基于并行处理的Hopfield联想记忆模型学习算法

LEARNING ALGORITHM FOR HAMS BASED ON PARALLEL PROCESS

本文在约束优化学习算法的基础上,提出了一种采用并行处理技术的Hopfield联想记忆模型学习策略.这种学习算法能有效地节约计算时间,提高学习的成功率.同时,本文给出一种多余吸引子的清除算法.计算机实验结果证明了算法的有效性.

This paper presents a parallel learning Algorithm for Hopfield Associative memories(HAMS), which is based on Constrained Minimization. The algorithm efficiently reduces the computational'time and failure rate of learning. Another algorithm is given to remove spurious states. Several computer simulations show advantages and efficiencies of the algorithm.