Min-max 模糊 Hopfield 网络的一个基于容错性的学习算法

在一定的条件下证明了Min－max模糊Hopfield网络吸引子的吸引域随着连接权矩阵的减小而增大，在此基础上，设计了一个基于容错性的学习算法。在一定的意义下，用该学习算法得到的矩阵最小，从而此时系统具有最优容错性。最后用实例验证了结论。

A Learning Algorithm Based on Fault-tolerance for Min-max Fuzzy Hopfield Networks

In the paper, we prove under some conditions, that the attractive basin of the attractorof min-max fuzzy Hopfield network increases when the connected weighted matrix decreases. In accordance with this conclusion,we design a learning algorithm based on the fault-tolerance of the network,the matrix obtained by the learning algorithm is minimum, consequently, the fault-tolerance of the network is optimal. Finally, the example demonstrates our conclusions.