Implementation of a vector-based cellular automata model for simulating land-use changes

Abstract

In the last few years, cellular automata (CA) models have been increasingly used to simulate the complex land-use changes. However, the traditional regular raster-based CA models are sensitive to the cell size and the neighborhood configuration used in the models, which restrict its ability to simulate the real world. By representing space as irregular shape and size geographic objects and defining artificial neural network as transition rule, a vector-based CA model is constructed and applied to the simulation and prediction of land-use changes. By taking the north branch of Yangtze River estuary as an example, studies illustrated that the vector-based CA model can make full use of artificial neural network to obtain the variable space parameters and simplify the land-use transfer rule. It is concluded that the vector-based CA model produces more realistic spatial patterns than those generated by a raster-based CA model.