E-VRP / E-VRPTW Instances
Benchmark instances for the Electric VRP:
Schneider, Stenger & Goeke (2014)
The reference benchmark for the E-VRPTW, introduced in [Schneider, Stenger & Goeke 2014]. It is built from the Solomon VRPTW instances, augmented with a set of recharging stations (21 stations in the base set) and battery/energy parameters. The full benchmark comprises clustered (C), random (R) and mixed (RC) instances of 100 customers.
A small sample is provided locally; the complete set is available from the sources linked below.
| Instance | Family | Customers | Stations | File |
|---|---|---|---|---|
| c101_21 | Clustered | 100 | 21 | TXT |
| c201_21 | Clustered | 100 | 21 | TXT |
| r101_21 | Random | 100 | 21 | TXT |
| rc101_21 | Mixed | 100 | 21 | TXT |
See the format description for details on the file layout and the energy parameters.
File format
Each file lists, in whitespace-separated columns, one node per line: StringID Type x y demand ReadyTime DueDate ServiceTime. The node Type is d (depot), f (recharging station) or c (customer). Trailing lines give the vehicle capacity, battery capacity, energy consumption rate and recharging rate.
External sources
- jmanzolli/E-VRPTW — CPLEX data and model for the Schneider et al. benchmark.
- fuvidani/e-vrptw — E-VRPTW solver and instances.
- VRP-REP — the VRP problem repository, which hosts additional E-VRP datasets and best-known solutions.