The double chaotic particle swarm optimization with the performance avoiding local optimum

The research with respect to Particle Swarm Optimization is concentrated in improving their performance on avoiding local maxima. Since standard Particle Swarm Optimization does not perform well in many cases, we propose double chaotic particle swarm optimization algorithm based on logistic map. This chaotic movement has good randomness and ergodic statistics property of chaos sequence. We propose to use chaotic sequence to initialize the particle positions, laying a solid foundation for the diversity of search particle swarm. The improved strategies in the algorithm increase the premature stagnation judgment that the new particles are added into a new region making changes in the trajectory of particles, which help algorithm escaping from local optima effectively and reduce invalid iterations. This strategies result in greatly improving the convergence of the algorithm, accuracy and speed optimization. On the other hand, the proposed algorithm requires very little number of particles and few iterations to fully meet the theory test function optimization. The results of the simulation show the good performance of the optimization algorithm.