地理研究 ›› 2002, Vol. 21 ›› Issue (2): 155-162.doi: 10.11821/yj2002020003

• 论文 • 上一篇    下一篇

基于GIS的细胞自动机模型与人地关系的复杂性探讨

刘继生1, 陈彦光2   

  1. 1. 东北师范大学地理系,长春130024;
    2. 北京大学城市与环境学系,北京100871
  • 收稿日期:2001-07-25 修回日期:2001-10-12 出版日期:2002-04-15 发布日期:2002-04-15
  • 作者简介:刘继生(1955-),男,吉林省东丰县人,东北师范大学地理系教授,博士。主要从事地理分形和地理 系统的空间复杂性研究,发表学术论文60余篇。
  • 基金资助:

    国家自然科学基金资助项目(49771035)

GIS-based cellular automata models and researches on spatial complexity of man-land relationship

LIU Ji sheng1, CHEN Yan guang2   

  1. 1. Department of Geography, Northeast Normal University, Changchun 130024, China;
    2. Department of Urban and Environmental Sciences, Peking University, Beijing 100871, China
  • Received:2001-07-25 Revised:2001-10-12 Online:2002-04-15 Published:2002-04-15

摘要:

探讨人地非线性关系可以揭示地理系统空间复杂性的许多简单本质 ,具有重要的理论意义和实践价值。以地理信息系统 (GIS)为技术支持 ,建立以细胞自动机 (CA)为核心的综合集成模型 ,将成为人地关系复杂性研究的主要方向。建议在CA GIS集成技术的基础上 ,一方面引入人工神经网络、进化算法、多重智能体系 (MAS)以及后现代数学分析方法 ,另一方面面向具体研究对象开发包括自然、人文地理要素在内的集成区域动力学模型 ,据此发展智能化综合集成模型体系 ,形成人地关系复杂性的模拟研究实验室。本文提出了开发智能化CA GIS模拟方法的初步设想。

关键词: 人地关系, 空间复杂性, 细胞自动机, 分形理论, 人工神经网络, GIS

Abstract:

It is of significance to research into nonlinear relationships between human systems and physical environment in order to explore spatial complexity, but it is of difficulty to do so in the past because of lack of effective research means and methodologies. If we develop integrated models including cellular automata(CA)which are based on geographical information system (GIS), we will be able to solve many complex problems on man land relationship. We suggest that geographers should construct a kind of metasynthetical technical systems being composed of GIS as geographical databases, GIS based CA models, integrated regional dynamic models, and other related CA based models. For the sake of making the CA models based on GIS intelligent to certain extent, artificial neural networks, which can be used to distinguish and compare between the simulation patterns and digital maps concerned, and genetic algorithms, and even genetic programming, which can be used to design and train the weights of neural nets, should be adopted into the systems. Going a step further, the multiagent systems (MAS)and the postmodern mathematics of fractal and chaos, the latter of which take the form of softwares by being compiled into some programming, should be comprised in the systems. So geographers will have an artificially intellectualized meta synthetical systems consisting of a number of integrated models based on integrated models which make it easy and convenient to expore spatial complexity on man land relationship.

Key words: man land relationship, spatial complexity, cellular automata, fractal theory, artificial neural nets, geographical information system