A robust optimization model for R&D project portfolio risk response decision-making considering risk interaction and regret perception

Abstract

This paper presents the robust optimization model to select project portfolio risk response strategies (PPRRSs) considering risk interaction and resource allocation fairly. Firstly, the Design Structure Matrix (DSM) is used to analyze the project portfolio risk interaction considering project interdependency. Meanwhile, the improved PageRank algorithm is used to assess risk considering risk interaction and expected loss. Then, a robust optimization model for multi-objective project portfolio risk response decision-making is established. The model has dual objectives of maximizing the weighted risk response effect and minimizing the average regret value of each risk subject in project portfolio. Considering the uncertainty of risk response costs, the polyhedral uncertainty set is introduced to convert the robust model into a linear programming model. Finally, the model is applied to actual cases. The method and model established in this paper can help the project portfolio decision maker to identify the key risks efficiently and choose the most robust risk response strategy set combatting uncertainty. At the same time, it can avoid the centralized allocation of resources in the project portfolio, and each risk subject can obtain relatively fair risk response resource.