人工蜂群优化非鲁棒路径时滞故障测试生成算法

摘要

针对数字电路中非鲁棒路径时滞故障测试时间长，故障覆盖率较低的问题，提出了人工蜂群优化的测试生成算法。该算法首先应用电路转换法则把数字电路转换成为其等效电路，然后用Hopfield神经网络构建等效电路单固定故障的约束电路，并得到能量函数，再应用人工蜂群优化算法计算能量函数的最小值以得到等效电路单固定故障的测试矢量，最后根据对应关系得到原电路非鲁棒路径时滞故障的测试矢量对。在ISCAS’85国际标准电路上的实验结果表明该算法故障覆盖率能够达到98％，并且平均测试生成时间明显减小。%The test generation algorithm based artificial bee colony optimization is proposed in this paper , be-cause the test generation time is long and faults coverage is low for non-robust path delay fault faults in digital circuits.This algorithm changes digital circuit into equivalent circuit according to circuit switching rule firstly , then constructs the constraint circuit for the single stuck-at fault equivalent circuit using Hopfield neural net-works and obtains the energy function .The test vectors for the single stuck-at fault in the equivalent circuit can be obtained by solving the minimum of energy function of the constraint circuit based on artificial bee colony op -timization method .Finally the test vectors pair for non-robust path delay fault in the original digital circuit can be obtained according to correspondence relation .The experimental results on ISCAS ’85 international standard circuits demonstrate the fault coverage can reach 98% and the average test generation time decreased signifi-cantly.