Ranking method · Brans, 1982
PROMETHEE
Preference Ranking Organization METHod for Enrichment of Evaluations (I & II)
Compares alternatives pairwise with preference functions, so small differences don't get over-rewarded.
How it works
- For every pair of alternatives and every criterion, translate the performance difference into a preference degree (0–1) using a per-criterion preference function (usual, U-shape, V-shape, level, linear, or Gaussian).
- Aggregate the weighted preference degrees into a matrix π(a, b).
- Compute the leaving flow Φ⁺ (how strongly a beats the others), entering flow Φ⁻ (how strongly it is beaten), and net flow Φ = Φ⁺ − Φ⁻.
- PROMETHEE I gives a partial ranking from Φ⁺/Φ⁻; PROMETHEE II ranks completely by net Φ.
- The GAIA plane (a PCA of the per-criterion flows) shows which criteria agree, conflict, and favor which alternatives.
Use it when
- Differences below some threshold shouldn't count indifference and preference zones matter.
- Criteria are measured in incomparable units and you want to avoid aggressive normalization.
- You want the GAIA visual to communicate the trade-off structure.
Watch out for
- Choosing preference functions and their q/p/s thresholds per criterion requires thought.
- Net flow (PROMETHEE II) can hide incomparabilities that PROMETHEE I would expose.
Parameters
Per criterion: a preference function plus q (indifference), p (preference) or s (Gaussian midpoint) thresholds.
Cite the method
Brans, J.-P., & Vincke, P. (1985). A preference ranking organisation method: The PROMETHEE method for MCDM. Management Science, 31(6), 647-656.
@article{brans1985preference,
author = {Brans, Jean-Pierre and Vincke, Philippe},
title = {A preference ranking organisation method: The {PROMETHEE} method for {MCDM}},
journal = {Management Science},
volume = {31},
number = {6},
pages = {647656},
year = {1985},
doi = {10.1287/mnsc.31.6.647}
}
Try PROMETHEE on your own data
Upload a spreadsheet or type your alternatives in every calculation step is traced.
Open the studio