Review of the State of the Art in Multi-Objective Optimization in Agile Software Development Projects

Abstract

In agile software project management, it is common to face complex decisions with multiple conflicting objectives, such as minimizing delivery time and cost while maximizing product quality and stakeholder satisfaction. Multi-objective optimization offers a systematic approach to balancing these conflicting criteria, seeking compromise solutions (Pareto optimal) rather than a single optimal plan. This article presents a review of the state of the art (2020-2025) on the application of multi-objective optimization techniques in agile software development projects. The databases reviewed were mainly IEEE, ACM, Springer, Elsevier, and Arxiv. The most used evolutionary algorithms were systematically analyzed: NSGA-II, NSGA-III, SPEA2, MOEA/D, some swarm metaheuristics, and exact methods such as integer and goal programming. Their most common applications in agile environments (version planning, task assignment, and requirement prioritization) were identified, as well as the most cited tools and the main challenges: uncertainty, scalability, stakeholder preferences, and industrial adoption. The results indicate that multi-objective evolutionary algorithms, such as NSGA-II and variants, dominate recent literature, being successfully applied to problems such as the Next Release Problem and project scheduling, although hybrid and exact approaches are emerging to improve the quality of solutions. In conclusion, multi-objective optimization is consolidating itself as a rigorous framework to support decision-making in agile projects, with proven benefits in case studies.