基于新变异算子的改进粒子群优化算法

An Improved Particle Swarm Optimization Algorithm Based on New Mutation Operators

粒子群优化算法(PSO)是一种基于群体智能的优化算法。本文在介绍PSO算法基本原理和流程的基础上,分析了该算法在处理一些复杂问题时容易出现的早熟收敛、收敛效率低和精度不高等问题,提出了一种基于新变异算子的改进粒子群优化算法(NMPSO)。NMPSO算法将产生的变异粒子与当前粒子进行优劣比较,选择较优的粒子,增强了种群的多样性,有效地避免算法收敛早熟。用5个常用基准测试函数对两种算法进行对比实验,结果表明:新提出的NMPSO算法增强了全局搜索能力,提高了收敛速度和收敛精度。

Particle swarm optimization （PSO） is an optimization algorithm based on swarm intelli- gence. Based on introducing PSO＇s theory and flow, this paper analyzes the phenomenon that it suffers from premature convergence, longer search time and lower precision when dealing with complex problems. An improved particle swarm optimization algorithm based on new mutation operators（NMPSO） is presented. The mutation operator is compared with the current particles, and the better one will be selected. So the diversity of population is improved, which can help the algorithm avoid premature convergence efficiently. The comparative simulation results on five benchmark functions verify that NMPSO improves PSO＇s global search capability, convergence rate and precision.