Research On Modification Of Particle Swarm Optimization Algorithm Based On Control

As one kind of swarm algorithm with profundity intelligence background, particle swarm optimization algorithm(PSO) has characters of can be applied in a flexible way, achieved easily and cooperative search,is an effective way to solve complex optimization problems. Therefore, it has both theory and application significance to analysis evolution characters and inner rules of particle swarm optimization algorithm and improve its optimizing performance.This article based on control thoughts; discussed particle swarm optimization algorithm from aspects of deterministic control, stochastic control, process optimization. The specific research content was as follows:firstly introduces fundamentals of particle swarm optimization algorithm, this article systematically reviews theory research status quo of algorithm and research results of algorithm based on control thoughts at home and abroad. Because the limited convergence process of PSO can be seen as a control problem, this paper advances particle swarm optimization algorithm based on the Model-free adaptive control. On the basis of basic particle swarm optimization algorithm, it takes the current optimal position as set value and uses the Model-free adaptive control to realize control of searching process. A verdict of simulation indicated, introducing the Model-free adaptive controller makes the convergence rate improved obviously and searching precision improved to some extend. Based on the analysis of deterministic system, this article considers inherent stochastic characteristics of algorithm, redefines PSO algorithm to be discrete uncertain time-delay system and advances PSO algorithm based on Delay Dependent Robust Control for Systems with Uncertainties(DURC-PSO). The simulation test result shows that DURC-PSO algorithm has good performance at searching precision. Meanwhile, consider the rationality of introducing controller and construct the hybrid algorithm on the basis of previous work Moreover, simulation conclusion proves the hybrid algorithm is efficiency. This article analysis characteristics of searching process and introduce process optimization thought into PSO algorithm. Moreover, analysis the optimal setting problem of control on swarm flight and defines the setting rules of process optimization. Thereby, formulates a preliminary particle swarm optimization algorithm based on process optimization. The simulation test result validates feasibility and effectiveness of the method.