Investigating Multi and Many-Objective Metaheuristics to Support Software Integration Testing

 In spite of the fact that Search-Based Software Testing (SBST) is a very appealing field today, there are few studies that deal with software integration testing and, even so, most of these works are not truly related to the generation of test cases to this testing level. In this paper, we present a method, InMeHy, which aims at investigating the use of metaheuristics to derive integration test cases based on C++ source code. A graph is created based on the code which represents the integration of several classes of the application. Multi and Many-Objective metaheuristics (Evolutionary Algorithms) were considered to generate integration test cases and were assessed via three quality indicators. Results show that the traditional Indicator-Based Evolutionary Algorithm (IBEA) turned out to be the best out of four algorithms evaluated, including newer Many-Objective strategies such as Nondominated Sorting Genetic Algorithm-III (NSGA-III).