基于混合混沌优化法的Hopfield网学习算法

A Learning Algorithm of Hopfield Neural Network Based on CTSA

提出了一个基于混合混沌优化法的Hopfield网学习算法。通过对Hopfield网权值不等式的处理,训练Hopfield网。利用混合混沌法的优点,即混沌的遍历性和禁忌搜索的“记忆性”和期望准则,有效地避免了局部最小解,克服了原Hopfield网学习的局限性,还能找到多个优化解。实验证明了该算法的有效性。

A learning algorithm of Hopfield neural network, which bases on Chaos and Tabu search optimization algorithm(CTSA) is presented. It forms the Hopfield neural network by dealing with inequality of weight of network. Using the virtue of CTSA, which includes the ergodicity of Chaos and the memory of Tabu search. This algorithm can avoid the local optimal solution effectively. It can also find several optimal solutions.The experimental results demonstrate the effectiveness of the algorithm.