Wu Deng, Huimin Zhao, Yinglian Luo and Xiumei Li
Ant colony optimization algorithm, fuzzy neural network, chaos theory, parameter optimization, traction motor, PID controller
Fuzzy neural network (FNN) has a strong self-learning and self-tuning function, but there exists the insufficiency in selecting parameters. So ant colony optimization (ACO) algorithm and chaos theory are introduced into the FNN in order to propose an optimized FNN (CACOFNN) model. In the proposed CACOFNN model, a switching function is designed for the connections between neurons to transform the structure and parameters of FNN into a simple function optimization problem. The fuzzy grid partition is used to determine the optimal rules and structure of FNN. Then the chaos theory is introduced into the ACO algorithm to propose a chaotic ACO (CACO) algorithm for optimizing the initial parameters of FNN. Finally, the proposed CACOFNN model is applied to optimize the parameters of PID controller of traction motor. The results show that the CACOFNN model takes on the faster approximation control objectives, more excellent performance and convergence.
Important Links:
Go Back
