Simulation And Decision Mechanism On Land Use Change Based On Households' View

As one of the core of global environmental change research, land use change was being concerned by scholars in different fields. Use of MAS model to simulate the process of micro-individual's decision-making and to explore the mechanisms to adoption of individual to the natural and socio-economic environmental change, MAS model will be one of the important trend to simulate the households land use behavior.In macro-scale, summary the land use change in spatial and temporal of YinZhou in MiZhi County, and analysis of land use change under the impacts of policies and the mechanism of the overall households' responding to land use change. In micro-scale, use multi-agent model, the use of BDI (Belief-Desire-Intension) model and the reinforcement learning model construction of the model of households land use behavior, from the individual households and the overall household's level of behavior separately, to explore the changes in land use decision-making mechanism. Moreover, the models pass the inspection and verification. The conclusions of this paper were as follows:(1) During 1986 to 1996, arable land had the increasing trend, while the woodland and grassland in the decrease trend; with the returning farmland to forests and grasslands of policy. During 1996 to 2006, the larger decrease in arable land, woodland and grassland area increased.The policy of returning farmland to forest can be used to explain the land use change occurred in the beam, which slop is greater than 25 degree. However, the arable land beside WuDing river were affected on the household's land use behavior.(2) BDI model based on individual household's behavior. Explore the affect of market to households, and the interaction between same type households. It mainly used continue function other than 0 or 1 value to reflect the land use change under the affected of two factor in above, and it suitable for the reality.(3) Reinforcement learning model based on the overall household's behavior. Study on the affected of decision-making of household's land use by the factors of policy, market, households themselves, the interaction in households and soil fertility degree. Reinforcement learning model express the probability of household planted certain crop with probability of quantitative methods by the continue function. Moreover, it explained the mechanism of decision-making of household's transformation to space clearly by using the correlation between household possessing land and probability number.(4) BDI model and Reinforcement learning model had the different in the factors of research on subject, consideration about factors, the weight of factors impact on households, the result of household's decision-making to the space. The result showed that there had little changes in the overall crop area, but each farmer's crop type had lager change.