帮助 本站公告
您现在所在的位置:网站首页 > 知识中心 > 文献详情
文献详细Journal detailed

基于多智能体的生态用地格局演化研究
Research of Ecological Land Pattern Evolution Based on Multi-agent

导  师: 杨木壮

学科专业: 120405

授予学位: 硕士

作  者: ;

机构地区: 广州大学

摘  要: 改革开放以来,我国经济迅猛发展,一些发展较快的地区土地利用发生了剧烈变化,由此带来的问题是大量具有维持区域生态平衡和保证生态环境质量的生态用地因为城市的扩张而不断退化。为此,如何遏制生态用地被挤占的态势,加强生态用地的保护刻不容缓。本文采用元胞自动机与多智能体相结合的方法研究肇庆市鼎湖区近20年来的生态用地格局演变规律,并模拟预测未来10年的演变情况。元胞自动机模型通过Logistic回归模块实现,利用1991年和1999年的历史数据挖掘几个空间变量和因变量之间的关系,从而确定logistic方程中各自变量的系数;多智能体模块部分考虑了政府土地利用总体规划和居民对城市绿地的需求。为了使模型更符合现实,还加入了约束模块和随机模块,约束条件是对一些限制区域进行识别,随机因素是为了体现土地利用变化过程中的不确定性因素。在上述工作基础上,根据研究需要和实际情况,对生态用地进行重新界定,建立符合本研究的生态用地分类系统。然后建立生态用地解译标志,利用ENVI4.3软件对肇庆市鼎湖区的TM遥感影像进行解译分类,结合当地土地利用数据和社会经济统计数据,运用ArcGIS10中ModelBuilder模块综合构建模型。该模型的最终确定还需要根据肇庆市鼎湖区的实际情况,反复试验,并最终获取最优的模型参数组合。经过精度评定符合要求之后利用模型模拟鼎湖区2020年的生态用地演变状况,并对其进行分析。模拟结果表明,在鼎湖区的生态用地格局演变过程中,其生态用地转为非生态用地多数是集中在中东部片区,这一方面是因为自然条件的限制,另一方面归功于政府的规划调整,使得大部分的生态用地保持连片。从生态用地的格局上看,1991年到2009年生态用地整体格局未发生大的变化,尽管如此,其局部数量的变化还是很明显的。通过对2020年的模拟预测可知,如果按照现有规划发展,其生态用地转变大多集中在中东部,并且在转变数量上比之前两期都比较明显,说明其用地压力很大。 Since the reform and opening up, China's economy develop rapidly. The land use insome regions where developed rapidly changed dramatically. It makes a lot of ecological landwhich maintains the regional ecological balance and ensures the quality of the ecologicalenvironment degenerate constantly. So, how to curb the trend of ecological land diverted andstrengthen the protection of ecological land is without delay. This text research the evolutionof ecological land pattern of the past20years in Dinghu, and simulate the situation ofecological land pattern evolution in the next10years using the model combining cellularautomata and multi-agent. Logistic regression model module is a function of cellularautomata, excavate the relationship between the dependent variable and several spatialvariables by historical data from1991and1999, and then determine the coefficient of theindependent variables in the logistic equation. The multi-agent module takes the governmentland use plan and the residents needs to urban green space into account. Also joined theconstraint modules and random modules in order to make the model more realistic, theconstraint can identify a number of restricted areas, the random factors can reflect theuncertainties in land-use change process. This article integrate the four parts, first,re-definition ecological land,determine the ecological land classification system compliancewith this study, then create the interpret signs of ecological land, classify the TM images ofDinghu with ENVI4.3software, build the model using ArcGIS ModelBuilder modulecombined with local land use data and socio-economic statistical data. The finaldetermination of the model need to test repeatedly, ultimately obtain the optimal combinationof model parameters. After a precision assessment make the model meet the requirements ofecological land evolution situation, Simulated and analyze the process of ecological landpattern evolution of Dinghu in2020. The results show that its ecological land converted tonon-ecological land most concentrated in the east-central. On the one hand, it is because ofthe limitations by natural conditions, on the other hand it due to the government's planningand adjustment, making the most of the ecological land to maintain contiguous. The overallpattern change of ecological land from1991to2009is not evident. Despite, its local changesin the number is still very evident. Through the simulation and prediction for2020,we knowif according to planning and development, ecological land use change are mostlyconcentrated in east-central and the number of change are more than the previous two period.The land pressure highlighted.

关 键 词: 生态用地 元胞自动机 多智能体 鼎湖区

分 类 号: [X24 F301]

领  域: [环境科学与工程] [经济管理]

相关作者

作者 魏兴琥
作者 温建明
作者 孙有发
作者 郭承军
作者 龚建周

相关机构对象

机构 中山大学地理科学与规划学院
机构 广东工业大学
机构 广州番禺职业技术学院
机构 中山大学
机构 中山大学管理学院

相关领域作者

作者 覃剑
作者 王逢文
作者 崔书琴
作者 吴大磊
作者 杨荷卿