Navigational path optimization methodology for wing-diesel hybrid ship based on improved Double Deep Q-Network

Wing-diesel hybrid ships can demonstrate quantifiable fuel efficiency gains by using wing-sails’ thrust to replace part of the fuel propulsion during navigation. But ships can be affected by factors such as wind, waves, currents and shore wall effects during navigation, thus necessitating precise pa...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Ocean engineering Ročník 341; s. 122543
Hlavní autori: Wang, Cong, Huang, Lianzhong, Ma, Ranqi, Wang, Kai, Sheng, Jinlu, Ruan, Zhang, Zhang, Rui, Cao, JianLin
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier Ltd 01.12.2025
Predmet:
Double Deep Q Network algorithm
Fuel Consumption Prediction Model
Reward Function
Route planning
Wing-diesel hybrid ship
Double Deep Q Network algorithm
Fuel Consumption Prediction Model
Route planning
Reward Function
Wing-diesel hybrid ship
ISSN:0029-8018
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Wing-diesel hybrid ships can demonstrate quantifiable fuel efficiency gains by using wing-sails’ thrust to replace part of the fuel propulsion during navigation. But ships can be affected by factors such as wind, waves, currents and shore wall effects during navigation, thus necessitating precise path planning methodologies to ensure both economic efficiency and operational safety. Therefore, this study develops an improved Double Deep Q Network (DDQN) algorithm for optimizing the navigation path planning of the wing-diesel hybrid ship. First, a ship fuel consumption prediction model based on XGBoost algorithm and a ship motion model are established. Then, the reward function is improved by incorporating fuel consumption, heading, distance, time, and collisions as rewards or penalties. Afterward, DDQN is introduced to learn the action-reward model, and the learning results are used to control the ship's movement. By pursuing higher reward values, the ship can autonomously find the optimal low-fuel consumption path. Experimental verification was conducted on the target voyage of the “New Aden” in the Arabian Sea. The results demonstrate that the proposed method effectively enhances the energy efficiency of wing-diesel hybrid ships, reducing fuel consumption and the Energy Efficiency Operational Indicator (EEOI) by approximately 6.35 % and 7.98 %, respectively. •A cooperative optimization method for the wind-assisted hybrid ship is proposed.•Improved Double Deep Q-Network is designed.•XGBoost algorithm is used for predicting ship fuel consumption.•Reward function is improved to optimize operational efficiency.•Interpolation method is adopted to align meteorological data with ship data.
ISSN:0029-8018
DOI:10.1016/j.oceaneng.2025.122543