SELF-ADAPTIVE INTEGER AND DECIMAL MUTATION OPERATORS FOR GENETIC ALGORITHMS

摘要：

Evolutionary algorithms are affected by more parameters than optimization methods typically. This is at the same time a source of their robustness as well as a source of frustration in designing them. Adaptation can be used not only for finding solutions to a given problem, but also for tuning genetic algorithms to the particular problem. Adaptation can be applied to problems as well as to evolutionary processes. In the first case adaptation modifies some components of genetic algorithms to provide an appropriate form of the algorithm, which meets the nature of the given problem. These components could be any of representation, crossover, mutation and selection. In the second case, adaptation suggests a way to tune the parameters of the changing configuration of genetic algorithms while solving the problem. In this paper two new selfadaptive mutation operators; integer and decimal mutation are proposed for implementing efficient mutation in the evolutionary process of genetic algorithm for function optimization. Experimentation with 27 test cases and 1350 runs proved the efficiency of these operators in solving optimization problems.

关键词： Evolutionary algorithm Genetic algorithm Function optimization Mutation operator Self-adaptive mutation operators Integer mutation operator Decimal mutation operator Fitness evaluation and analysis

作者: Ghodrat Moghadampour

作者单位: Vaasa University of Applied Sciences, Wolffintie 30, 65200 Vaasa, Finland

会议类型: 国际会议

会议名称: 13th International Conference on Enterprise Information System(第13届企业信息系统国际会议 ICEIS 2011)

会议地点: 北京

会议语种:英文

页码: 2651-2658

在线出版日期: 2011-06-08（万方平台首次上网日期，不代表论文的发表时间）