数字电路的神经网络模型研究

Neural Network Model For Digital Circuits

研究数字电路的Hopfield神经网络模型,说明了这种模型的原理及建立方法,并指出这种模型存在的不足,然后研究电路的最优神经网络模型,给出了该模型的相关定义,讨论了任意数字电路对应的最优神经网络的存在性.这些建模方法可应用于电路的测试、逻辑验证等领域,能提高相关技术的计算效率.

The Hopfield neural network model and optimal neural network model for digital circuits are studied.For a logic circuit which has m input and n output nodes,the circuit can be represented by an optimal neural network, i e the circuit is characterized by the network energy function that has global minima only at the neuron states consistent with the logic functions of the circuit and that the neural network consists of m+n neurons It is proved that there exist optimal neural networks for an arbitrary digital circuit The optimal neural network model will have wide applications in circuit test and verification