The four publications within the cumulative part of this thesis focus on the layout planning of manufacturing plants. After the systematic literature review in publication I to get a comprehensive overview of ML and especially RL approaches for FLPs, a software library to formalise the FLP as MDP, a proof of concept for optimising layouts with RL, and a generalisation study are presented as design artefacts in publications II through IV. The results presented in this work demonstrate that the underlying formulation of the FLP as an MDP is valid, and that selected RL algorithms are capable of learning strategies for sequential actions that minimise a facility layout from a given initial machine positioning setup.