人工蜂群优化非鲁棒路径时滞故障测试生成算法  

Non-robust path delay faults test generation algorithm based on artificial bee colony optimization

作　　者：赵莹[1] 孟祥[1] 李艳娟[2] 王一帆[1] 赵彦超[1] 

机构地区：[1]北华大学电气信息工程学院,吉林吉林132021 [2]东北林业大学计算机学院,黑龙江哈尔滨150040

出　　处：《重型机械》2015年第4期18-22,共5页Heavy Machinery

基　　金：国家自然科学基金青年基金(61300098);吉林省教育厅"十二五"科学技术研究项目(吉教科合字2011第121号);吉林市科技计划项目(201414006)

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.

关 键 词：非鲁棒路径时滞故障 神经网络 人工蜂群算法 能量函数 

分 类 号：TN407[电子电信—微电子学与固体电子学]