[关键词]
[摘要]
针对粒子群算法易陷入局部最优问题, 从数学角度分析了粒子群算法易陷入局部最优的理论原因, 提出一种 自适应混沌变异粒子群算法, 对陷入局部最优的粒子产生变异, 增加算法的遍历性、种群的多样性, 以跳出局部最优 解, 用来解决水库优化调度问题。与现有算法相比, 自适应混沌变异粒子群算法计算快, 稳定性强, 既避免了粒子群 算法陷入局部最优, 同时在一定程度上又保证了收敛性。
[Key word]
[Abstract]
The pa rticle sw arm o pt imizatio n( PSO) alg or ithm can easily fall into the local o ptimizat ion. I n this paper , the theo retic reasons of local o pt imization wer e analy zed f rom the mathematical perspect ive, and a new self2ada ptive chaotic mutatio n PSO al2 g or ithm w as pro posed and applied to a hy dr opow er statio n. The results showed t hat the new alg or ithm can mut ate the particles which fall int o the local o ptimization, incr ease the alg or ithm erg odicity and sw arm diver sity, and find the global o ptimizat ion so2 lut ion, w hich is useful fo r the o pt imal operat ion o f reservo ir . Compa red w ith the current alg or ithms, the self2adaptiv e chaotic mutation PSO alg or ithm has fast computation and str ong stability , avo ids the local o pt imization, and ensures the co nv erg ence.
[中图分类号]
[基金项目]
国家科技支撑计划项目( 2011BAD20B01) ; 山东省科技发展计划项目( 2011GGB01138)