The Research of PSO Algorithms with Non-linear Time-decreasing Inertia Weight

Abstract

The particle swarm optimization (PSO) firstly proposed by Eberhart and Kennedy, is a computational intelligence technique. The inertia weight is an important parameter of PSO algorithm. In this paper, we designed 2 nonlinear time-decreasing inertia weight to use in GCPSO algorithm. At last a series of experiment is performed to test the performance of GCPSO with different inertia weight function. for most case, The result indicates the nonlinear time-decreasing inertia weight, especially the convex nonlinear time-decreasing inertia weight has a better performance than linear time-decreasing inertia weight and constant inertia weight.