A circle chaos random search strategy particle swarm optimization with its application

Abstract

As a popular optimization, particle swarm optimization has attracted many researcher's attention and also been successfully applied to solve many real world problems. As adopting the strategy of history experience, the convergence rate of PSO is faster than other evolutionary algorithms. But the individuals of PSO are easily be trapped into local optimum, which causes the algorithm could not guaranty confidence of the final results. In this paper, for further enhance the global search ability of PSO, based on comparing different chaos based random strategies, we propose a circle chaos based mutation strategy into PSO algorithm. First, compared with other operators, since the circle chaos based operator could generate more even results, we propose the circle operator based random strategy to balance the global and local search of our PSO algorithm. Second, for attracting individuals to precisely search in a promising region, a new historic experience based elite array is proposed to lead the search direction of population. Then the algorithm could achieve more precise results. For further proving the effectiveness of our algorithm, we apply the algorithm into PID fuzzy logic controller design. Compared with other EA algorithms, the experimental results demonstrate our methods achieve better performance on convergence rate and precise of results.