作　　者： (梁心雨）; (郭彤）; (孟祥海）;

机构地区： 哈尔滨工业大学交通科学与工程学院,哈尔滨150090

出　　处： 《交通信息与安全》 2017年第4期20-28,35,共10页

摘　　要： 对国内宏观道路交通安全模糊综合评价方法进行改进,重点是改进了模糊层次分析法中由两两比较判断矩阵确定权重的方法,即提出了基于三角模糊数的权重计算方法。基于交通事故绝对指标及人、车、路等关联因素指标,确定出了9个评价指标,构建了宏观交通安全评价体系。引入三角模糊数来描述各指标权重。与传统模糊层次分析法中用1~9个整数及其倒数确定权重的方法相比,三角模糊数表达的是一个区间的概念,给定可能性区间的上限、下限及取值可能性最大的中限值,从而得到整个区间范围内不同参数的可能性,更好的保留了模糊评价过程中的有效信息,评价结果更为合理。给出了基于隶属函数的评判矩阵确定方法和最终的模糊综合评价方法。并以国内31个省、自治区、直辖市近10年的统计数据为例,进行方法应用并对宏观交通安全状况进行分析。分析评价结果可知,基于三角模糊数权重算法的交通安全评价方法更为准确合理,从2005年至2014年的10年间,安全状况为差的地区比例从87%降低至6.5%,整体的交通安全状况有了极大提高。 In order to improve a method on fuzzy comprehensive evaluation of road traffic safety in China. The focus is paid to determining the weight by a pairwise comparison matrix in fuzzy analytic hierarchy （FAHP.） process. A weigh-ting method based on triangular fuzzy numbers （TFN.） is proposed. First , 9 indices of accident rate arc determined based on an absolute index of traffic accidents and related factors, including people,vehicles,roads,etc. Second, a triangular fuzzy number is expressed as an interval concept. Unlike the traditional FA HP that used 1 - 9 integers and their reciprocal for weights determination, the upper and lower limits of the probability interval） and the maximum possible value of the whole range arc set. Thereby the possibility of different parameters in the whole range is obtained. The effective informa-tion in the process of fuzzy evaluation is better reserved, so that evaluation results arc more reasonable and accurate. Based on membership function,a judgment matrix and final fuzzy comprehensive evaluation method is developed. Taking the statistical data of 31 provinces, autonomous regions, and municipalities in China as a ease study, the proposed method is applied to analyzing the macroscopic situations of traffic safety. According to the results, the proportion of areas with poor safety has been reduced from 87% to 6.5% during the past 10 years from 2005 to 2014 , which reflects a great im-provement of traffic safety in China.

关 键 词： 交通安全 模糊综合评价 三角模糊数 指标权重 隶属函数