What is it about?

We develop a methodology for the semi-automatic search for consensus building in group decision making in a local context (one criterion). The methodology permits to analyse if there is a consensus between the actors involved in the decision-making process. Besides, if there is a lack of agreement among them, it allows the identification of the reasons of that lack of agreement, and to build different scenarios from which negotiation processes can be inititated in order to reach consensus among all decision makers.

Featured Image

Why is it important?

The methodology is based in the Analytic Hierarchy Process (AHP) which is one of the most used approaches to analyse decision making processes with a finite number of alternatives and decision criteria. The metodology uses Bayesian models to describe the individual judegment elicitation processes and allows the analysis of the extent to which it is possible to aggregate the information provided by each decision maker without losing the consistency of judgements and calculating the priorities vector associated with compared alternatives. All this gives a greater scientific rigor to the group decision making processes without losing the flexibility of AHP.


This paper is a first step in the development of statistical and computer tools to reach large consensus in group decision making problems using AHP. Currently, we are extending the framework of the paper to the analysis of decision problems with a hierarchy of decision criteria and a high number of decision makers. We think that our results will help to carry out a scientific analysis of decision problems where the decision makers express their opinions through the Web, which will result in more realistic and representative decision-making processes.

Facultad de Economía y Empresa. University of Zaragoza

Read the Original

This page is a summary of: Consensus Building in AHP-Group Decision Making: A Bayesian Approach, Operations Research, December 2010, INFORMS,
DOI: 10.1287/opre.1100.0856.
You can read the full text:



The following have contributed to this page