论文部分内容阅读
针对多芯片组件 (Multi- chip Module)互连测试的特点 ,用构造互连结构的神经网络状态表 ,求解满足状态表所有状态的能量函数方程组的一组解来获得神经网络模型参数的方法 ,为互连网络的正常状态和几类典型的故障状态分别建立了对应的二值 Hopfiled神经网络模型 ,并验证了模型的正确性。
Aiming at the characteristics of multi-chip module interconnection test, a method to obtain the parameters of neural network model by solving a set of solutions to the system of energy functions satisfying all states of the state table by using the neural network state table that constructs the interconnection structure , A corresponding binary Hopfiled neural network model is established respectively for the normal state of the interconnection network and several typical fault states, and the correctness of the model is verified.