RRVRP Instance

We adopt the trip decomposition method to transfer the trip to arc demand so as to propose a vehicle flow formulation for the RRVRP. In the computational experiments, we use 40 randomly-generated small-scale instances and the 20 benchmark instances provided by Bodin et al. (2000). To randomly generate the 40 small-scale instances, we abstract the transportation network on a 60×60 grid where the nodes denote the depot (on node (30, 30)), the disposal facility (on node (20, 40)) and customers. The small-scale instances are classified into four classes with 10 test problems (with 5, 10, 15 and 20 trips) in each class. The four classes differed in the mix of Trip type 1, 2, 3 and 4.本文采用行程分解法将行程转化为弧段需求，以此为RRVRP构建车辆流建模公式。在计算实验中，我们使用了40个随机生成的小规模算例，以及Bodin等人（2000）提出的20个基准算例。为生成这40个小规模算例，我们在60×60的网格上抽象得到交通网络，网络中的节点分别代表仓库（位于节点(30, 30)）、处置设施（位于节点(20, 40)）以及客户。该小规模算例被划分为四个类别，每个类别各包含10个测试问题，分别对应5、10、15和20个行程。四个类别的差异在于行程类型1、2、3、4的组合比例不同。