|
|
|
Session TB4 - Tournées de véhicules III / Vehicle Routing III
Day |
Tuesday, May 06, 2003 |
Room |
Hélène-Desmarais |
President |
Teodor Gabriel Crainic |
Presentations
15:30 |
A Threshold Accepting Metaheuristic for the Vehicle Routing Problem with Time Windows |
|
Olli Bräysy, SINTEF Applied Mathematics, Optimization, P.O. Box 124, Oslo, Norway, 0314
Jean Berger, Defence Research Development Canada, Decision Support Technology Section, 2459, boul. Pie-XI nord, Val-Bélair, Québec, Canada, G3J 1X5
Mohamed Barkaoui, Defence Research Development Canada, Decision Support Technology Section, 2459, boul. Pie-XI nord, Val-Bélair, Québec, Canada, G3J 1X5
Wout Dullaert, University of Antwerp, Ufsia-Ruca Faculty of Applied Economics, Prinsstraat 13, 2000 Antwerp, Belgium
We present a Threshold Accepting metaheuristic for the vehicle routing problem with time windows. The suggested metaheuristic is used to improve results obtained with a recent genetic algorithm and multi-start local search; this illustrates its value as a post-processor. The results for the 356 well-known benchmark problems indicate that our method is efficient and competitive.
|
15:55 |
Information Management in a Cooperative Method for the Vehicle Routing Problem with Time Windows |
|
Alexandre Le Bouthillier, Université de Montréal, C.R.T., C.P. 6128, Succ. Centre-ville, Montréal, Québec, Canada, H3C 3J7
Teodor Gabriel Crainic, Université du Québec à Montréal, C.R.T., C.P. 6128, Succ. Centre-ville, Montréal, Québec, Canada, H3C 3J7
In this presentation on Cooperative Search, we will show the influence of incorporating in the selection mechanism of a data warehouse a preference for good solutions with a maximum number of different arcs from an average pattern of utilized arcs in elite solutions. A new dynamic selection function specific to each search method (tabu search, evolutionary algorithm) will allow different schemes of solutions selection from the data warehouse. These mechanisms will provide greater diversity in solution exploration on P processors. Results will be presented on standard and extended benchmark problems of VRPTW.
|
|