Modified Particle Swarm Optimization Combined with Trigonometric Function and Variable Neighborhood Search

Based on the inertia weight and acceleration coefficients of trigonometric function and variable neighborhood search, a modified particle swarm optimization is discussed. The modified PSO with the parameters of the trigonometric function has powerful global exploration at the beginning and ending of the evolution, and strongly local exploitation in the interim of run. Through analyzing the behavior of the PSO, three novel heuristic neighborhood search rules are suggested to enhance the robustness and convergence. The proposed approach demonstrates its superiority in convergence, robustness and solution quality with the experiment tested on the benchmark multimodal function.