树枝形专用线非直达车流取送的三阶段综合协调优化

Integrated approach with three stages for taking-out and placing-in shunting of non-through wagons flow on branch-shaped siding

本文针对一类树枝形铁路专用线非直达车流取送问题,考虑装卸站能力、调机牵引定数、调机走行和货车最晚取回时间等限制,构建以总调运成本最小化为目标的数学模型。鉴于模型特点,本文设计取送作业顺序集合生成-取送作业顺序集合优化-调机自适应分配三阶段融合求解策略。该方法首先利用模型起作用约束组生成初始取送作业顺序集合,进而利用IBH-GA启发式实现装卸站间货车取送作业顺序优化,最后给出基于批次时间窗-走行时长的调机自适应分配策略。本文通过设计实验场景对所提方法进行过程验证,并利用不同规模问题对算法进行测试对比和性能评估。

The railway freight transport with strictly scheduled-based operation modes mainly deliveries the bulk commodity and undertakes the long-haul transport task.The railway freight terminal is a huge and complex system which consists of the marshalling yard,the handling station and the railway sidings.The railway freight terminal is well recognized as one of the most important and critical elements for enhancing the competitiveness of the railway freight transportation chain.Therefore,the performance of railway freight terminal is crucial for the transportation chain effectiveness and must be closely optimized.Taking-out and placing-in shunting of wagons is an important work for the large railway freight terminal.Their efficiency is directly related to the wagons turnover,the goods delivery,and the completion of the transport production quota.The operators of railway terminal must think how to arrange reasonably placing-in and taking-out shunting of wagons.In the whole process of placing-in and taking-out shunting,the local wagons will be decoupled from the inbound trains,delivered to the relevant handling station for loading or unloading,retrieved back from the handling station after the loading and unloading tasks are finished,and coupled onto the outbound trains.Depending on the layout of handling stations in the railway terminal,we can divide them into two types:the radial siding network and the branch-shaped siding network.In a radial siding network,after the shunting engine has sent a wagon group to one handling station,it must go back to the marshalling yard before running to the next handling station.For a branch-shaped siding network,the shunting engine doesn′t need to return to the marshalling yard before placing-in another wagon group.In this paper,we discuss the placing-in and taking-out shunting of non-through wagons flow on branch-shaped siding.The takingout and placing-in shunting of non-through wagons on branch-shaped siding is formulated as a mixed integer programming model which considers some constraints on the handling station capacity,the engine′s traction number,the engine running time and the deadline of retrieving wagons.The objective function makes minimization of the total shunting cost which is composed of the engine utilization cost,the wagon utilization cost and the detention time cost of wagons.As an NP problem,using traditional method for solving the model is difficult and inefficient.According to the characteristics of the model,the integrated approach with three stages is given for generating the placing-in and taking-out wagons sequence set,optimizing the placing-in and taking-out wagons sequence and allocating adaptively the engine.This approach is abbreviated as IATS in order to facilitate problem statement.Firstly,a generating initial solution procedure with active constraints group is obtained to generate an initial set of placing-in and taking-out wagons sequences.The set of placing-in and taking out wagons sequence is given.Some active constraints which denote the placing-in and taking-out wagons order,the handling station capacity,the engine′s traction number,the engine running time and the deadline of retrieving wagons are adopted to update the initial set of placing-in and taking-out wagons sequences.Secondly,an improved bat heuristic with genetic algorithm is proposed to optimize the placing-in and taking-out wagons sequence.The heuristic is abbreviated as IBH-GA.Let the initial placing-in and taking-out wagons sequences be the implement object.Some control parameters of iteration procedure are given.And the implement object and initial iteration population are provided.An iteration procedure embedding bat algorithm and genetic algorithm is developed.The local searching and parameters updating based on bat algorithm are presented.Then the swapping implement based on genetic algorithm is provided.These implement objects meeting the constraints are selected and are collected into the feasible iteration population.The objective function value of all these implement objects in feasible iteration population are calculated and then the optimal implement object can be obtained from them.Thirdly,the engine adaptive allocation approach with the batch time window and the running time is obtained.The service time window of placing-in and taking-out wagons sequences by IBH-GA is calculated.According to the service time window,the shunting engines are allocated to the corresponding placing-in and taking-out wagons sequences.And then according to the engine working time load,the dividing batch scheme is adjusted.To evaluate the performance of the approach proposed,we generate randomly the three sets of experiments with the different amount of the task during the whole shunting horizon and the same branch-shaped siding composed of 12 handling stations.Three sets of experiments are oriented to compare the performances of the IA-TS approach with the GA-EAAP approach and BA-EAAP approach.In the first set of contrast experiment,there are 12 item tasks for the placing-in and taking-out local wagons group.In the second set of contrast experiment,there are 20 tasks.In the third set of contrast experiment,the amount of the task increases to 32.Finally,we evaluate the performance of the IA-TS in terms of the three traditional measures of the average objective function,the average computer execution time and the algorithm improvement rate.The results show that the proposed algorithm performs outstanding in the three evaluation measure.