A Local Search Particle Swarm Optimization with Dual Species Conservation for Multimodal Optimization

摘要

This work presents a new optimization technique called dual species conservation particle swarm optimization (DSPSO) for finding multiple optima (global or local) of multimodal functions. The basis of the proposed algorithm is repeatedly using species conservation and hill-valley detecting mechanism to refine the species set. To improve the balance between exploration and exploitation of the standard Particle Swarm Optimization (PSO), a local search around found optima strategy is adopted in PSO. The performance of DSPSO is validated on a set of widely used multimodal benchmark functions. Numerical results show that the proposed technique is effective and efficient in finding multiple solutions of selected benchmark.