Urban simulation Using Neural Networks and Cellular Automata for Land Use Planning

Cellular automata models consist of a simulation environment represented by a gridded space (raster), in which a set of transition rules determine the attribute of each given cell taking into account the attributes of cells in its vicinities. These models have been very successful in view of their operationality, simplicity and ability to embody logics- as well as mathematics-based transition rules in both theoretical and practical examples. Even in the simplest CA, complex global patterns can emerge directly from the application of local rules, and it is precisely this property of emergent complexity that makes CA so fascinating and their use so appealing. The calibration of CA1 models is very difficult when there is a large set of parameters. In the proposed model, most of the parameter values for CA simulation are automatically determined by the training of artificial neural network. In this paper ANN2 based CA urban growth simulation and prediction of Esfahan over the last four decades succeeds to simulate specified tested growth years at a high precision level. Some real data layer have been used in the ANN-CA simulation training phase such as 1990 while others used for testing the prediction results such as 2001. Next step includes running the developed ANN- CA simulation over classified raster data for forty years in a developed.An ArcGIS extension has been developed to define