Research on Chaotic Variable Velocity Limit Particle Swarm Optimization with Modified Random Benchmark

The standard Particle Swarm Optimization(PSO) will be easily trapped by the local optimization of the optimization problem. To overcome this shortage, the Chaotic Variable Velocity Limit Particle Swarm Optimization(CVVL-PSO) is proposed. It shrinks the velocity limitation with iteration going on. And it changes the inertia weight in the pattern of exponential decreasing. The algorithm introduces the modified tent series into the velocity updating process of particle. The modified random benchmark with random local optima and random radius is proposed. The global optimized point is generated randomly. And it is applied to compare with the classic benchmarks to testify the searching ability of several PSOs. In the numerical experiment, the results show the searching ability and wider feasibility of the CVVL-PSO.