Improved Particle Swarm Optimization based on Chaotic Cellular Automata

In this paper, a new improved Particle Swarm Optimization (PSO) combined with Chaotic Cellular Automata (CCA) has been proposed. PSO is sensitive to initial conditions and values like other stochastic search algorithms. In the proposed method, features of chaotic Pseudo Random Number Generator (PRNG) are used to move particles in the problem space. This factor leads to the appropriate random behavior of particles in the space that is capable of high exploitation ability and also changes in the coefficient inertia w with big steps moving from converging prematurely and falling in local minimum. In the proposed method, by combining small steps by CCA, that has high exploitation ability and with a large step changes in the coefficient inertia w the high exploration ability leading to balance in the random behavior of the algorithms. Proposed method display good performance for searching the problem space compared with other algorithms.