GRASP with adaptive memory programming for finding good starting solutions to the multicommodity capacitated network design problem
Chapter in Scopus
- Additional Document Info
- View All
The capacitated multicommodity network design problem presents a major challenge. At the present time, there exists no exact algorithm that can solve large instances, common in several applications, in a reasonable period of time. In these cases, heuristic methods are being used to find near optimal solutions. Some heuristic procedures need initial solutions or a set of initial solutions to start their search. A GRASP is proposed that gives solutions of acceptable quality. In this paper we adapt the GRASP technique including memory features to the mentioned problem, and the results are compared to the same GRASP without these features. © 2003 Springer Science+Business Media New York.