Exploiting Spatio–temporal Data for the Multiobjective Optimization of Cellular Automata Models
The increased availability of remotely sensed spatio-temporal data offers the chance to improve the reliability of an important class of Cellular Automata (CA) models used for the simulation of real complex systems. To this end, this paper proposes a multiobjective approach, based on a genetic algorithm, which can present some significant advantages if compared with standard single-objective optimizations. The method exploits the available temporal sequences of spatial data in order to produce CAs which are non-dominated with respect to multiple objectives. The latter represent, in different metrics, the level of agreement between the simulated and real spatio-temporal processes. The set of non-dominated CAs proves to be a valuable source of information about potentialities and limits of a specific CA model structure.
KeywordsCellular Automaton Multiobjective Optimization Cellular Automaton Cellular Automaton Model Cellular Automaton Model
Unable to display preview. Download preview PDF.
- Yeh, A.G.O., Li, X.: Simulation of development alternatives using neural networks, cellular automata, and GIS for urban planning. Photog. Eng. & Remote Sensing 69, 1043–1052 (2003)Google Scholar
- Goldstein, N.C.: Brains vs. brawn comparative strategies for the calibration of a cellular automata based urban growth model. In: Proceedings of the 7th International Conference on GeoComputation (2003)Google Scholar
- Pareto, V.: Cours d’Economie Politique. Volume I, II. F. Rouge, Lausanne (1896)Google Scholar
- Project Gigalopolis, NCGIA (2003), http://www.ncgia.ucsb.edu/projects/gig/