Certified Lower-Bound Orientation Search under Cross-Stage Objective Mismatch in Agricultural Coverage Path Planning

Résumé fourni par la source

CLBOS reproducibility package v2.1.0 This release provides the reproducibility package accompanying: Certified Lower-Bound Orientation Search under Cross-Stage Objective Mismatch in Agricultural Coverage Path Planning ## What is new in v2.1.0 This release adds manuscript-facing reproducibility assets for additional evaluation and analysis: - the equal-scope end-to-end runtime comparison between CLBOS-MST and the fixed Top-10 strategy on the frozen stratified 18-field subset;- the first-level component decomposition of the 25 pure cross-stage mismatch cases, including the trade-off between additional working-swath distance and reduced non-working travel distance;- updated reproducibility mappings, metadata, and package integrity records. This release does not modify: - the CLBOS algorithm;- the admissible lower-bound formulation;- the original 90-field held-out benchmark;- the predefined finite orientation domain;- the planner assumptions used for certification. ## Previous updates in v2.0.0 Version 2.0.0 extended the reproducibility package with additional manuscript-facing evidence: - added the pre-specified fixed-Top-10 transfer evaluation and corresponding CLBOS transfer results;- added the controlled minimum turning radius analysis;- added the 60-field secondary analysis of field difficulty and certification behavior;- added the stratified 18-field five-repeat runtime evaluation;- aligned figure reproduction assets with the manuscript;- refreshed package metadata, documentation, and publication-path sanitisation. ## Reproduction Core benchmark results: python3 scripts/reproduce_numbers.py Additional manuscript-facing results: python3 scripts/reproduce_v2_added.py Package integrity verification: python3 scripts/verify_package.py ## Reproducibility scope The package contains the data, scripts, and derived results required to reproduce the manuscript-reported numerical analyses, including: - the 90-field held-out benchmark evaluation;- exhaustive and CLBOS-MST comparison results;- fixed-budget Top-10 comparisons;- transfer and sensitivity analyses;- equal-scope runtime measurements;- cross-stage mismatch component decomposition. ## Interpretation boundary The fixed Top-10 strategy remained exact in all tested transfer domains. CLBOS is not claimed to be uniformly cheaper than a successful fixed budget strategy. Instead, CLBOS provides instance-specific finite-domain stopping evidence under the stated planner semantics and admissibility assumptions. ## Version history Previous public version-specific release: v2.0.0 DOI: 10.5281/zenodo.21920938 Concept DOI: 10.5281/zenodo.21920937 The version-specific DOI for v2.1.0 is assigned by Zenodo upon publication.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.