【作者】 胡海龙；

【导师】 曾永年；

【作者基本信息】 中南大学 ， 地图制图学与地理信息工程， 2011， 硕士

【摘要】 快速的城市化进程严重的影响着城市生态环境质量,为了缓解城市生态环境压力急需合理配置城市生态用地。传统的空间布局选址模型没有充分考虑政府政策、方针的影响,其选址结果的可操作性不够。多智能体能够通过模拟现实世界中利益方的博弈、协商过程,做出合理的决策,大大增加了选址方案的可行性。本文根据生态安全格局理论和城市可持续发展观念,提出了多智能体与蚁群算法结合选址模型,并应用于长沙市城市生态用地的选址,为城市生态用地选址提供了新的方法,并为长沙市生态选址提供了决策参考。本文的主要研究内容和结论如下：(1)设计了多智能体与蚁群算法结合选址模型。在多智能体中构建了政府、市民和开发商三类Agent,政府Agent通过生态安全格局分析协调经济发展与生态环境的关系；市民Agent中通过引入蚁群算法提高模型的运行效率。(2)利用ANN_CA模型,动态模拟了长沙市城区历史和未来的城市空间形态。结果表明选取的ANN_CA模型能够比较好的模拟、预测长沙市城区的空间形态,预测精度达到了74.85%。(3)利用多智能体与蚁群算法结合选址模型和传统的简单选址方法,在参考2006年长沙市城市空间形态下,做了选址对比试验。得出的结论是：多智能体选址模型选址结果比传统的简单选址方法更科学、合理,蚁群算法的引入使模型运行时间由简单选址方法的51.29S减少为22.37S,运行效率有了显著提升(4)基于2006、2010和2020年长沙市城市空间形态,做了三组选址对比试验。试验结果表明,在考虑城市未来格局情况下决策城市生态用地选址,能够避免城市建设与环境保护的冲突。更多还原

【Abstract】 Rapid urbanization impacts on the quality of urban ecological environment seriously. In order to alleviate the urban ecological environment pressure, we need to allocate the urban ecological land reasonably. The traditional method does not consider the influence of government policy, so the results of spatial layout of urban ecological land are not enough operable. By simulating the game and consultative process of the interested party in the real world, the multi-agent system makes reasonable decision which increases the feasibility of site selecting significantly.This paper focuses on a site selecting model integrating multi-agent system and ant colony algorithm from the view of ecological security and urban sustainable development. The site selecting model is used for the site selecting of ecological land in Changsha. This model provides a new method for the site selecting of urban ecological land and the policy-making reference for Changsha ecological land selecting. The main research content and conclusion are as follows:(1) Designing a site selecting model integrating multi-agent system and ant colony algorithm.Constructing three kinds of agent that are government, resident and developer agent respectively.Government agent coordinates the relationship of economic development and ecological environment by analyzing ecological security pattern. Resident agent enhances the model operating efficiency through introducing ant colony algorithm.(2) Simulating urban space shape of Changsha city history and future dynamically by using ANN_CA model. The result indicated that the ANN_CA model can better simulate and forecast the spatial shape in Changsha city. The precision of prediction achieved 74.85%.(3) Through using the site selecting model integrating multi-agent system and ant colony algorithm and the traditional simple site selecting model, a comparative trial is performed in refers to 2006 Changsha city spatial shape. The conclusion is:the site selecting model results with multi-agent system model is more scientific and reasonable than traditional simple site selecting method. The model introduced ant colony algorithm makes the running time from 51.29 S to 22.37S.(4) Based on 2006,2010,2020 Changsha city spatial shape, three groups of site selecting comparative trial are performed. The test result indicates that the policy-making city ecological land selecting under the circumstances of considering the city pattern in the future can avoid the conflict of urban construction and environmental protection.更多还原