Derivation of a generalized PSO model and convergence analysis

摘要

The convergence of Particle Swarm Optimization (PSO) is an important and difficult problem. In this article, a new generalized PSO analysis model (γ-GPSO) is offered with discretization technique to second-order analog system of PSO; Order one or second stability areas of parameters is confirmed Using dynamic random analysis method; Stability region, the contour of the spectral radius, the mean, variance and covariance track graphics is drawn by Matlab and Better reveal the inherent rules of the PSO for reasonable configuration of parameters and the design of efficient PSO algorithm.