Mathematical model of software development project team composition optimization with fuzzy initial data

Abstract

Increasingly frequent changes in demand for products, reducing product life cycles, changes in the business environment during a crisis, the innovative nature of projects, the complexity of forecasting external and internal conditions, the impact of the human factor lead to increased uncertainty and inability to plan team activities with a given degree of accuracy. In this regard, the subject matter of the article is the task of creating an adaptive project team that can work effectively in the mentioned above conditions. This task is especially relevant for the sphere of software development. The sphere is dynamic and characterized by frequent changes in product requirements, technologies, working conditions, and restrictions on project implementation. Agile approaches are used to manage such projects, which can help the team respond to uncertainties and frequent changes. To date, there are many agile approaches to project management, but the issue of selecting team members in such approaches is insufficiently covered. Therefore, this work formalizes the task of deciding on the selection of software development team members, considering the uncertainty and subjectivity of the information that affects the selection of candidates for the team. The task of the work is to create a decision-making model based on the use of the mathematical apparatus of fuzzy sets and methods of operations research. Such a model should allow considering the uncertainty of estimates of project requirements and the level of competence of team candidates. The result is a mathematical model of a two-criterion constrained optimization problem. The first objective function is aimed at finding a team composition that maximizes the maximum competencies of its members. The second criterion is aimed at forming a team with the maximum sum of competencies for all indicators, considering the weight of each indicator. The first constraint assumes that at least one team member meets the competency requirements expressed by a specific indicator. Additionally, it is required that the available time fund of the team members allows the project to be completed on time. It considers the limitation on the salary of the team. Conclusions. Solving the problem in accordance with the proposed mathematical model will allow making a team as readily as possible to meet the existing and new requirements for the project staff. The last circumstance is especially important when implementing a software development projects.