NSF/ESRC Agenda Setting Workshop on Agent-Based Modeling of Complex Spatial Systems: April 14-16, 2007 Evaluating Agent-Based Spatial Models

I am interested but worried about the development of this relatively new class of models which tend to fight against very long standing principles in science that seek for simplicity in abstraction and application. Agent-based models (ABM) have developed for spatial systems through advances in computing that enable distinct objects to be defined with respect to their behaviours which in turn suggest that much more highly disaggregate systems can be represented than hitherto. The fact that ABMs are defined with respect to behaviour implicitly means that such models are temporally dynamic and this too increases the richness of these models in terms of the variability of the systems that are being simulated. Spatial data too is being acquired at ever finer spatial and temporal scales and is making possible the development of richer models structures of which ABMs are typical.