An Improved Restricted Random Testing Algorithm And Its Application In Object-oriented Software Testing

Software testing is regarded as an effective method to ensure software quality,it is an indispensable part in the software development.With the increasing demand for software,there has been more and more software development technologies.Object-oriented technology is the most widely used software development technology because of it's unique advantage with reusability and scalability.However,characteristics of object-oriented technology also bring a lot of new problems for software development,which has the most obvious influence on software testing.At present,Random Testing(RT)is the most commonly used method in object-oriented software testing.Experiments show that RT reduces the time overhead of selecting test cases and reduces the cost of software development,but ithas poor failure-detection effectiveness.In order to improve the failure-detection effectiveness of RT,T.Y.Chen et al.proposed an improved RT algorithm,named Adaptive Random Testing(ART).Various ART algorithms have been implemented for software testing.Restricted random testing(RRT)is one algorithm of ART by exclusion.RRT defines an exclusion region around each previously executed test case,and selects test cases only from outside all exclusion regions.RRT has a higher failure-detection effectiveness than RT;However,it has a very high time overhead in the process of test case selection process.This means that it takes too much time when using the RRT for software testing,and it will takes a lot of manpower and resources to develop software.Therefore,it is necessary to improve the time effectiveness of RRT algorithm.In this paper,an alternative method of RRT,named RRTtp,has been proposed.In order to prove the feasibility,this paper conducts quantitative experiments for RRT and RRTtp.And the application of the two algorithms in object-oriented is realized.The main research work is decsribed as follows.1.In order to reducing time overhead of RRT,this paper proposed a method named RRTtp algorithm.RRTtp defines an exclusion region and optional regionaround each previously excuteds test case.Then all of the optional region are compared to pick out the largest one,and the next test case is randomly generated in the region.First,all the excuted test cases are sorted out in order to geting the size of the optional region between the two test cases;then the largest optional region is picked out;At each comparison step,it is necessary to update the starting and ending points of the largest optional region to randomly generate next test case.2.The application of RRT and RRTtp algorithm in object-oriented is realized.When testing a numerical program,the distance between test cases is a numerical difference.However,when testing object-oriented program,the distance between test cases can not be calculated directly.This paper proposed a distance form of test case;then the distance between test cases is defined as the sum of the dynamic distance and the static distance after analyzing the structure of the test case;finally the distance measure of test cases is completed through combined with RRT and RRTtp algorithm,and the application of RRT and RRTtp algorithm in object-oriented is also realized.3.A test prototype system based on RRT algorithm and RRTtp algorithm is designed and implemented.This system designs the numerical program testing module and the non-numerical program testing module based on the application of RRT and RRTtp in different programs.The numerical program testing module mainly includes a algorithm execution module and the result analysis module;The non-numerical program testing module includes a class diagram input module,test case distance calculation module,algorithm execution module and result analysis module.