Dynamic Data Objects Based On Particle Swarm Optimization Algorithm Recursive Estimation Problem

Particle Swarm Optimization algorithm is a kind of group intelligent algorithm used for optimal. The advantages of the algorithm are simple recursion formula, fast convergence and high efficiency. However, with regard to recursive estimation problems,which widely exist in adaptive filter, corrector, controller real-time parameter control, basic PSO is not fit for processing dynamic data real-time.This thesis, based on the particle swarm optimization algorithm, inspired by recursion algorithm, this paper applies recursive particle swarm optimization algorithm to solve optimization problems in dynamic data environments. Recursive particle swarm optimization algorithm calculates the new solution based on the previous solution and the current input, output observation data, according to the simple recursive formula. Through this continuous recursive calculation to get a satisfactory solution. The algorithm inherites recursion algorithm's advantage being convenient for dynamic data object in recursive optimization problems and the particle swarm optimization algorithm's superiority of fast convergence speed, good real-time, and strong adaptive ability in nonstationary environments. What is more, recursive particle swarm optimization algorithm is fit for real-time estimation of time varying parameter.In this article, the recursive particle swarm optimization algorithm and recursive least square algorithm performance are compared in the simulate experiments of optimization problems of adaptive FIR filter optimal weight coefficient. Recursive particle swarm optimization algorithm is proved to be advantage. And its application in optimization of the parameters of the radial basis function (RBF) network simulation system is simulated. The simulation results verify the feasibility of recursive particle swarm optimization algorithm.