Simulation of city development using cellular automata and agent based integrated model — A case study of Qingdao city

Qiang Fu,Jun Song,Feng Mao,Wensheng Zhou

DOI: https://doi.org/10.1109/GEOINFORMATICS.2010.5567860

2010-01-01

GEOINFORMATICS

Abstract:This paper first developed an integrated model of urban land use changing using cellular automata (CA) and agent technique. In this model, every cell not only has the current state information of the cell, but also has the land use changing related information such as policy and natural condition information; the agents with different roles are used to process the information contained in the cell and then decide the state of the next time step of the cell. For instance, the natural condition information analysis agent is in charge of processing natural condition information; the policy information analysis agent is in charge of processing policy information; surrounding sensor agent is in charge of sensing the state of the cells in the neighborhood; and the logical decision agent makes the logical decision basing on the analytical results processed by the agents mentioned above, and decides whether the cell change its state in the next time step. This model can be extended according to the increasing understanding of the development mechanism of city which is known as the complex giant system. Thanks to the addition of multi-role agents, the model has a global vision of the focused area and thereby the model has complexity and flexibility in processing which can effectively simulate urban city development. Qingdao city has changed to an important seaport, tourist, information technology and ocean science city in China from a small fishery village in the past over 100 years. The city has experienced three main phases of development, and the driving forces behind each development are different. In this paper, the model is used to simulate the three development scenes in the Qingdao city history. In the simulation, the driving forces in each phase are quantified and implemented, and the agents with different roles are assigned specific algorithms correspondingly. The simulated results are compared with the maps or remote sensing. Although the quality of the fine compa- - rison of the simulated results and the real data is affected by the poor real data quality because the real data of Qingdao City before 1949 is hard to obtain, the results show that the model has a good capability to simulate the city's development with the selected parameters of Qingdao city. The model has a good potential to analyze the driving forces related to city development and predict the city development under different scenes.