论文部分内容阅读
针对传统粒子群优化(PSO)算法在处理复杂函数优化问题时容易陷入局部最优、迭代后期收敛速度慢的问题,提出一种具有成长特性的粒子群优化算法(GPPSO).该算法根据人类成长的特性被分为3个阶段:前期阶段,为速度更新公式增加叛逆项,以降低进入早熟收敛的概率;中期阶段,为平衡全局与局部的搜索,通过对粒子群信息的整合,为速度更新公式添加平衡项;后期阶段,在速度更新公式中去除速度项,充分利用前期粒子进化得到的经验进行局部寻优.同时给出成长阶段划分的两个依据.运用典型函数进行测试,实验表明该算法对于提高收敛性能具有明显优势.
Aiming at the problem that traditional Particle Swarm Optimization (PSO) is easy to fall into the local optimum when dealing with the problem of complex function optimization and the convergence speed is slow in the late iteration, a Particle Swarm Optimization (GPPSO) algorithm with growth features is proposed. Is divided into three stages: the early stage, the recursive term is added to the velocity update formula to reduce the probability of entering the premature convergence. In the middle stage, in order to balance the global and local search, through the integration of particle swarm information, Formula to add the balance term; in the later stage, the velocity term is removed in the velocity update formula, and the partial optimization is made by using the experience obtained from the particle evolution in the early stage.At the same time two criteria for the division of the growth phase are given.Using the typical function to test, The algorithm has obvious advantages for improving the convergence performance.