【摘要】 为了解决大面积的VLSI(超大规模集成电路)电路制造过程中因缺陷而造成的成品率低的问题,可以采用降阶和冗余两种VLSI阵列重构方法,这两种方法都属于NP-完全问题。该文是通过冗余修复方法来解决这一问题的。该文基于Hopfield神经网络为模型,将阵列的重构问题转化为矛盾图的最大独立集问题。通过Hopfield神经网络的能量函数方程进行求解,求得合理的补偿通道来完成问题的求解。实验分析表明该方法是简单有效的。更多还原

【Abstract】 In order to solve the problem of low yield caused by the defects in VLSI manufacture,we usually employ two approaches for redundant VLSI array reconfiguration or degradable VLSI array reconfiguration.Both approaches are proved NP-hard.The paper will get a redundant repair method to solve the problem.It’s based on the model of Hopfield neural network,changing the problem of array reconfiguration into the problem of maximum independent set of contradiction graph.Through the energy function of Hopfield neural network,it will be easy to find reasonable path to solve the problem.It is proved by experiment to be a simple and useful method.更多还原