In this paper, we focus on understanding the joint problem of container ship route generation and consolidation center selection, two important sub-problems influencing the effectiveness of the liners shipping industry, which addresses the ship-routing problem. Two different metaheuristics procedures are presented that both consist of two stages: a solution construction phase (either nearest neighborhood with greedy randomize and Clark and Wright with greedy randomize selection) and a solution improvement phase, based on local search. Both metaheuristics are compared in terms of quality of solution, robustness analysis and computing time under variety of instances, ranging from small to large. A thorough comparison evaluation uncovers that both metaheuristics are close-to-each other. An argument in favor of the nearest neighborhood with greedy randomize approach is that it produces better performance than Clark and Wright configuration. Additionally, through sensitivity analysis, we investigate and test two hypotheses in this paper. |
