Genetic algorithms with age structure

被引：17

作者：

N. Kubota

T. Fukuda

机构：

[1] Department of Mechanical Engineering,

[2] Osaka Institute of Technology,undefined

[3] 5-16-1 Omiya,undefined

[4] Asahi-ku,undefined

[5] Osaka 535,undefined

[6] Japan,undefined

[7] Department of Micro System Engineering,undefined

[8] Nagoya University,undefined

[9] 1 Furo-cho,undefined

[10] Chikusa-ku,undefined

[11] Nagoya 464-01,undefined

[12] Japan,undefined

来源：

Soft Computing | 1997年 / 1卷 / 4期

关键词：

Key words Genetic Algorithm; Age Structure; Optimization; Knapsack Problem;

D O I：

10.1007/s005000050017

中图分类号：

学科分类号：

摘要：

This paper deals with genetic algorithms with age structure. Evolutionary optimization methods have been successfully applied to complex optimization problems, but the evolutionary optimization methods have a problem of bias in candidate solutions due to genetic drift in search. To solve this problem, we propose the introduction of age structure into genetic algorithms as a simple extension. In nature, an individual is removed from a population when the individual reaches lethal age. Therefore, genetic algorithms with age structure (ASGA) can maintain the genetic diversity of a population by removing aged individuals from the population. First, we conduct simple simulations of two subpopulations considering the age structure. Next, we apply the ASGA to a kanapsack problem. Finally, we discuss the optimal parameters for the age structure of the ASGA. These simulation results indicate that the ASGA can control selection pressure by aging process and relatively maintain the genetic diversity of a population.

引用

页码：155 / 161

页数：6