A Confrontational Bare Bones Particle Swarm Optimization Algorithm

Abstract

Optimization problems are common and important in artificial intelligence. Population-based methods are usually used in solving these problems. However, in recent years, optimization problems become complicated and high dimensional. Traditional methods are difficult to present effective results when searching in the high dimensional hyper-cube. To cross these shortcomings, a novel confrontational bare bones particle swarm optimization (CBBPSO) algorithm is proposed in this paper. Different from traditional population-based algorithms, the global worst particle is also recorded by the CBBPSO. A new confrontation operator is proposed. The confrontation operator offers every particle a second choice. In each iteration, each particle can move to the global best particle or the global worst particle. To verify the optimization ability of the proposed method, five famous benchmark functions from CEC2005 are used in the experiment. Experimental results show that the CBBPSO can solve high-dimensional optimization problems.