摘要
邻域因子是城市动态变化的重要驱动因子,该文提出了动态邻域约束思想,在借助蚁群优化(ACO)算法提取城市用地转换规则的基础上,结合元胞自动机(CA)模型构建了基于动态邻域约束的ACO-CA城市动态模拟模型,实现了对城市用地的动态模拟,并以重庆市沙坪坝区为例,设计不同方案验证了该模型的有效性。研究结果显示:当采用动态邻域方案时,总的Kappa系数比静态邻域方案高1.70%;城市用地的Kappa系数比采用静态邻域方案时的模拟精度高出6.37%。研究结果表明:构建的基于动态邻域思想的ACO-CA模型能够有效模拟城市用地的动态变化;采用动态邻域约束条件时,尽管算法的复杂度有所增加,但与静态邻域约束方案相比,城市用地模拟精度要高,且更符合城市发展演变规律。
Neighborhood factor is one of the main driving factors for the dynamic change of urban land use.A ACO-CA model based on dynamic neighborhood constraint was constructed in this paper,which combined ant colony optimization algorithm（ACO）with cellular automata（CA）model.ACO here was used to extract mutual conversion rules between land use types.Taking Shapingba District in Chongqing City as an example,two schemes were designed：the scheme of dynamic neighborhood factors and the scheme of static neighborhood factors.The results showed that total Kappa index is 82.16% when the dynamic neighborhood factors is considered,which the simulation accuracy is 1.70% higher than static neighborhood factor;otherwise the KAPPA index of urban land use is 6.37% higher than that of the static neighborhood factor scheme.The research shows that ACO-CA model based on dynamic neighborhood theory can effectively simulate the dynamic change of urban land use.When the dynamic neighborhood constraints are used,although the complexity of the algorithm is increased,the simulation accuracy of urban land use is higher than that of the static neighborhood constraints,and it is more in line with the law of the evolution of urban development.
出处
《地理与地理信息科学》
CSCD
北大核心
2016年第3期74-80,88,F0003,共9页
Geography and Geo-Information Science
基金
重庆市教育科学技术研究项目(KJ1400420)
重庆市应用开发计划重点项目(cstc2014yykfB30003)
关键词
蚁群优化算法
动态邻域
元胞自动机
转换规则
城市动态模拟
ant colony optimization algorithm
dynamic neighborhood
cellular automata
conversion rules
urban dynamic simulation
