Chaotic particle swarm optimization with golden section strategy

Particle swarm optimization (PSO), as one of the meta-heuristic optimization approaches, has the advantages of few adjustable parameters and fast searching process, but it still has the demerits of low convergence precision and easily falling into local optimal. Therefore, a chaotic particle swarm optimization with a golden section updating strategy (abbreviated as CPSOG) is proposed in this paper. To begin with, the chaotic mapping method is employed to produce high-quality initial particles for the swarm. Next, a method based on the inertia weight strategy is utilized to balance the global and local search ability via the velocity and position formulas for particles. Subsequently, a golden section update strategy is adopted to refine the obtained optimal solution, thereby, enhancing the convergence ability of the swarm. In addition, an adaptive Levy flight updating mechanism is used to achieve better global convergence. At length, both the simulation results of CPSOG method on 6 classical benchmark test functions and the convergence performance are detailed, which show that our proposal has much better robustness and stability through circumventing the disadvantages of canonical PSO algorithm.