Accés ràpid intranet

Més informació...

a a a



Dominance-based Rough Set Approach to Multiple Criteria Decision Support


Roman Slowinski

Professor/a organitzador/a


University of Technology and Systems Research Institute


24-04-2009 11:00


For scientific support of decisions one needs a more or less explicit model of the decision problem. The model relates the decision to the characteristics of the alternatives expressed in terms of evaluations on the considered criteria. The usual models, such as those used in Multi-Attribute Utility Theory, Analytic Hierarchy Process or ELECTRE methods, require information about conditions and parameters of aggregation of multiple criteria evaluations of alternatives (e.g. substitution rates for value functions of Multi-Attribute Utility Theory, pairwise comparisons of alternatives in terms of intensities of preferences for Analytic Hierarchy Process, attribute weights and several thresholds for ELECTRE methods) which is not easily definable. Moreover, these models process the information in a way that is not intelligible for the decision maker (DM), such that (s)he cannot see what is the exact relation between the provided information and the final recommendation. Consequently, very often the decision model is perceived by the DM as a black box whose result has to be accepted because the analyst's authority guarantees that the result is ?right?. In this context, the aspiration of the DM to find good reasons to make decision is frustrated. This motivates the search of a more transparent methodology in which the relation between the original information and the final recommendation is clearly shown. Dominance-based Rough Set Approach (DRSA) supplies such a transparent methodology searched for, and therefore this approach can be seen as a glass box. DRSA requires an input information in terms of examples of decisions and builds a decision model in terms of easily understandable ?if..., then...? statements, such as ?if the maximum speed of the car is at least 200km/h and its price is not greater than 50,000 $, then the car is at least attractive?. These decision rules are very meaningful for several reasons: they are based on the ordinal properties of considered criteria only, they are easily understandable by the DM, even when (s)he is not an expert of multiple criteria decision analysis, and they give clear argumentation for justifying the decision. We present DRSA applied to the following decision problems: multiple criteria sorting (also called ordinal classification), multiple criteria choice and ranking, decision under uncertainty and interactive multiobjective optimization.


Laboratori 231