Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | English |
| Published: |
Zenodo
2025
|
| Subjects: | |
| Online Access: | https://doi.org/10.5281/zenodo.17640906 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p>This work introduces Geo-first TSP, a new geometric-prior heuristic for the Euclidean<br>Traveling Salesman Problem (TSP). The method integrates convex-hull extraction, direction-<br>biased candidate edge generation, spatial neighborhood filtering, and randomized geometry-<br>guided constructive tours, followed by 2-opt local search.<br>The algorithm is motivated by the hypothesis that injecting global geometric structure into<br>randomized initial tours improves the quality of 2-opt local minima. Experiments on several<br>TSPLIB instances demonstrate that Geo-first TSP significantly outperforms classical construc-<br>tive heuristics and approaches the quality of multi-start 2-opt, sometimes within 0.03% of the<br>best obtained solution.<br>The preprint documents the method, reports empirical results, and outlines conceptual links<br>to computational geometry, probabilistic heuristics, and candidate-set design in local search.<br>Code and scripts are openly provided for reproducibility.</p>