Mathematical model of social behavior for the allocation of members in software development teams

摘要

The allocation of members in software development teams has a direct impact on the software project's success. Several approaches have been proposed to deal with this task, mainly using optimization techniques to manage the cost, scope, and time constraints, where experience and backgrounds represent important resources; however, when there is no previous teamwork between the candidates the problem might escalate. For software development team formation different variables are considered from the individual profiles, along with team characteristics such as team size, members' geographic location, or team roles; while it can be expected that the team interaction influences its members' performance. In this paper, a mathematical model is introduced as a means to understand how individuals evolve as a result of their interaction, presenting an overview of team behavior as a whole. The model simulates the individuals clustering according to similarities, and the variation of the individual states as a result of their interaction. In the model, adjustable parameters include the number of individuals, the number and degree of their characteristics with an expected probability of manifestation, the degree of similarity among them for clustering, and the probability of clustering. The model is applied to an approach based on interactive styles and the Team Software Process methodology team roles, simulating their interactive style degree adaptability over time.