前向神经网络在图论码译码中的应用

The Application of the Feedforward Neural Network in Decoding to Graph Theoretic Code

构造了两个分别用于图论码的硬判决译码和软判决译码的前向神经网络，并证明所构造的前向神经网络能够实现图论码的最小汉明距离译码和最小软距离译码。最后给出采用所构造的前向神经网络进行一个图论码的硬判决译码和软判决译码的计算机模拟结果。

Two feedforward neural networks for hard-decision decoding and soft-decision de-coding to graph theoretic code were constructed.The fact that the constructed networks canfinish minimum Hamming distance decoding and minimum soft distance decoding to graphtheoretic code was proved.