Inertia Weight Particle Swarm Optimization with Boltzmann Exploration

摘要

This paper proposes a novel inertia weight particle swarm optimization (IWPSO) algorithm with Boltzmann exploration (BPSO). In allusion to the blindness in traditional IWPSO search process, we introduce the Boltzmann exploration strategy to adaptively tune the weights of individual and social cognition terms in velocity update equation, aiming to balance the exploration and exploitation in search process. The proposed algorithm can guide particles searching for the most promising region in search space and adjust the weights adaptively. Eight typical multi-modal functions are used to validate the proposed algorithm. The experimental results show that our algorithm consistently outperforms inertia weight PSO (IWPSO), constriction factor PSO (CFPSO), unified PSO (UPSO), adaptive fuzzy PSO (AFPSO), quadratic interpolation PSO (QIPSO), and dynamic multi-swarm PSO (QMSPSO).