多智能体与蚁群算法结合选址模型：长沙市生态用地选址

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,特别是空间布局优化。传统的空间布局选址方法没有考虑政府政策、方针的影响,其选址结果的科学性、可行性不够。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新模型。新模型对多智能体在生态用地选址中的应用以及和蚁群算法的结合上进行了首次尝试。为了验证模型,以国家确立的"资源节约型、环境友好型"两型社会建设实验区的重要组成部分长沙市城区为研究区域,在调研和收集整理的GIS数据以及经济社会统计数据的基础上,利用新模型解决了长沙市城区新增一块生态用地的选址问题。实验结果表明,提出的结合模型能够比较好的解决城市生态用地选址问题,比传统的简单选址方法更科学、合理。同时,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,运行效率有了很大的提高。

Integration of a Site Selection Model with the Multi-Agent System and the Ant Colony Algorithm and Its Application to Changsha

Rapid urbanization significantly impacts the quality of the urban environment. In order to alleviate the urban environment pressure, urban ecological land should be reasonably allocated. In particular, the spatial layout of the urban ecological land should be optimized, which can provide a new method for planning decision-making. Traditional methods do not consider the influence of government policy; therefore results of spatial layout of urban ecological land seem to be far from satisfactory and feasible. This paper focuses on a site selection model integrating the multi-agent system and the ant colony algorithm from the point view of ecological security and urban sustainable development. Application of the multi-agent system in the layout of ecological land and the method of integrating the multi-agent system with the ant colony algorithm were first carried out in this model, in which urban ecological land was reasonably selected by jointly taking the influence of governmental policies into account and coordinating the relationship between urban economical development and environmental protection. Three types of agent were designed in the model, i.e., government agent, resident agent, and developer agent. The government agent is the most important agent, which plays a major role in the site selection of urban ecological land. The three types of agent finally determine the location of urban ecological land through dice game and negotiation. The ant colony algorithm was utilized to improve the efficiency of this model. The ant colony algorithm can help the resident agent rapidly identify the satisfactory site of urban ecological land. To verify the model, an application of selecting a new ecological land site in Changsha City, Hunan Province, was carried out based on the model. Changsha City is the important component of the test area where resource saving and environment-friendly society is being constructed. Data used in this study include GIS data and basic economical and social statistic data. Results show that the model can provide solutions for the site selection of urban ecological land. Compared with the method of brute-force search, selection results from the integration model seem more reasonable. The simple brute-force search located an area close to Orange Island Bridge which has been already a critical development zone of Changsha City. Meanwhile, the proposed approach can improve computational efficiency to a certain degree.