Q-Learning-Based Sailing Speed Optimization for Ocean-Going Liners Under the EU ETS: Considering Shipper Satisfaction

With the formal inclusion of the shipping industry in the European Union Emissions Trading System (EU ETS), the speed optimization of ocean-going container ships must simultaneously balance operating costs, incorporating carbon emission costs and shipper satisfaction with transportation timeliness. Taking ocean-going container liner routes as the research object, this paper establishes a ship navigation resistance model based on meteorological and hydrological conditions, and constructs a route segmentation mechanism and a ship fuel consumption model on this basis. The spatially differentiated carbon accounting rules of the EU ETS are introduced, a fuzzy membership function is adopted to quantify shipper satisfaction, and a Q-learning-based solution algorithm for ship speed optimization that balances operating costs and shipper satisfaction is designed. Numerical experiments on a 20,150 Twenty-foot Equivalent Unit (TEU) container ship demonstrate that the proposed framework reduces total operating costs by 5.56%, EU ETS carbon compliance costs by 18.72%, and total voyage carbon emissions by 11.01% compared with the conventional constant-speed strategy. Meanwhile, the algorithm can spontaneously form an optimal speed strategy adapted to meteorological conditions and policy rules. Through parameter sensitivity analysis, this paper further extracts management implications for liner-operating companies.