A multi-objective mean-semivariance model for project selection using reinvestment and synergy under uncertainty

Project selection problems with imprecise parameters is one of the hot spots that have attracted many scholars' interest. In this paper, we first propose a multi-objective mean-semivariance model to solve the multi-objective project selection problem considering reinvestment and synergy between projects with different investment and operation periods by applying the uncertainty theory. The investment outlay and profit are treated as uncertain variables with an uncertainty distribution function that are determined based on experts' evaluations. The objectives are to maximize the expected value of uncertain net present value (NPV) and to minimize its risk (semivariance). The difference between the models taking variance and semivariance as the risk of NPV is compared, and the effects of reinvestment and synergy on project selection are analyzed. We next propose new binary versions of Jaya and Rao algorithms and develop multi-objective binary meta-heuristic algorithms binMOJaya, binMORao1 and binMORao2 to solve the proposed model. The performances of the proposed three algorithms are demonstrated through comparison with other well-known binary multi-objective algorithms for the 10 problems including large-scale problems. Finally, the validity of proposed model is illustrated through an example problem using the zigzag uncertainty distribution. Computational experiments have shown that the proposed uncertain model explains reinvestment and synergy well, and that the proposed algorithms, especially the binMORao2 algorithm, are very suitable for solving the proposed model.