Priority rules are often used in production planning and control for sequence planning of production orders to optimise production efficiency based on
key figures such as order throughput time, machine utilisation or production output. Compared to priority rules, reinforcement learning agents are adaptive and can adjust to dynamic production conditions. This offers enormous potential for optimising production control. This study shows the successful implementation of a reinforcement learning agent that is trained with the Proximal Policy Optimisation algorithm. The agent is used to adjust the sequence of transport orders with the aim of achieving space-efficient production with short lead times. The results of the simulation study show an improvement in lead time and space efficiency compared to conventional priority rules.