Novel PROMETHEE method with dependent criteria

For how broad and popular the field of multi-criteria decision making (MCDM) is, relatively little space has been devoted to modelling the dependencies between criteria so far. While it is true that the assumption of absolute independence between evaluation factors may be too simplifying. Most of the popular MCDM methods do not allow such natural relationships to be included. The exceptions are often not very user or computationally friendly. For example, substantially more complex algorithms have to be used, or many additional inputs are required from decision-makers (not to mention the difficulties with setting their values in practice). The aim of this paper is to provide a new extension of the PROMETHEE ranking method allowing the inclusion of interdependencies between criteria and preserving the main benefits of the original algorithm (computational simplicity and versality). The proposed method is demonstrated on several numerical examples which reveal the usefulness of the extension.