A hierarchical interval outranking approach with interacting criteria

Complex decision-making problems can be conveniently decomposed in more manageable sub-problems through a hierarchical approach. Often, there are interaction effects between sub-criteria or elementary criteria in the hierarchy. In this paper, we propose a generalization of the recently published interval outranking approach to deal with this kind of problems. This new method allows an easy setting of weights (including interaction weights), and other model parameters as interval numbers. The proposal also allows to set outranking relations associated with each non-elementary criterion. On this background, the INTERCLASS-nB and INTERCLASS-nC multi-criteria ordinal classification (multi-criteria sorting) methods are generalized to cope with interacting and hierarchical criteria; these methods can suggest assignments of actions to ordered classes for each non-elementary criterion. Finally, an application of the classification methods is illustrated by evaluating research and development projects.