Determining the most desirable portfolio of construction projects in a power system is a multiple-attribute decision-making problem. In traditional portfolio decision-making methods

the attribute values of indices are described as deterministic ones

and the optimal portfolio is selected from single projects ranked with the attribute values of the indices. However

it is not appropriate to represent some indices as deterministic quantities

and the optimal portfolio should not be selected based on the evaluations of single projects because of the possibility of having correlations among some single projects. Given this background

a method for computing the gains/losses and their probabilities of the indices with deterministic

interval

probabilistic

fuzzy types of data is first developed for the portfolio of construction projects. In order to deal with the impacts of psychology and risk preference of the decision-makers

the well-established prospect theory is next introduced to the portfolio decision-making problem. Then

the differentiations of weights are calculated for the indices and the optimal portfolio determined by employing the VIKOR(VIseKriterijumska Optimizacija I Kompromisno Resenje)method

which can avoid the disadvantage of compensating the bad indices with good ones in traditional weighting methods. Finally

a sample example for a provincial power system is employed to demonstrate the feasibility and efficiency of the proposed method. This work is supported by National Natural Science Foundation of China(No. 51477151).