Computational Technology Resources - CCP

Keywords: crossover, mutation, population size, large truss structures, value encoding, binary encoding.

Summary

There are many parameters affecting the performance of genetic algorithms. Among these parameters, the important ones are the probabilities of crossover and mutation operators, population sizes, and encoding types.

Probabilities of crossover and mutation operators affect the result of the corresponding problem. Population size affects the run-time of corresponding problem. Encoding types are another important factor in encoding of design variables from the run-time and computer memory points of view. But, no references have been found in the literature studying and comparing the effects of these parameters from the run-time and computer memory perspective.

In this paper, the effects of different probabilities of crossover and mutation operators, population sizes, binary encoding and value encoding are studied. For this aim, a genetic algorithm program coded in FORTRAN is run for large truss structures. This program includes displacement, stress and stability constraints. The results of value encoding and binary encoding are compared with each other to show the advantages and disadvantages of them. Also, the run-times of the solutions of truss problems and the computer memories used in the value and binary encodings are presented.

For the purpose of the study, 60-bar [1] and 940-bar [2] truss structure examples taken from the literature are solved. Generally, the ratio of crossover is taken as higher than 0.80 and the ratio of mutation is taken lower than 0.010. Variations of crossover and mutation probabilities are studied for the 60-bar truss structure.

It is concluded that the algorithm coded by using value encoding requires less run-time and computer memory when compared with the algorithm coded by using binary encoding. It is also concluded that when the high ratio of crossover and low ratio of mutation are used, the result will be close to the optimum solution.

References

1: H.J.C. Barbaso and A.C.C. Lemonge, "A New Adaptive Penalty Scheme for Genetic Algorithms", Information Sciences, 156, 215-251, 2003. doi:10.1016/S0020-0255(03)00177-4
2: M.R. Ghasemi, E. Hinton and R.D. Wood, "Optimization of Trusses Using Genetic Algorithms for Discrete and Continuous Variables", Engineering Computations, 16(3): 272-301, 1999. doi:10.1108/02644409910266403

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 £62 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 87 PROCEEDINGS OF THE NINTH INTERNATIONAL CONFERENCE ON THE APPLICATION OF ARTIFICIAL INTELLIGENCE TO CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING Edited by: B.H.V. Topping Paper 12 Effects of Some Parameters in a Genetic Algorithm for Large Truss Structures T. Dede¹, S. Bekiroglu¹ and Y. Ayvaz² ¹Graduate School of Natural and Applied Sciences, ²Department of Civil Engineering, Karadeniz Technical University, Trabzon, Turkey doi:10.4203/ccp.87.12 purchase the full-text of this paper Full Bibliographic Reference for this paper T. Dede, S. Bekiroglu, Y. Ayvaz, "Effects of Some Parameters in a Genetic Algorithm for Large Truss Structures", in B.H.V. Topping, (Editor), "Proceedings of the Ninth International Conference on the Application of Artificial Intelligence to Civil, Structural and Environmental Engineering", Civil-Comp Press, Stirlingshire, UK, Paper 12, 2007. doi:10.4203/ccp.87.12 Keywords: crossover, mutation, population size, large truss structures, value encoding, binary encoding. Summary There are many parameters affecting the performance of genetic algorithms. Among these parameters, the important ones are the probabilities of crossover and mutation operators, population sizes, and encoding types. Probabilities of crossover and mutation operators affect the result of the corresponding problem. Population size affects the run-time of corresponding problem. Encoding types are another important factor in encoding of design variables from the run-time and computer memory points of view. But, no references have been found in the literature studying and comparing the effects of these parameters from the run-time and computer memory perspective. In this paper, the effects of different probabilities of crossover and mutation operators, population sizes, binary encoding and value encoding are studied. For this aim, a genetic algorithm program coded in FORTRAN is run for large truss structures. This program includes displacement, stress and stability constraints. The results of value encoding and binary encoding are compared with each other to show the advantages and disadvantages of them. Also, the run-times of the solutions of truss problems and the computer memories used in the value and binary encodings are presented. For the purpose of the study, 60-bar [1] and 940-bar [2] truss structure examples taken from the literature are solved. Generally, the ratio of crossover is taken as higher than 0.80 and the ratio of mutation is taken lower than 0.010. Variations of crossover and mutation probabilities are studied for the 60-bar truss structure. It is concluded that the algorithm coded by using value encoding requires less run-time and computer memory when compared with the algorithm coded by using binary encoding. It is also concluded that when the high ratio of crossover and low ratio of mutation are used, the result will be close to the optimum solution. References 1 H.J.C. Barbaso and A.C.C. Lemonge, "A New Adaptive Penalty Scheme for Genetic Algorithms", Information Sciences, 156, 215-251, 2003. doi:10.1016/S0020-0255(03)00177-4 2 M.R. Ghasemi, E. Hinton and R.D. Wood, "Optimization of Trusses Using Genetic Algorithms for Discrete and Continuous Variables", Engineering Computations, 16(3): 272-301, 1999. doi:10.1108/02644409910266403 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 £62 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions