Neural network PID control is an important topic in control theory with its extensive uses and applications. BP algorithm is a very effective learning algorithm that is usually used in the process of neural network PID control. While these algorithms have good nonlinear mapping ability, generalization ability and fault tolerance, they also have some shortcomings, such as slow convergence, easy to fall into local minimum and forget the samples. In order to overcome these shortcomings, we introduce a mixed idea that use genetic algorithm to optimize neural network: Firstly, use genetic algorithm to optimize the weights of the neural network so that the range of the weights can be reduced. Then, use BP algorithm to solve the problem precisely so that it can avoid falling into local minimum and have a quick convergence. In this paper, we'll study the neural network neural network in single-variable system and multi-variable system respectively, according to the mixed idea.My thesis is organized as follows:Chapterl, Introduced the background of this topic and some related basics. Specifically we mainly introduced the history of neural network, some concepts and theory involved in this paper, such as artificial neurons, the classification and study of neural network, etc. In addition, we also introduced some basics about genetic algorithm.Chapter2, Introduced the neural network PID control of single-variable system, in this chapter, we employed the mixed idea introduced above, and got a satisfactory result.Chapter3, Firstly, we popularized the method used in chapter2 to multi-variable system. Similarly, we got a satisfactory result. Then, we studied the improved PID neural network multi-variable control that based on the genetic algorithm and got a satisfactory result.Chapter4, Summarized this paper and propose the problem for further research. |