基于CR-BDI模型的农户作物种植行为模拟——以陕西省米脂县姜兴庄为例

运用模拟平台模拟农户土地利用行为,对揭示农户土地利用变化机制和调整土地利用决策具有重要意义。本文选择生态脆弱区一乡一业、一村一品典型村陕西省米脂县姜兴庄为例,以信念—愿望—意图模型(BDI)为基础,构建有限理性能力与资源(CR-BDI)模型。基于NetLogo平台对CR-BDI模型和传统BDI模型的模拟情况进行对比分析,结果表明：①CR-BDI模型更适于微观尺度农户土地利用决策研究。2014年该模型面积失误率为6.78%,比传统BDI模型低8.48%,空间准确率比传统BDI模型高10.10%,2015年CR-BDI模型和传统BDI模型的整体空间正确率分别为78.8%和69.2%;②CR指数有利于体现农户的有限理性,与实际种植决策较为符合;③利用NetLogo平台有利于直观再现农户土地利用行为,揭示土地利用变化的微观机理,可作为研究土地利用变化互动机理的良好平台。

Simulation of households' planting behavior based on a CR-BDI model: Case study of Jiangxingzhuang Village of Mizhi County in Shaanxi Province

At present, the research of agent-based modeling (ABM), which considers farming households' bounded rational behavior and uses simulation platform to simulate farming households' land use behavior, has become a focus of land use and land cover change simulation. Based on the beliefs-desires-intention (BDI) model, this study constructed a bounded rationality-resources (CR-BDI) model. A case study was conducted for Jiangxingzhuang Village of Mizhi County in Shaanxi Province. To verify the effectiveness of the CR-BDI model, a comparison between CR-BDI and BDI modeling results was performed. A number of observations and conclusions can be drawn: (1) Compared to the BDI model, a more realistic representation of farming households' decision-making process was provided by the CR-BDI model. The simulation accuracy of the CR-BDI model of 2014 is 93.22%, 8.48% higher than that of the traditional BDI model. The spatial distribution accuracy of the CR-BDI model output is 10.1% higher than that of the traditional BDI model. The simulated results of all crops (except scallion) in the CR-BDI model were of higher accuracy than in the BDI model. In 2015, the spatial accuracy of the CR-BDI model output is 78.8%, 9.6% higher than that of the traditional BDI model. (2) The indicator of CR is suitable for reflecting the bounded rationality of farming households, and matches the actual planting decision of farming households better. (3) The NetLogo platform is useful for representing farming households' land use behavior and provides a new way to research land use and land cover change.