Adaptive Discretization in Online Reinforcement Learning. (arXiv:2110.15843v3 [stat.ML] UPDATED)

Discretization based approaches to solving online reinforcement learning
problems have been studied extensively in practice on applications ranging from
resource allocation to cache management. Two major questions in designing
discretization-based algorithms are how to create the discretization and when
to refine it. While there have been several experimental results investigating
heuristic solutions to these questions, there has been little theoretical
treatment. In this paper we provide a unified theoretical analysis of
tree-based hierarchical partitioning methods for online reinforcement learning,
providing …

arxiv online reinforcement learning reinforcement reinforcement learning