Agent-based spatial models applied to agriculture: a simulation tool for technology diffusion, resource use changes and policy analysis

This paper presents a spatial multi-agent programming model, which has been developed for assessing policy options in the diffusion of innovations and resource use changes. Unlike conventional simulation tools used in agricultural economics, the model class described here applies a multi-agent/cellular automata (CA) approach by using heterogeneous farm-household models and capturing their social and spatial interactions explicitly. The individual choice of the farm-household among available production, consumption, investment and marketing alternatives is represented in recursive linear programming models. Adoption constraints are introduced in form of network-threshold values that reflect the cumulative effects of experience and observation of peers' experiences. The model's economic and hydrologic components are tightly connected into a spatial framework. The integration of economic and hydrologic processes facilitates the, consideration of feedback effects in the use of water for irrigation. The simulation runs of the model are carried out with an empirical data set, which has been derived from various data sources on an agricultural region in Chile. Simulation results show that agent-based spatial modelling constitutes a powerful approach to better understanding processes of innovation and resource use change. (C) 2001 Elsevier Science BN. All rights reserved.