基于多智能体系统模型的中小城镇用地扩展模拟

Simulation of small and medium-sized urban land expansion based on multi-agent system model

为了解决传统方法在复杂系统中不能妥善地处理人在参与城市扩展中的问题,借助遥感技术获取研究区多个时相的土地利用图,从城镇用地扩展的驱动力入手,构建了由政府Agent、开发Agent和利用Agent以及土地利用层组成的多智能体模型。对常用的多智能体仿真平台进行比较,选取了Repast平台作为仿真平台,实现了对现有城镇用地格局的模拟,以现有年份的城镇用地扩展现状与模型结果进行对比检验模拟的精度,在精度可行的情况下,通过模型分情景对未来城镇空间的格局做出了预测,为政府决策提供参考。

In order to solve the problem that traditional methods can’t be used to manage people’s participation in urban expansion in complex systems, land use maps of multiple phases in the study area are obtained by remote sensing technology. Starting from the driving force of urban land expansion, a multi-agent model composed of government agent,development agent, user agent and land use layer is constructed. Compared with the common multi-agent simulation platform, repast is selected as a simulation platform, the platform realizes the simulation of the existing urban land use pattern. The accuracy of the simulation is tested by comparing the current situation of urban land expansion in the existing years with the model results. When the accuracy is feasible, the future urban spatial pattern is predicted through the model scenarios, which provides a reference for government decision-making.