Utilizing drones for delivery is an effective approach to enhancing delivery efficiency and lowering expenses. However, to overcome the delivery range and payload capacity limitations of drones, the combination of trucks and drones is gaining more attention. By using trucks as a flight platform for drones and supporting their take-off and landing, the delivery range and capacity can be greatly extended. This research focused on mixed truck-drone delivery and utilized reinforcement learning and real road networks to address its optimal scheduling issue. Furthermore, the state and behavior of the vehicle were optimized to reduce meaningless behavior, especially the optimization of truck travel trajectory and customer service time. Finally, a comparison with other reinforcement learning algorithms with behavioral constraints demonstrated the reasonableness of the problem and the advantages of the algorithm.
Deep Reinforcement Learning for Truck-Drone Delivery Problem
2023
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Metadata by DOAJ is licensed under CC BY-SA 1.0
Truck-Drone Delivery Optimization Based on Multi-Agent Reinforcement Learning
DOAJ | 2024
|Genetic-Algorithm-Aided Deep Reinforcement Learning for Multi-Agent Drone Delivery
DOAJ | 2024
|