Article ID: | iaor20172918 |
Volume: | 11 |
Issue: | 6 |
Start Page Number: | 1047 |
End Page Number: | 1056 |
Publication Date: | Aug 2017 |
Journal: | Optimization Letters |
Authors: | Gendron Bernard, Hanafi Sad, Mjirda Anis, Todosijevi Raca, Mladenovi Marko |
Keywords: | heuristics: local search, combinatorial optimization, programming: travelling salesman, heuristics, transportation: water |
In this paper, we present two general variable neighborhood search (GVNS) based variants for solving the traveling salesman problem with draft limits (TSPDL), a recent extension of the traveling salesman problem. TSPDL arises in the context of maritime transportation. It consists of finding optimal Hamiltonian tour for a given ship which has to visit and deliver products to a set of ports while respecting the draft limit constraints. The proposed methods combine ideas in sequential variable neighborhood descent within GVNS. They are tested on a set of benchmarks from the literature as well as on a new one generated by us. Computational experiments show remarkable efficiency and effectiveness of our new approach. Moreover, new set of benchmarks instances is generated.