Deep reinforcement learning based optimization of automated guided vehicle time and energy consumption in a container terminal
Darius Drungilas, Mindaugas Kurmis, Audrius Senulis, Žydrūnas Lukošius, Arūnas Andziulis, Jolanta Janutėnienė, Marijonas Bogdevičius, Valdas Jankūnas, Miroslav Vozňák
Abstract
The energy efficiency of port container terminal equipment and the reduction of CO2 emissions are among one of the biggest challenges facing every seaport in the world. The article presents the modeling of the container transportation process in a terminal from the quay crane to the stack using battery-powered Automated Guided Vehicle (AGV) to estimate the energy consumption parameters. An AGV speed control algorithm based on Deep Reinforcement Learning (DRL) is proposed to optimize the energy consumption of container transportation. The results obtained and compared with real transportation measurements showed that the proposed DRL-based approach dynamically changing the driving speed of the AGV reduces energy consumption by 4.6%. The obtained results of the research provide the prerequisites for further research in order to find optimal strategies for autonomous vehicle movement including context awareness and information sharing with other vehicles in the terminal.