基于GECM与CA+ANN模型的土地资源优化配置与模拟

土地资源优化配置是土地科学研究的热点。论文在前人研究成果的基础上,以多目标动态规划方法中的逐步宽容约束(GECM)方法为基础,以社会、经济和生态效益为系统目标,设置8个方面的GECM约束条件,对广州市花都区2020年土地利用结构进行优化配置研究,并将土地利用结构优化配置结果作为约束条件,结合人工神经网络与元胞自动机(ANN+CA)模型对研究区土地利用空间布局进行模拟。研究结果表明:①GECM是一种较好的土地利用结构优化方法,花都区2008年建设用地规模(23 097 hm²)接近于2020年的合理用地水平(28 800 hm²);②ANN+CA可以实现土地资源的空间配置与模拟,通过GECM与ANN+CA的耦合进行土地资源优化配置的结果比较符合花都区的实际情况,模拟的方案更具合理性,能实现区域土地资源优化配置的时间与空间的统一。

Simulation on Optimized Allocation of Land Resource Based on GECM and CA+ANN Model

Optimal allocation of land resources is becoming a hot spot in land scientific research.Based on the results of previous studies,the authors build up the model of the Gradual Easy Constraint Method(GECM) with the multi-objective dynamic programming method.This model aims on social,economic and ecological benefits,and sets up eight aspects as GECM constraints to research on the optimized allocation of land use structure in Huadu District in 2020.Then through the results of structural optimization as constraints,we can use the artificial neural networks and cellular automata(ANN+CA) model to simulate and predict the change of land use from 2008 to 2020.The results indicate that: 1) GECM is a good method on land use structure optimization research,the built-up area in 2008(23097 hm2) will reach the rational level in 2020(28800 hm2),and this district must select the direction of land intensive use in the future years.2) CA+ANN can achieve the target of spatial allocation and simulation on land resources.Coupling with GECM and ANN+CA models,the research results are more suitable and reasonable to the actual situation in Huadu District.It also can achieve the unity of time and space in optimal allocation of regional land use.