Constrained consistency enforcement in AHP

Abstract

Decision-making in the presence of intangible elements must be based on a robust, but subtle, balance between expert know-how and judgment consistency when eliciting that know-how. This balance is frequently achieved as a trade-off reached after a feedback process softens the tension frequently found between one force steadily pulling towards (full) consistency, and another force driven by expert feeling and opinion. The linearization method, developed by the authors in the framework of the analytic hierarchy process, is a pull-towards-consistency mechanism that shows the path from an inconsistent body of judgment elicited from an expert towards consistency, by suggesting optimal changes to the expert opinions. However, experts may be reluctant to alter some of their issued opinions, and may wish to impose constraints on the adjustments suggested by the consistency-enforcement mechanism. In this paper, using the classical Riesz representation theorem, the linearization method is accommodated to consider various types of constraints imposed by experts during the abovementioned feedback process.