A Similarity-Based Approach to Aggregation

This paper gives an approach to the aggregation of a pair of values a and b in [0,1] assuming that there is an indistinguishability operator or similarity defined in this interval. The basic idea is to understand the aggregation of two numbers a and b as a value lambda that has the same similarity with respect to a than to b or more precisely, the degrees of equivalence of lambda with a and with b must coincide. Some interesting aggregation operators are obtained from natural indistinguishability operators associated to t-norms which are ordinal sums