Authors:
Angel A. Juan
1
;
Daniel Riera
1
;
David Masip
1
;
Josep Jorba
1
and
Javier Faulin
2
Affiliations:
1
Open University of Catalonia, Spain
;
2
Public University of Navarra, Spain
Keyword(s):
Vehicle routing problem, Hybrid algorithms, Heuristics, Simulation, Decision support systems.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence and Decision Support Systems
;
Enterprise Information Systems
;
Industrial Applications of Artificial Intelligence
;
Strategic Decision Support Systems
Abstract:
The aim of this work is to present a simulation-based algorithm that not only provides a competitive solution for instances of the Capacitated Vehicle Routing Problem (CVRP), but is also able to efficiently generate a full database of alternative good solutions with different characteristics. These characteristics are related to solution’s properties such as routes’ attractiveness, load balancing, non-tangible costs, fuzzy preferences, etc. This double-goal approach can be specially interesting for the decision-maker, since he/she can make use of this algorithm to construct a database of solutions and then send queries to it in order to obtain those feasible solutions that better fit his/her utility function without incurring in a severe increase in costs. In order to provide high-quality solutions, our algorithm combines a CVRP classical heuristic, the Clarke and Wright Savings method, with Monte Carlo simulation using state-of-the-art random number generators. The resulting algorit
hm is tested against some well known benchmarks and the results obtained so far are promising enough to encourage future developments and improvements on the algorithm and its applications in real-life scenarios.
(More)