关键词: 反向传播（BP）神经网络, 变异灰色算子, 多群体协同, 混沌理论, 泵控马达系统

Abstract: Aiming at how to improve BP network disadvantages that is easy to fall into local minimum and slow convergence, a multi group with variation of grey operator Collaborative Chaotic Particle Swarm Optimization algorithm （GMMCCPSO） is proposed. The gray mutation operator is applied to multi-group cooperative particle swarm main group  in order to avoid the phenomenon that is premature local convergence of the main group. In order to enhance the capacity of local search from group, the chaos theory is introduced from each group. Using improved particle swarm optimization algorithm to optimize BP neural network weights and thresholds, the shortcoming of the BP network that is easy to fall into local minima and slow convergence of shortcomings is effectively improved, and it also greatly improves its mapping capabilities. By the study of controlling the pump motor system simulation of MATLAB, the results show that the improved PSO-BP network effectively improves the shortcomings of the system about the ability to identify the presence of sudden load, the oscillation and the response speed being poor of the system.

Key words: Back Propagation（BP） neural network, mutation gray operator, multi-group collaboration, chaos theory, pump controlled motor system