摘要

This paper examines decision makers' bounded rationality in consensus building, and proposes a direct consensus framework based on cumulative prospect theory to solve group decision making problem with eight kinds of preference representation structures. This framework consists of two processes: selection process and consensus process. In selection process, we first transform different preference structures into uniform individual preference vectors using different transformation functions. Then, according to the individual preference vectors and gain and loss matrices, we determine reference points using preference-approval structures and the prospect value matrix. Third, we design an optimization model to obtain decision weights, and further get the prospect values of alternatives. In consensus process, we define a consensus measure by applying cumulative prospect theory, design a feedback adjustment to make consensus better, and output satisfying comprehensive prospect value after several iterations if necessary; then, the final optimal solution is obtained. At last, an example is provided to illustrate the effectiveness of framework.