ACTUM –tool for automatic class testing using meta-heuristics

Software testing is the most expensive phase in terms of time, effort and cost. Manual generation of test cases to satisfy identified testing requirement is labour intensive and challenging. The main challenge in testing remains to identify an efficient and minimal pool of test cases those can meet a said adequacy criterion. In this advanced technology era, fault free software with less delivery time is the utmost requirement. With this vision, in past multiple different types of research is done towards automation of test-data generation with a sole aim to perform effective software testing with reduced effort and time. Maximum of the work in this field is dedicated to testing of procedural programs. Testing applications developed in object-oriented paradigm have different challenges and techniques adopted for procedural programs are not suitable. The work in this paper discusses a mechanism to automate class testing and presents Automatic Class Testing Using Meta-heuristic (ACTUM), a tool that provides a framework for test-case generation of classes automatically. The implementation of ACTUM is based on a hybrid algorithm applying evolutionary meta-heuristics. Evolutionary technique genetic algorithm and its variants are applied to search test cases that can perform data flow based structural testing of class under test. An experiment is conducted to evaluate ACTUM with test case generation for def-use coverage of eight java classes. Result analysis of the experiment outcomes is also presented.