Metraipolis › archive › dc_bad63f3e2847

DCL-GPGLS: Dynamic Curriculum Learning for Genetic Programming Guided Local Search in Large-Scale Vehicle Routing — arXiv:2609.24105

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.24105
License (as archived)
CC0-1.0
Archived
2026-10-09T22:40:06.953Z
SHA-256 of the archived text
4dc75f666c119ac0f27a819d6077437fc8b58deb1f26e2dd01c8693ef630740e
Stable URL of this page
https://metraipolis.com/documento/dc_bad63f3e2847

How to cite

DCL-GPGLS: Dynamic Curriculum Learning for Genetic Programming Guided Local Search in Large-Scale Vehicle Routing — arXiv:2609.24105 — arXiv (en), https://arxiv.org/abs/2609.24105 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bad63f3e2847 (Metraipolis archive dc_bad63f3e2847, sha256 4dc75f666c119ac0f27a819d6077437fc8b58deb1f26e2dd01c8693ef630740e)
BibTeX
@misc{dcbad63f3e2847,
  title = {DCL-GPGLS: Dynamic Curriculum Learning for Genetic Programming Guided Local Search in Large-Scale Vehicle Routing — arXiv:2609.24105},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.24105},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bad63f3e2847; sha256 4dc75f666c119ac0f27a819d6077437fc8b58deb1f26e2dd01c8693ef630740e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24105. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Genetic Programming Guided Local Search (GPGLS) uses genetic programming to evolve utility functions for guided local search in large-scale vehicle routing problems (LSVRPs). Evaluating every GP individual on every training instance at every generation is expensive, so GPGLS is usually trained on small instance batches. Existing curriculum-based GPGLS orders these batches mainly by instance size. Adaptive Curriculum Learning GPGLS (ACL-GPGLS) improves training efficiency by adapting when the search moves between fixed curriculum stages, but the instance difficulty order remains predefined. We propose DCL-GPGLS, which estimates the difficulty of each training instance from the current population's solution quality and updates the estimates during evolution. Each generation then receives a batch near a scheduled difficulty level, with a correction that limits repeated selection of the same instances. Experiments on a fixed training-test split of the CVRPLIB X set show that DCL-GPGLS achieves the best observed average rank and mean test cost among six training policies. It obtains the lowest mean cost on 36 of 65 unseen test instances and is significantly better than the static feedback-derived curriculum, matched in total evaluator calls, on 6 instances, with no significant difference on the remaining 59.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.NE).
Categorie arXiv: Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2609.24105 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-09T22:23:32.459Z dalla pagina https://arxiv.org/catchup/cs.NE/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).