基于种群规模动态减小的混合微粒群优化算法研究

Research on Hybrid PSO Algorithm Based on Dynamic Decrease of Population Size

针对基本微粒群优化(PSO,particle swarm optimization)算法存在早熟、易陷入局部极值等缺点,提出了一种改进的PSO优化算法。该算法分为全局搜索和局部搜索两个阶段。在全局搜索阶段采用基本PSO算法快速收缩搜索范围;在局部搜索阶段将PSO算法与模拟退火(SA,simulated annealing)算法结合,通过产生部分变异微粒确保算法能够跳出局部极值。同时为提高搜索效率,动态地减少种群规模。仿真结果表明,该算法具有较好的优化性能以及较高的执行效率。

An improved PSO algorithm is proposed for the disadvantages in the basic PSO algorithm such as premature convergence and easily trapping into local maxima.The improved algorithm is divided into overall and local search steps,in the first of which,basic PSO algorithm is used to decrease search space;in the second,the SA algorithm＇s thinking is injected to generate some worse particles and improve the search performance.In the same time,to heighten the search efficiency,the population size is dynamically reduced.The simulation results show that the improved PSO algorithm has better optimization performance and higher execution efficiency.