Learning Mechanisms and Local Search Heuristics for the Fixed Charge Capacitated Multicommodity Network Design
In this paper, we propose a method based on learning mechanisms to address the fixed charge capacitated multicommodity network design problem. Learning mechanisms are applied on each solution to extract meaningful fragments to build a pattern solution. Cycle-based neighborhoods are used both to generate solutions and to move along a path leading to the pattern solution by a tabu-like local search procedure. Within this concept, the method integrates important mechanisms such as intensification and diversification. Experimental results show that the proposed algorithm is effective for large structured instances with several commodities.
Keywords: Adaptive memories, Tabu search, fixed charge capacitated multicommodity network design, Meta-heuristics, Cycle-based neighborhoods
Download Full-Text
ABOUT THE AUTHORS
Ilfat Ghamlouche
Faculté des Sciences Économiques et de Gestion, Université Libanaise Beirut, Hadath C.P: 6573-14 , Lebanon
Teodor Gabriel Crainic
Département de management et technologie Université du Québec à Montréal and CIRRELT, Université de Montréal Montréal, Québec, Canada
Michel Gendreau
CIRRELT, Université de Montréal Montréal, Québec, Canada
Ihab Sbeity
Faculté des Sciences, Université Libanaise Beirut, Hadath C.P: 6573-14 , Lebanon
Ilfat Ghamlouche
Faculté des Sciences Économiques et de Gestion, Université Libanaise Beirut, Hadath C.P: 6573-14 , Lebanon
Teodor Gabriel Crainic
Département de management et technologie Université du Québec à Montréal and CIRRELT, Université de Montréal Montréal, Québec, Canada
Michel Gendreau
CIRRELT, Université de Montréal Montréal, Québec, Canada
Ihab Sbeity
Faculté des Sciences, Université Libanaise Beirut, Hadath C.P: 6573-14 , Lebanon
