计算机与现代化 ›› 2013, Vol. 1 ›› Issue (9): 8-12.doi: 10.3969/j.issn.1006-2475.2013.09.002

• 人工智能 • 上一篇    下一篇

基于遗传算法和模糊神经网络的PID控制器参数优化方法

周 頔   

  1. 四川文理学院,四川 达州 635000
  • 收稿日期:2013-05-08 修回日期:1900-01-01 出版日期:2013-09-17 发布日期:2013-09-17

Parameters Optimization Method of PID Controller Based on Genetic Algorithm and Fuzzy Neural Network

ZHOU Di   

  1. Sichuan University of Arts and Science, Dazhou 635000, China
  • Received:2013-05-08 Revised:1900-01-01 Online:2013-09-17 Published:2013-09-17

摘要: 针对传统的PID控制器参数优化需要被控对象精确数学模型问题,利用不需要被控对象数学模型的模糊控制理论和神经网络的自适应和自学习的能力以及遗传算法的全局优化能力,提出一种基于遗传算法、模糊控制理论和神经网络相结合的PID控制器参数优化方法。该方法首先利用十进制编码对遗传算法进行编码,然后集中优化模糊神经网络参数和结构,接着再用优化的模糊神经网络确定PID控制器参数,获得模糊神经网络PID控制器。最后通过达州钢铁集团扎钢厂中央空调控制系统实际应用,结果表明该优化方法具有较强的抗干扰能力和鲁棒性。

关键词: 遗传算法, 模糊控制理论, 神经网络, 优化, PID控制器

Abstract: In allusion to the disadvantages of the traditional parameters optimization method of PID controller for needing the precise mathematic model, the adaptive ability and learning ability of fuzzy control theory and neural network and the global optimization ability of genetic algorithm are used. A parameters optimization method of PID controller based on genetic algorithm, fuzzy control theory and neural network is proposed. The parameters and structure of fuzzy neural network are comprehensively optimized by using genetic algorithm based on the decimal coding. Then the optimized fuzzy neural network is used to compute the three parameters of PID controller in order to obtain the optimization PID controller. Finally, the optimization method is applied in the central air conditioning control system of Dazhou Iron & Steel Group, the simulation results show that the optimization method takes on the strong anti-interference ability and robustness.

Key words: genetic algorithm, fuzzy control theory, neural network, optimization, PID controller

中图分类号: