Automatic Generation of Desk and Chair Layouts for Cafes using PCGRL

Abstract

PCG (Procedural Content Generation) is the automatic generation of content through computation, and PCGML(Procedural Contents Generatiaon via Machine Learning) is a method that uses machine learning, while PCGRL(Procedural Contents Generatiaon via Reinforcement Learning) is a method that uses reinforcement learning. In this research, reinforcement learning was performed by having an agent to place chairs and desks in a cafe in a 3D virtual space and feeding their evaluations back to the agent. As a result, a method to control the various arrangement of desks and chairs in a café by means of a reward function is established.