A genetic algorithm for task allocation in collaborative software developmentusing formal concept analysis

Abstract

Software development is no longer an isolated or localized task but a collaborative process with well coordinated contributions from personnel across the globe. Such an approach boosts productivity, but also poses challenges that must be met. One of them is to formally analyze the realms of software development tasks and the teams that are commissioned to perform them to derive the full set of conceptual units that describe these domains in terms of the needed proficiencies. Then, the best possible matching between the cohesive task-sets and the inter-coordinating teams must be obtained. In this paper, we present a model for Collaborative Software Development that addresses these issues. We employ Formal Concept Analysis to generate the concept lattices in the domains of tasks and teams in terms of various skills. We employ Genetic Algorithm, a meta-heuristic that stochastically scans the search space in a guided manner to generate the best possible pairings between task concepts and team concepts. Results show that this approach forms cohesive task sets, identifies sets of homogeneous teams and produces optimum task-team mappings that gives high skills utilization and provides a basis for coordinated and reliable operation. The GA yields a range of non-inferior solutions giving wide scope of tradeoff between various objectives.