Speaker
Description
The PROMETHEE family is a well-known class of multicriteria outranking methods. It has been applied in various fields such as computer science, environmental management, finance, health care, etc. PROMETHEE methods first compute pairwise preferences from an evaluation table (based on inter and intra-criteria parameters) and then aggregate them to derive flow scores. Those scores are then used to produce partial or complete rankings of alternatives.
Recent works have studied properties of the PROMETHEE II net flow scores to better understand the resulting rankings. This includes axiomatic characterizations of the net flow scores, the development of indicators to assess the ranking quality (or information loss) or conditions under which rank reversal is not possible. However, the structural properties of the aggregation from pairwise preferences to net flow scores remain insufficiently explored.
This contribution extends this line of research by interpreting the PROMETHEE II aggregation step as a linear map. Based on this perspective, we derive new properties of PROMETHEE II and its net flow scores using common principles of linear algebra. We also propose a new indicator for ranking quality that complements existing measures of information loss.
| Keywords | Outranking methods; PROMETHEE II; Linear algebra; Ranking quality |
|---|---|
| Participation in the SNSF Early-Stage Researcher Support and Award Scheme | I am submitting a regular presentation and would like to participate in the SNSF Early-Stage Researcher Support and Award Scheme. |