针对软件故障定位问题,提出一种新型故障定位技术SPRank(可疑参数排名).该方法利用测试数据构建操作剖面模型,并提出输入参数谱度量程序各个参数的可疑度;在此基础上,利用优化的思想搜索程序的可疑参数和与之相关的可疑语句,进一步缩小程序故障的检索范围,从而提升故障定位准确率.实验表明SPRank方法能有效地提升三种经典高效的故障定位方法的故障定位准确率.
Aiming at fault localization, a technique SPRank (suspicious parameters rank) was pro- posed. Besides traditional information sources, the input data involved in test suite could contribute to localize program faults by converting itself into a kind of operational profile (OP) with optimizational principles. With the proposed input parameter spectrum, the built OP could be used to identify some suspicious parameters, which formed the basis of the proposed SPRank method to improve the efficiency of fault localization. Experimental studies show that the SPRank method could effectively improve the fault localization efficiency of three state-of-the-art SBFL techniques.