13 New Instances for the Mixed Capacitated Arc Routing Problem with Intermediate Facilities under Time Constraints (MCARPTIF) (Road Cleaning Problem)

Dataset Overview

This dataset includes 13 instances specifically designed for evaluating the performance of vehicle routing algorithms in road cleaning problems. Existing benchmark instances in the literature predominantly focus on waste collection applications, which are not directly suitable for testing the unique characteristics of the road cleaning problem addressed in this study. To fill this gap, we have utilized a real-world dataset from the city of Zibo and a synthetic dataset modeled after the well-known lpr dataset.

1. Zibo Road Cleaning Dataset (zb)

The Zibo dataset (zb) represents a portion of the actual road network in Zibo, consisting of three depots, 319 nodes, and 962 arcs. We partitioned this road network into three sub-networks—zb1, zb2, and zb3—based on the distance between the arcs and depots, as well as the operational capabilities of the fleet. This transformation allowed us to treat the problem as a single-depot problem for each sub-network. The zb dataset serves as a realistic benchmark for the road cleaning problem.

2. lpr-rc Dataset

To further test the heuristics, we generated the lpr-rc dataset using a generation method inspired by the lpr dataset (Belenguer et al., 2006). The lpr-rc dataset consists of ten instances (labeled lpr-rc-1 to lpr-rc-10) of varying sizes, with nodes ranging from 27 to 268 and arcs ranging from 88 to 952. Each instance features a depot fixed in the top-left corner of the road network. Infrastructure facilities (IFs), including sewage disposal stations and water replenishment stations, are uniformly distributed across the network. The number of IFs is scaled based on the network size, adding complexity to the routing tasks.

Dataset Format

The dataset is provided in Excel format, with each instance containing two sheets:

Road Network: This sheet includes the unique road segment codes, the codes for the endpoints of each road segment, and the latitude and longitude coordinates for these points. Both road segment codes and endpoint codes are unique identifiers.

Intermediate Facilities (IFs): This sheet provides the unique codes for each facility and their respective latitude and longitude coordinates. The facilities represent points where vehicles can perform water replenishment or sewage disposal.

Vehicle Fleet Information

While the dataset does not impose specific vehicle parameters, it is designed to support the use of heterogeneous fleets. The following information describes a typical fleet of medium-sized mechanized cleaning vehicles, which we used in our study:

The fleet is homogeneous, and we define the following parameters:

Q_w: Freshwater tank capacity = 8000 L

Q_s: Sewage tank capacity = 2000 L

R_w: Freshwater consumption rate = 83.33 L/min

R_s: Sewage collection rate = 22.22 L/min

\lambda_w: Duration for water replenishment at IFs = 0.25 h

\lambda_s: Duration for sewage disposal at IFs = 0.16 h

V_w: Service speed = 10 km/h

V_d: Dead-heading speed (traveling without service) = 30 km/h

This fleet configuration enables a realistic simulation of road cleaning operations and can be adjusted based on real-world vehicle specifications as needed.

Availability

The 13 instances, including the zb and lpr-rc datasets, are now publicly available on Zenodo. These instances are designed to facilitate further research in road cleaning and related vehicle routing problems.

Reference

Belenguer, J.M., Benavent, E., Lacomme, P., Prins, C., 2006. Lower and upper bounds for the mixed capacitated arc routing problem. Computers & Operations Research 33, 3363–3383. doi:10.1016/j.cor.2005.02.009.