Uncertainty-Permeable Analytics for the Software Release Planning Problem

Software development processes present a number of challenges due to the impact of uncertainty. In this context, the problem of deciding which requirements will be implemented in the next k releases is known as the Software Release Planning Problem. This article presents a novel fuzzy model for Software Release Planning that accounts for the penetration and exploitation of uncertainty related to the effort required to implement each requirement. The solution is a Consonant Family of Plans, which allows for the determination of confidence measures regarding the release in which each requirement will be implemented. The model is tested on an example instance, and a “what-if” analysis is conducted, demonstrating how the solution responds to updates in parameters and the arrival of new data on work completed. This information constitutes a flexible prescriptive analysis that supports decision-making, and the paper shows how the solution responds as the context changes. Also, an initial computational study is carried out. The results show that the solution’s usefulness is maintained over time as new information becomes available, which poses new challenges and encourages further research.