2E-VRP-SBS instances

In considering route optimization at a series of express stages from pickup to delivery via the intercity linehaul, we introduce the two-echelon vehicle routing problem with satellite bi-synchronization (2E-VRP-SBS) from the perspective of modeling the routing problems of two-echelon networks. The 2E-VRP-SBS involves the inter-satellite linehaul on the first echelon, and the pickups from senders to origin satellites (i.e., satellites for cargo collection) and deliveries from destination satellites (i.e., satellites for cargo deliveries) to receivers on the second echelon. The 2E-VRP-SBS integrates satellite bi-synchronization constraints, multiple vehicles, and time window constraints on the two-echelon network and aims to find cost-minimizing routes for various types of trucks. Small-scale instances. Considering the computing abilities of CPLEX 12.4 indicated by some computational trials, there are one depot, three origin satellites, and three destination satellites on the first echelon. The number of senders or receivers served by an origin satellite or a destination satellite is assumed to be the same, and the number of senders or receivers served by a satellite is 5, 6, 7, 8, or 9. Each small-scale instance is denoted by num_os-num_ds-num_gc-sn, where num_os or num_ds denotes the number of origin satellites or destination satellites, num_gc is the number of senders or receivers served by a satellite, sn is the instance sequence for the same num_os, num_ds, and num_gc, and sn is 1, 2, 3, 4, 5, 6, 7, 8, or 9. We designed the large-scale instances by referring to data provided by a logistics company in China. Referring to the locations of 17 distribution-centers (DCs), which are located in 17 prefecture-level cities in Shandong province, we designed the nodes on the first echelon as follows. First, it is assumed that each DC can be used as the depot, i.e., the depot location is the same as one of the DC locations. The large-scale instances are distinguished by the location of the depot. Second, there are 17 origin satellites with different locations, and the location of an origin satellite is the same as the location of a DC. Third, there are 17 destination satellites with different locations. The location of a destination satellite is the same as the location of a DC, and the location of an origin satellite should not be the same as the location of a destination satellite, if there are inter-satellite linehaul demand between the origin satellite and the destination satellite. Generally speaking, we designed 17 large-scale instances that are distinguished by the location of the depot. The number of senders or receivers served by an origin satellite or a destination satellite is assumed to be the same. The number of senders or receivers served by a satellite is 120. The sending-receiving relationships between senders served by an origin satellite and receivers served by destination satellites are randomly generated.

针对城际干线运输下揽收至派送的多阶段快递路径优化问题，本文从两阶段网络路径问题的建模视角出发，提出了带卫星双向同步的两阶段车辆路径问题（two-echelon vehicle routing problem with satellite bi-synchronization，2E-VRP-SBS）。

该2E-VRP-SBS包含两个层级的运输作业环节：第一层级为卫星间干线运输，第二层级涵盖两类核心作业——从寄件人至始发卫星（货物集散卫星）的揽收环节，以及从目的卫星（货物配送卫星）至收件人的派送环节。

2E-VRP-SBS整合了卫星双向同步约束、多车辆调度约束以及两阶段网络的时间窗约束，目标是为各类货运车辆求解总成本最小化的配送路径。

小规模测试实例：结合部分计算试验所验证的CPLEX 12.4计算性能，设定第一层级包含1个场站、3个始发卫星与3个目的卫星。假设单个始发卫星或目的卫星所服务的寄件人、收件人数量一致，且该数量取值为5、6、7、8或9。每个小规模测试实例的命名规则为`num_os-num_ds-num_gc-sn`，其中`num_os`与`num_ds`分别代表始发卫星与目的卫星的数量，`num_gc`为单个卫星所服务的寄件人、收件人数量，`sn`为相同`num_os`、`num_ds`及`num_gc`配置下的实例序号，取值范围为1至9。

大规模测试实例：本文参考国内某物流公司提供的真实运营数据设计了大规模测试实例。首先，参照山东省17个地级市内17个配送中心（distribution-centers，DCs）的地理位置，对第一层级的节点进行如下设计：其一，默认每个配送中心均可作为场站，即场站位置与某一配送中心的位置重合，大规模测试实例以场站的位置进行区分；其二，设置17个位置各异的始发卫星，其位置与某一配送中心的位置一一对应；其三，设置17个位置各异的目的卫星，其位置同样与某一配送中心的位置一一对应，且当始发卫星与目的卫星间存在卫星间干线运输需求时，二者的位置不得重合。综上，本文共设计了17个以场站位置区分的大规模测试实例。假设单个始发卫星或目的卫星所服务的寄件人、收件人数量统一为120。始发卫星所服务的寄件人与目的卫星所服务的收件人间的收发匹配关系均为随机生成。