MARCOS
AvailableMeasurement of Alternatives and Ranking according to COmpromise Solution. Stević et al. (2020)
The idea
The dashed rows are imaginary: the best and the worst value of every criterion. K- is a bar against the anti-ideal, K+ against the ideal; f(K) on the right.
Show as table
| Alternative | Weighted total S | K⁻ | K⁺ | Utility f(K) |
|---|---|---|---|---|
| Ideal AI | 1.0000 | |||
| A8 | 0.9309 | 1.982 | 0.931 | 0.8094 |
| A19 | 0.9084 | 1.934 | 0.908 | 0.7899 |
| A18 | 0.9043 | 1.925 | 0.904 | 0.7863 |
| A3 | 0.9001 | 1.917 | 0.900 | 0.7827 |
| A2 | 0.8986 | 1.913 | 0.899 | 0.7813 |
| A22 | 0.8865 | 1.888 | 0.887 | 0.7708 |
| A4 | 0.8769 | 1.867 | 0.877 | 0.7624 |
| A5 | 0.8763 | 1.866 | 0.876 | 0.7620 |
| A9 | 0.8694 | 1.851 | 0.869 | 0.7560 |
| A20 | 0.8666 | 1.845 | 0.867 | 0.7535 |
| A13 | 0.8554 | 1.821 | 0.855 | 0.7438 |
| A1 | 0.8485 | 1.807 | 0.849 | 0.7378 |
| A11 | 0.8436 | 1.796 | 0.844 | 0.7335 |
| A14 | 0.8427 | 1.794 | 0.843 | 0.7327 |
| A21 | 0.8399 | 1.788 | 0.840 | 0.7303 |
| A15 | 0.8296 | 1.767 | 0.830 | 0.7214 |
| A16 | 0.8214 | 1.749 | 0.821 | 0.7143 |
| A7 | 0.8186 | 1.743 | 0.819 | 0.7118 |
| A10 | 0.8105 | 1.726 | 0.811 | 0.7048 |
| A23 | 0.8028 | 1.709 | 0.803 | 0.6980 |
| A17 | 0.7987 | 1.701 | 0.799 | 0.6945 |
| A6 | 0.6264 | 1.334 | 0.626 | 0.5446 |
| A12 | 0.6227 | 1.326 | 0.623 | 0.5415 |
| Anti-ideal AAI | 0.4696 |
Every dot lies on one line through 0, so no two alternatives can swap places.
Show as table
| Alternative | SAW (Linear) score S | MARCOS f(K) |
|---|---|---|
| A8 | 0.9309 | 0.8094 |
| A19 | 0.9084 | 0.7899 |
| A18 | 0.9043 | 0.7863 |
| A3 | 0.9001 | 0.7827 |
| A2 | 0.8986 | 0.7813 |
| A22 | 0.8865 | 0.7708 |
| A4 | 0.8769 | 0.7624 |
| A5 | 0.8763 | 0.7620 |
| A9 | 0.8694 | 0.7560 |
| A20 | 0.8666 | 0.7535 |
| A13 | 0.8554 | 0.7438 |
| A1 | 0.8485 | 0.7378 |
| A11 | 0.8436 | 0.7335 |
| A14 | 0.8427 | 0.7327 |
| A21 | 0.8399 | 0.7303 |
| A15 | 0.8296 | 0.7214 |
| A16 | 0.8214 | 0.7143 |
| A7 | 0.8186 | 0.7118 |
| A10 | 0.8105 | 0.7048 |
| A23 | 0.8028 | 0.6980 |
| A17 | 0.7987 | 0.6945 |
| A6 | 0.6264 | 0.5446 |
| A12 | 0.6227 | 0.5415 |
Data: Truck drivers (Stević and Brković 2020).
MARCOS adds an imaginary ideal row (the best value of every criterion) and an anti-ideal row (the worst) to the table, and scores each alternative by its weighted total against both.
When to use it
- All values are positive amounts.
- Results should read as closeness to the ideal, distance from the worst.
- Results should be comparable with FUCOM, CRITIC or AHP then MARCOS studies.
Algorithm
1Add ideal and anti-ideal rows
2Normalize against the ideal
Applied to the AI and AAI rows as well.
3Weighted sums
S_aai and S_ai are the same sums for the two imaginary rows; with weights summing to 1, S_ai = 1.
4Utility functions
Both are the same number for every alternative, so they are shown once.
5Final utility
Rank by decreasing f(K); f(K) = c S with one c for the whole table, so the order is that of SAW (Linear).
Used together with
Why these pairs
- FUCOM weights, then MARCOS: Stević & Brković (2020), the example used here.
- AHP weights, then MARCOS: Madenoğlu (2020), the second test example.
- CRITIC is the weighting most often paired with MARCOS (Demir et al. 2024, from a search snippet).
- Same ranking as SAW with the Linear option, always; only the scale of the scores differs.
Reference example
Details
Stević, Ž.; Brković, N. (2020). A Novel Integrated FUCOM-MARCOS Model for Evaluation of Human Resources in a Transport Company. Logistics 4(1), 4. doi:10.3390/logistics4010004
Published example, reproduced by the tests: 23 drivers, five criteria (two cost), FUCOM weights (Tables 2, 4, 5). All 23 f(K) values match the printed 3 decimals (largest gap 0.0006, from the paper's rounded divisor 0.470) and so do all 23 ranks. The paper applies the weights of C3 and C4 swapped against its own FUCOM figure; the test uses them as applied and records that the figure's weights move 9 ranks. Also tested: Madenoğlu (2020), six suppliers, 16 criteria, AHP weights, all ranks.
Checked by this site's test suite.
Sources
Show 6 sources
- Stević et al. (2020). Sustainable supplier selection in healthcare industries using a new MCDM method: Measurement of alternatives and ranking according to COmpromise solution (MARCOS). Computers & Industrial Engineering 140, 106231. doi:10.1016/j.cie.2019.106231Paywalled; not opened. The formulas follow the first author's open restatement (Stević & Brković 2020, Eqs. 1-12), checked against Madenoğlu (2020).
- Stević, Pamučar, Puška & Chatterjee (2020), Computers & Industrial Engineering 140, 106231 (not opened) doi:10.1016/j.cie.2019.106231
- Stević & Brković (2020), Logistics 4(1), 4 doi:10.3390/logistics4010004
- Madenoğlu (2020), Economics Business and Organization Research 2(2), 99-120 dergipark.org.tr/tr/download/article-file/1432780
- Demir, Chatterjee, Kadry, Abdelhadi & Pamučar (2024), Decision Making: Applications in Management and Engineering (search snippet only, not read) doi:10.31181/dmame7220241137
- pyDecision 5.1.8, marcos_method github.com/Valdecy/pyDecision