Agent-Based Modeling for Evaluation of Crop Pattern and Water Management Policies

Abstract

The objective of the present paper is to propose a framework for development of an optimal cropping pattern aimed at ground water recovery using an agent based approach. In the proposed agent-based model (ABM), the agents’ learning from each other as well as their self-learning from their own behavioral feedback was studied through simulation of the behavior of agricultural agents using fuzzy inference system (FIS). Moreover, the agents’ behavior were determined using linear programming in order to maximize the farmers’ income. The governmental agent regulated the interactions between agricultural and environmental agents by imposing its policies in the form of scenarios. The efficiency of the presented methodology was evaluated using hydrological data of Najaf Abad region, located in Iran’s central plain, on the basis of three hydrological scenarios (wet, normal, and dry) subject to governmental policy of aquifer recovery. The results showed that in a normal scenario with current groundwater withdrawal, the water level reduced by an average of 0.18 m per year. In contrast, the water level increased by an average of 0.48 m under aquifer recovery scenario. Furthermore, despite 17% reduction in water rights of agricultural agents in the study area, the total long-term agricultural income declined only by less than 4%, and in the planning horizon, their average annual income associated with these two management scenarios were estimated at 123.5 and 119 million US dollars, respectively.