专家控制与神经元控制结合算法的研究  

On a Combined Algorithm with Expert Control and Neuron Control

作　　者：王永黎 张晓兰[1] 徐桂英[1] 涂健[1] 

机构地区：[1]华中理工大学自动控制工程系

出　　处：《华中理工大学学报》1994年第8期35-37,共3页Journal of Huazhong University of Science and Technology

摘　　要：提出了一种改进的神经元控制算法，此算法采用专家控制与神经元控制相结合的方法，有效地解决了神经元网络控制时学习速度较慢、动态响应过程长、克服扰动不力等问题，具有较强的鲁棒性。控制性能优于经典ＰＩＤ方法。典型工业过程对象的仿真表明了本算法的可行性和有效性。An improved neuron control algorithm is presented,in which the proportional coeffi-cient of neuron(K)is chosen according to error feedback by the expert control idea,Nonlin-ear transformation is used to strengthen the neuron control sensitivity to small deviation. The learning rate,dynamic response as well as the anti-disturbance ability have all beengreatly improved;The new algorithm has strong robustness and the performance is found tobe better than that of the classical PID method.Based on the simulations carried out on typi-cal 2 nd order plus pure lag object,the effectiveness and feasibility of the algorithm are veri-fied.

关 键 词：专家控制 神经元控制 算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]