基于蝙蝠算法的PID参数整定  

BA-based PID Tuning Method

作　　者：吕磊[1] 章国宝[1] 黄永明[1] 

机构地区：[1]东南大学自动化学院,南京210096

出　　处：《控制工程》2017年第3期548-553,共6页Control Engineering of China

摘　　要：建立在PID参数整定是一种优化问题这一本质上,尝试使用一种新的群体智能算法(蝙蝠算法)整定PID参数。为了通过仿真实验验证这种PID参数整定方法的可行性,文中选取被控系统的ITAE指标作为蝙蝠算法的目标函数,对于越界粒子采取边界缓冲墙策略。以五种常见过程控制系统模型为被控对象,分别使用Ziegler-Nichols方法,粒子群算法和蝙蝠算法获得PID控制器的参数,并对比了这三种方法所得到PID控制器的闭环系统性能,以及粒子群算法和蝙蝠算法的运行效率。实验结果表明,蝙蝠算法不仅在获得的PID控制器性能上优于Ziegler-Nichols方法和粒子群算法,并在算法运行结果的稳定性和对初始种群分布的依赖性方面优于粒子群算法。Based on the fact that PID parameter tuning is a problem of global optimization, this paper tried to combine a new swarm intelligence algorithm （bat algorithm） with PID parameter tuning, to find a new PID tuning method. In order to prove the practicability of this method through simulation experiments, this paper adopted ITAE criteria as the objective function of the bat algorithm and used the tactic of boundary wall buffer to deal with the bats across the boundaries. This paper used the bat algorithm, particle swarm optimization, and Ziegler-Nichols method to tune the PID parameters of five different control system models. Additionally, this paper compared the closed-loop system＇s performance of these three methods and the efficiency of the bat algorithm and particle swarm optimization. The experimental results show that the closed-loop system＇s performance of the bat algorithm is preferable to particle swarm optimization and Ziegler-Nichols method, the bat algorithm is more independent from the initial population than particle swarm optimization, and the change of the bat algorithm results in different experiments is smaller than particle swarm optimization.

关 键 词：PID控制 蝙蝠算法 PID参数整定 粒子群算法 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]