The SR-GCWS hybrid algorithm for solving the capacitated vehicle routing problem

  • Authors:

  • Angel A. Juan;Javier Faulin;Rubén Ruiz;Barry Barrios;Santi Caballé

  • Affiliations:

  • Dep. of Computer Science and Telecommunication, Open University of Catalonia, Rambla Poblenou, 156, 08018 Barcelona, Spain;Dep. of Statistics and Operations Research, Public University of Navarre, Campus Arrosadia, 31006 Pamplona, Spain;Grupo de Sistemas de Optimización Aplicada - ITI, Universidad Politécnica de Valencia, Camino de Vera, s/n, 46021 Valencia, Spain;Dep. of Statistics and Operations Research, Public University of Navarre, Campus Arrosadia, 31006 Pamplona, Spain;Dep. of Computer Science and Telecommunication, Open University of Catalonia, Rambla Poblenou, 156, 08018 Barcelona, Spain

  • Venue:
  • Applied Soft Computing
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The capacitated vehicle routing problem (CVRP) is a well known problem which has long been tackled by researchers for several decades now, not only because of its potential applications but also due to the fact that CVRP can be used to test the efficiency of new algorithms and optimization methods. The objective of our work is to present SR-GCWS, a hybrid algorithm that combines a CVRP classical heuristic with Monte Carlo simulation using state-of-the-art random number generators. The resulting algorithm is tested against some well-known benchmarks. In most cases, our approach is able to compete or even outperform much more complex algorithms, which is especially interesting if we consider that our algorithm does not require any previous parameter fine-tuning or set-up process. Moreover, our algorithm has been able to produce high-quality solutions almost in real-time for most tested instances. Another important feature of the algorithm worth mentioning is that it uses a randomized constructive heuristic, capable of generating hundreds or even thousands of alternative solutions with different properties. These alternative solutions, in turn, can be really useful for decision-makers in order to satisfy their utility functions, which are usually unknown by the modeler. The presented methodology may be a fine framework for the development of similar algorithms for other complex combinatorial problems in the routing arena as well as in some other research fields.