Illustrative example: a representative scenario, not a customer story
Water Quality Assessment
Evaluate water quality across multiple sampling stations using physicochemical and biological indicators. Grey relational analysis handles incomplete monitoring data while multi-criteria ranking identifies priority areas for intervention.
The decision problem
Six monitoring stations along a river are ranked to decide where remediation money goes first. The monitoring data is incomplete: some parameters are recorded as ranges, some stations have gaps in certain months, and a laboratory reports a value below the detection limit instead of a number. Deleting those rows would drop precisely the stations under suspicion.
What the decision matrix looks like
The typical shape of a study like this. Your own matrix can be larger or smaller; nothing here is fixed.
6
Alternatives
5
Criteria
Incomplete and interval measurement
Uncertainty type
| # | Criterion |
|---|---|
| C1 | Dissolved oxygen |
| C2 | Biochemical oxygen demand |
| C3 | Nitrate |
| C4 | Heavy metal concentration |
| C5 | Turbidity |
Which method, and why
Grey TOPSIS works on grey numbers, which carry a lower and an upper bound instead of a single point, so incomplete records enter the computation without being invented. SWARA turns the environmental team's ordering of the indicators into weights. CODAS separates stations with nearly identical scores by combining Euclidean and Taxicab distance, which matters when the top of the list is tightly packed.
What you get out
Ranking and scores
Every alternative with its score, its rank and the intermediate matrices that produced them.
Weight sensitivity
One criterion weight moves at a time, so you see exactly where the leader changes.
Method agreement
When more than one ranking method is run, the orders are compared with Spearman rho and Kendall W.
Report and citation
PDF, DOCX and XLSX output carrying the seminal source of every method used.
Run this flow on your own data
Load the matrix, choose the weighting and ranking methods, read the sensitivity.