一种改进的基于遗传算法的测试调度方法

An Improved Test Scheduling Method of So C Based on Genetic Algorithm

提出了一种改进的基于遗传算法的So C测试调度方法,通过该方法可以有效地优化测试总线的划分,合理调度各个IP核以实现并发测试,能够有效地缩短芯核测试时间。该算法把测试调度问题的可行解集用种群表示,逐代演化产生出越来越好的近似解。详细分析了该算法过程,对2002年国际测试会议(ITC’02)所提供的So C国际基准电路进行测试调度实验,实验结果表明,此算法比传统的整数线性规划(ILP)和遗传算法的结果要好。

This paper proposes an improved So C test scheduling method based on genetic algorithm,which can efficiently optimize the division of testing bus,reasonably schedule each core to realize parallel test,and efficiently shorten the test time of core.Representing feasible solution of test scheduling with population,based on ＂survival of the fitness＂,and beginning with the initial population,this algorithm evolves by generation to produce approximation solution. Utilizing international reference circuit provided by International Test Conference 2002（ ITC＇02）,we execute the test scheduling experiment. And the results suggest that this algorithm be superior to conventional integer linear programming（ ILP） algorithm and conventional genetic algorithm.