The Future for Cellular Automata in Geographic Research

Research work in this area is focused on a broad range of interests. Development of new applications is still quite active, with new cellular automata models popping up for a variety of phenomena of geographic interest, with an associated growth in their use as test beds for theory, practice, and policy.

Researchers have also begun to focus on extending the basic idea of cellular automata from mathematics and computer science for geographic applications. Work in specifying cells, lattices, and neighborhoods through GI systems is particularly active, with recent advances in the use of graphs, Voronoi polygons, irregular cells, and 2.5D lattices.

Cellular automata are naturally allied with remote sensing, GI systems, and the dataware associated with them. Various attempts have been made to develop cellular automata engines within GI systems and to build GI systems functionality into cellular automata. Recent work, however, has focused on the mutual links between cellular automata and GI science.

Research into the connections between cellular automata and agent automata is central to the current research agenda. The two are popularly confused, even though the distinction on geographic grounds is reasonably straightforward: cells do not actually move within their lattices and engage in action by proximity, while agents can move freely in vector spaces and can engage in action at a distance. Tool kits and methodologies that support both approaches are just beginning to be built.

Issues surrounding calibration and validation of cellular automata models are chief among challenges facing future research in this area. Existing work has an overarching focus on patterns and their role in benchmarking cellular automata models, but research into the role of processes in validation and calibration is comparatively underdeveloped.