Validation data sheet.

Collapse risk assessment is an important basis for the prevention and control of geological disasters in mountainous areas. The existing research on collapse hazard is less, and there is still no further advancement in the evaluation of collapse hazard for the traditional indicator assignment method for the diversification of the assignment results of the indicators and the comprehensive evaluation method that cannot consider the ambiguity and randomness of the indicator data at the same time. In this paper, we utilize the respective advantages of the linear programming theory and the cloud model from the prevention and control point of view, and evaluate the collapse samples. Firstly, the weight interval of evaluation index is determined by improved analytic hierarchy process, entropy weight method and coefficient of variation method. Secondly, the linear programming algorithm is used to select the specific weight of each collapse sample when the risk is the largest in the interval. Finally, a comprehensive evaluation model of cloud model is constructed to determine the risk level of collapse. In this paper, 20 collapse samples counted by predecessors in G4217 Wenchuan-Lixian section are taken as research cases. The evaluation results of 20 collapse samples are compared with other evaluation methods and field survey conditions to prove the reliability and rationality of the method. The evaluation results show that 13 of the 20 collapse samples are extremely dangerous, 2 are highly dangerous, 4 are moderately dangerous, and 1 is lowly dangerous. Among them, the extremely dangerous collapse samples account for 65% of the total number of collapses. Compared with other methods, this method is more in line with the actual situation.崩塌风险评估是山区地质灾害防治的重要依据。现有崩塌灾害相关研究相对匮乏，传统指标赋值法存在指标赋值结果离散化的问题，且综合评价方法无法同时兼顾指标数据的模糊性与随机性，因此崩塌灾害评价领域仍未取得进一步进展。本文从防灾减灾视角出发，结合线性规划理论（linear programming theory）与云模型（cloud model）的各自优势，对崩塌样本展开评价。首先，采用改进层次分析法（improved analytic hierarchy process）、熵权法（entropy weight method）与变异系数法（coefficient of variation method）确定评价指标的权重区间。其次，借助线性规划算法（linear programming algorithm），在权重区间内求解各崩塌样本对应的风险最大化最优权重。最后，构建云模型综合评价模型，以确定崩塌灾害的风险等级。本文以G4217线汶川-理县段前人统计的20组崩塌样本作为研究案例，将20组崩塌样本的评价结果与其他评价方法及野外实地勘测结果进行对比，以验证本文所提方法的可靠性与合理性。评价结果显示，20组崩塌样本中，13组为极高危险等级，2组为高危险等级，4组为中危险等级，1组为低危险等级。其中极高危险等级崩塌样本占总样本量的65%。相较于其他评价方法，本文方法更贴合实际工况。