A guided tour of multiple criteria sorting models and methods

Multiple criteria sorting methods assign objects into ordered categories while objects are characterized by a vector of n attributes values. Categories are ordered, and the assignment of the object is monotonic w.r.t. to some underlying order on the attributes scales (criteria). Our goal is to offer a survey of the literature on multiple criteria sorting methods, since the origins, in the 1980’s, focusing on the underlying models. In the first part of the paper, we start by recalling two main models, one based on additive value functions (UTADIS) and the other one an outranking relation (Electre Tri). Then we draw a (structured) picture of multiple criteria sorting models and the methods designed for eliciting their parameters or learning them based on assignment examples. In a second part of the paper, we attempt to provide a theoretical view of the field and position some existing models within it. We then discuss issues related to imperfect or insufficient information.