基于演化计算的Hopfield神经网络多用户检测模型研究

人们已经对Hopfield神经网络应用于多用户检测进行了大量研究，但Hopfield神经网络多用户检测模型存在局部最优和稳定性差的问题一直没有得到有效解决。演化计算的高度并行与自组织、自适应和自学习等特点正好能有效解决稳定性差与局部最优的问题。因此，文章在考虑随机扰动变量的前提下，将演化计算结合进Hopfield神经网络多用户检测模型中。理论推导证明，该方法切实可行，预计有很好的应用前景。

A Study of Hopfield Neural Network Multi-user Detection Model Based on Evolutional Calculation

Multi-user Detection Model based on Hopfield neural network is extensively studied, but two problems about local optimal solution and stability at the Hopfield neural network Multi-user Detection Model have not been effectively solved. Evo-lutional Calculation is just capable to solve the two problems based on its characteristics including highly parallel, self-organi-zation, self-adaptive and self-study etc. At the paper, under thinking about random variables perturbation, the evolutional cal-culation is combined in the Hopfield neural network Multi-user Detection Model. A theory deduction proves that the method is effective and can be extensively applied.