Effective Test Case Minimization and Fault Detection Capability Using Multiple Coverage Technique

摘要

Software testing is the primary source to check whether the software is working correctly in the given environment under all possible inputs. The testing process requires generation of large amount of test cases which is a time consuming process and resources need to be involved. As the software development takes place, new test cases are added to the pool of existing test cases to test new features in the software or to capture any newly discovered fault. Due to this continuous addition of test cases, regression testing tends to grow in size. The multiple test cases may exist which may test the same feature of requirements. Large test suite size means expensive execution, both in terms of computing resources as well as human resource. In order to limit the size of the test suites, Test Suite Minimization techniques were used into software testing as it produces a set of optimized test cases which need not be reduced, hence reducing the amount of resources involved in the generation and reduction process of test cases. One such optimization technique is Multiple Coverage Criteria Test Suite Minimization, a heuristic algorithm, which identifies redundancy in a given test suite based on multiple coverage criteria for example function, branch coverage of given test cases. The limitation of Regression Testing is that it sometimes throws those test cases which might be useful for the testing process of another same kind of software. In this research we propose a fault detection technique to overcome this limitation by selectively retaining some additional test cases from the rejected pool. We implemented our approach by modifying the above existing heuristic algorithm for test suite minimization. Our experiments show that our approach can significantly improve the fault detection capability of Regression Testing without severely affecting the extent of suite size reduction.