摘要
为综合客观地评价铁路事故的严重性,利用模糊数学理论和改进熵权-TOPSIS法对铁路事故定级问题展开研究.通过分析既有铁路事故定级方法中存在的不足,根据铁路事故评价指标的界定形式,设计能够反映评价指标与事故等级间模糊关系的隶属度函数以构建二者之间的隶属度矩阵.通过介绍熵权法的指标赋权过程,分析并指出熵权法存在极化指标重要性的缺陷,进而利用灰靶理论从指标贡献的角度对熵权法做了改进.最后,根据构建的隶属度矩阵和改进熵权法求得的指标权重值,利用基于绝对贴近度的TOPSIS法对铁路事故进行模糊评价.以6个铁路事故案例验证了改进熵权-TOPSIS法及有关理论的合理性,结果分析表明:改进熵权-TOPSIS法比传统熵权-TOPSIS法更能综合客观反映铁路事故的严重性.
In order to objectively and comprehensively evaluate the severity of railway accident,fuzzy mathematic theory and improved entropy weight-TOPSIS method are employed to study the problem of railway accident classification.By analyzing the shortcomings existed in active method of railway accident classification,the membership functions that can reflect the fuzzy relationship between evaluation indexes and accident classification are designed to further construct the membership matrix according to the definition form of the evaluation indexes of railway accident.This paper introduces the index weighting process of entropy weight method,analyzes and points out the defect of the importance of polarization index in entropy weight method,and then improves the entropy weight method from the perspective of index contribution by using grey target theory.According to the constructed membership matrix and the index weight resulted by the improved entropy weight method,TOPSIS method based on absolute closeness is finally used to evaluate the railway accident fuzzily.Six cases of railway accident are exhibited to verify the reasonability of improved entropy weight-TOPSIS method,as well as the related theories.The result shows that the improved entropy weight-TOPSIS method can reflect the severity of railway accidents more comprehensively and objectively than the traditional methods.
作者
张正坤
朱昌锋
ZHANG Zheng-kun;ZHU Chang-feng(School of Traffic and Transportation,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《兰州交通大学学报》
CAS
2021年第1期31-38,共8页
Journal of Lanzhou Jiaotong University
基金
国家自然科学基金(61364028)
教育部人文社科规划基金(18YJAZH148)
甘肃省自然科学基金(18JR3RA125)。
关键词
铁路事故定级
改进熵权法
TOPSIS法
灰靶理论
模糊评价
railway accident classification
improved entropy weight
TOPSIS method
grey target theory
fuzzy evaluation
