Application of Particle Swarm Optimization Algorithm in Computer Neural Network

Abstract

Particle swarm optimization algorithm is a new intelligent optimization algorithm, which is based on the social activities of biological groups. Because the particle swarm optimization algorithm is random and uncertain, the theoretical basis of the algorithm is still not complete, for example, there are some problems, such as the premature convergence, slow convergence speed and so on. Particle swarm optimization algorithm theory analysis and improvement has become the focus and hotspot of the algorithm. Therefore, based on the theory of the traditional PSO algorithm analysis based on, combined with algorithm for biological prototype features, improved PSO algorithm is proposed for the convergence of the algorithm. Because the traditional BP artificial neural network learning algorithm is an optimization algorithm based on gradient information, algorithm is easy to fall into local optimum defects. Therefore, in this paper, the improved particle swarm optimization algorithm is used in neural network training. This study not only can improve the theoretical basis of the PSO algorithm, but also provide a reference for the application of PSO algorithm in artificial neural network.