非对称边界变异的分层并行量子粒子群算法

摘要

This paper proposed a new multilevel parallel quantum-behaved particle swarm optimization algorithm based on asymmetrical boundary variation(MQPSOBV).This algorithm is designed to solve the problem of crossing the predefined boundary that standard PSO algorithm may often suffers.By introducing the concept of multilevel,the algorithm set the asymmetrical effective regions separately,as well as the layers parameters.And they operated in parallel.So it occurred variation near the respond boundary and maintainred the swarm among the effective region if the swarm was beyond the asymmetrical boundary.Test resuits show that the proposed algorithm effectively overcomes the shortcomings of standard PS0.Its accuracy and the performance of global search for optimal solutions have been greatly improved.Therefore,it provides a perfect application in the practice.%针对粒子群算法在非对称可行性区间经常发生越界的问题,提出了一种非对称边界变异的分层并行量子粒子群算法(MQPSOBV).该算法中引入分层思想,将粒子非对称可行性区间分层设置和并行运行；当粒子越界时,对越界粒子在非对称上下边界进行相应变异,从而使算法完全控制粒子越界行为,有效地克服了粒子群算法的缺陷.测试结果表明,该算法在精度和全局搜索能力方面有了很大的提高,具有一定的实际应用价值.