基于耦合地理模拟优化系统GeoSOS的农田保护区预警

依据农田保护区规划与城市发展模拟相耦合的新思路，开展农田保护区预警研究。以广东番禺为实验区，首先利用基于多智能体的空间优化配置模型(AgentLA) 自动生成农田保护区范围。AgentLA模型能避免产生分散、破碎的空间格局，有利于保护区的管理和维护。通过地理模拟优化系统GeoSOS中的神经网络CA模型对研究区2025 年的城市发展格局进行三个情景的模拟：低速增长情景、基准情景和高速增长情景。最后利用GIS空间分析方法将预测结果与农田保护区相结合，提取农田保护与城市扩张产生冲突的区域。总体上，冲突的严重程度随着空间上城市扩张强度的增大而加剧，高速增长情景中冲突区域的面积较大。冲突区域的存在一方面说明了农田保护区需要相应的法律、法规支持，否则可能会因城市扩张而被侵占；另一方面也反映了农田保护与地区发展之间的冲突。因此，可以根据所提取的冲突区域面积大小、空间位置等特征，采取某种形式补偿，以此取得农田保护与地方利益之间的平衡。

Coupling Geosimulation and Optimization (GeoSOS) for Zoning and Alerting of Agricultural Conservation Areas

Rapid urban expansion in China has caused the significant loss of agricultural land. Specific efforts are made by the country to protect its scarce and valuable agricultural land. The State Council promulgated the Ordinance for the Protection of Primary Agricultural Land in 1994 to promote the zoning of agricultural conservation areas. However, the agricultural land still underwent a rapid loss due to the urgent demand of local economic development. The conflict between protection and local development calls for solution. This paper provides a new thought of integrating urban development into the zoning of agricultural conservation areas, and Panyu district in Guangdong province is selected for a case study. The proposed Agent-based Land Allocation (AgentLA) model is used to automatically generate the spatial distribution of agricultural conservation areas. Compared to conventional methods, the resulted pattern from AgentLA is more compact that favors routing maintenance of the conservation areas. Moreover, the future development of the region is predicted using Geographical Simulation and Optimization System (GeoSOS), a platform for geosimulation. Finally, GIS provides the function to combine the zoning result and the predicted development pattern in order to recognize the spatial distribution of the conflict areas.