基于CBDI的农户耕地撂荒行为模型构建及模拟研究——以陕西省米脂县冯阳坬村为例

目前,建立于复杂性科学基础上的多智能体模型（MAS）已成为模拟研究土地利用变化的重要方法。论文在MAS理论之上,构建一个基于Category-Belief-Desire-Intention（CBDI）决策结构的MAS模型,结合NetLogo软件平台,模拟分析农户Agent土地利用行为,并将结果与Belief-Desire-Intention（BDI）模型进行比较。论文以陕西省米脂县冯阳坬村为例进行研究,分析结果表明：1）通过对基于CBDI的MAS模型验证（模型精度为87.19%）,相比于传统的BDI决策模型精度有所提高;2）NetLogo平台能够展示农户决策的过程,直观地显示MAS模型的模拟结果;3）不同类型农户耕地撂荒的主要原因有所不同,该村种植户撂荒的主要因素为地理环境认知的差异和其他类农户的影响。有待进一步解决的问题：1）不同类Agent相互影响及其量化;2）高精度基础数据的获取和梯田DEM的引入。

Modeling and Simulating Land Abandonment Behavior ofFarmer Households Based on the CBDI

Recently, Multi-agent system (MAS), based on the theory of the complexity science, is one of the important methods in the study of LUCC. In this paper, a Category-Belief-Desire-Intention (CBDI) decision-making model based on MAS was constructed. The paper takes the Fengyanggua Village in Mizhi County, Shaanxi Province as an example. The land abandonment behaviors of farmer households were simulated and analyzed based on the NetLogo platform. The Belief-Desire-Intention (BDI) model was also applied. The results of CBDI model and BDI model are compared. Several conclusions were drawn: 1) Compared to the BDI model, the CBDI model (simulation accuracy is 87.19%) improves the accuracy of simulation. 2) The decision making process of farmer households can be shown on NetLogo platform, and the simulation results can give an explicit spatial result. 3) Though the reasons of farmers’ abandoning the land are different, the main reasons are the differences of geographical cognitive and the effects of other farmer households. In the future, some issues should be resolved: 1) the interaction among different kinds of farmer households should be focused and quantified; 2) highly accurate data resources are needed in order to improve the model precision, such as the terrace DEM.