The Algorithm Study For Abstract Test Case Prioritization Techniques

With the increasing complication of the system,the software may be affected by many factors such as system configuration,user inputs,internal events,etc.At the same time,these parameters have same or different number of optional values or options.Depending on the user's different input and the choice of these values will form an exponentially increasing combination space.To this end,testers need to selectively select a portion of the test cases for testing,while ensuring that errors in the tested program can be effectively discovered.Random testing is the easiest method,but its error detection rate is not good.At this point,it is very urgent to find an algorithm that can prioritize test case set based on combined space to improve the error detection rate.The prioritized sorting algorithm based on combined space is to find those test cases that can locate and repair software system errors earlier.The study found that such test cases have a feature that always covers the most uncovered combinations.Among them,R.C.Bryce et al.proposed an Interaction Test Suite Prioritization(ITSP),which has a significant error detection effect.Further research found that the prioritizing interaction test suites using repeated base choice coverage is one of the best algorithms in the current test algorithms.The main idea of the algorithm is to fix a suitable combination coverage strength t,each time selecting the test case that covers the most uncovered t-dimension combination.After analyzing the algorithm,this paper proposes two improvements for the shortcomings.Then verify the performance of the two algorithms.Finally,based on these two algorithms and FICBP algorithm,a test prototype system is implemented to verify and compare the performance of three algorithms.The main work done in this paper is as follows:1?Aiming at the problem of FICBP algorithm in selecting high-level combined coverage,although the error detection rate is guaranteed but the time overhead is too large,prioritizing interaction test suites using repeated base choice coverage is proposed.The core idea of FICBPR algorithm is to repeat 1-dimension combination coverage.FICBPR always chooses the test cases which cover the most uncovered1-dimensional combinations.If the 1-dimensional combination is completely covered,the FICBPR no longer considers the coverage of the previously selected test case,but repeats the above process.The experimental results show that the proposed FICBPR algorithm greatly reduces the time while maintaining error detection.Furthermore,in many cases,the method can achieve similar results to the FICBP algorithm based on 2-dimension combination coverage,and has more stable results than FICBP at high combined coverage strength.2 ? Based on the common shortcomings of FICBPR algorithm and FICBP algorithm,the prioritizing interaction test suites based weight algorithm(FICBPW)is proposed.All the algorithms proposed so far,once all the combinations based on the combined coverage strength t are completely covered,the coverage of the combined space is blanked,and the process of covering the combined space is restarted.Therefore,the information of the previous combined coverage is lost.Thus,the result obtained each time is only a local optimal solution rather than a global optimal solution.In response to this problem,the concept of weight is proposed.By accumulating the coverage of the combined space,the test case with the smallest total weight is selected each time until the end.The experimental results show that the FICBPW algorithm proposed in this paper has higher error detection rate than FICBPR in the case of similar time overhead,and obtains similar results with FICBP algorithm based on combined coverage strength of 3 and 4.3?Design and implement a sorting system based on the interaction of test cases combined coverage.The prototype system includes an algorithm selection module,an experimental parameter setting module,an execution module,and an experimental result output module.The effectiveness of the algorithm is verified by different procedures.