A Research On Fault Location Based On Combinatorial Testing

Software testing is not only an important computer science,but also a very critical process of the software development.It can be said that,to some extent,software testing determines whether the software project success or not.Software testing includes black-box testing and white-box testing.In order to precisely test whether a system has errors or not,the software testing method must test all possible inputs(test cases).However,when the number of SUT's test cases is very huge,testing all possible test cases is impossible.Therefore,the combinatorial testing appears naturally.Combinatorial testing is a testing technique which is widely used in black-box testing.It is designed to detect interaction errors triggered by interaction between parameters in a system.It is helpful for the programmer to find the cause of the error quickly and improve the development efficiency.Therefore,it is very important to research the fault location method based on combinatorial testing.Fault location is a very critical process of the software development,and determines whether the software project success or not.A number of adaptive fault location methods based on combinatorial testing have been proposed.This paper aims to use the experimental data of a batch of TCAS Boolean Test Set as the experimental object,and use the additional test cases,recall,precision and f-measure to compare and verify the efficiency of the adaptive fault location method based on combinatorial testing.In addition,this paper presents a novel adaptive fault location method named BELF,which is based on combinatorial testing.This method is suitable for the system where each parameter only has two values.A large number of experimental results show that the efficiency of BELF is better than the existing fault location methods.