VDSP instances

These are the 69 instances used in the article: B. Domínguez-Martín, I. Rodríguez-Martín, J.J. Salazar-González, 2017. An exact algorithm for a Vehicle-and-Driver Scheduling Problem. Computers & Operations Research, vol. 81, pp.247-256. https://doi.org/10.1016/j.cor.2016.12.022. The instance name gives information about the number of nodes, the number of vehicles of each type available at the depots, and the class of the instance. For example, n16-1-1-1-1a.dat is an instance with 16 nodes, including the depots, one vehicle of each type at each depot, and the class of the instance is a. These data files have the same format as the classical CVRP instances from the literature, but some new parameters have been added. That is: The first line contains the name of the instance. The next lines contain: TYPE = ‘CVRP’ because the instances are CVRP instances adapted to our problem. DIMENSION = the number of nodes, including the depots EDGE_WEIGHT_TYPE = ‘EUC_2D’ because we use the Eucliden distance for all our instances. The next lines contains the capacity of the vehicles of type 1 (CAPACITY Y) and the capacity of the vehicles of type 2 (CAPACITY X). VEHICLES1 and VEHICLES 2 denotes the number of vehicles of type 2 available at first and second depot, respectively. DRIVERS1 and DRIVERS2 denotes the number of vehicle of type 1 available ar first and seconds depot, respectively. The lines after NODE_COORD_SECTION contain the following information: n x y where n=customer number (1 and the last number correspond to the depots), x=x coordinate, y=y coordinate. The lines after DEMAND_SECTION show the following information: n d where n=customer number (0 and the last number correspond to the depots), d=customer demand(0 for the depots and 1 for the customers) Finally, DEPOT_SECTION1 and DEPOT_SECTION2 indicate that the first and the last node represent the first and the second depot, respectively.

本文所用的69个算例均源自B. Domínguez-Martín、I. Rodríguez-Martín、J.J. Salazar-González于2017年发表于《计算机与运筹学（Computers & Operations Research）》第81卷第247至256页的论文《一种车辆与驾驶员调度问题的精确算法（An exact algorithm for a Vehicle-and-Driver Scheduling Problem）》，其DOI链接为https://doi.org/10.1016/j.cor.2016.12.022。

算例名称包含节点总数、各车场可用的各类型车辆数量以及算例类别三类核心信息。例如，`n16-1-1-1-1a.dat`代表该算例包含16个节点（含车场），每个车场各配备1台各类型车辆，且算例类别为a。

此类数据文件沿用了经典文献中容量约束车辆路径问题（Capacitated Vehicle Routing Problem，CVRP）算例的格式，但新增了若干参数，具体格式说明如下：
1. 首行为算例名称；
2. 后续行依次定义以下参数：
- `TYPE = "CVRP"`：因本算例均为适配本文研究问题的CVRP变体；
- `DIMENSION`：节点总数（含车场）；
- `EDGE_WEIGHT_TYPE = "EUC_2D"`：表示所有算例均采用二维欧氏距离作为边权重；
- 随后两行分别指定类型1车辆的载货容量（CAPACITY Y）与类型2车辆的载货容量（CAPACITY X）；
- `VEHICLES1`与`VEHICLES2`分别代表第一、第二车场可用的类型2车辆数量；
- `DRIVERS1`与`DRIVERS2`分别代表第一、第二车场可用的类型1车辆数量（原文此处存在笔误，将“vehicles”写为“vehicle”，且将“at”写为“ar”、“second”写为“seconds”）；
3. `NODE_COORD_SECTION`后的行格式为`n x y`，其中`n`为客户节点编号（节点1与最后一个节点对应车场），`x`、`y`分别为节点的横、纵坐标；
4. `DEMAND_SECTION`后的行格式为`n d`，其中`n`为客户节点编号（原文标注为节点0与最后一个节点对应车场，疑似笔误），`d`为客户需求值（车场需求为0，客户节点需求为1）；
5. `DEPOT_SECTION1`与`DEPOT_SECTION2`分别表明第一个节点与最后一个节点对应第一车场与第二车场。