Uma abordagem de apoio à decisão para formação de múltiplas equipes em projetos ágeis de software.

Abstract

Agile methodologies have emerged as a form of management and development, based on an incremental approach to satisfy customers demands and their projects. Scrum is a framework for management and agile development of software projects, centered on the team concept, aiming to deliver business value. One of the greatest challenges, not only for software companies, but also for industry as a whole, is focused on team formation. Traditionally, software project team formation is a human based non automated process, susceptible to errors. The process complexity becomes even greater when we consider the distribution of people, within the same organization, in several teams, with different demands of competence and experience, which is called multiple teams formation problem. The overall objective of this research is to design a decision support approach to form multiple teams for agile software projects, specifically, those following Scrum, from the reallocation of human resources available in the company. The proposed approach uses genetic algorithm to suggest teams, derived from technical profiles generated during the execution of the company software projects. The validation was performed with a historical database generated from real world data from 12 software development projects and 52 different developers. The approach was executed and validated in 13 different scenarios created from the database. Afterward, it was evaluated with the support of four agile project managers and as result it reached an average of 86.4 % of Precision. In addition, an average of 75 % of acceptance was obtained.