基于佳点集构造的改进量子粒子群优化算法

Improved quantum particle swarm optimization based on good-point set

针对粒子群优化算法易出现早熟收敛及局部搜索能力不足的特点,提出一种改进的量子粒子群优化算法(IQPSO)。该算法在量子粒子群优化算法(QPSO)的基础上,引入佳点集初始化量子的初始角位置,提高初始种群的遍历性;在粒子角速度位置更新中,采用混沌时间序列数,促使粒子跳出局部极值点;为避免粒子陷入早熟收敛,在算法中加入变异处理。仿真实验结果表明:与标准粒子群优化(SPSO)算法和量子粒子群优化(QPSO)算法比较,提出的算法具有快速的收敛能力、良好的稳定性,其优化性能有较明显的提高。

In order to solve the problems of premature convergence and poor local search on particle swarm optimization （PSO） algorithm, an improved quantum particle swarm optimization（IQPSO） approach was proposed. Based on quantum particle swarm optimization algorithm （QPSO）, good-point set was introduced to the approach to initialize initial angle of quantum position, to improve ergodicity of initial population. To make particle jump out of local extreme value point, the chaotic time series numbers were used to update particle velocity. To prevent particle from premature convergence, mutation process was added in the approach. The simulation experiment results show that the improved algorithm has rapid convergence, good stability and it gives better performance than standard particle swarm optimization （SPSO） and quantum particle swarm optimization （QPSO）.