Project Portfolio Selection Using a Combination of Fuzzy Analytic Hierarchy Process and A Modified Integer Linear Programming

The selection of project portfolio is the main problem in project-based organizations. In project selection, optimized decision to maximizing utilization of limited resources and achieving organizational goals by considering restrictions are important. Therefore, to facilitate project portfolio selection in project-based organizations, an efficient model is proposed that is able to consider limitations, goals, and priorities of the organization in selecting the optimal combination of projects. The proposed model is a hybrid and flexible one which, including four main steps. First, with initial screening, to save time and money by using non-compensatory method, some of the poor candidate options are deleted. Then, due to multi-criteria nature of the problem, qualitative and quantitative criteria, and uncertainty of data in real world, the projects are ranked using fuzzy analytic hierarchy processes and uncertain judgments of decision-makers. To generate project baskets, integer linear programming (ILP), backpack model, is employed modifying the coefficients of the objective function of each project to preserve multi-criteria rating of projects, and then evaluate and rank the baskets. To validate the proposed model, we have carried out selection of project portfolio in the oil refining industry.