摘要
利用数据包络分析(DEA)方法进行效率计算时,若有新增参评样本则下阶段计算时旧样本的效率计算结果可能会随之改变,针对此问题,引入反向传播神经网络(BPNN),基于含偏好序的DEA-CCR模型效率计算结果进行训练,构建DEA-BPNN效率计算模型,在不影响旧样本阶段计算结果的前提上计算新增样本效率。以杭深线厦深段高速列车运行图为例,根据运行线技术指标,利用DEA-BPNN高速列车运行线效率计算模型对2021年第1季度和第2季度厦深段运行图的本线高速列车进行效率计算,研究结果表明,第1季度厦深段本线列车平均效率为0.946,而第2季度本线列车平均效率达到了1.048,说明2021年第2季度厦深达速调图后运行线效率相较上一季度有较大提升。
When using the data envelopment analysis(DEA) method for efficiency calculation, if there are newly added samples, the efficiency calculation results of the old samples in the next stage of calculation may be changed accordingly. To solve this problem, the back propagation neural network(BPNN) is introduced,train based on the efficiency calculation results of the DEA-CCR model with preference order, construct the DEA-BPNN efficiency calculation model, and calculate the efficiency of the new sample without affecting the calculation results of the old sample stage. Take the high-speed train operation diagram of the XiamenShenzhen section of the Hangzhou-Shenzhen line as an example, use an DEA-BPNN train operation line efficiency calculation model according to the technical indicators of the operation line, and compare the highspeed trains of the Xiamen-Shenzhen section in the first quarter and second quarter of 2021. Carrying out efficiency calculations, the research results show that the average efficiency of intra-line trains on the XiaShenzhen section in the first quarter was 0.946, while the average efficiency of intra-line trains reached 1.048in the second quarter. The higher operating efficiency indicates that the train operation lines in the second quarter of 2021 have improved their efficiency compared with the previous quarter after the speed adjustment of the Xiamen-Shenzhen section.
作者
陈泽文
张杰
CHEN Zewen;ZHANG Jie(Southwest Jiaotong University,School of Transportation and Logistics,Chengdu 610031,China;Research and Development and Training Center of Railway Train Operation Drawing Compilation,Southwest Jiaotong University,Chengdu 610031,China;National and Local Joint Engineering Laboratory for Integrated Transportation Intelligence,Chengdu 610031,China)
出处
《综合运输》
2022年第6期96-103,共8页
China Transportation Review
基金
国家重点研发计划资助(2017YFB1200702)
国家自然基金项目(52072314)
四川省科技计划项目(2020YFH0035,2020YJ0268,2020YJ0256,2021YFQ0001,2021YFH0175)
中国国家铁路集团有限公司科技研究开发计划课题(2019F002)
成都市科技项目(2019-YF05-01493-SN,2020-RK00-00036-ZF,2020-RK00-00035-ZF)
北京局集团公司科技研究开发计划课题(2021BY02)。
关键词
列车运行图
列车运行线效率
数据包络分析
BP神经网络
偏好序
Train operation diagram
Train operation line efficiency
Data envelopment analysis
BP neural network
Preference order
