论文部分内容阅读
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Net Logo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)Net Logo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。
At present, the multi-agent model (MAS) based on complexity science has become an important method to simulate the change of land use. Based on MAS theory, this paper builds a MAS model based on the Category-Belief-Desire-Intention (CBDI) decision structure and simulates the land use behavior of the farmer Agent by Net Logo software platform. The results are compared with Belief-Desire-Intention BDI) model for comparison. The results of the study show that: 1) Through the CBDI-based MAS model validation (model accuracy of 87.19%), compared with the traditional BDI decision model accuracy has been increased ; 2) The Net Logo platform can demonstrate the decision-making process of farmer households and intuitively display the simulation results of MAS model; 3) The main reasons for the shortage of cultivated land by different types of farmers are different. The main factors of farmers’ abandonment in this village are geographical cognition Differences and the impact of other types of peasant households. Problems to be further solved: 1) the interaction between different agents and their quantification; 2) the acquisition of high-precision basic data and the introduction of terraced DEM.