.png){kind=icon}
Return
Learning-based dispatching system with trajectory optimisation for autonomous mining transportation
.png){kind=icon}
PRE
AI
DOI：10.1080/17480930.2025.2486316.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This study addresses this challenge by formulating fleet management as a multi-agent reinforcement learning (MARL) problem and integrating a Deep Q-Network (DQN) dispatching system with Dynamic Programming (DP) velocity optimisation. Utilising an event-driven mining simulator based on real mine parameters, the proposed approach achieves a 10% reduction in average energy consumption per ton and increases total production by 3,000 tons over an 8-hour simulation compared to conventional baseline methods. It ensures more equitable task distribution and improved resource utilisation. Additionally, this policy maintains robust performance across varying fleet sizes (20-25 trucks) without retraining, ensuring consistent efficiency in dynamic operational environments.
Keywords:
Energy consumption
fleet management
learning-based method
mining and sustainability
transportation efficiency
word
Journal
IF:
2.6
Papers:
67
Citations:
1.2K
Organization
Cited Papers
ShareSaveShareSaveShareSaveShareSaveThe Use of a Machine Learning Method to Predict the Real-Time Link Travel Time of Open-Pit Trucks
OAAI
OAAIShareSaveShareSavePrediction of fuel consumption of mining dump trucks: A neural networks approach.png)
APPLIED ENERGY
APPLIED ENERGYIF11
PREAI
ShareSaveShareSave
View more