Incorporating Cellular Automata simulators as analytical engines in GIS

Abstract

Contemporvy GIS suffer from a variety of problems. These include poor ability to handle dynamic spatial models, poor performance for many operations, and poor handling of the temporal dimension. Cellular Automata (CA) are dynamic mathematical systems based on discrete time and space. While CA have many similarities to GIS, and indeed excel in many areas in which GIS are deficient, they cannot themselves be considered as GIS since they lack the necessary sophisticated Capabilities in data input, data storage and retrieval, and data output. However, the advantages of CA in data analysis and modelling are significant. In order to exploit fully the advantages of each system ‐ GIS and CA ‐ we have developed a methodology and prototype system for coupling GIS and CA, in which the CA serves as an alternative analytical engine for the GIS. In the prototype system, spatial, temporal (i.e. time series), and spatio‐temporal filters as well as spatial diffusion operators are developed. In addition, the integrated system provides a flexible framework for the programming and running of dynamic spatial models. We conclude that CA represent a viable alternate analytical engine for GIS and provide increased ability for dynamic spatial modelling within GIS.

Modelling sustainable urban development by the integration of constrained cellular automata and GIS

“…The development of CA within GIS greatly enhances the ability of dynamic spatial modelling within GIS. Other researches also support that integrated CA-GIS approaches can enhance the current poor modelling capability of GIS (Park and Wagner 1997 ).…”

“…The limitations of contemporary GIS include its poor ability to handle dynamic spatial models, poor performance for many operations, and poor handling of the temporal dimension (Park and Wagner 1997 ). In coupling GIS with CA, CA can serves as an analytical engine to provide a¯exible framework for the programming and running of dynamic spatial models.…”

Loose-coupling a cellular automaton model and GIS: long-term urban growth prediction for San Francisco and Washington/Baltimore

“…In a recent paper, Park and Wagner ( 1997 ) implemented a tight coupling of several cellular automaton models (including ours) within Idrisi using the Cellang CA language. In the context of the our cellular model, GIS served at least three important roles, none of which could be called tight coupling.…”

“…This function alone favours the use of loose coupling. In spite of e orts to build cellular modelling functions into GIS directly ( Takeyama and Couclelis 1997;Park and Wagner 1997) and the suitability both of speci® c GIS packages (e.g. GRASS) and of control languages (e.g.…”

The Challenges to Coupling Dynamic Geospatial Models

“…For example, a fine grain wind model at 1 meter resolution and 1 second timesteps cannot be simply generalized to a 1km grid cell at 1 day intervals, or vice versa. One potential solution to this problem is using Cellular Automata (CA) as components of spatial models, as they are standardized in their spatial grain, and their temporal resolution can be made flexible, depending on the differences of the temporal patterns of each of the models (Park and Wagner 1997).…”