Mitigating Dimensionality in 2D Rectangle Packing Problem under Reinforcement Learning Schema
This work addresses packing optimization for logistics and manufacturing, but it is incremental as it builds on existing RL methods without surpassing heuristic benchmarks.
The paper tackled the two-dimensional rectangular packing problem by applying Reinforcement Learning with a reduced state-action representation and UNet+PPO, achieving performance comparable to the MaxRect heuristic.
This paper explores the application of Reinforcement Learning (RL) to the two-dimensional rectangular packing problem. We propose a reduced representation of the state and action spaces that allow us for high granularity. Leveraging UNet architecture and Proximal Policy Optimization (PPO), we achieved a model that is comparable to the MaxRect heuristic. However, our approach has great potential to be generalized to nonrectangular packing problems and complex constraints.