A state-based fitness function for the integration testing of object-oriented programs

Testing object-oriented program is quite challenging task due to the nature of its features like inheritance and polymorphism. In practice test case generation is the most laborious and resource consuming process in software testing hence generating test data for object-oriented programs is even more challenging and effort demanding. Object-oriented evolutionary testing aims at automating test case generation process using evolutionary strategies like Genetic Algorithm. Evolutionary testing gives a helping hand to the software testers to speed up the process and to reduce the amount of project resources. The existing approaches provide sound platform to the researchers to take the work further to meet the challenges that object-oriented paradigm has introduced. One of the limitations in existing approaches is that they combine branch distance of object's state variables with local variables that leaves no guidance for the search process whether object has gained desired state or not. We propose a state-based fitness function for the evolutionary testing of object-oriented programs that can solve object's state problem by evaluating object's state as an independent segment of overall test case fitness. Our initial experiments show that by separating object's state evaluation, search gets better guidance to prevent object's state problem.