基于关系数据库的CA模型扩展和时空模拟实验研究

The Study on Extended Cellular Automata Based on Relational Database and the Spatio-temporal Simulating

具备时空计算特征的元胞自动机 (CA)模型与地理信息系统 (GIS)集成在地理过程模拟方面具有很大的优势 ,但标准CA在邻居规则等方面的严格定义制约了CA对真实世界的模拟能力。在分析了元胞空间关系和元胞邻居描述的基础上 ,该文提出元胞邻居不仅存在几何空间上的邻接形式 ,还存在空间上不邻接但属性上相关的邻居形式。传统思路的模型扩展很难完全解决CA的局限性 ,基于关系数据库来定义和选择元胞邻居是一种有益的理论尝试。以时空为载体的文化革新扩散系统具有多空间扩散形式和遗传特征 ,相对于标准CA基于关系数据库的CA扩展模型可以更真实地对其进行时空模拟 ,笔者以服装传播为典型案例做实验研究 ,取得了比较满意的实验结果。认为基于关系数据库的CA扩展模型比较适用于具有多空间扩散形式的复杂系统的过程模拟。

The integrity of GIS and CA with the ability to calculate time and space have the great advantage of simulating geographical process. But standard CA has some restriction in cellular shape and neighborhood and neighbor rules, which restricts the CA's ability to simulate real world of culture innovation diffusion. The paper puts forward that cellular neighbor's type has not only adjacency of geometrical space but also neighbor of non-space because of relation in attribution based on cellular space relation and neighbor description. It is impossible to settle CA's restriction completely based on traditional idea, and it is significative academic experiment to definite and select cellular neighbor based on relational database. The culture innovation diffusion system has a great deal of forms of space diffusion and inherited character, so the extended CA based on RD can simulate it more effectively than normal CA do. The paper carried through an experiment on fashionable dress's diffusion, and acquired satisfied experimental result. In conclusion, the extended CA is an effective means in simulating complex system with multiplicative forms of space diffusion.