Saved in:
Bibliographic Details
Main Authors: Yang, Hyukjun, Lim, Han-Dong, Lee, Donghwan
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.06837
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914457602490368
author Yang, Hyukjun
Lim, Han-Dong
Lee, Donghwan
author_facet Yang, Hyukjun
Lim, Han-Dong
Lee, Donghwan
contents The problem of solving Markov decision processes under function approximation remains a fundamental challenge, even under linear function approximation settings. A key difficulty arises from a geometric mismatch: while the Bellman optimality operator is contractive in the Linfty-norm, commonly used objectives such as projected value iteration and Bellman residual minimization rely on L2-based formulations. To enable gradient-based optimization, we consider a soft formulation of Bellman residual minimization and extend it to a generalized weighted Lp -norm. We show that this formulation aligns the optimization objective with the contraction geometry of the Bellman operator as p increases, and derive corresponding performance error bounds. Our analysis provides a principled connection between residual minimization and Bellman contraction, leading to improved control of error propagation while remaining compatible with gradient-based optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2604_06837
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Contraction-Aligned Analysis of Soft Bellman Residual Minimization with Weighted Lp-Norm for Markov Decision Problem
Yang, Hyukjun
Lim, Han-Dong
Lee, Donghwan
Machine Learning
The problem of solving Markov decision processes under function approximation remains a fundamental challenge, even under linear function approximation settings. A key difficulty arises from a geometric mismatch: while the Bellman optimality operator is contractive in the Linfty-norm, commonly used objectives such as projected value iteration and Bellman residual minimization rely on L2-based formulations. To enable gradient-based optimization, we consider a soft formulation of Bellman residual minimization and extend it to a generalized weighted Lp -norm. We show that this formulation aligns the optimization objective with the contraction geometry of the Bellman operator as p increases, and derive corresponding performance error bounds. Our analysis provides a principled connection between residual minimization and Bellman contraction, leading to improved control of error propagation while remaining compatible with gradient-based optimization.
title Contraction-Aligned Analysis of Soft Bellman Residual Minimization with Weighted Lp-Norm for Markov Decision Problem
topic Machine Learning
url https://arxiv.org/abs/2604.06837