Routing short-haul trucks under the uncertainties of travel time and service time

This dataset supports research on optimizing battery electric vehicle (BEV) fleet dispatching in last-mile freight logistics under uncertainty. It accompanies the study on the Electric Vehicle Routing Problem with Backhauls and Time Windows under Travel Time and Service Time Uncertainty (EVRPBTW-USUT), which extends previous research by incorporating a backhauling strategy and modeling uncertainty in travel and customer service times. The dataset consists of 60 benchmark instances, derived from the well-known EVRPTW dataset, with varying customer sizes and backhaul proportions, enabling robust evaluations of routing strategies. Additionally, a real-world dispatching dataset from a full-service supply chain company in San Bernardino County, California, is included to validate the approach in practical applications. Each instance is provided in CSV format, with detailed solutions recorded in Excel files. These datasets support the development and benchmarking of optimization algorithms, p..., , # Routing short-haul trucks under the uncertainties of travel time and service time

Author: Dongbo Peng ([[e-mail]](mailto:[e-mail]))

#### 1.1 Vehicle Routing Problem with Backhauls Benchmark Dataset Description

Folder: Section_3_1_BKS_dataset_and_results

The file \"TV_33_instances\" contains a total of 33 problem instances in CSV format, original introduced by (Toth & Vigo, 1997), with customer sizes ranging from 21 to 100. Each instance is named according to the number of customers and the proportion of linehaul customers. For example, the file \"eil22_50.csv\" represents an instance with 22 nodes, where 50% of the customers are linehaul customers. Specifically, in each instance, the columns \"node_id\", \"type\", \"x\", \"y\", \"demand\", \"Q\", \"k\", \"L\", and \"B\" represent the node ID, service type (where -1 denotes the depot node, 0 denotes a linehaul customer, and 1 represents a backhaul customer), the x and y coordinates of the node, the customer demand, vehicle capacity (Q), ...,