The Analytical Hierarchy Process

General problem-solving methodology useful for complex multi-criteria decisions with variables that do not have exact numerical consequences.

Particularly useful when the decision-maker can provide a logical structure to the problem, but does not have objective data to evaluate alternatives (strategic decision making, for example).

- Helps structure complex problem(s) in an intuitive manner.
- Uses "relative" judgments that are often easier to make than absolute judgments.
- Enables managers to more fully and carefully consider the alternatives.
- Allows manager to include qualitative aspects of a problem situation.
- Has been "validated" in a number of tests.
- Has been widely applied in industry.

Decompose a problem into hierarchical levels using the following building blocks:
1. Overall Goal (subgoals) to be attained (e.g., best location for store).
2. Criteria and Subcriteria (e.g., store traffic, growth of neighbourhood, etc.).
3. Scenarios (e.g., growing neighbourhood, stable economy, etc.).
4. Alternatives (e.g., suburban shopping mall, stand-alone store, downtown).

Establish priorities concerning alternatives in the hierarchy. This is done by comparing alternatives in a pair-wise manner to assess their relative importance, relative preference, or relative likelihood with respect to criteria at the immediately higher level in the hierarchy. The evaluation is made using either a verbal scale, numerical scale, or a rating scale.

Assess the consistency of the judgments using consistency ratio.

Synthesize priorities at each level to obtain overall priority of each alternative with respect to the primary goal. The total priority for each alternative is the degree to which that alternative fits all the criteria, subcriteria, and scenarios.

Use sensitivity analyses to explore the alternatives.
1. Reciprocal condition

If A is 5 times more comfortable than B, then B is 1/5th as comfortable as A.

2. Comparability within limit