基于元胞自动机的滑坡时空预测模型研究

项目名称： 基于元胞自动机的滑坡时空预测模型研究

项目编号： No.41301588

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 黎华

作者单位： 武汉理工大学

项目金额： 25万元

中文摘要： 滑坡系统是开放的复杂巨系统，传统的数理方法无法对滑坡时空特性做出比较准确的模拟和预测。元胞自动机是时空离散的动力学模型，特别适合空间复杂系统的动态模拟和预测。将元胞自动机模型和地理信息系统集成，构建集滑坡稳定性分析模型和滑动预测模型为一体的滑坡时空预测模型，为滑坡时空动态模型提供新的研究思路。采用基于案例推理的方式建立案例库，然后利用元胞自动机自动从案例库中智能获取转换规则，结合滑坡元胞邻域的转换概率，最终预测各个滑坡元胞的转换概率，据此分析出滑坡危险性的区划范围。在分析不稳定性区划的基础上，进行滑坡的滑动过程动态预测和模拟。采用坡度、坡向、地形特征参数来简化滑坡流动的影响因子，将逻辑回归模型运用到元胞自动机模型中来获取转换规则，预测滑坡滑动的空间范围，最终成果在地理信息系统中图形化显示。滑坡时空预测模型的建立，为滑坡的系统研究提供新的解决方案，为滑坡的综合治理提供可视化的辅助决策。

中文关键词： 元胞自动机；滑坡；逻辑回归；神经网络；

英文摘要： Landslide is an open complex system, the traditional mathematical methods fail to make more accurate prediction and simulation of landslide model.Cellular automata is spatial-temporal discrete dynamics model, which is suitable for simulating and predicting the spatial dynamic system.The systematic and integrative model is constructed to analyse the landslide stability and predict the landslide sliding track based on the cellular automaton model and geographical information system.We establish case database based on Case-Base Reasoning method, and then use the cellular automata to achieve the translation rules from the case database, combined with transition probability of the landslide cellular neighborhood, ultimately determine the transition probability of the every landslide cellular and predict the landslide hazard zoning range.According to the zoning range, we can model and predict the landslide sliding process and track.We adopt the slope,aspect,terrain feature parameters to simplify the influence factors of landslide flow.Then the logistic regression model was applied to cellular automata model to obtain the transformation rules.The landslide sliding range was obtained and was displayed in the geographic information system.The dynamic model provides new solution for the landslide research and provides

英文关键词： cellular automata；landslide；logistic regression；neural network；