| As the basis of human survival and development,arable land is an important resource and condition for ensuring human food security.The acceleration of urbanization has increased the degree of human intervention in the change of arable land use.Therefore,the agent of arable land use and its decision-making behavior should be taken seriously.This article mainly starts with the relationship between the theory of multi-agent and cultivated land use decision,and takes the Changsha-liuyang-pingjiang region of Hunan Province as a research area,the main subjects affecting cultivated land use are divided into three categories,namely government agents,enterprise agents,and farmer agents.Through field investigations,using SPSS and other statistical software for preliminary analysis,to clarify the behavior decision-making rules and characteristics of each agent,the relationship between agents and their interaction with the environment are sorted out,and the basic framework of agent decision-making rules and models is established.Through the Net Logo modeling platform,a multi-agent cultivated land use simulation model is constructed from the bottom to top to simulate and predict the change of cultivated land use in Changsha-liuyang-pingjiang area,and explore the micro-decision mechanism and action process of macro land use change,so as to provide reference basis for rational planning and land use,and promote the protection of cultivated land and sustainable development of land use.The main conclusions of this article are as follows:(1)The agent and its decision rules are defined.This paper defines three types of agents: government,enterprise,and farmer households.It discusses the choice tendencies,values,and interests of these three types of agents when they conduct land use behaviors,and analyzes the factors that affect the agents’ decision-making.The study concluded that the factors that affect the use of cultivated land by agents include natural,economic,and social factors,that is,natural factors such as elevation and slope;distance variables,density variables,and other factors;individual attributes,family attributes,economic attributes,and land attributes of farmers,by sorting and analyzing the questionnaire,we can know that the relationship between farmers’ decision-making tendency and their own attributes is closely related,then the most important factors were obtained by using Logistic regression model analysis in SPSS platform.By analyzing the interrelationships among the three types of agents,the behavior decision rules of various types of agents are designed.(2)A multi-agent decision-making model for cultivated land use was constructed.Based on the above research,a multi-agent model framework is constructed.By analyzing the factors that influence behavioral decisions and their interrelationships,and the interrelationship between the behavior of the subject and the environment,after designing the behavioral decision rules of agents and agents,A multi-agent interaction model was established from the bottom up,combining with the environmental characteristics of its range of activities,to construct a simulation model of cultivated land use change in the research area.(3)Model application based on NetLogo platform.Based on Net Logo software,taking the Changsha-liuyang-pingjiang region of Hunan Province as a research example and according to the established multi-agent model,the complex process of cultivated land use change is simulated and the cultivated land use change trend is analyzed and predicted.the model operation results were obtained and verified.The verification results showed that the accuracy of forest land,cultivated land,grassland,construction land,water area and unused land are 91.15%、88.09%、91.89%、83.12%、81.32%、74.03%,and the overall Kappa coefficient value is 0.8058,indicating that the simulation results of the model are good and the reliability is high,which is suitable for carrying out the decision-making simulation research on cultivated land use change.Therefore,the change of cultivated land in the study area in 2030 was simulated.The results showed that the land use changed roughly according to the expected trend,that is,the cultivated land continued to decrease,but the policy impact was controlled to a certain extent.The urban land continued to expand and occupied a large amount of forest land. |
