AAAI Conference Demonstration (2026)
Sunghoon Hong, Junseok Park, Whiyoung Jung, Deunsol Yoon, Woohyung Lim, Soonyoung Lee, Kanghoon Lee
Abstract
Industrial automation in smart logistics and factories requires simulation platforms that support rapid environment building before costly physical deployment. Yet existing tools often require substantial expertise, complex setup, and long configuration times, hindering agile prototyping. We present RAPID, a simulation platform with two components: layout design, which enables intuitive visual configuration of factory layouts, and behavior simulation and validation, which allow users to attach behavior models and evaluate system performance. RAPID lowers the entry barrier to industrial simulation, letting users apply existing behavior models or trained reinforcement learning (RL) agents to new layouts with minimal effort. This approach lets practitioners prototype facilities in minutes rather than weeks and gives researchers a standardized environment for benchmarking multi-agent RL and coordination algorithms. By combining rapid design with simulation-based validation, RAPID accelerates automation development from concept to implementation.