Advanced optimization by progressive mapping search method of PSO and neural network

Abstract

This paper proposes a novel optimization approach by the fusion of the Progressive Mapping Search Method (PMSM) and the Neural Network (NN) aided Particle Swarm Optimization (PSO) that can obtain the global optimal solutions easily and speed up the overall search procedure. The PMSM merged with the NN and PSO has an important role as the navigation when the PSO is searching all the areas in order to acquire the optimum. It can help to improve the search capability of the original PSO method. That is, the PMSM together with the NN and PSO is trained to capture the PSO-searched solutions. To verify and demonstrate the effectiveness of our technique, we use a total of four test functions. The PMSM strategy employed in our paper is faster than the traditional PSO algorithm in all these four test functions. We also apply this new optimization scheme in the AVR (Automatic Voltage Regulator) system of the thermal power plant, which has resulted in faster and more stable responses. © 2013 Springer International Publishing.