細胞自動機(cellular automata，CA)是由資訊科學領域的Turing Machine所衍生出來的空間演化模式。CA的組成要素包括：網格空間、演化規則和離散的演化時段，在某一特定時間上，網格空間上每一個單元有明確的狀態，而這個狀態依據演化規則，伴隨離散的時間間隔而做演化。CA是一種機制、一種模式、一種觀點，在地理現象的演化模擬上有許多應用潛力。CA的網格空間及演化規則，和網格式地理資訊系統有許多共通之處，因此可以和地理資訊系統密切結合。本研究屬於探索性研究，針對CA和地理資訊系統的結合和潛在的問題進行測試研究。本研究利用GRASS地理資訊系統作為開發環境，成功地發展出CA指令，並得以和現有的數值地理資料庫結合。本研究進一步利用隨機產生的樣本資料來測試CA模式的解析度和演化規則等問題。研究成果除了確立CA在技術面的可行性外，也凸顯CA在實際應用上所需要面臨的相關問題，提供未來研究參考。

Cellular automata (CA) is a spatial model derived from the Turing Machine developed by the computer science. A CA is composed of three major components: a cellular space, a set of transition rules, and a time space in a discrete sequence. Each cell remains in a certain state for each discrete time, which is subject to change based on the states of its neighboring cells according to s set of transition rules. CA stands for a mechanism, a model, a perspective, and a tool. The cellular space and the transition process of CA are very similar to those of a raster GIS, which suggests a great potential for the integration of these two subjects. This research takes an exploratory approach to study the integration of CA with GIS. A CA function has been developed within the GRASS GIS software. Furthermore, randomly created data are used to test the factors of spatial resolution, transition rules, and cycles on the results of transition. This research demonstrates a promised integration of CA and GIS. On the other hand, it also reveals the factors that undermine the reliability of CA modeling which suggests area for future research.