Project portfolio selection with scheduling: an evolutionary approach

Abstract

This paper addresses the project portfolio selection problem enriched with scheduling. Thefactors considered for project scheduling are: the planning time horizon and the negative impact of theproject completion time on the total profit. The objective is to select a subset of projects which maximizesthe discounted total gain by late completion time, respecting resource constraints and without exceedingthe time horizon. A mixed integer linear programming model was formulated and compared to two recentmodels addressing the same problem. To show the potential of the model on the large scale we used ametaheuristic based on the genetic algorithm Non-dominated Sorting Genetic Algorithm II. We showexperimentally the benefits of our proposal and leave open the possibility of its study applied on a largerscale in future works.