Niching particle swarm optimizer with entropy-based exploration strategy for global optimization

Abstract

As a kind of evolutionary algorithms, particle swarm optimization is famous for its simplicity and efficiency in optimization. However, for complex problems, PSO is prone to be trapped into the local optima. To address this issue, a particle swarm optimizer with niching strategy and entropy-based exploration strategy (PSO-NE) is proposed in this paper. To be specific, on one hand, a distance based niching strategy and the competitive learning strategy are adopted to design the exploitation in PSO-NE; on the other hand, the exploration in PSO-NE is achieved by an entropy based exploring strategy. With such kind of designs, the exploitation and exploration in PSO-NE can be dependently adjusted, which is beneficial for balancing these two factors. To validate the effectiveness of the proposed algorithm, extensive experiments have been conducted based on 28 benchmarks from CEC’ 2013. The proposed algorithm shows its competitive performance with comparing to six other typical variants of PSO.