Opposition-Based Bare Bone Particle Swarm Optimization

Bare bone particle swarm optimization (BPSO) is a simple approach for solving optimization problem. However, this population-based algorithm also suffers premature problem for some complex problems, especial for high-order dimensional, nonlinear problems. This paper presents a new approach to enhance BPSO's searching capability. The proposed opposition-based bare bone particle swarm optimization (OBPSO) employs opposition learning strategy to extend the exploration capability such that avoiding get stuck on local optimum. A set of six benchmark functions is applied for numerical verification. Experimental results confirm the strength of the proposed approach, based on comparison with PSO and original OBPSO. It is seen that OBPSO outperforms PSO and BPSO both in solution accuracy and convergent rate.