Land-Use Decisions in Developing Countries and their Representation in Multi-Agent Systems

Publikations-Art

Zeitschriftenbeitrag (peer-reviewed)

Autoren

Schreinemachers, P., Berger, T.

Erscheinungsjahr

2006

Veröffentlicht in

Journal of Land Use Science

Band/Volume

1/1

DOI

10.1080/17474230600605202

Seite (von - bis)

29-44

Schlagworte

sustainability

Recent research on land use and land cover change (LUCC) has put more emphasis on the importance of understanding the decision-making of human actors, especially in developing countries. The quest is now for a new generation of LUCC models with a decision-making component. This paper deals with the question of how to realistically represent decision-making in land use models. Two main agent decision architectures are compared. Heuristic agents take sequential decisions following a pre-defined decision tree, while optimizing agents take simultaneous decisions by solving a mathematical programming model. Optimizing behaviour is often discarded as being unrealistic. Yet the paper shows that optimizing agents do have important advantages for empirical land use modelling and that multi-agent systems (MAS) offer an ideal framework for using the strengths of both agent decision architectures. The use of optimization models is advanced with a novel three-stage decision model of investment, production, and consumption to represent uncertainty in models of land use decision-making.