Choose the Right Multi-Criteria Decision-Making (MCDM) Method
Multi-criteria decision-making (MCDM) is a family of approaches for evaluating choices against several, sometimes competing, criteria. The right method depends on what your data represent, how criteria relate, and whether you need to rank alternatives, understand cause-and-effect, or measure efficiency. This guide helps you shortlist a method; follow its method page for the detailed procedure.
Quick guide: which MCDM method should you consider?
| Ask yourself | Methods to explore | Why |
|---|---|---|
| Are judgments qualitative, vague, or uncertain? | Fuzzy AHP, Fuzzy TOPSIS, Fuzzy VIKOR | Fuzzy scales represent imprecise expert assessments. |
| Do you need to derive criterion weights from pairwise judgments? | AHP or ANP | AHP structures a hierarchy; ANP models interdependence. |
| Do you already have weights and need to rank alternatives? | TOPSIS, VIKOR, SAW, PROMETHEE | These compare alternatives using performance across criteria. |
| Are you evaluating the relative efficiency of units? | DEA | DEA compares inputs and outputs across decision-making units. |
| Do criteria influence one another? | DEMATEL or ANP | DEMATEL maps influence; ANP carries dependencies into priorities. |
All MCDM methods, grouped by family
Pairwise weighting and network methods
- Analytic Hierarchy Process (AHP) — derives criterion weights and ranks alternatives through structured pairwise comparisons in a hierarchy.
- Fuzzy Analytic Hierarchy Process (Fuzzy AHP) — applies fuzzy judgments when experts cannot express comparisons precisely.
- Analytic Network Process (ANP) — prioritizes criteria and alternatives when elements have feedback or interdependence.
- Fuzzy Analytic Network Process (Fuzzy ANP) — extends network-based prioritization to uncertain expert judgments.
Influence analysis
- Decision-Making Trial and Evaluation Laboratory (DEMATEL) — identifies and visualizes cause-and-effect relationships among criteria or factors.
- Fuzzy DEMATEL — maps those influence relationships when assessments are linguistic or imprecise.
Alternative ranking and compromise methods
- Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) — ranks alternatives by closeness to an ideal solution and distance from a negative ideal.
- Fuzzy TOPSIS — applies ideal-solution ranking when ratings or weights are uncertain.
- VIKOR — seeks a compromise ranking when criteria conflict and a balanced solution is needed.
- Fuzzy VIKOR — uses compromise ranking with fuzzy evaluations and uncertain preferences.
- Simple Additive Weighting (SAW) — scores each option by adding normalized criterion values multiplied by their weights.
- Fuzzy Simple Additive Weighting (Fuzzy SAW) — adapts weighted additive scoring to imprecise ratings or weights.
- Preference Ranking Organization Method for Enrichment Evaluations (PROMETHEE) — ranks alternatives using preference functions that express how decision-makers compare performance.
- Weighted Aggregated Sum Product Assessment (WASPAS) — combines additive and multiplicative scoring for multi-criteria alternative assessment.
Efficiency measurement
- Data Envelopment Analysis (DEA) — measures relative efficiency of comparable units with multiple inputs and outputs.
Criteria weighting
- Step-wise Weight Assessment Ratio Analysis (SWARA) — derives criterion weights from expert ranking and stepwise judgments.
At-a-glance comparison
“Small/medium/large” describes typical decision-model size and effort, not a strict software limit. Suitability depends on study design and data.
These labels are starting points, not fixed rules: a study with many criteria may still use a simple ranking method if its assumptions fit. Before selecting a technique, check that alternatives are comparable, define whether each criterion is a benefit or a cost, and decide how weights will be obtained. Normalize measurements that use different units when the chosen method requires it. Document expert judgments, data sources, and any sensitivity checks so readers can understand how robust the ranking is. When two methods produce different orders, compare their assumptions and explain why one better answers your research question.
| Method | Data | Typical criteria | Complexity | Common use |
|---|---|---|---|---|
| AHP / Fuzzy AHP | Exact / fuzzy | Small–medium | Medium | Weighting, vendor choice |
| ANP / Fuzzy ANP | Exact / fuzzy | Small–medium | High | Interdependent priorities |
| DEMATEL / Fuzzy DEMATEL | Exact / fuzzy | Small–medium | Medium | Cause-and-effect analysis |
| TOPSIS / Fuzzy TOPSIS | Exact / fuzzy | Several | Low–medium | Ranking alternatives |
| VIKOR / Fuzzy VIKOR | Exact / fuzzy | Several | Medium | Compromise solutions |
| SAW / Fuzzy SAW | Exact / fuzzy | Several | Low | Transparent weighted scores |
| DEA | Exact | Multiple inputs/outputs | Medium–high | Efficiency benchmarking |
| SWARA | Exact judgments | Small–medium | Low | Expert-based weighting |
| PROMETHEE | Exact preferences | Several | Medium–high | Preference-based ranking |
| WASPAS | Exact | Several | Low–medium | Combined additive/product scoring |
Frequently asked questions
What is the difference between AHP and ANP?
AHP assumes the decision can be organized as a hierarchy, with criteria treated as independent across levels. ANP uses a network structure to account for feedback and dependencies among criteria or alternatives.
Which method is better for uncertain data?
Fuzzy variants such as Fuzzy AHP, Fuzzy TOPSIS, Fuzzy VIKOR, and Fuzzy DEMATEL can represent vague linguistic judgments. Choose based on whether you need weighting, ranking, compromise, or influence analysis.
Which method should I use to rank alternatives?
TOPSIS, VIKOR, SAW, PROMETHEE, and WASPAS can rank options. Consider how you want to represent preferences and trade-offs, and check whether criterion weights are already available.
Is DEA an alternative-ranking method?
DEA primarily evaluates relative efficiency among comparable decision-making units using multiple inputs and outputs; its purpose differs from ranking products or policies against weighted criteria.
Can I combine MCDM methods?
Yes. A study may use one method to derive weights and another to rank alternatives, provided the combination fits the research question and is clearly justified.
Ready to apply your method?
Once you have selected an approach, explore the OnlineOutput MCDM products to find a tool for your analysis.