基于克隆选择算法的Hopfield网络容量提升方法

Improvement of Hopfield Network Capacity Based on Clonal Selection Algorithm

Hopfield网络容量大小对网络模式识别正确率有重要影响。为进一步提升Hopfield的网络容量,提出了一种基于克隆选择算法优化Hopfield网络容量的方法。首先将克隆选择算法引入到Hopfield网络中,以Hopfield网络的初始输入作为克隆选择算法中的抗原;然后随机产生权值矩阵作为克隆选择算法的初始抗体;最后依据克隆选择算法对初始抗体进行克隆、交叉、变异,根据亲和力的大小选择出网络的优化权值,以提升Hopfield网络容量。将上述方法应用于含噪声的样本识别,实验结果表明:与传统的Hopfield网络相比,所提出的方法能有效地提升Hopfield网络的容量。为提高Hopfield神经网络的记忆容量提供了一种新的思路。

The capacity of Hopfield network is a very important factor for the accuracy of network pattern recognition. In order to further boost network capacity of Hopfield, optimized Hopfield network capacity based on clonal selection algorithm is proposed. Firstly, clonal selection algorithm is introduced to Hopfield network. Initial input of Hopfield network serves as antigen of clonal selection algorithm. Then, weight matrix generates randomly as initial antibody of clonal selection algorithm. Finally, initial antibody is cloned, crossed and varied according to clonal selection algorithm. Network optimization weight is selected according to appetency to improve Hopfield network capacity. The above method is applied to identify the samples containing noise. The results show that compared with traditional Hopfield network, the method can effectively improve Hopfield network capacity and provide a new thought for improving memory capacity of Hopfield neural network.