The Research Of Ability Of Fault-Detecting Between Combinatorial Test And Random Test

Combinatorial test is popular with testers because of its good performance on the ability of fault-detecting.And random test gets lots of attention from researchers due to its simplicity.So recent years,with people paying more and more attention to ways of software test,comparisons about effectiveness and efficiency among software test methods have been the topic that some researchers are working on.Choosing the right test strategy can not only improve the efficiency of test,estimate the quality of software,but also reduce the time and cost of developing.So from this point,we can see the necessity of comparing strategies of test.This paper aims to study the ability of fault-detecting about combinatorial test and random test as an experimental way,and evaluate the effectiveness of this two test methods,which can allow us to understand the principle of combinatorial test and random test better.Contents are as follows:(1)Knowledge and techniques about random test are various,but there are few reviews on them.This paper helps readers to review the developing process of random test by summarizing the basic knowledge and theories,techniques and methods,and original models of random test tools.(2)Under the situation of boolean-specification test,according to the common 10 types of mutation,mutants are got from the 20 expressions which are extracted from TCAS(Traffic Collision Avoidance System).This paper treats these mutants as experimental subjects,and compares the ability of fault-detecting of combinatorial test and random test,and analyzes the formula of the lower bound of combinatorial test from the point of theory,and points out the defects.(3)This paper analyzes the existing formula of lower bound of random test,and proves it is not accurate and existence of error between the theoretical value and the practical one,and the error is related to the amount of MFS.