Skip to content

Ari2018 (TDVRP)

The TDVRP layer of Ari2018 is the same curated 160-instance subset with the time windows removed: same demands, capacities, fleet sizes, service times, and the same shared IGP arrival-time-function sidecars. The Duration objective and canonical checker contract are unchanged. No previously published solutions exist for this variant; all BKS come from KAYROS heuristic runs on Grid5000.

Licensing note: as for the TDVRPTW layer, MAMUT-authored curation artifacts are under the MIT License, while underlying third-party material is not relicensed.

The family currently contains 160 curated instances over 4 sizes, each with a Duration BKS.

Data

Instances and best-known solutions: /benchmarks/tdvrp/ari2018/ on the website. Repository: https://github.com/ANR-MAMUT/MAMUT-routing-TDVRP-Ari2018.

Repository README

Ari2018 TDVRP benchmark family (Arigliano-generator TD travel times, curated VRP variant)

Satellite benchmark repository of MAMUT-routing (ANR MAMUT, ANR-22-CE22-0016), mounted there as benchmarks/TDVRP/Ari2018. Self-contained: instances, canonical arrival-time-function (ATF) sidecars and best-known solutions ship together.

What this family is — and what it is not

The raw data is the classic TD-TSPTW benchmark of the Arigliano et al. lineage: a single uncapacitated vehicle, no demands, no fleet size, no service times; per instance a symmetric distance matrix, integer time windows, and IGP time-dependent travel speeds (73 speed zones, 3 speed profiles shared by all arcs; congestion depth Delta ∈ {70, 80, 90, 95, 98}, traffic patterns A/B, TW-width parameter β·100 ∈ {0, 25, 50, 100}, three TW classes A/B/C, ten instances each; sizes 15/20/30/40 customers — a 4800-instance full factorial).

This canonical family is the VRP variant curated by Onyr (2024–2026): demands, vehicle capacities, fleet sizes and gaussian service times were generated once by the legacy TDVRPTW-benchmarks pipeline and are pinned here verbatim (the generation was unseeded, so the values — recorded per instance in metadata.legacy_attribute_source — are the authoritative definition); time windows removed (TDVRP twin). It is therefore not comparable with published TD-TSPTW results on the underlying raw files.

Curated subset (deterministic)

The family ships 160 instances, not the full 4800: one per structural cell (n, Delta, pattern, width), chosen by a rule that is a pure function of the raw tree and the legacy BKS store — candidates holding a legacy best-known solution win first, remaining ties are decided by ascending SHA-256 of the canonical instance name (e.g. sha256("Ari-C3-pA-d98-w50")). Instance names Ari-<Class><id>-p<Pattern>-d<Delta>-w<Width> under n=<customers> carry all structural parameters. The full factorial raw trees remain available upstream for anyone needing the complement.

Format

  • <Name>.vrp.json — MAMUT shape: depot 0, mandatory display coordinates (deterministic stress-layout embedding of the symmetric raw distance matrix; quality in metadata.coordinates_method), horizon [0, depot due date], td block with the sidecar's storage-independent sha256.
  • <Name>.atf.json.gz — canonical ground truth: one arrival-time NDCPWLF per arc, exact IGP consolidation (breakpoints exactly where departure or arrival crosses a speed-zone boundary, each re-evaluated by the exact forward Ichoua loop; the last zone's speed extends beyond the horizon). All sidecars are gzipped in this satellite.
  • <Name>.bks.Duration.json — best known solution under the Duration objective; costs are always the canonical checker's output.

Population pipeline: populate_td_ari_vu v1 (curation tooling maintained outside this repository; links will be published with the benchmark paper).

Provenance and licensing

Underlying raw data: the Arigliano et al. TD-TSPTW benchmark generator distribution. MAMUT-authored curation artifacts under the MIT license; this notice does not relicense the underlying third-party benchmark definitions.