论文部分内容阅读
为了保持神经网络在优化计算求解过程中结构不被改变,以迟滞混沌神经元和迟滞混沌神经网络为研究对象,提出了一种基于滤波跟踪误差的控制策略来实现神经元/网络的稳定控制.采用该控制策略,在不改变非线性特性发生机理的情况下,神经元/网络可实现函数优化计算问题的求解.所设计的控制律包含两部分:一部分是系统进入滤波跟踪误差面时的等效控制部分,另一部分为确保系统快速进入滤波跟踪误差面的控制部分.采用Lyapunov方法对神经元/网络的控制进行了稳定性证明.根据待寻优函数直接求得神经元的控制律,在该控制律的作用下,神经元/网络可逐渐稳定到优化函数的极值点,从而实现优化问题的求解,仿真实验结果验证了该控制方法在优化计算中的可行性和有效性.
In order to keep the structure of neural network unchanged in the process of optimization and calculation, a new control strategy based on filter tracking error is proposed to realize the stable control of neuron / network by using delayed hysteretic chaotic neuron and delayed hysteretic chaotic neural network as research objects. Using this control strategy, the neuron / network can solve the problem of function optimization calculation without changing the mechanism of nonlinear characteristic.The control law is designed to contain two parts: one is when the system enters the filter tracking error surface, etc. Efficiency control part and the other part is to ensure that the system quickly enter the control part of the filter tracking error surface.The Lyapunov method is used to prove the stability of the neuron / network control.According to the function to be optimized, Under the control law, the neuron / network can be gradually stabilized to the extreme point of the optimization function to solve the optimization problem. The simulation results verify the feasibility and effectiveness of the proposed control method in optimization calculation.