Improving Particle Swarm Optimization of Its Convergence

Abstract

There are few controllable parameters in the particle swarm optimization algorithm and its value is fixed, it has some problems such as premature convergence and falling into local optimal solutions. In this paper, an improved particle swarm optimization algorithm is proposed. By dynamically adjusting the inertia weight and introducing a behavior learning factor to update the particle's velocity vector and the particle has more movement diversity in each iteration. Through the demonstration of experimental data, compared with the random weight particle swarm optimization algorithm and the linear decreasing weight particle swarm optimization algorithm, the improved particle swarm optimization algorithm effectively accelerates the convergence speed of the algorithm, improves the optimization performance of the algorithm, and has good convergence and optimization results, the number of convergence iterations is reduced by at least 38.8%, and the accuracy of the convergence accuracy of the optimal solution and the average solution all is improved by at least 28.9%.