Multi-step ranking of alternatives in a multi-criteria and multi-expert decision making environment

Abstract

In this work we introduce a decision model, in the form of a recursive aggregation algorithm, that attempts to mimic a multi-step ranking process of a set of alternatives in a multi-criteria and multi-expert decision making environment. The main idea is rather intuitive. Each alternative is initially assigned a list of values, representing the group experts’ opinion about the extent to which the alternative satisfies a set of given criteria. Then the values for each alternative are combined with the weighted mean operator according to vectors of weights, one for each decision maker in the group. These weights express the personal judgement of the decision makers about the relative importance of the individual criteria. Consequently, a new vector of values is obtained for each alternative. These new values are combined again with the weighted mean operator taking into account the different degrees of influence each decision maker accepts from the rest of the group. The latter aggregation step is repeated again and again for each alternative until a consensus is attained.