On the estimation of a satisficing model of choice using stochastic multicriteria acceptability analysis

This paper addresses the task of estimating the kind of aspirations that are most likely to have given rise to an observed partial rank ordering of alternatives. The proposed approach uses a Tchebycheff goal programming formulation to serve as a descriptive model of choice and stochastic multicriteria acceptability analysis to generate candidate aspiration vectors and estimate the true aspiration vector from this candidate set. A simulation experiment is used to assess the accuracy of the estimated aspiration levels in a variety of problem contexts. The approach can perform well if strong performance is demanded on a small subset of the attributes and the rank order that is observed is sufficiently detailed.