Investigating the impact of developer productivity, task interdependence type and communication overhead in a multi-objective optimization approach for software project planning

Abstract

One of the most important activities in software project planning involves scheduling tasks and assigning them to developers. Project managers must decide who will do what and when in a software project, with the aim of minimizing both its duration and cost. However, project managers often struggle to efficiently allocate developers and schedule tasks in a way that balances these conflicting goals. Furthermore, the different criteria used to select developers could lead to inaccurate estimation of the duration and cost of tasks, resulting in budget overruns, delays, or reduced software quality. This paper proposes an approach that makes use of multi-objective optimization to handle the simultaneous minimization of project cost and duration, taking into account several productivity-related attributes for better estimation of task duration and cost. In particular, we focus on dealing with the non-interchangeable nature of human resources and the different ways in which teams carry out work by considering the relationship between the type of task interdependence and the productivity rate of developers, as well as the communication overhead incurred among developers. The approach is applied to four well-known optimization algorithms, whose performance and scalability are compared using generated software project instances. Additionally, several real-world case studies are explored to help discuss the implications of such approach in the software development industry. The results and observations show positive indications that using a productivity-based multi-objective optimization approach has the potential to provide software project managers with more accurate developer allocation and task scheduling solutions in a more efficient manner.