因素空间-未确知测度多重识别模型的构建及应用

Construction and Application of Factor Space-Unascertained Measure Multiple Identification Model

为了解决围岩稳定性影响因素众多且包含大量未确知信息的问题,并识别围岩系统危险因素,本文利用因素空间的概念提取功能及未确知数学理论处理不确定信息的优势,建立了因素空间-未确知测度模型.首先,利用定量因素和定性因素构建因素空间,将属性识别转化为概念提取;然后,以区间数表述识别规则,利用分段函数构造测度函数;接着,赋予因素变动的组合权重,通过因素合成实现因素空间降维;最后,采用置信度准则识别5个工作面的围岩稳定性,验证模型可靠性及适用性.研究结果表明:该模型识别结果与其他模型识别结果及现场勘查结果一致,可分析影响因素在识别过程中的作用,识别围岩系统危险因素,提出有针对性的整治措施,为类似工程提供值得借鉴的经验.

Many factors affect the surrounding rock stability,and some factors contain much uncertain information.To solve these problems and identify risk factors in surrounding rocks,a factor space-unascertained measure model was proposed through extracting concepts in factor spaces and the advantage of unascertained mathematics in processing uncertain information.Firstly,construct a factor space using quantitative and qualitative factors,and transform attribute identification into concept extraction.Secondly,the identification rules were expressed by interval numbers,and piecewise functions were used to construct measure functions.Then,combination weights were assigned to factors,and the dimensionality reduction of factor space was accomplished by factor synthesis.Finally,the surrounding rock stability was identified by the confidence criterion to verify the reliability and applicability of the model.The results show that the identification results of this model are consistent with other models and on-site surveys.It cannot only analyze the role of influencing factors in identifying,but also identify dangerous factors and propose targeted remediation measures.This model will provide valuable experience for attribute identification under the influence of multiple factors.