The uncertainty in maritime transportation dramatically impacts liner shipping efficiency; to improve operational stability, timely recovery of vessel schedule in liner shipping service is essential after disruptions; however, recovery actions usually cause vessel carbon emissions to increase. In order to reduce the effect of disruption and carbon emissions of liner shipping, in this study, we develop a mixed-integer nonlinear programming mathematical model that contains three recovery strategies and considers the costs of the vessel’s voyage in liner shipping service, with emphasis on taking into account the carbon emission and container cumulative delay costs. An adaptive genetic algorithm (RLSGA) is developed to solve the model based on reinforcement learning and simulated annealing algorithm. Numerical experiments are conducted on a real liner shipping route to investigate the effectiveness of the model and method. The results indicate that compared with the CPLEX, the proposed RLSGA can provide a more effective vessel schedule recovery solution after disruption occurrences within 20 seconds of all test cases and has a less than 1% deviation rate. Implementing the model in the liner shipping service can lead to a reduction in carbon emissions, cost savings on voyages, and an obvious decrease in the number of containers affected by disruptions. Moreover, this work also analyzes the potential of clean fuel on vessel schedule recovery and obtains managerial insights. The research findings can provide a meaningful reference for liner shipping companies in operation decision-making.
Schedule Disruption Recovery in Liner Shipping Service Based on a Reinforcement Learning-Enabled Adaptive Genetic Algorithm
IEEE Transactions on Intelligent Transportation Systems ; 25 , 12 ; 21622-21633
01.12.2024
8193767 byte
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Disruption Recovery for a Vessel in Liner Shipping
British Library Online Contents | 2015
|Disruption recovery for a vessel in liner shipping
Online Contents | 2015
|