Multi-Criteria Decision Analysis¶
Method family — instantiates Objective Weighting Governance
Provides structured methods for comparing alternatives across multiple criteria.
Where a single meeting produces one team's weights, Multi-Criteria Decision Analysis (MCDA) is the library of formal methods the field has built for eliciting weights and aggregating criteria — Analytic Hierarchy Process, TOPSIS, ELECTRE, MAUT, outranking, goal programming, and their kin. Its defining idea is that it is a family of interchangeable techniques, each with an explicit theory of how importance is measured and how criteria combine, not a single formula. Choosing an MCDA method is choosing a stance: whether weights are derived from pairwise comparisons or direct rating, whether criteria compensate freely or are held apart, whether the output is a full ranking or a filtered shortlist. It supplies the mathematical machinery — including methods where a score means distance to an ideal rather than a weighted total — that other mechanisms merely assume.
Example¶
A municipal water utility must pick one of five wastewater-treatment technologies. The criteria — capital cost, operating cost, energy use, effluent quality, and footprint — trade off against each other in ways no one can eyeball. The engineering team reaches for MCDA rather than a hand-built spreadsheet. They use the Analytic Hierarchy Process to derive weights: instead of asking "how important is energy use, 1–10?", AHP asks a series of pairwise questions[1] — "energy use versus operating cost: which matters more, and by how much?" — and computes weights from the answers, flagging when the pairwise judgments are internally inconsistent. Then they run the alternatives through TOPSIS, whose score is defined as how close each option sits to an ideal best-and-worst profile rather than a simple weighted sum. The interpretation rule is set by the method: TOPSIS yields a closeness ranking, so the team reads the output as an ordering to shortlist from, not as an absolute quality figure. Two methods, two explicit theories, one defensible comparison.
How it works¶
- Pick the elicitation method. Pairwise (AHP), swing weighting, direct rating, or trade-off analysis — each turns human judgment into weights with a different theory and a different consistency check.
- Pick the aggregation method. Weighted sum, distance-to-ideal (TOPSIS), or non-compensatory outranking (ELECTRE) — the choice decides whether strong criteria can rescue weak ones.
- Define what the output means. The method fixes the interpretation: a cardinal score, an ordinal ranking, or a dominance shortlist — and the analyst commits to reading it only that strongly.
- Check method-internal consistency. Many MCDA methods carry their own diagnostics (e.g. AHP's consistency ratio) that reject incoherent judgments before they propagate.
Tuning parameters¶
- Method choice — the master dial. Compensatory methods reward all-round options; non-compensatory ones protect against a fatal weakness. Match it to whether tradeoffs should be allowed.
- Elicitation mode — pairwise versus direct rating. Pairwise resists anchoring and exposes inconsistency but is cognitively heavy for many criteria.
- Normalization scheme — how raw criterion values are scaled before combining. Different schemes can reorder results, so it is a substantive choice, not a formatting one.
- Number of criteria — more criteria capture nuance but dilute weights and multiply elicitation burden.
- Scoring scale — ordinal, interval, or ratio; richer scales enable more methods but demand more defensible measurement.
When it helps, and when it misleads¶
Its strength is rigor and range: it gives contested multi-objective choices a defensible, replicable structure, and its diversity means the analyst can pick a method whose assumptions actually fit the decision — including protecting a criterion from compensation when that matters.
Its failure mode is method-induced artifact: some MCDA techniques exhibit rank reversal, where adding or removing an irrelevant alternative flips the order of the others, so a result can be an artifact of the method rather than a fact about the options. The classic misuse is method-shopping — trying techniques until one produces the preferred winner. The guarding discipline is to fix the method and its normalization before seeing results, and to treat any ranking that a plausible alternative method overturns as fragile rather than final.
How it implements the components¶
objective_weight— its elicitation methods (AHP, swing weighting) are formal procedures for deriving the relative-importance numbers themselves.weight_setting_process— it supplies the structured, theory-backed procedures for how weights are elicited and made coherent, with built-in consistency checks.score_interpretation_rule— each method fixes what its output authorizes (a cardinal score, a closeness ranking, a dominance shortlist), constraining how strongly the result may be read.
It computes and ranks but does not police whether its measured inputs still track the real objectives, nor trace who wins and loses: proxy_alignment_check and decision_impact_trace belong to Weighted Scoring Model; MCDA is the method family, its twin is the deployed scorecard that runs the impacts.
Related¶
- Instantiates: Objective Weighting Governance — supplies the formal method library for eliciting weights and comparing options.
- Consumes: Deliberative Weight-Setting Session — the human judgments its elicitation methods formalize often come from a facilitated session.
- Sibling mechanisms: Weighted Scoring Model · Weighted Sum Objective · Weight Sensitivity Sweep · Ranking Stability Report
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: The mechanism elicits weights, scores alternatives, applies a declared aggregation method, and checks sensitivity to produce a ranking or dominance result.
Nearest alternative: Decision, Gate & Allocation — A decision maker may act on the result, but the mechanism itself supplies comparative analysis rather than committing the allocation.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Operations Research
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Multi-criteria decision analysis is an established operations-research method family for evaluating alternatives under plural objectives.
Related originating lineages:
- Economics & Finance — Utility theory and welfare aggregation provide important formal roots.
- Organizational & Management Science — Strategic decision practice operationalizes criteria weighting with stakeholders.
Review resolution: Both independent reviews agree on primary origin operations_research; reconciliation resolves secondary fields (alternate_origin_disagreement, origin_mode_disagreement, domain_reach_disagreement). Alternate origins retained (economics_finance, organizational_management) are the union of reviewer-supported formative lineages with explicit rationales, not a list of later application domains. Present-day breadth is represented separately as domain_reach=multi_domain; origin_mode=cross_disciplinary_synthesis records the historical relationship among lineages. Confidence is conservatively reconciled to high, and encyclopedia_synthesis=false preserves either reviewer's finding that the encyclopedia generalized the mechanism.
Review outcome: Reconciled after independent review; high confidence.
References¶
[1] Saaty, T. L. The Analytic Hierarchy Process: Planning, Priority Setting, Resource Allocation. McGraw-Hill International Book Company (1980). Derives criterion weights in AHP from pairwise judgments of relative importance rather than standalone ratings. registry ↩