提出了基于一类最大T -模运算的模糊联想记忆连接权矩阵神经网络学习算法 ,并且给出了严格的理论证明 ,该算法成功地解决了多模糊模式对的存储问题。对于给定的模糊模式对 ,若这些模式对存在连接权矩阵 ,则该算法很容易求出它们的最大权矩阵 ,如果这些模式对不能够用一个连接权矩阵来存储 ,应用本文的算法 ,可以应用尽可能少的连接权矩阵来存储 ,从而可以有效地减少存储空间。 A learning algorithm based on a kind of maximal T - modular operator for fuzzy associative memory connected matrix is ​​proposed, and a rigorous theoretical proof is given. The proposed algorithm solves the storage problem of multi - fuzzy pairs successfully. For a given pair of fuzzy patterns, if these pairs have a connection matrix, then the algorithm can easily find their maximum weight matrix. If these patterns are not able to be stored by a connection matrix, the algorithm in this paper can be used Use as little connection matrix as possible to store, which can effectively reduce the storage space.