IA-GPS : uma ferramenta baseada em algoritmos de busca para alocação de equipes em projetos de software

Abstract

The planning of a large project involves many variables, uncertainties and conflicting objectives, being a very complex task. Given that the manual process may lead to low quality solutions, recently some papers have proposed the use of Artificial Intelligence algorithms, in particular metaheuristics, to help the software project manager in the planning task. This form of modeling falls within the scope of Search-Based Software Engineering and the planning problem addressed in this work is known as Software Project Scheduling Problem whose goal is to allocate employees to project tasks, in order to optimize project cost and duration. This work presents IA-GPS, an application integrated with jMetal that simplifies the use of heuristic algorithms, providing a more user-friendly interface for software engineers to analyse the possible solutions for the problem, generated by the algorithms. IA-GPS provides managers with possible employee allocations for tasks and timelines, which can be viewed in a variety of ways. The application was developed in Java and was validated with synthetic case studies, extracted from the literature.