【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。更多还原

【Abstract】 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.更多还原