BWM
AvailableBest-Worst Method. Rezaei (2015)
The idea
Filled marks place each criterion by the answers about the best criterion, hollow marks by the answers about the worst; the dot is the weight, in times the worst criterion's weight; here every answer stays within the consistency threshold (CR 0.14, threshold 0.30), so the workbench would not warn.
Show as table
| Criterion | From the best | To the worst | Weight |
|---|---|---|---|
| Price | 8 | 8 | 0.4308 |
| Quality | 4 | 4 | 0.2462 |
| Safety | 2.6667 | 2 | 0.1538 |
| Comfort | 2 | 4 | 0.1231 |
| Style | 1 | 1 | 0.0462 |
Data: Buying a car (Rezaei 2016).
BWM compares the most important criterion with all others and all others with the least important (1‑9).
When to use it
- Subjective weights with fewer questions than AHP.
- Between 3 and 9 criteria.
- One decision maker, or a group agreed on one pair of vectors.
Algorithm
1Two comparison vectors
B is the best (most important) criterion and W the worst, on a 1‑9 scale; a_BW must be the same number in both vectors.
2Linear model (default)
Rezaei (2016): each absolute value becomes two linear inequalities, giving a linear program with a unique optimal weight vector; solved here in exact closed form via Ratandhara & Kumar (2026) with v_W* = 1 without an external LP solver, and xi close to 0 means high consistency.
3Non-linear model (not offered)
Rezaei (2015): the original non-linear minimax model where xi* is unique but inconsistent comparisons produce weight intervals rather than a single vector; this site computes the unique Rezaei (2016) linear model instead.
4Input-based consistency ratio
Liang, Brunelli & Rezaei (2020): no optimization needed, CR^I = 0 when a_BW = 1, and full consistency means a_Bj a_jW = a_BW for every j.
Used together with
Why these pairs
- BWM then TOPSIS or VIKOR are the recurring partners: of 124 BWM publications up to January 2019, 83 integrate BWM with other methods (Mi et al. 2019). Example: You et al. (2017), BWM-TOPSIS.
- BWM then VIKOR, for example Gupta (2018) on air transport.
- BWM then CoCoSo or MABAC in recent studies.
- BWM then MABAC, for example Telli & Ayçin (2021) on teacher selection.
- BWM and AHP are competing subjective methods; Rezaei (2015) compares them.
Reference example
Details
Rezaei, J. (2016). Best-worst multi-criteria decision-making method: Some properties and a linear model. Omega 64, 126-130. doi:10.1016/j.omega.2015.12.001
Published examples (Tables 2-4 (input), Sections 3 and 4 (output)): car purchase, five criteria, best = price, worst = style. All three examples are verified: consistent (Table 2, xi^L = 0), slightly inconsistent (Table 3, xi^L = 0.0109, largest difference 4.6e-5) and higher-inconsistency (Table 4, xi^L = 0.0615, shown on this page and in the workbench). The site computes the linear model only.
Checked by this site's test suite.
Sources
Show 9 sources
- Rezaei (2015). Best-worst multi-criteria decision-making method. Omega 53, 49-57. doi:10.1016/j.omega.2014.11.009The 2015 paper introduces the non-linear minimax model. The linear (LP) model with a unique solution is Rezaei (2016), Omega 64, 126-130, which this site solves in exact closed form without an LP solver via Ratandhara & Kumar (2026). The input-based consistency ratio is from Liang, Brunelli & Rezaei (2020).
- Rezaei (2015), Omega 53, 49-57 doi:10.1016/j.omega.2014.11.009
- Rezaei (2016), Omega 64, 126-130 doi:10.1016/j.omega.2015.12.001
- Liang, Brunelli & Rezaei (2020), Omega 96, 102175 doi:10.1016/j.omega.2019.102175
- Ratandhara & Kumar (2026), IMA Journal of Management Mathematics, dpag027 doi:10.1093/imaman/dpag027
- Mi, Tang, Liao, Shen & Lev (2019), Omega 87, 205-225 doi:10.1016/j.omega.2019.01.009
- You et al. (2017), Sustainability 9(12), 2329 doi:10.3390/su9122329
- Gupta (2018), Journal of Air Transport Management 68, 35-47 doi:10.1016/j.jairtraman.2017.06.001
- Telli & Ayçin (2021), BWM-MABAC teacher selection, TroyAcademy 6(2), 733-750 doi:10.31454/troyacademy.899397