What Is MCDM?

MCDM, or multi-criteria decision making, is a branch of operations research concerned with structuring and solving decision problems in which several conflicting criteria must be weighed at once. It is also written MCDA, for multi-criteria decision analysis, with the analysis label often reserved for the discursive, workshop-based practice and the decision-making label for the algorithmic side. A typical problem supplies a finite set of alternatives, a set of criteria on which those alternatives are scored, and a decision maker whose preferences determine how the criteria trade off against one another. Selecting a supplier by price, lead time, and defect rate, or siting a substation by cost, land use, and outage exposure, are canonical examples.

The field grew out of utility theory and mathematical programming in the 1960s and 1970s, drawing on Bernard Roy's work on outranking in France and on decision analysis in the United States. Its core insight is that there is usually no alternative that is best on every criterion, so the analyst's task is to make the trade-offs explicit rather than to hide them in a single objective function.

Value Measurement and Weighting

One family of methods builds an explicit value or utility function and aggregates weighted criterion scores into a single index. Multi-attribute value theory and multi-attribute utility theory formalize the conditions under which such an aggregation is valid. The analytic hierarchy process, introduced by Thomas Saaty, decomposes a problem into a hierarchy of goal, criteria, and alternatives, then derives weights from pairwise comparisons using the principal eigenvector of a reciprocal comparison matrix, with a consistency ratio flagging incoherent judgments. Weight elicitation is the most contested step in the whole field, since small changes in weights can reorder the alternatives.

Outranking and Distance-Based Methods

A second family avoids a single value scale. Outranking methods such as ELECTRE and PROMETHEE compare alternatives pairwise, accumulating concordance and discordance evidence and permitting incomparability when the evidence conflicts. Distance-based methods such as TOPSIS instead rank alternatives by their geometric closeness to an ideal solution and their distance from an anti-ideal one. Hybrid pipelines are common in practice: an integrated AHP and TOPSIS methodology for software selection uses pairwise comparison to set criterion weights and then applies the distance measure to rank candidates. Comparative studies, including a survey of multi-criteria decision-making methods prepared for the US Army Research Laboratory, examine how far these procedures agree on the same data.

Uncertainty and Method Selection

Criterion scores are frequently imprecise, so fuzzy set, grey system, and rough set extensions of the main methods replace crisp values with intervals or membership functions. Stochastic multicriteria acceptability analysis samples weight spaces to report how often each alternative ranks first rather than committing to one weight vector. Rank reversal, in which adding or removing an alternative changes the relative order of the others, remains a known pathology of several normalization schemes. Because different methods can produce different orderings from identical inputs, a comparative analysis of MCDM methods and routine sensitivity testing are treated as part of a defensible study rather than optional extras.

Applications

MCDM has applications in a wide range of disciplines, including:

• Energy planning, covering technology portfolios, renewable site selection, and grid investment
• Supply chain management, particularly supplier evaluation and facility location
• Environmental management, including watershed planning and land use allocation
• Software and systems engineering, for architecture, tool, and requirements prioritization
• Healthcare policy, where benefit-risk assessment of treatments involves several outcome measures
• Transportation planning, in route, mode, and infrastructure project selection