摘要
为克服rand/1和best/1两种变异策略存在的缺陷,提出分工差分进化算法.该算法结合rand/1变异策略全局搜索能力强和best/1变异策略局部搜索能力强、收敛速度快的特点,在进化过程中对个体进行分工,优秀个体选择best/1策略承担开发任务,一般或较差个体选择rand/1变异策略承担探索任务,通过个体分工负责从而提高算法性能.对典型函数的测试结果证明,新算法能够大大提高算法的收敛速率和全局搜索能力.
Differential Evolution algorithm with division of labor is proposed in this paper to overcome the drawbacks of rand/1 and best/1 mutation strategies. Combining with the properties of good local searching ability fast convergence speed of best/1 mutation strategy and the properties of good global searching ability of rand/1 mutation strategy, the algorithm employs the ideal of division of labor and give different individual different tasks. Excellent individuals choose best/1 to be charged with exploitation tasks and the other individuals choose rand/1 to be charged with exploration tasks. The algorithm＇s performance is improved through the cooperation of individuals with different tasks. Experiments on benchmark functions show that the proposed algorithm can improve the convergence speed and the global searching ability greatly.
出处
《小型微型计算机系统》
CSCD
北大核心
2009年第7期1302-1304,共3页
Journal of Chinese Computer Systems
基金
国家装备重点项目(EP-030093-1)资助
关键词
差分进化算法
变异策略
分工
全局搜索
differential evolution algorithm
mutation strategy
division of labor
global search
