关键词: 粒子群优化, 惯性权重, 位置变异, 全局搜索, 局部搜索

Abstract: Standard Particle Swarm Optimization（SPSO） easily leads to premature convergence in optimizing high-dimensional functions，to overcome this shortcoming，a New Particle Swarm Optimization algorithm with Position Mutation（NPSO-PM） is proposed.To balance the ability of local search and global search of PSO，NPSO-PM alternately uses random inertia weight and inertia weight of SPSO in possibility；to enhance the population diversity and to restrain premature convergence of PSO，the position of particles，which meet certain conditions，mutates in a method that can make particle escape from local areas effectively in each iteration.Finally，5 benchmark functions on 60 and 90 dimensions simulation experiments show that proposed algorithm has high convergence precision and overcomes premature convergence effectively.

Key words: Particle Swarm Optimization（PSO）, inertia weight, position mutation, global search, local search