摘要
为了避免局部最优与收敛速度慢的问题,在粒子群算法中引入"竞争"与"拉伸"技术,提出一种新的粒子群算法——竞争-拉伸粒子群算法。用3个基准函数对标准粒子群算法、局域粒子群算法、提出的新算法等同时进行测试,对比结果表明,提出的算法不仅具有较高的收敛速度,而且能有效地进行全局搜索。
In order to avoid the problem of partial optimization and slow-footed convergence, the technology of competition and stretch is introduced and a new PSO-competition-stretch PSO is proposed. Three datum functions are employed to simultaneously test standard PSO, partial PSO and the new PSO proposed. The result show that the new arithmetic not only has the high convergence speed but also can conduct global search effectively.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第22期5802-5804,5809,共4页
Computer Engineering and Design
基金
陕西省自然科学基金项目(98X11)
陕西省教育厅重点科研计划基金项目(00JK015)