A study on the anonymity of pairwise comparisons in group decision making

Pairwise comparisons between criteria/alternatives are a well-established methodology for group decision making. Both in the literature and in real-world applications, it is common practice to average the opinions of various experts expressed as pairwise comparisons to find a compromise solution. This paper dwells on the inverse problem: given a decision maker who knows his own preferences and the aggregate ones, we introduce and study some optimization problems to help him estimate, in the form of intervals, the preferences of the other decision makers. Since the possibility of estimating the other participants’ preferences violates the requirement of anonymity, this paper also reports the results of some numerical simulations examining the relation between the number of participants in the decision process and their anonymity. In addition to numerical simulations with randomly generated data, this manuscript reports the results of experiments on a dataset of preferences collected in a real-world survey.