Which is the appropriate triangular distribution to employ in the modified analytic hierarchy process?

The analytic hierarchy process (AHP) is a widely used tool. It is employed in multicriteria decision making and is used to deal with complex problems. Traditionally in the process, decision makers produce a single numerical value indicating their preference between a pair of objectives. As an extension, to encapsulate uncertainty, the decision maker is allowed to select parameters (minimum, mode and maximum) from a triangular distribution. This is referred to as the modified AHP. Our paper investigates what precise form the triangular distribution should take and suggests that the current practice is not ideal.