ContainerGym: A Real-World Reinforcement Learning Benchmark for Resource Allocation

TL;DR

ContainerGym is a new benchmark derived from a real-world industrial resource allocation problem, designed to evaluate reinforcement learning algorithms on complex, realistic decision-making tasks with varying difficulty levels.

Contribution

It introduces a versatile, real-world inspired benchmark for reinforcement learning that captures practical challenges and enables thorough evaluation of algorithms like PPO, TRPO, and DQN.

Findings

Baseline results show limitations of PPO, TRPO, and DQN on the benchmark.

The benchmark can be configured for different problem complexities.

Statistical analysis highlights specific weaknesses of existing algorithms.

Abstract

We present ContainerGym, a benchmark for reinforcement learning inspired by a real-world industrial resource allocation task. The proposed benchmark encodes a range of challenges commonly encountered in real-world sequential decision making problems, such as uncertainty. It can be configured to instantiate problems of varying degrees of difficulty, e.g., in terms of variable dimensionality. Our benchmark differs from other reinforcement learning benchmarks, including the ones aiming to encode real-world difficulties, in that it is directly derived from a real-world industrial problem, which underwent minimal simplification and streamlining. It is sufficiently versatile to evaluate reinforcement learning algorithms on any real-world problem that fits our resource allocation framework. We provide results of standard baseline methods. Going beyond the usual training reward curves, our results and the statistical tools used to interpret them allow to highlight interesting limitations of well-known deep reinforcement learning algorithms, namely PPO, TRPO and DQN.