Concluding comments

The paper has pursued two closely connected purposes. The first has been to provide a theoretical framework, based on coherence constraints, for a technique of multicriteria analysis with which to convert partial pre-order into a total pre-order. The second has been to use this framework to evaluate the alteration produced in the observed binary comparisons in order to eliminate incomparabilities.

I believe that this latter aspect, which is of greater practical relevance, may of use to decision-makers for re-examination of both the data and the criteria used, and that it may also be utilised to evaluate the reliability of the total pre-order yielded by the multicriteria method.

I intend to illustrate this aspect in a subsequent study which analyses a concrete case. Here, however, it has seemed opportune to restrict the treatment to a numerical example with already-known and immediately verifiable data.

By Dr Elvio Mattioli