提出一种改进的粒子群优化算法——基于全局劣汰策略的混合粒子群优化算法(GTPSO)。GTPSO在保持PSO算法快速收敛的基础上,以郭涛算法(GuoA)的寻优机制确保种群的多样性和算法的坚韧性。数值计算结果表明,对于高维(维数≥10)复杂非凸多峰函数的数值优化问题,GTPSO算法的计算结果均优于GuoA算法和粒子群优化算法。
This paper proposed a modified PSO : hybrid particle swarm optimization based on global inferior-substitution strate- gy(GTPSO). GTPSO not only kept the virtue that the convergence velocity of PSO was faster, but also had the diversity and stronger of GuoA. The numerical results show that GTPSO is more superior to PSO and GuoA for complex and non-convex multimodel function optimizaion problems,