Multicriteria

Multicriteria refers to problems or analyses that involve more than one criterion for evaluation. In decision making and optimization, it recognizes that choices must be assessed along several dimensions that may conflict, such as cost, quality, and risk. The goal is to identify alternatives that perform well across the criteria or to understand trade-offs between conflicting objectives.

Multicriteria decision analysis (MCDA) is a field within operations research that aims to rank, compare, or select alternatives while incorporating stakeholder preferences. Multicriteria optimization treats the problem as a mathematical program with multiple objective functions to be optimized simultaneously, often requiring Pareto efficiency concepts and constraint handling.

Key concepts include criteria (attributes), alternatives, and preference information. Criteria may be qualitative or quantitative and require normalization. Preferences can be expressed as weights, value functions, or outranking relations. Aggregation rules combine criteria into an overall judgment, while sensitivity analysis assesses robustness of results.

Common methods include additive models like weighted sum, multivariate extensions; pairwise comparison methods such as AHP; similarity-to-ideal methods like TOPSIS; outranking approaches such as ELECTRE and PROMETHEE. Some methods are interactive, iteratively eliciting preferences.

Applications span engineering design, environmental planning, energy and transportation, finance, and public policy, where decisions must balance multiple objectives and stakeholder views.

History and issues: MCDA emerged in the mid-20th century and has evolved with advances in decision theory, data availability, and computation. Key concerns include subjectivity in weighting, combining heterogeneous data, and ensuring robustness through sensitivity analysis.