A Multi-Stage Approach for Evaluating and Selecting a Project Portfolio Based on Data Envelopment Analysis Under Uncertainty

Abstract

A project portfolio is a collection of ongoing or future projects within an organization, and its optimal selection—as one of the critical decisions in project management—contributes to optimal resource allocation and improved organizational performance. Accurate evaluation of projects for inclusion in the project portfolio is essential; due to various factors and uncertainties such as project duration, modeling these uncertainties is necessary to achieve optimal decision-making. In this study, uncertainties are modeled using an interval-valued fuzzy valuation method, which enhances the precision of the decision-making process. The present research proposes a three-stage method for optimal project portfolio selection under uncertain conditions. In the first stage, projects are scored using the Marcus multi-criteria decision-making method while considering uncertainties. In the second stage, project efficiency is evaluated through cross-efficiency data envelopment analysis while accounting for uncertainty. In the third stage, a multi-objective mathematical model is designed and solved for selecting the optimal project portfolio, which—beyond optimizing cross-efficiency scores and multi-criteria decision-making—maximizes the overall profit of the portfolio while satisfying budget, time, and workforce capacity constraints. The results indicate that combining these methods significantly improves decision-making accuracy and enables optimal project selection considering organizational constraints and objectives in complex and uncertain environments.