Adaptive Random And Combinatorial Testing Based On Elite Search Strategy

As an important means of software quality assurance,software testing plays an indispensable role in the modern society where software products are widely used.Facing increasingly complex software systems,combined Testing is a common test case generation method.However,the popular combination test method is usually used for t-way coverage by greed or combined mathematics to fully detect the failure of the input configuration interaction in the software system,which is often complicated and not conducive to the promotion of the algorithm.Random Testing(RT)is widely used in industry because of its simplicity and easy to implement,but it is also criticized by many people for its poor testing effect caused by its strong randomness.As an enhanced version of Random Testing,the excellent performance of Adaptive Random Testing(ART)in failure detection has been gradually pursued by researchers in recent years.It is an important issue in the field of test case generation to design a test case generation method for non-numerical large complex software with simple and excellent test effect.For most ART methods,the input field of the tested object is usually numerical,so the Euclidean distance method is usually used in the distance calculation.For such non numerical programs in Combinatorial Testing,Hamming distance or other similarity measurement methods are generally replaced.In this kind of ART,ARTsum has been paid more attention by researchers in the industry because of its linear time complexity and strong failure detection ability.However,compared with common combination Testing methods such as AETG and IPO,the ARTsum algorithm is still lack of t-way coverage,which leads to the need of more test cases for t-way Testing of the programs under test.In view of this situation,this thesis proposes ARCTsum algorithm.ARCTsum algorithm designs a special data structure-multi-dimensional coverage matrix,which records the set of tested cases and guide the next test case generation.With the addition of elite search strategy,the candidate set of ARCTsum has a high level of tway coverage.And the distance calculation method based on t-way can effectively screen the best candidate use cases.In order to verify the effectiveness and efficiency of the algorithm,this thesis compares ARCTsum with the popular ART and combinatorial Testing algorithms,and finally collects and analyzes the experimental data.The experimental results show that ARCTsum algorithm is not weaker than AETG algorithm in most scenarios,and ARCTsum is significantly better than AETG algorithm in time cost.Furthermore,ARCTsum is significantly better than the popular ART algorithm in failure detection and t-way coverage.And ARCTsum has the advantage of linear time complexity,which is second only to RT and ARTsum methods.In short,the ARCTsum algorithm greatly improves the test effect of the algorithm while ensuring the excellent performance of the algorithm through three effective optimization strategies.The research method can provide some reference for the academic research and industrial application of random Testing and combination Testing.