基于改进D-S的高速公路改扩建交通安全风险评估模型

Traffic Safety Risk Assessment Model for Expressway Reconstruction and Expansion Based on Improved D-S Evidence Theory

为提高高速公路改扩建工程交通安全风险评估结果的确定性和准确性,建立了基于改进D-S证据理论的相关风险评估模型。首先建立包含24个影响因素的三层级评估指标体系;然后利用云模型(Cloud Model,CM)求出定性指标的基本信度赋值(Basic Probability Assignment,BPA),利用高斯隶属度函数求出定量指标BPA;接着,通过层次分析法确定各评估指标的权重,进而对各指标BPA进行加权;利用D-S证据理论融合加权后的BPA,归一化处理后得到改扩建工程交通安全风险状态评估结果。最后,为验证模型的准确性,选取沪陕高速公路平潮至广陵段高速公路改扩建工程作为实例进行交通安全风险评估。评估结果显示,实例工程的低风险水平隶属度最大,为0.6615,表明该实例总体处于低风险水平,与现有资料和现实情况吻合。同时发现,基于CM、AHP及D-S证据理论的评估模型对各评估指标进行量化、加权、融合后所得到的风险等级隶属度和不确定性有所区别,能更均衡地表示风险的隶属度,量化后的安全风险状态评估结果具有更好的准确性,解决了指标体系中模糊定性指标难以量化表征及指标差异化权重赋值的难题。

In order to improve the certainty and accuracy of traffic safety risk assessment results of expressway reconstruction and expansion project,a related risk assessment model based on improved D-S evidence theory was established.Firstly,a three-level evaluation index system including 24 influencing factors was established.Secondly,the Basic Probability Assignment(BPA)of the qualitative index was calculated using Cloud Model(CM),and the BPA of quantitative index was calculated using Gaussian membership function.The weight of each evaluation index was determined by the Analytic Hierarchy Process(AHP),and BPA was weighted for each index.D-S evidence theory was used to integrate the weighted BPA,and the normalized treatment was used to obtain the traffic safety risk state assessment results of the reconstruction and expansion project.Finally,in order to verify the accuracy of the model,the highway reconstruction and expansion project of Pingchao to Guangling section of Shanghai-Shaanxi Expressway was selected as an example to evaluate the traffic safety risk.The evaluation results showed that the membership degree of low risk level was 0.6615 and was the largest,indicating that the case was at the low risk level,which is consistent with the existing information and reality.The result also showed that the risk level membership and uncertainty obtained by the evaluation model based on CM,AHP and D-S evidence theory after quantization,weighting and fusion of each evaluation index were different,which could represent the risk membership more balanced,and the security risk status evaluation results after quantization had better accuracy.The difficulty of quantifying the fuzzy qualitative index and assigning different weights to the index system was solved.