Agent-based simulation of stakeholders' negotiation regarding land development scenarios using a fuzzy Analytic Hierarchy Process

This paper describes an agent-based model (ABM) that was developed to simulate the negotiation process of stakeholders over different land development scenarios in the Elbow River watershed in southern Alberta, an area subject to considerable urbanization pressure due to its proximity to the fast growing City of Calgary. In this initial phase of development, the stakeholders represented as agents include representatives of a Municipal District and two non-profit organizations. The modeling framework contains three main components: a web interface designed to facilitate the interactions of the users with the system, a PostgreSQL database in which data regarding the stakeholders’ preferences are stored that uses PostGIS plugin for spatial functionalities, and an ABM developed in Java that accesses Repast Simphony libraries to simulate the negotiation process among stakeholders. The negotiation starts by a development plan being submitted by a user (stakeholder) through the web interface. The ABM module receives this plan and conducts the negotiation process in a step-wise manner. In each time step, the agents move the proposed plan to a new location and evaluate that location based on their stored criteria, using GIS tools provided by Repast Simphony and PostGIS. The limits of the search space are specified by the plan proposer. As the decision makers are often not willing to assign crisp numerical values to the relative importance of their criteria, in this model the users compare their preferences using linguistic expressions. Due to the uncertainty and fuzzy nature of such comparisons, the agents make use of a fuzzy Analytic Hierarchy Process (AHP) to prioritize their criteria and to bring fuzziness into the pairwise comparison of AHP. In the final step, the ABM investigates the highly ranked locations of all agents to find a common area which is outputted as the result of the negotiation. The proposed simulation model facilitates the interactions of stakeholders who have different perspectives regarding potential land development scenarios in the watershed and allow them to reach an acceptable agreement considering their own preferences along with other stakeholders’ preferences.