基于微分进化算子和混沌扰动的量子粒子群算法

Im proved quantum-behaved particle swarm optim ization algorithm based on d ifferential evolu tion operator and chaos d isturbance

在分析量子粒子群算法搜索特性的基础上,引入了微分进化算子以及混沌扰动,提出了基于微分进化算子和混沌扰动的量子粒子群算法:在粒子搜索的过程中,以一定的概率对粒子的每一维执行微分进化操作,以增加粒子的随机性,减少粒子因为多样性的缺失而陷入局部最优的概率;并根据混沌的随机性和遍历性在性能较差的部分粒子中加入混沌扰动,从而增加粒子局部搜索精度,提高算法的优化性能.通过多个测试函数的试验对比表明,文中提出的算法是可行且有效的.

By analyzing characters ofQuantum-behaved Particle Swarm Optim ization A lgorithm（QPSO）,an improved algorithm called CQPSO-DE is proposed in this study.The algorithm introduces the ideas of d ifferential evolution（DE） operator and chaos d isturbance.During search procedure,every d imension of particles executes the DE operator accord ing to a certain probability.This can increase randomness and avoid local optimums.Fur-thermore,in order to enhance the particles′search accuracy,applying the randomness and ergod icity of chaos d isturbs the worse part of particles.At last,some benchmark functions are used to test performance of the algo-rithm.Computing results demonstrate the effectiveness and feasibility of the CQPSO-DE algorithm.