Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 93
Edited by:
Paper 124

Global Optimization using Particle Swarm Optimization and a Comparison with Evolutionary Algorithms and an Artificial Immune System

M. Szczepanik1, A. Poteralski1, W. Kus1 and T. Burczynski1,2

1Department for Strength of Materials and Computational Mechanics, Silesian University of Technology, Gliwice, Poland
2Institute of Computer Science, Computational Intelligence Department, Cracow University of Technology, Poland

Full Bibliographic Reference for this paper
M. Szczepanik, A. Poteralski, W. Kus, T. Burczynski, "Global Optimization using Particle Swarm Optimization and a Comparison with Evolutionary Algorithms and an Artificial Immune System", in , (Editors), "Proceedings of the Tenth International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 124, 2010. doi:10.4203/ccp.93.124
Keywords: optimization, particle swarm optimizer, artificial immune system, evolutionary algorithm, topology optimisation, finite element method.

Recently, non-gradient optimization methods such as swarm [1], immune [2] and evolutionary algorithms [3] have been widely applied in mechanics, and in structural optimization. The main feature of these methods is to simulate biological processes. The evolutionary methods are based on the theory of evolution. The artificial immune system is based on the mechanism discovered in a biological immune system. The swarm algorithms are based on models of animal social behaviour: moving and living in the groups. The main advantage of the methods is the fact that these approaches do not need any information about the gradient of the fitness function and provide a good probability of finding the global optimum. The main drawback of these approaches is the long time of calculations. So the choice of the fastest method seems to be quite important, therefore their comparison for the multi-modal mathematical functions (Rastrigin, Griewangk, Goldstein-Price, Branin) have been presented. The tests of the particle swarm optimizer, the artificial immune system and evolutionary algorithms comparing an average number of the objective function evaluation, for selected mathematical functions, show that the particle swarm optimizer was better then the others in the case of all the tested functions and also for the engineering optimization problem presented.

An effective tool of topology optimization of the structures has been presented. Using this approach shape, topology and material optimization is performed simultaneously. The main feature of the proposed optimization method is the swarm distribution of the material in the structure changing its material properties. This process leads to the elimination of part of the material from the structure and as a result the new shape and the topology of the structure emerges. The application of interpolation surfaces (hypersurface) reduces the number of the design variables and shortens the time of the computation. The application of the finite element code MSC NASTRAN in this method enables the optimization of complex mechanical systems. Numerical examples confirm the efficiency of the proposed optimization method and demonstrate that the method based on swarm computation is an effective technique for solving computer aided optimal design problems.

J. Kennedy, R.C. Eberhart, "Swarm Intelligence", Morgan Kauffman, 2001. doi:10.1007/0-387-27705-6_6
S.T. Wierzchon, "Artificial Immune Systems, theory and applications", EXIT, (in Polish), 2001.
Z. Michalewicz, "Genetic Algorithms + Data Structures = Evolutionary Programs", Springer Verlag, Berlin, 1992.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £145 +P&P)