Kalpa Publications in Civil Engineering

Volume 1, 2017, Pages 239–247

ICRISET2017. International Conference on Re-
search and Innovations in Science, Engineering
&Technology. Selected papers in Civil Engineering

Applications of Harmony Search

Algorithm in Structural Engineering
Gunjan Chauhan 1,3, Vishal Patel2,3, Vishal Arekar2,3

1
Research Scholar, 2 Assistant proffesor
3
Structural Engineering Department,B.V. M. Engineering College,V. V. Nagar, India
[email protected], [email protected], [email protected]

Abstract
Harmony search Algorithm (HSA) is mostly preferred for solving optimization
problems and gives optimum solution of the problems. . It is based on improvisation of
harmony in music process where musicians improvise their instruments’ pitch by
searching for a aesthetically pleasant harmony. As the musicians in improvisation process
try to find the best harmony in terms of aesthetics, the decision variables in optimization
process try to be the best solution in terms of objective function. In the present work, the
harmony search method is studied with an attempt to use it to solve various structural
optimization problems. Harmony search can be more effective than some of the
optimization available right now like genetic algorithms, particle swarm algorithm, ant
colony algorithm, gravity search algorithm etc. The programming language used in this
work is Visual Basic-macro excel. The programs for harmony search algorithm is
developed in macro and their reliability is checked by verifying it with various
mathematical optimization problems.

1 Introduction
Optimization is a method which gives best solution for a given problem satisfying all its
requirements. Every optimization problem has three components. First one is the objective function.
Second the side constraints followed by behavioral or functional constraints. The objective function is
the criteria for comparing various alternative solution available for a given optimization problem.
Objective function is a mathematical expression formed by variables. Variables are the entities for a
given optimization problem. The variables are bounded within the lower and upper range up to which
they can vary. These are called side constraints of the given problem. Other than side constraints the
variables must also satisfy certain functional requirements. These requirements are called functional
constraints. Let’s take example of optimization of a rectangular beam. The objective is to minimize
cross section of the beam. Hence, objective function can be formed as b*d. Where “b” and “d” are the
breadth and depth of the beam. Thus “b” and “d” are called variables of the problem. Suppose breadth

C.D. Modhera, G.J. Joshi, D. Soni, I.N. Patel, A.K. Verma, L.B. Zala, S.D. Dhiman, D.R. Bhatt, J.M.
Rathod, B.C. Goradiya, M.S. Holia and D.K. Patel (eds.), ICRISET2017 (Kalpa Publications in Civil
Engineering, vol. 1), pp. 239–247
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

should be between 200mm to 350 mm only. And depth should not be more than 3 times breadth. These
are called side constraints. The deflection of the beam should not be more then span/250. This is called
functional constraint in context of given problem.
There are various methods available for optimization. Conventional methods require more
computational time and also it is very tedious job. So various heuristic methods are developed which
are based on natural principles. These methods are widely adopted since; they are reliable in context of
conventional methods.
Different Metaheuristics methods available are:
 Simulated Annealing
 Genetic Algorithms
 Ant Colony Optimization
 The Harmony Search Method,
 Gravitational Search
 Particle Swarm Optimization
 Bee Colony Optimization
 Bacterial Forging Optimization

Metaheuristics method utilizes nature’s concept into optimization problems. Still studies are going
on to develop new optimization techniques based on natural phenomenon.
Zong Woo Geem and Joong Hoon Kim in 2001 [1], developed a new heuristic algorithm inspired
by improvisation of music players, named Harmony Search (HS). Harmony Search algorithm can solve
a different continuous- variable problem as well as complex problems, and gives better result than other
existing heuristic methods in specified optimization problems.
S. Gholizadeh, A. Barzegar [2], proposed a modified harmony search (MHS) algorithm for size and
shape optimization of structures. This MHS algorithm uses elitism. In fact, the MHS is a multi-layered
upgradation of the HS and in each level a new harmony memory is created with the help of the
information of the previous levels. The numerical results show that use of MHS not only better solutions
can be found but also a noticeable reduction in computational time may be achieved.
Sachin A. Patil, D. A. Patel [3], gave an improved harmony search algorithm (IHSA). In IHSA few
disadvantages of HSA are removed and so that performance of HSA is improved. In IHSA Harmony
search parameters par and bw are improved at every iteration. The IHSA showed faster convergent then
HSA.
Moacir Kripka, Deise Boito [4], carried out research on Optimization of Reinforced Concrete
Frames by Harmony Search Method. This work dealt with the problem of optimization of reinforced
concrete building frames, following the requirements of the Brazilian standard NBR 6118 (ABNT
2007), and using the Harmony Search optimization method. To the examples analyzed, the optimization
method was quite efficient in minimizing structural cost. The software has been an important tool for
pre-sizing of reinforced concrete elements.
Mohamed F. El-Santawy and A. N. Ahmed [5], introduced a new a new Multi-Objective
Evolutionary technique. The new method uses Harmony Search optimization to Chaos search. The well-
known Fitness Sharing method is adopted to employ the size of the external archive used by the
technique during search. The proposed method is applied to Structural optimization which is one of the
most challenging areas in Multi Objective Optimization.

240
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

2 Harmony Search Method

The harmony search algorithm is a newly developed optimization algorithm developed by Geem,
Kim and Loganathan in 2001. It is based on an analogy to musical improvisation of jazz, where
musicians try to find, through repeated iterations, the aesthetically pleasant harmony given (optimum
solution to a problem). Iterations are called improvisations or practice. Variables analogs to musical
instruments. Values for variables are the notes of instruments. Every solution is called harmony.
Through repetitive random search technique, a new harmony vector is generated which replaces worst
harmony in harmony memory. The analogy between harmony search and optimization problem is
shown in table 1.
Optimization problem
Harmony search
Musical Instruments Components of the system being optimized
Notes: sets of possible tones to be design variables: set of possible design
played by musicians values for system components
Chord: specific collection of notes design: specific collection of design
variables
Harmony: synthesis of all notes objective function: synthesis of all design
variables
Goal: maximize harmony goal: extremize objective function
Process: trial and error, improvise process: improvise on designs by trial
new harmonies and try them out and error
Table 1: Analogy between harmony search and optimization problem

In general, mathematical optimization problems, harmony search uses the harmony memory (HM)
to store the solutions, and the harmony memory size (HMS) gives the number of solutions stored in the
harmony memory. The harmony memory is initialized by assigning random values to every design
variable from its possible range of values. With the help of initial HM, the value of the objective
function for every harmony is calculated and the harmony with the worst objective function value is
determined. In each iteration a new harmony is improvised and its objective function is calculated, if it
gives a better objective function value, it replaces the worst harmony in the harmony memory and the
harmony memory is updated. The previous procedure is again repeated till a termination criterion is
met which is normally maximum number of iterations or until no further improvement of harmony or
design performance can be achieved.
The following figure (fig.1) shows the flowchart of harmony search algorithm.

241
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

Figure 1: Flow chart of harmony search method

For the purpose of study of harmony search method, a visual basic macro program was developed
in excel and a mathematical optimization problem was solved using the developed program. The
problem is shown below:
Minimize,
f (X) = 5.358547 * (x3) 2 + 0.8356891 * x1 * x5 + 37.293239 * x1 - 40792.141
subject to,
g1= 85.334407 + 0.0056858 * x2 * x5 +
0.0006262 * x1 * x4 - 0.0022053 * x3 * x5
g2= 80.51249 + 0.0071317 * x2 * x5 +
0.0029955 * x1 * x2 + 0.0021813 * (x3) 2
g3= 9.300961 + 0.0047026 * x3 * x5 +
0.0012547 * x1 * x3 + 0.0019085 * x3 * x4
where,
0 ≤ g1 ≤ 92; 90 ≤ g2 ≤ 110; 20 ≤ g3 ≤ 25
and,
78 ≤ x1 ≤ 102; 33 ≤ x2 ≤45; 27 ≤ x3, x4, x5 ≤ 45
The results are tabulated below:

242
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

Particular Present work Literature

x1 78.073 78.000
x2 34.545 33.000
x3 30.114 29.995
x4 44.009 45.000
x5 33.828 36.775
F -30814.0 -30664.9
g1 91.883 91.999
g2 98.903 98.840
g3 19.571 19.991
Table 2: Comparision between present work and literature

By comparing the results, it is seen that harmony search method is an effective method in finding
out optimum solution of a given mathematical problem.

3 Applications in Structural engineering

Various Visual basic macro program was developed for harmony search method with improved par
and bw parameters. Exterior penalty function method is used to construct an unconstrained optimization
problem. The developed programs are applied to various structural optimization problems. The
problems are as follows:

3.1 Two Bar Truss Problem

The objective is to design of the truss for minimum volume. The optimization problem for three
variables y (vertical distance between B and C in m), x1 (area of AC in cm2) and x2 (area of BC in m) is
constructed as follows:
Minimize, f = x1 (16+ y2)1/2 + x2 (1+ y2)1/2

Subject to,
σAC <=100 MPa
σBC <=100 MPa
Where,
σAC =20 (16+ y2)1/2 ÷ x1 y
and σBC =80 (1+ y2) 1/2÷ x2 y

x1, x2 ≥ 0 and 1m ≤ y ≤ 3m

243
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

Figure 2: Two bar truss problem

The best solution is obtained as {x1, x2, y} = {0.4474, 0.8935, 1.9977} with a minimum volume of
0.03996 m3. Mohamed F. El-Santawy and A. N. Ahmed [5], developed a chaotic harmony search
technique and used it to find the optimum solution for same problem. The range for optimum volume
in their technique is between 0.0375 to 0.0537 m3. By comparison it is seen that the proposed program
gives optimum value within the range suggested by El- Santawy and A. N. Ahmed.

3.2 Pressure Vessel Design

A cylindrical vessel is to be optimized for minimization of cost. The objective is to minimize the
total cost, which includes the cost of material, forming and welding.

Figure 3: Pressure Vessel design

The design variables are : Ts (thickness of the shell, x1), Th (thickness of the head, x2), R (inner
radius, x3) and L(length of cylindrical section of the vessel, not including the head, x4). Ts and Th are
integer multiples of 0.0625 inch, which are the available thickness of rolled steel plates, and R and L
are continuous. The problem can be formulated as follows,
Minimize,
f= 0.6224 * x1 * x3 * x4 + 1.7781 * x2 * ((x3)2) + 3.1661 * ((x1)2) * x4 + 19.84 * ((x1)2) * x3
subject to,
g1 = -x1 + 0.0193 * x3
g2 = -x2 + 0.00954 * x3
g3 = (-(22 / 7) * ((x3)2) * x4) - ((4 / 3) * (22 / 7) * ((x3)3)) + (1296000)
g4 = (x4 - 240)
g5 = (-x1 + 1.1)
all gj ≤ 0, j= 1, 2, 3, 4, 5
and,
1.10 ≤ x1 ≤ 6.1875
0.6 ≤ x2 ≤ 6.1875
40 ≤ x3 ≤ 80
20 ≤ x4, x5 ≤ 60
After running the macro, the optimum solution obtained is tabulated in table 3.

244
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

Particular Present study Wu and chow(2001)

x1 1.125 1.125
x2 0.625 0.625
x3 58.219 58.1978
x4 44.037 44.293
F 7200.202 7207.775
g1 -0.001 -0.00178
g2 -0.07 -0.06979
g3 -0.249 -0.191
g4 -195.963 -195.707

Table 3: Comparision between present work and literature

Wu and chow developed a program to find the optimum solution of given problem using the GA
algorithm. The comparison between their results and present study is shown in table 3.
From the table of comparison, it is seen that the proposed macro program gives a better solution
than proposed by Wu and Chow.

3.3 Optimum design of R.C. concrete beam

The problem consists of finding optimum rectangular section of a given simply supported beam.
The program takes data from the user like, grade of concrete, grade of steel, permissible bounds for
breadth and depth of beam.
The side constraints are given by the user and other constraints are that, the deflection should be
within permissible limits and max amount of reinforcement should not exceed 4 %. This program
follows the procedure suggested by IS: 456,2000 for finding out the reinforcement required both in
tension as well as compression.
The objective function is
Minimize f = b * d
Where, b= breadth of the beam
d= depth of the beam
subject to,
bl < b < b u
dl < d < d u
Asc and Ast < 4 % of b*d
Def < span / 250

245
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

Figure 3: Design of R.C. Beam (excel sheet)

After running the macro program, the user will be able to see min cross section, amount of
compression and tension steel required and type of the beam i.e. singly reinforced or doubly reinforced
beam.

4 Conclusion
Harmony search algorithm is relatively new meta-heuristic algorithm which is very important in
optimization problems. It has been applied to divergent problems in recent years and results are very
encouraging when compared to other meta-heuristic algorithms. HSA is very flexible to implement
simple to apply.
Compared to gradient based optimization algorithms the harmony search method requires lesser
mathematical requirements and does not require initial value setting of decision variables. HSA uses
random searches so derivative information is also not needed.
This paper describes the application of harmony search algorithm in structural engineering by
dealing with the programs related to two bar truss problem, pressure vessel design and design of R.C.
Beam etc. In this paper the algorithm used is a modified one by improving HS parameters at each
iteration. The results show that the improved algorithm shows better solution requiring fewer iteration
cycles.
There still needs improvement in the existing program, as there are chances of getting local optimum
rather than global optimum. There is possibility of hybridizing the harmony search algorithm with other
existing meta heuristics like, GA, PSO, SA etc. The research needs to be done in this field.

References
SIMULATION 76:2,60-68 ©2001, Simulation Councils Inc. ISSN 0037-5497/01 Printed in the United States of
America
INTERNATIONAL JOURNAL OF OPTIMIZATION IN CIVIL ENGINEERING Int. J. Optim. Civil Eng., 2011;
3:485-494
International Journal of Engineering Science and Innovative Technology (IJESIT) Volume 2, Issue 1, January 2013
11th World Congress on Structural and Multidisciplinary Optimisation 07th -12th, June 2015, Sydney Australia
INTERNATIONAL JOURNAL OF COMPUTING SCIENCE, VOL. 1, NO. 3, April 2012 ISSN (Online): 2164-
1374, ISSN (Print): 2164-1366, Published online April, 2012
Engineering optimization: theory and practice, fourth Edison by S. S. Rao, John willy and sons, INC.
Klarbring, A. (2008). An introduction to structural optimization (Vol. 153). Springer
Yang X.S., “Harmony Search as a Metaheuristic Algorithm”, in: Music-Inspired Harmony Search Algorithm: Theory
and Applications (Editor Z. W. Geem), Studies in Computational Intelligence, Springer Berlin, vol. 191, pp. 1-14.
IS : 456, 2000
Hasancebi O, Erdal F, Saka MP, “An adaptive harmony search method for structural optimization”. J Struct Eng 1:72.
Lee KS, Geem ZW. A new meta-heuristic algorithm for continuous engineering optimization: harmony search theory
and practice, Computer Methods in Applied Mechanical Engineering, 2005, Vol. 194(36-38), pp. 3902-3933.
Optimization of Examination Timetable Using Harmony Search Hyper-Heuristics (HSHH) , Jayashree Rankhambe et
al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 5 (5) , 2014, 6719-6723
A Harmony Search with Adaptive Pitch Adjustment for Continuous Optimization, Chukiat Worasucheep,
International Journal of Hybrid Information Technology Vol. 4 No. 4, October, 2011
IMPROVED MUSIC BASED HARMONY SEARCH FOR SOLVING JOB SHOP SCHEDULING PROBLEMS,
Hymavathi Madivada, International Journal of Programming Languages and Applications (IJPLA) Vol.3, No.3, July
2013International Journal of Programming Languages and Applications (IJPLA) Vol.3, No.3, July 2013

246
Applications of Harmony Search Algorithm G. R. Chauhan, V. B. Patel and V. A. Arekar

A NEW MUSIC-INSPIRED HARMONY BASED OPTIMIZATION ALGORITHM. THEORY AND

APPLICATIONS, I. Kougias.
Hybrid Discrete Harmony Search Algorithm for Scheduling Re-Processing Problem in Remanufacturing, Kaizhou
Gao.
Geem, Z. W. (2005). Harmony search in water pump switching problem. Advances in natural computation (pp. 751-
760) Springer 25.
Greblicki, J., & Kotowski, J. (2009). Analysis of the properties of the harmony search algorithm carried out on the
one dimensional binary knapsack problem. In Computer Aided Systems Theory-EUROCAST 2009 (pp. 697-704).
Wang, L., Mao, Y., Niu, Q., & Fei, M. (2011). A multi-objective binary harmony search algorithm. In Advances in
Swarm Intelligence (pp. 74-81). Springer Berlin Heidelberg.
Wang, L., Yang, R., Xu, Y., Niu, Q., Pardalos, P. M., & Fei, M. (2013). An improved adaptive binary Harmony
Search algorithm. Information Sciences, 232, 58-87.
Topping, B. H. V. (1983). Shape optimization of skeletal structures: a review. Journal of Structural Engineering,
109(8), 1933-1951.
Hasançebi, O., & Erbatur, F. (2002). Layout optimisation of trusses using simulated annealing. Advances in Engineering
Software, 33(7), 681-696.

247