边坡稳定性的聚类未确知综合识别方法及应用

影响岩石边坡稳定性的因素众多且关系复杂，且存在大量未确知信息，很难用简单的方法进行分析判断，借鉴工程类比的思想，采用聚类与未确知测度相结合的方法进行研究。针对边坡工程问题环境的复杂性，以大量的历史数据为训练样本，通过动态聚类分析，求得其分类中心。针对大量的未确知信息，利用未确知测度方法对其进行评价，提出一种分析边坡稳定问题的新方法。研究表明，该算法可以对边坡的稳定状态进行预测，正确率在90%以上，为比较合理快速地分析边坡稳定分析方法提供了一条新的途径。

Study of rock slope stability based on clustering uncertained measurement complicated algorithm

For the big disaster of landslide, it is important to study the stability of slopes. For the influence factors of rock slopes are numerous and their relationships are very complicated; it can not be solved by traditional methods; so generally based on engineering analogy, the clustering methods are used widely. For the complicated environment influence of slopes, this clustering is a complicated uncertained measurement optimization problem, and can not be solved by traditional methods very well. So, here the clustering uncertained measurement complicated algorithm has been introduced into slope engineering field for the first time. Based on this method, one new method for study stability of rock slope is proposed. Based on analyzing the data of slope examples, using clustering uncertained measurement complicated algorithm, the stability of rock slopes can be estimated. The engineering application can prove that, this new algorithm can automatically sort the slope samples, and the validity is more than 90%, so it is a very practical and new method for slope stability analysis.