R&D Project Portfolio Selection with Fuzzy Data Envelopment Analysis

Abstract

R&D investments are becoming increasingly important in the developing world. Companies withlimited resources should make the most favorable investments for their own strategies. It is crucial that theseinvestments are transferred to the right projects. It is difficult to make decisions in an environment where thereare technical difficulties as well as uncertainties. At this point, it is necessary to decide which projects should bedone and which projects should not be done. In this study, project portfolio selection that seeks a systematicsolution to this decision, is covered. To solve this problem, data envelopment analysis that can evaluate theparameters without the need to build precedence relationship, is used. Parameters were set after a detailedresearch. Vagueness that is associated with difficulty of making precise judgment, was included in the model byintroducing linguistic variables. Ambiguity that characterizes the situation where there are two or morealternatives, is defined with triangular fuzzy sets and α cut method. Different models are constructed fordifferent extreme cases to solve the ambiguity. The models provide the optimal value regardless of the α value.A sample dataset of 30 projects is created to test the models and observe the results. Optimal parameters weightsare found in the models. Full pairwise comparisons are considered while examining the interdependencies.These parameters weights are recalculated according to interdependencies. Using these weights, the efficiencyscore of each project is calculated for each model. Projects are prioritized for different strategies by usingdecision making under uncertainty.