Discretization and Representation of a Complex Environment for On-Policy Reinforcement Learning for Obstacle Avoidance for Simulated Autonomous Mobile Agents

Abstract

In recent years, the demand for digitalization, automation, and smart systems in the airline industry has accelerated. Furthermore, due to the ongoing global pandemic as of 2022, airlines are faced with the challenge of offering flexibility in both cargo and passenger capacity. Studies show that the use of smart products and autonomous agents are expected to play a key part in the digital transformation of the logistics industry. This paper aims to examine the current state-of-the-art in multi-agent systems and reinforcement learning with special interest in intelligent baggage handling systems. How to simplify, implement, and simulate a system of autonomous baggage carts as a software model in order to examine congestion situations will be the main topics of this paper. Furthermore, how the findings from the software model may be applied to real-world scenarios related to Industry 4.0, and baggage handling will also be discussed.