Multi-objective optimization of test sequence generation using multi-objective firefly algorithm (MOFA)

摘要

Software testing is one of the essential parts of the software development life cycle. In software industry, the testing cost can be approximately 50% of the total cost of a software project so efficient ways of testing software are crucially important in reducing costs, time and effort. There are two major methods of software testing; black-box testing (focuses only what the software can do) and white-box testing (tests the internal structure of the software under consideration thoroughly and the ultimate goal is to write test cases that force the program coverage.) For program coverage, identification of suitable paths is one of the major software testing problems. These test paths are known as test sequences. Generation of automated and effective test sequences is also a very difficult task in software testing process. In the proposed work, the problem “Test Sequence Generation” is considered as a multi-objective optimization problem by having two objectives to be optimized simultaneously, Oracle Cost, and Path Priority. In real time environment, there are many constraints which have to be fulfilled when dealing with an effective testing. So, such test sequences that meet multiple objectives simultaneously are generated in order to reduce the testing efforts. To solve this problem a recently developed algorithm “Multi-Objective Firefly Algorithm (MOFA)” is used. The problem “Test Sequence Generation” is first implemented by Firefly Algorithm and later by using MOFA-considering the problem as Multi-Objective Optimization Problem. The proposed technique implementing test sequences with multiple (two) objectives and its results are presented.