摘要
遗传算法中,部分局部最优个体常常会抑制种群中其他个体的成长,使进化停滞,结果陷入局部最优.对此,提出一种基于多样化成长策略的遗传算法,通过采用迁移杂交、多态变异和群体突变等方法,对种群中不同适应度的个体给予不同的进化成长策略,以保持个体间的多样性,从而增强算法的寻优能力.对改进算法的算法复杂度进行了分析,并通过算例表明,多样化成长策略的各个部分均对改善寻优结果发挥了作用.
In genetic algorithms, some local optimum individuals often inhibit the development of other individuals, so the evolution stagnates with a result of local optimum. Therefore, a genetic algorithm based on diversified development strategy is proposed, which assigns different individuals of different fitness with different development strategy, so as to diversify the population, and enhance the algorithm＇s optimizing ability through migrated crossover, divergent mutation, collective saltation. The complexity of the algorithm is analyzed. Benchmark example shows that all parts of the strategy play roles for optimum search.
出处
《控制与决策》
EI
CSCD
北大核心
2009年第12期1801-1804,共4页
Control and Decision
基金
国家自然科学基金项目(60474059)
国家863计划项目(2006AA04Z160)
关键词
遗传算法
多样化成长
迁移杂交
多态变异
群体突变
Genetic algorithm
Diversifed development
Migrated crossover
Divergent mutation
Collective saltation
