【作者】 麦家骥；

【导师】 陈峰；

【作者基本信息】 上海交通大学 ， 管理科学与工程， 2012， 硕士

【摘要】 本文研究随机需求的越库(Cross Docking)调度的建模和算法。越库调度是指物流的任何中间点只实现收发货的功能而没有货物存储与订单获取的做法。越库调度一般分为两个阶段,从供应商取货到越库中心的过程为第一阶段,从越库中心(Cross Docking Center)将货物配送到客户的过程为第二阶段。越库调度按照运输方式划分,主要有直送(Direct Delivery)和循环取料(Milk Run)两类。越库调度问题属于车辆路径问题的一种。需求的随机性,是实际调度中最经常发生,而且也最难以避免的,也是随机调度问题中较常见和被研究得最多的一种,但是在越库调度方面,需求的随机性还未得到较多的研究。本文在对越库调度、运输方式和随机车辆路径问题进行分析研究后,结合越库调度问题的特性和运输方式的区别,对不同运输方式情况下的越库调度的随机需求进行了分析并提出了合理的假设,并给出了详细的需求生成方法。根据不同的运输模式,本文研究了两类随机需求的越库调度问题: (1)两阶段直送的随机需求越库调度的建模与算法。首先对问题进行描述并提出问题的假设,并在此基础上提出了两阶段补偿模型及求解问题的样本均值近似算法(SAA算法),同时提出了两阶段的启发式算法。最后通过在小规模情况下进行决策准确性和平稳性的数值实验验证了SAA算法和启发式算法的有效性。(2)直送-循环取料的随机需求混合越库调度的建模和算法。首先描述问题并提出问题的假设,提出了两阶段补偿模型及求解问题的SAA算法,并提出了基于预留容量的两阶段启发式算法。小规模数值实验的结果表明了算法的有效性,而大规模的数值实验给出了启发式算法的最佳参数选择策略。本文的研究旨在可以为企业越库物流的实际运作提供决策支持,同时希望可以将研究成果真正运用到实际中,为越库物流的调度管理提供理论基础和有用的算法。更多还原

【Abstract】 This thesis studies the modeling and heuristics of cross docking scheduling problem with stochastic demand. Cross docking is an operation that moves goods directly from receiving point to shipping point without storage. There are two stages in cross docking scheduling: the first stage in which goods are collected from suppliers to cross docking center by vehicles, the second stage in which goods are sent from cross docking center to customers. Cross docking scheduling can be classed into two classes by the transportation model: direct delivery and milk run. Cross docking scheduling problem is a kind of vehicle routing problem.The randomness of demand always occurs in practical scheduling and is difficult to avoid. Also it’s widely studied in stochastic scheduling. However, it’s seldom studied in cross docking scheduling. This thesis analyzes cross docking, transportation models and vehicle routing problems. The stochastic demands in cross docking systems with different transportation models are analyzed. Reasonable assumptions and data generating methods are introduced.According to different transportation model, this thesis studied two kinds of cross docking scheduling problems with stochastic demand:(1) The modeling and heuristics of two-stage direct directly cross docking scheduling with stochastic demand. First, the problem is described and assumptions are raised. A two-stage recourse model, a SAA based heuristic and a two-stage heuristic are presented. According to the numerical experiment about solution effectiveness and steady in small scale, the SAA heuristic and two-stage heuristic are proved to be effective.(2) Modeling and heuristics of direct directly - milk run mixed cross docking with stochastic demand. First, the problem is described and assumptions are raised. A two-stage recourse model, a SAA based heuristic and a capacity reservation based two-stage heuristic are presented. According to the numerical experiment about solution effectiveness and steady in small scale, the SAA heuristic and two-stage heuristic are proved to be effective. And the best parameter settings are given by the large scale numerical experiment.This thesis aims to provide decision support for the practical cross docking operation in logistics companies. Also, the research aims to be applied in the practical operation and provide the theory support and useful heuristic for the scheduling management in cross docking system.更多还原