Particle Swarm Optimization with Dynamic Inertia Weight and Mutation
Xuedan Liu, Qiang Wang, Haiyan Liu, Lili Li · 2009
The Particle Swarm Optimization (PSO) plunges into the local minimum easily. In order to overcome this shortcoming, we propose an improved PSO algorithm with the features of linearly decreasing of inertia weight and the re-initialization of the particle when it gets stagnated. The improved PSO is a local PSO and its topology is wheels. From the experimental results of three non-linear testing functions and a problem with non-convex solution space, it is obvious that the improved PSO algorithm greatly enhances the rate of global convergence.