论文部分内容阅读
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小于0.8 s。
Aiming at the problem that the time delay of non-robust path in digital circuit is tested for a long time and the coverage of fault is low, a test generation algorithm based on artificial bee colony optimization is proposed. Firstly, the circuit is converted into its equivalent circuit by using the circuit conversion rule, and then the Hopfield neural network is used to construct the constraint circuit of the single circuit with fixed faults. The energy function is obtained by using artificial bee colony optimization algorithm Value to get the equivalent single-circuit fault test vector fixed, and finally get the original circuit non-robust path delay fault test vector pair. The experimental results on the ISCAS’85 international standard circuit show that the fault coverage of this algorithm can reach 98% and the average test generation time is less than 0.8s.