Research of Particle Swarm Optimization algorithm based on Nelder-Mead simplex and its application on partial discharge parameter recognition

Particle Swarm Optimization (PSO), firstly presented in 1995, is mainly used in high-dimension optimization. Despite its wide use, PSO have a disadvantage of prematurity. This paper put forward a new Particle Swarm Optimization algorithm based on Nelder-Mead simplex algorithm to overcome the PSO´s nature of prematurity and precision problems. Nelder-Mead simplex algorithm is hybrid into the process of PSO. The test results of high dimensional Griewank function show that this novel algorithm is efficient to solve high-dimension optimization problem with a balance of convergence and precision. Finally, an example of partial discharge parameter recognition shows this novel algorithm have advantage to solve these type of problem.