Optimization of object-oriented testing using firefly algorithm

Abstract

Software testing is an integral part of software development. Testing ensures that software is correctly doing what it was designed for and not terminating abruptly. To overcome the challenges associated with exhaustive software testing and to tackle the problems imposed by object-oriented paradigm of software development, a new approach of testing has been used in recent years called model based testing. Model based testing relies on creating test suites from the formal model of application behaviour. Hence, the main aim is to optimize the software process for creation of only effective test cases and omit the redundant ones. Nature-inspired soft computing techniques provide a meta-heuristic approach to achieve the same. In this study, a meta-heuristic firefly algorithm has been proposed for state chart testing of five benchmarks of object-oriented application and optimized test cases have been generated for them. After that, the results have been compared with already existing nature-inspired ant colony optimization algorithm. The inference shows that there is no redundancy in the generated test cases via firefly algorithm whereas test cases generated by ant colony optimization algorithm are highly redundant. Also, firefly algorithm generates reduced test cases as compared to the ant colony algorithm.