中文摘要：

针对交通安全风险评估过程中众多指标难以定量判断,而往往采用专家经验进行评估的情形,为了更好地将专家的定性描述转变为定量分析,提出了一种基于信息熵与D-S证据理论的评估方法。首先利用德尔菲法与模糊分析法相结合的方法进行专家排序和基于熵理论的＂盲度＂分析,构建了结构熵权法来确定评估指标的权重;然后利用信息熵确定的权重,引入模糊子集来描述证据空间中的事件,从而建立了一种基于模糊子集的D-S证据理论合成方法。最后,以西安市轨道交通运营安全风险为例进行了安全风险评估。实例分析表明,用该理论方法计算得到的交通安全风险评估结果与实际基本吻合。

英文摘要：

This paper is aimed at introducing a new approach to as sessing the traffic safety risk. As is known, it is of great difficulty to quantify the number of indexes in the process of traffic safety risk as sessment, which may account for the need to ask for the experts to apply their wealth of experience. In spite of this, the paper intends to integrate the quantitative analysis and the qualitative analysis into an improved risk assessment model based on the information entropy and D -S evidence theory in an integrated manner. In doing so, first of all, we have attempted to bring about Delphi method and Fuzzy method to make the output source as well as ＂the blind degree＂ anal ysis firmly founded on the entropy theory so as to reflect the expert opinion and attach more attention to the factors of uncertainty or unexpectedness. What is more, the structure entropy method should be resorted to for more comprehensive account of the index weights. Sec ondly, we have been trying to establish a fuzzy subset for depicting the events to be assessed in space and use the information entropy and D- S evidence theory. Furthermore, we have made up for the part of the traditional D-S evidence theory in an attempt to reduce the fac tors of uncertainty in determining the assessed results. And, the last of all, we have accomplished Xi＇an rail operation as an example to verify the method, in which we have selected a total 16 indexes from human factors, equipment and facilities factors, environmental factors and management factors in the traffic safety risk analysis of Xi＇ an rail, whose process has been given in a detailed way in the paper. The assessment results of our sample study shows that the method is not only of great reliability and warranty in determining weights, but also convenient in reasoning process and of confident processing ca pacity in vague and uncertain information. Thus, the results we have gained from the traffic safety risk assessment proves well in accord with the actual situation for appreciation.