Deep Reinforcement Learning Based Real-Time Open-Pit Truck Dispatching System

Open-pit mines are highly dynamic and uncertain environments with complex interactions between haulage and loading equipment on a shared road network. Truck fleet management systems play a crucial role in providing real-time assignment of trucks for bulk material transportation in mining operations to ensure efficient utilization of the mine equipment assets and achieve the different mining and ore processing targets. In this research, we propose developing a Deep Reinforcement Learning (DRL) based truck dispatching system for open-pit operations using a Double Deep Q-Learning algorithm. A discrete event simulation model of the open-pit truck and shovel environment is developed to capture uncertainties throughout the equipment operating cycle and train the DRL truck dispatching system to learn a stochastic control policy. A case study is presented in an iron ore deposit where the trained agent manages to learn a robust dispatching policy to achieve the ore and waste mining targets and maintain the metal concentration of the ore feed to the processing plants within a desired range