【摘要】 针对货车编组问题,采用半分离式两阶段0-1线性规划模型对各阶段联合求解,对局部最优解采用调度时序图可视化表述.首先,对无、有调车辆分离,无调车采用启发式安排.有调车推峰顺序可以转化为零件加工问题,以驼峰总工作量最大、等待时间最小为目标建立模型I.列车解体时间与解体方向数成正比增长,但在未确定具体解体方案时无法确定(即模型I的独立),通过在模型II中对解体时间模糊化来处理两步独立的缺陷,从而达到两步规划的连续特性.车辆新编,决策变量属于多维结构,通过将多维稀疏变量转化为一维序列,有效解除其稀疏特性,形成二维决策变量建立规划模型II直接求解.其次,通过仿真创建模拟数据,运用主模型求解,得到了驼峰是编组站主要瓶颈的结论.最后,我们还对铁路资源的紧缺性、编组效率建模给出了较详细改进措施.更多还原

【Abstract】 Considering the problem of freight train formation,this paper adopts semi-sparated two-stages of linear 0-1 programming model for combined solution of each stage.Besides,in allusion to local optimal solution,we express it by sequence diagram.First,according to the split of the non-resorting or shunting trains,the former can be used heuristic arrangement.As to the later,it′s sequence of pushing train stock to hump may be transformed into part maching case.We regard the max of total hump workloads and the min of waiting time as objective functions,then establish ModelⅠ.Trains′ disintegration time is rising by disintengration direction number′s increasing.However,it can′t be confirmed under the condition of no concrete disintegration scheme.(That is the independence of Model I.) Through fuzzification of disintegration time in Model II,we deal with the defects of independeces of Model I and Model II to reach continuous characteristic of two models′ planning.Trains′ new edition and decision variable are multidimensional structures.This paper transforms multi-dimensional sparse variables to one-dimensional series,which relieves the sparseness,then can get weo-dimensional decision variable.Afterward,build planning Model II for direct solution.Secondly,we gets the conclusion that the hump is the main bottleneck of railway marshalling.The method is establishing simulated data by simulation.In the end,we give out more detailed improvement measures for the scarcity of railway resources and the modeling of marshalling efficiency.更多还原