September 25, 2024
Automated Guided Vehicle Systems (AGVS) represent a new and challenging area in logistics,
with high performance demands driven by the exponential growth of global product traffic. This
paper proposes a novel approach to address AGVS challenges using deep reinforcement learning
algorithms as an alternative to traditional methods.
Classical approaches typically rely on
predefined vehicle movement rules, whereas our method derives these movement rules through a
trial-and-error reinforcement learning strategy. Our approach has demonstrated effective results
in relatively small areas with a limited number of agents.
For more details: View article document
