论文部分内容阅读
原始粒子群优化算法(PSO)和各种改进方法存在着参数取值固定、收敛精度低等问题.为此,提出一种采用抽样策略的粒子群优化算法(SS-PSO).通过拉丁超立方抽样(LHS)策略更新粒子速度和位置,以加快收敛速度;提出一种基于随机采样的最优位置修正方法,以微调全局最优;提出“双抽样”LHS局部搜索方法,以提高收敛精度.与其他新近提出的两个算法进行对比,结果显示SS-PSO在一定程度上提高了算法的性能.
Particle Swarm Optimization (PSO) and various improved methods have some problems such as fixed parameter values and low convergence accuracy, so a particle swarm optimization algorithm (SS-PSO) with sampling strategy is proposed.Through the Latin Hypercube The sampling (LHS) strategy updates the particle velocity and position to speed up the convergence rate. An optimal position correction method based on random sampling is proposed to fine tune the global optimum. A “double sampling” LHS local search method is proposed to improve the convergence Accuracy.Compared with other two newly proposed algorithms, the results show that SS-PSO improves the performance of the algorithm to a certain extent.