Constrained episodic reinforcement learning in concave-convex and knapsack settings
This work addresses constrained reinforcement learning for scenarios requiring theoretical guarantees and practical performance, though it appears incremental by extending beyond linear constraints and single-episode settings.
The paper tackles constrained episodic reinforcement learning by proposing an algorithm for settings with concave rewards and convex constraints, as well as hard knapsack constraints, and demonstrates that it significantly outperforms previous approaches in existing environments.
We propose an algorithm for tabular episodic reinforcement learning with constraints. We provide a modular analysis with strong theoretical guarantees for settings with concave rewards and convex constraints, and for settings with hard constraints (knapsacks). Most of the previous work in constrained reinforcement learning is limited to linear constraints, and the remaining work focuses on either the feasibility question or settings with a single episode. Our experiments demonstrate that the proposed algorithm significantly outperforms these approaches in existing constrained episodic environments.