Train blocking plan and shipment path (TBSP) are key parts of China’s freight operation plan, and the integrated optimization problem of them has attracted more scholars’ attention in recent years. In China’s existing practice, shipment path and train blocking plan are determined in order. In this paper, a reverse-order two-stage solution method is proposed to solve the TBSP problem, which is opposite to the existing decision order. The first stage determines train blocking plan, that is, which stations to build blocks and the distribution of shipments to blocks. The second stage determines the physical path of each block, and then shipment path is determined. The corresponding two-stage integer programming model is established. A series of numerical experiments show that reverse order method balances solution quality and computational efficiency, and has advantages over both the existing positive order method and the method of establishing an integrated model.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

A Reverse-Order Two-Stage Method for the Train Blocking and Shipment Path Optimization Problem

  • Qining Guo,
  • Yixiang Yue,
  • Bo Zhang

摘要

Train blocking plan and shipment path (TBSP) are key parts of China’s freight operation plan, and the integrated optimization problem of them has attracted more scholars’ attention in recent years. In China’s existing practice, shipment path and train blocking plan are determined in order. In this paper, a reverse-order two-stage solution method is proposed to solve the TBSP problem, which is opposite to the existing decision order. The first stage determines train blocking plan, that is, which stations to build blocks and the distribution of shipments to blocks. The second stage determines the physical path of each block, and then shipment path is determined. The corresponding two-stage integer programming model is established. A series of numerical experiments show that reverse order method balances solution quality and computational efficiency, and has advantages over both the existing positive order method and the method of establishing an integrated model.