OPTFIN.ORGMultidisciplinary Research Laboratory for Operations Research and AI
Sign in

You are browsing anonymously. A verified sign-in is required before any personal or project data can appear.

IMPLEMENTED_NOT_INDEPENDENTLY_VERIFIEDTruck + drone

ALGORITHM ARTICLE

Best-improvement VND

Problem. Improve a truck tour using exact split evaluation.

Core idea. Search several neighborhoods and accept the best exactly reevaluated move.

Procedure

  1. Generate neighborhood moves.
  2. Evaluate every move with exact split.
  3. Repeat until no neighborhood improves.

Certificate. Every accepted move and final solution are reevaluated.

Data structures and API

from optfin_orlab import TruckDroneProblem

solver = TruckDroneProblem()
result = solver.improve_tour_vnd(...)

Complexity. Neighborhood- and round-dependent

Limits. A local optimum is not a global proof.

OPTFIN AUDIT CHECKS

What has to reconcile before this method is trusted.

  • DefinitionProblem, objective, inputs and output are explicit.
  • Data structureThe public API and canonical source path are identified.
  • CorrectnessEvery accepted move and final solution are reevaluated.
  • ComplexityNeighborhood- and round-dependent
  • Operational limitA local optimum is not a global proof.
  • ReproductionSource, executable test and evidence route remain linked.
View sourceOpen executable testReproduce evidence

LUNA ACADEMIC

Continue this algorithm in one research conversation.

Ask for the paper trail, executable test, source explanation, or a reproducibility plan for Best-improvement VND.

Open Luna for this algorithm

Back to the algorithm blog