Ranking
Z-PROMETHEE: Z-Number Preference Ranking Organization Method (II)
Nuriyev, M. · 2020
Overview
Z-PROMETHEE is PROMETHEE with two channels per cell ('how good is alternative a on criterion j' + 'how sure is the expert'). Reliability collapses to α via centroid, then stretches/shrinks the restriction TFN by √α (Kang transformation). After all cells are Z→TFN, classical PROMETHEE runs: for each pair (a, b) and each criterion j, compute the difference; pass through a step function (Type 4) that returns 0/0.5/1 based on q,p thresholds; weight by criterion importance; sum across criteria → π(a,b). Average π over all other alternatives → Φ+ (leaving, how much a outranks others) and Φ− (entering, how much a is outranked). Net Φ = Φ+ − Φ−; rank descending.
Strengths
- •Outranking respects intransitivity and non-compensation more than utility methods
- •Type-4 preference function captures both indifference and strict-preference thresholds explicitly
- •Z-numbers capture both vagueness and reliability of expert judgments
- •Net flow Φ produces clear total ordering: easier to communicate than partial ranking
Limitations
- •Sensitivity to q, p threshold choice: must be elicited carefully
- •Computationally O(m² × n): moderate scaling
- •Single dominant criterion can mask others if its weight is large (mitigated by careful weight elicitation)
- •PROMETHEE II loses incomparability information present in PROMETHEE I
Method assistant
Grounded explanations: it explains the method, it does not compute.
Edge cases
- •When q=p=0, Type-4 collapses to Type-1 Usual Criterion (strict 0/1 step)
- •When two alternatives are identical, Φ(a)=Φ(b)=0 → tie
- •When weights sum ≠ 1, engine auto-normalizes with warning
- •When K=1 expert, group aggregation step is identity
- •Sum of all Φ(a) is always exactly 0 (flow conservation): useful sanity check
Common pitfalls
- •Don't forget criterion direction (benefit/cost): flip differences for cost criteria
- •Don't confuse PROMETHEE I (partial) with PROMETHEE II (full ranking via net Φ)
- •Don't pick arbitrary q, p thresholds: elicit from decision-maker
- •Don't treat Z-numbers as regular fuzzy: Z→TFN via Eq.(6)-(8) first
Worked example
- 1.Step 1
- 2.Step 2
- 3.Step 3
- 4.Step 4
- 5.Step 5
- 6.Step 6
- 7.Step 7
- 8.Step 8
- 9.Step 9
How to cite
Nuriyev, M. (2020). Z-numbers Based Hybrid MCDM Approach for Energy Resources Ranking and Selection. International Journal of Energy Economics and Policy. https://doi.org/10.32479/ijeep.9950