Analytic Hierarchy Process

The Analytic Hierarchy Process originated from Thomas Saaty's work at the University of Pittsburgh, first published in 1980. Saaty developed AHP to help the U.S. State Department address arms control decisions — problems involving numerous stakeholders, conflicting objectives, and both tangible and intangible factors. The method proved so versatile that it quickly spread to business strategy, engineering, healthcare, and environmental planning.

AHP works through four main steps. First, the decision is structured as a hierarchy: the goal at the top, criteria and sub-criteria in the middle levels, and alternatives at the bottom. This decomposition forces decision-makers to explicitly identify all relevant factors and their relationships. Second, pairwise comparisons are performed at each level — every criterion is compared against every other criterion on a 1-9 scale (1 = equal importance, 9 = extreme importance of one over another). Third, priority vectors are computed from the comparison matrices using eigenvalue methods, producing numerical weights for each criterion. Fourth, alternatives are scored against each criterion, and the weighted scores are synthesized into an overall ranking.

A distinctive feature of AHP is its consistency check. Because pairwise comparisons can be contradictory (if A is twice as important as B, and B is three times as important as C, then A should be six times as important as C), AHP calculates a consistency ratio. If the ratio exceeds 0.10, the comparisons contain significant inconsistency and should be revised. This built-in quality control is unique among MCDA methods and helps prevent logically flawed evaluations.

AHP has been applied to thousands of documented decisions worldwide. The World Bank uses it for project selection, IBM used it for strategic planning, and the U.S. military uses it for systems acquisition. Its strength lies in handling both quantitative and qualitative criteria within the same mathematical framework — a vendor's cost (quantitative) and cultural fit (qualitative) are evaluated on the same scale through pairwise comparison.

One limitation of AHP is rank reversal: adding or removing an alternative can sometimes change the relative ranking of the remaining alternatives. Researchers have proposed various solutions, including using an ideal-mode AHP variant or combining AHP with other methods like TOPSIS. In practice, the most robust approach is running multiple MCDA methods and comparing their rankings.