A GRASP-based Heuristic for the Project Portfolio Selection Problem

The project portfolio selection (PPS) problem consists of constructing a project portfolio, that is, a selection of projects scheduled over a period of time using various, potentially conflicting, criteria and resource constraints. The PPS is a well-known problem, recurrently occurring in several applications, with a rich history of approaches for modeling it and a large number of techniques for solving it. In this work we present a model for the PPS problem based on a real-world situation of selection and scheduling of projects in the power generation industry. We also propose a heuristic, based on the metaheuristic GRASP, to solve the problem, and assess its quality and performance through computational experiments. We describe the implementation of a decision support system prototype for the PPS problem that uses the proposed heuristic and includes several usability features that may help decision makers through the selection process.