论文部分内容阅读
带时间参数的二叉判决图（TBDD）在电路的时滞故障测试中有着重要的应用价值，但其变量排序是用常规方法无法解决的一个优化问题．本文提出一种基于遗传算法的TBDD排序算法．用快速衡量值和TBDD节点数来计算个体的适应度，针对变量排序的特定问题，提出一种模板保序交叉方法．采用自适应的变异概率计算方法，并提出一个适合于TBDD排序问题的变异算法．实验结果表明较好地解决了TBDD的排序问题．
The binary decision diagram with time parameters (TBDD) is of great value in the circuit fault testing of time-delay. However, the variable ordering is an optimization problem that can not be solved by conventional methods. This paper presents a genetic algorithm based TBDD sorting algorithm. The fast fitness and the number of TBDD nodes are used to calculate individual fitness. Aiming at the specific problems of variable ranking, a template preserving and interleaving method is proposed. The adaptive mutation probability calculation method is adopted and a mutation algorithm suitable for the TBDD ordering problem is proposed. Experimental results show that the problem of sorting TBDD is well solved.