一种基于PSO算法的闭环辨识方法

闭环模型辨识一直是工业先进控制领域中的一个主要课题。而现在基于粒子群优化算法PSO的辨识,大多都是连续开环系统的辨识。离散闭环模型辨识在计算机控制、运算量等方面比连续开环系统的辨识有较大的优势。文中讨论了PSO的时变惯性权重算法与参数初值的设置和选择方法。通过仿真实验表明,PSO与最小二乘递推算法相比,在有效性和一致性方面,有着明显的优势。PSO算法是一种有效地解决优化问题的群集智能算法,它的突出特点是算法中需要选择的参数少,程序实现简单,并在种群数量、寻优速度等方面较其他进化算法具有一定的优势。该方法在实际项目的应用中取得了较好的效果,应用前景广阔。

An Identification Method of Closed-loop Based on PSO Optimization

A novel method was studied for discrete closed-loop identification based on particle swarm optimization （PSO） algorithm. The time-varying inertia weight PSO algorithm and parameter initial value setting and selecting methods were discussed. Through the simulation results show that, PSO is more advantage than the reeursive least squares algorithm in the aspects of effectiveness and consistency. PSO algorithm is an effective solution to optimization problem of swarm intelligence algorithm, the advantages of which are easy to implement, few parameters to adjust, small population size, quick convergence ability and so on. This method displays an approving effect in praeti- cal project and has an extensive foreground.