| Regarding to the importance of transition rules to cellular automata and thefact that they are always got from historical data, A research framework has been setcontaining four steps, analyzing, giving some premises, modeling and verifying, tostudy how to convert the knowledge got from historical data to transition rules(orparameters) in the time scale of cellular automata, which is a partial problem of theparallelism of real time and CA time.This paper aims at the construction of Xiamen Island's urban cellular automatamodel, and follow the structure of the framework step by step. Firstly, analysis theland-use changes in Xiamen Island so as to get some characteristics of it; secondly,give some premises and come with a solution of the problem of converting theknowledge got from historical data to transition rules(or parameters) in the timescale of cellular automata in this article; thirdly, with the help of the data comingfrom the analyzing stage, parameters of the CA model have been fixed; finally, usethe kappa coefficient and Moran index, to evaluate the modeling results accordingto the historical data, and choose the optimal parameters among the providedparameters based on the evaluation results.The evaluation of the modeling result shows that: urban CA model constructedin this article is sufficient to simulating the urban development of Xiamen Islandduring the period from 1989 to 2004. As the model is built in following the process ofanalyzing-giving some premises-solution modeling-verify the results, on thecondition that the results is found acceptable, the premises also are deemed ascorrect, as well as some theoretical conceptions, such as the relativity of CA time andthe stochastics of CA model. Finally, the goal of convert the knowledge got fromhistorical data to transition rules(or parameters) in the time scale of cellularautomata has been accomplished.
|
PDF Full Text Request