IMI presents research at LION19-conference in Prague

IMI presented a new research paper at the 19th Learning and Intelligent Optimization (LION) conference, which took place in Prague from 15 to 18 June. The paper was written as part of the NextGenerationEU-funded FlexTools project.
The paper entitled “Reinforcement Learning for AMR Charging Decisions: The Impact of Reward and Action Space Design” deals with battery charging decisions of Autonomous Mobile Robots (AMRs) in block storage systems.
The aim is to use reinforcement learning (RL) to learn intelligent strategies for the timing and duration of charging processes. A fine-grained simulation environment was used for training and evaluation.
The results show that RL-based loading decisions can significantly outperform conventional heuristic strategies, especially with regard to shorter service times and the availability of AMRs. The work thus makes an important contribution to the practical optimization of autonomous warehouse processes.