such as self-organization, chaos with a large degree of freedom, and learning and associative memories through nonlinear large-scale interactions among those elements. This department aims to clarify the fundamental principles and structures of these complex systems, analyze and extract useful information from this huge mass of information, and design suitable systems. We undertake education and research related to the clarification of solution structures for complex mathematical models by way of mathematical and numerical analysis, development of models and solution algorithms for chaos in complex dynamics, nonlinear phenomena such as pattern formation, and complex systems, as well as system control, intelligence, and self-organization.

29

Graduate School of Informatics

Department of Applied Analysis and Complex Dynamical Systems

An Invitation to the Department of Applied Analysisand Complex Dynamical Systems

YAMAMOTO Yutaka,
Department of Applied Analysis and Complex Dynamical Systems

Applied Analysis and Complex Dynamical Systems Science is a discipline which strives to explain complex phenomena that could not be understood or utilized fully on simplified mathematical models in the past by using various computers and analysis based on new concepts such as chaos and fractals. In particular, our department focuses on systems that are sensitive to strong nonlinearity, a high degree of freedom, large scales, and errors in order to fuse together engineering perspectives with scientific perspectives based on mathematics, numerical calculations, nonlinear physics, sample data control theory, etc. In other words, we hold in esteem the respective values of both science, which represents the pursuit of "reason," and engineering, which represents the art of "craftsmanship." Both in terms of our research and education, effort is made to mutually compensate for the drawbacks of each of these aspects to truly combine science and engineering. With faculty members who are also comprised of graduates from both science and engineering fields, the department has three divisions: Applied Analysis, Complex Dynamics, and Applied Mathematical Sciences. Each of these divisions operates under the keywords "applied analysis," "nonlinear physics," and "engineering" in their educational and research activities. More specifically, both basic and applied education and research is conducted primarily on fluids, elastic bodies, and neural systems that display strong large-scale or nonlinear characteristics, separating these into fields such as differential equation, numerical analysis, probability theory, fractal analysis, nonlinear dynamics, fluid dynamics, statistical mechanics, computational dynamics, control theory, and digital signal processing.

Graduate School of Informatics

30

Department of Applied Analysis and Complex Dynamical Systems

Department of Applied Analysis and Complex Dynamical Systems

Toward the Analysis and Synthesis of Varied Behavior of Complex Systems

31

Graduate School of Informatics

Department of Applied Analysis and Complex Dynamical Systems

Divisions and Groups

Division Applied Analysis Nonlinear Dynamics Complex Dynamics Nonequilibrium Dynamics Applied Mathematical Sciences Computational Mechanics Intelligent and Control Systems Group Applied Analysis Research and Education Topics Inverse and Ill-Posed Problems, Fractal Theory, Nonlinear Differential Equations, Numerical Analysis, Probability Theory, and Computational Mechanics Analysis of the Chaotic and Stochastic Behavior Exhibited by Fluids and Other Nonlinear Dynamical Systems Analysis of Nonlinear Dynamical Systems with Many Degrees of Freedom, Theoretical Neuroscience, and Network Science Computational Mechanics and Monte Carlo Simulations System Control, Digital Signal Processing, Modelling, and System Design Theory NISHIMURA Naoshi YAMAMOTO Yutaka FUNAKOSHI Mitsuaki Professor ISO Yuusuke KIGAMI Jun

Graduate Curriculum
Courses for the Master's Program Applied Analysis A Applied Analysis B Complex Dynamics A Complex Dynamics B Applied Mathematical Sciences A Applied Mathematical Sciences B Topics in Differential Equations A Topics in Differential Equations B Topics in Nonlinear Analysis A Topics in Nonlinear Analysis B Seminar on Applied Analysis I Seminar on Applied Analysis II Topics in Applied Analysis I Topics in Applied Analysis II Topics in Nonlinear Dynamics A Topics in Nonlinear Dynamics B Topics in Nonequilibrium Dynamics A Topics in Nonequilibrium Dynamics B Seminar in Complex Dynamics I Seminar in Complex Dynamics II Topics in Complex Dynamics I Topics in Complex Dynamics II Topics in Computational Mechanics A Topics in Computational Mechanics B Topics in Control Theory A Topics in Control Theory B Seminar in Applied Mathematical Sciences I Seminar in Applied Mathematical Sciences II Topics in Applied Mathematical Sciences I Topics in Applied Mathematical Sciences II Advanced Study in Applied Analysis and Complex Dynamical Systems I Advanced Study in Applied Analysis and Complex Dynamical Systems II Courses for the Doctoral Program Seminar on Applied Analysis and Complex Dynamical Systems (Advanced) Seminar on Applied Analysis (Advanced A & B) Seminar on Complex Dynamics (Advanced A & B) Seminar on Applied Mathematical Sciences (Advanced A & B)

Teaching Staff
Professors ISO Yuusuke; KIGAMI Jun; FUNAKOSHI Mitsuaki; NISHIMURA Naoshi; YAMAMOTO Yutaka Associate Professors TANAKA Hiroaki; FUJIOKA Hisaya; HINO Masanori; AOYAGI Toshio Senior Lecturers KUBO Masayoshi; MIYAZAKI Syuji; WAKANO Isao Assistant Professors KANEKO Yutaka; TUTU Hiroki; HARADA Kenji; FUJIWARA Hiroshi; NAGAHARA Masaaki

Graduate School of Informatics

32

Department of Applied Analysis and Complex Dynamical Systems

Applied Analysis
Applied mathematics seeks to go far beyond the mere application of pure mathematics to the solution of physical and mechanical problems. Through research into mathematical models of phenomena, applied mathematics creates new mathematics. Our Division teaches and carries out research in applied analysis where there is particular emphasis on analysis in fields of applied mathematics. We seek to improve our understanding of existing analytics and create new analytics for the 21st Century. To give specific examples of the kinds of research we do, we analyze mathematical models of physical and mechanical phenomena by applying mathematical and numerical analysis and stochastic theory to get a better understanding of both the analytical methods and the mathematical structure of the model, and to establish new analytical techniques. In this Division, the key words are nonlinear analysis and inverse problem analysis, and our full-time teaching staff constantly interacts with one another while they teach and conduct research.

Nonlinear Analysis and Inverse Problem Analysis

Analysis in the 21st Century
Faculty Members and Their Research Interests ISO Yuusuke (Professor) Numerical Analysis of (Partial) Differential Equations, Numerical Analysis of Ill-posed Problems, Analysis of Inverse Problems I carry out both mathematical and numerical analysis in research regarding the determination of unknown coefficients and other inverse problems, as well as boundary value problems and other forward problems with respect to partial differential equations that describe mechanical and physical phenomena. KIGAMI Jun (Professor) Fractal Analysis, Fractal Geometry I am interested in the mathematical theory of problems concerning heat and wave propagation in fractal concept models new models for the natural world. HINO Masanori (Associate Professor) Probability Theory, Stochastic Analysis on Infinite Dimensional Spaces and Fractal Sets I conduct research using stochastic theory and functional analysis to analyze path space and other infinite dimensional space as well as fractal sets (complex spaces). KUBO Masayoshi (Senior Lecturer) Inverse Problem Analysis, Numerical Analysis of (Partial) Differential Equations, Partial Differential Equations, Brain Model Mathematical Research I mathematically analyze partial differential equations that appear in mathematical physics and mathematically and numerically analyze the inverse problems found in these partial differential equations, where the unknown coefficients of these inverse problems are determined by observed data. WAKANO Isao (Senior Lecturer) Numerical Analysis of (Partial) Differential Equations, Partial Differential Equations, Mathematical Research in Fracture Mechanics I am involved in research into the mathematical and numerical analysis of partial differential equations in mechanics. Recently, I have carried out numerical calculations of problems in the field of fracture mechanics. FUJIWARA Hiroshi (Assistant Professor) Numerical Analysis of Ill-posed Problems, Design and Implementation of Multi-precision Arithmetic Environments I am involved in research concerning regularization methods and numerical analysis of multiple-precision calculations with the aim of solving inverse problems that occur in mechanics and geophysics and reconstructing them numerically.

33

Graduate School of Informatics

Department of Applied Analysis and Complex Dynamical Systems

Complex Dynamics
The dynamic behavior of systems that have a high degree of freedom and engage in nonlinear mutual interactions is both complex and richly diverse, and yet, many of these systems possess qualities in common with each other, such as a regular structure, and are fascinating subjects for research. In this Division, we use logical analysis and computer simulations to gain a better understanding of the complex behavior and control of these kinds of dynamic systems, as well as to clarify their common principles.

Nonlinear Dynamics
Studies of the complex behavior of fluids and other nonlinear dynamic systems
We aim to understand and control the various patterns of complicated behavior of nonlinear dynamical systems such as fluid systems, many-particle systems, and structural systems. In particular, we are trying to clarify, control and utilize in fluid systems and coupled dynamic systems the nonlinear behavior of chaos, synchronization, pattern formation, generation and interaction of nonlinear waves, interaction of vortices, and thermal convection. To achieve this aim, we conduct logical analysis based on the theory of nonlinear dynamic systems, reduction theory, and singular perturbation methods, and apply simulation techniques such as differential calculation, spectral methods, Monte Carlo techniques, and molecular dynamic methods. Our teaching and research also covers the development of techniques to analyse the reliability and risk concerning the failure of structural systems from the perspective of the application of stochastic process theory and stochastic field theory, and the use of computer simulations and theory relating to crystal growth and surface properties. Professor: FUNAKOSHI Mitsuaki, Associate Professor: TANAKA Hiroaki, Assistant Professor: KANEKO Yutaka

An efficient mixture of fluids is attained if two cylinders, one inside the other, are rotated slowly in opposite directions so that each fluid particle moves chaotically. This is called chaotic mixing.

Nonequilibrium Dynamics
From nonlinear and non-equilibrium physics to theories on networks as well as living and neural systems
Neural systems and social networks are nonlinear dynamic systems that have a high degree of freedom and develop as their dynamic elements change the structure of their interconnections. At first glance, these systems might appear to be completely different, but the invariant structure that they both have may be hidden. Mathematical models are useful tools to help us understand these systems.

Physical systems like fluids and chemical reactions are not the only cooperative phenomena that are made up of comparatively simple elements, and yet exhibit complex behavior and advanced functions that are impossible to predict from the individual elements alone. This same characteristic is also found in neural systems and social phenomena. For example, in neural systems, the mutual interaction between the basic elements called neurons all grouped together allows the neural system to acquire the advanced information processing functions of learning, memory and decision-making, or from a more ordinary perspective, they form the dynamic elements in a network (neurons, cities, people, etc.). The network structure and dynamic activity of the elements are simultaneously changing, and the network exhibits the ability to organize itself. Our research looks at these kinds of cooperative phenomena that have multiple elements, and focuses on reduction theory, rhythmic phenomena, and chaos theory from the perspective of nonlinear dynamics and non-equilibrium physics. (Associate Professor: AOYAGI Toshio, Senior Lecturer: MIYAZAKI Syuji, Assistant Professor: TUTU Hiroki)
Graduate School of Informatics

34

Department of Applied Analysis and Complex Dynamical Systems

Applied Mathematical Sciences

Many of the objects that we study in mechanics are large and complex, and often exhibit behavior that is uncertain and difficult to predict. Although the most important issue for us is to be able to accurately predict this behavior and control it, this is generally not easy to do. High-speed simulation techniques are essential to the prediction and analysis of large, complex systems; and in order for us to be able to control complex systems, the control method we use has to take into account the indefinite nature of the system and the uncertainty of its behavior i.e., the method of control has to be robust. In our Division, we develop and apply these techniques from the standpoint of applied mathematical science.

Computational Mechanics
Computer simulations: from the development of high-speed computation to the understanding of quantum systems
Numerical simulations are a powerful tool to help us solve many different kinds of problems in science and engineering. Computational mechanics, together with theoretical and experimental mechanics, is an effective method for investigating mechanical phenomena in engineering. Our Group is developing the Boundary Integral Equation Method (BIEM), a major technique in computational mechanics that is particularly effective in the analysis of waves and fractures. We are conducting research into fast BIEMs and their applications to large-scale problems. We are also working to solve periodic problems so we can apply the results to studies particularly, in optical science. We are carrying out research into simulations of macroscopic phenomena (quantum phase transitions, etc.) that are governed by quantum mechanics and the algorithms that are used in these simulations. (Professor: NISHIMURA Naoshi, Assistant Professor: HARADA Kenji)

Example of analysis using a time domain fast multi-pole boundary integral equation method (top) ; an algorithm to stochastically sample a quantum state (bottom)

Intelligent and Control Systems

Controlling systems and signal processing
We are surrounded by a vast number of systems in daily life: natural systems such as water cycles and the weather, and man-made systems ranging from satellites, aircraft and computer networks to robots, production systems, and small electronic devices. For all these systems to function properly, there has to be some kind of control mechanism at work. It is desirable that the control mechanisms of man-made systems work in a more sophisticated and intelligent manner. In our Group we carry out theoretical and advanced applied research into digital robust control, signal processing systems, network systems, and other areas to make control systems more advanced and intelligent, and improve voice processing as well as still- and moving-image processing. Professor: YAMAMOTO Yutaka, Associate Professor: FUJIOKA Hisaya, Assistant Professor: NAGAHARA Masaaki

A high-performance DA converter for CDs, designed using sampled-data control theory. It recovers and recreates sounds up to a frequency of 150 kHz and can create a sound quality close to that of the acoustics of a concert hall.

35

Graduate School of Informatics

Department of Applied Analysis and Complex Dynamical Systems

Departmental Activities
In cooperation with the Department of Mechanical Engineering and Science, the Department of Micro Engineering, and the Department of Aeronautics and Astronautics of the Graduate School of Engineering, as well as the International Innovation Center, the Department of Applied Analysis and Complex Dynamical Systems promoted education and research in the field of mechanical engineering at the Research and Education on Complex Functional Mechanical Systems New Developments in Mechanical Engineering Inspired by Complex Systems Science Center for research and education (FY2003-2007) as part of the 21st Century Center of Excellence (COE) Program. Based on the novel analytical techniques proposed in the new academic field of complex systems science, as well as our knowledge of the order-forming characteristics of complex systems and function creation arising from this, the Program aimed to develop phenomenon analysis, modeling, control, and function design for complex mechanical systems to create and establish a new field of learning known as complex systems mechanical engineering. Over the past five years, with the collaboration of researchers from the three departments at the Graduate School of Engineering, as well as the International Innovation Center and with the notion of fusing engineering and physical science in mind our department developed various activities. The Program aimed to give researchers the motivation and ability to pioneer new fields of research according to the principles of education and research laid out by the Department of Applied Analysis and Complex Dynamical Systems.

A view of open lectures held by our department as part of the 21st Century COE Program

Graduate School of Informatics

36