Data-Driven Brownian Reflection Control

Abstract

We study a data-driven reflection control problem for a Brownian model with unknown drift and volatility. We first propose a learn-then-optimize (LTO) algorithm: it estimates the policy-relevant parameter during exploration, plugs the estimate into the optimality equation, and exploits the resulting policy---achieving an O(√T) finite-time expected regret bound. We further propose two algorithms, adaptive-updating (AU) and full-history adaptive-updating (AU-FH), which continuously update the estimator and reflecting level, attaining an improved O(log T) regret bound. Notably, AU-FH algorithms leverages all historical data, yielding better performance in numerical simulations. Our analysis decomposes regret into exploration, transient, and learning components. Transient regret from nonstationarity is bounded by the time-integrated deviation of the transition semigroup from stationarity evaluated on the holding cost, which can be further bounded via a Foster-Lyapunov inequality for exponential convergence of the controlled reflected Brownian motion (RBM). For learning regret, we establish local regularity properties together with consistency and mean-squared error bounds for the estimator, which control the stationary cost gap between the learned and optimal reflection policies. In addition, we leverage the monotonicity of the moving boundary Skorokhod map to derive moment bounds for the AU algorithms' switching states, via pathwise comparison with fixed-boundary RBMs.

Necessary and sufficient conditions for optimal control of coupled forward-backward stochastic control systems driven by G-Brownian motion
Xiaojuan Li, Mingshang Hu
arXiv21 Oct 2026PreprintarXiv licenceFull textOptimisation & controlarXiv
Parametric estimation for stochastic wave equations driven by an infinite dimensional fractional Brownian motion
B. L. S. Prakasa Rao
arXiv3 Sep 2026PreprintCC BY 4.0Full textProbabilityarXiv
The Numerical Onsager-Machlup action functional for Euler discretized SDEs driven by fractional Brownian motion
Shanqi Liu, Qiqi Lian, Jinqiao Duan +1
arXiv28 Sep 2026PreprintarXiv licenceFull textDynamical systems Numerical analysis & scientific computingarXiv
Weyl Pseudo Almost Periodic Type Solutions to Semilinear Stochastic Evolution Equations Driven by Fractional Brownian Motion
Dimplekumar N. Chalishajar, Marko Kostic, Daniel Velinov
arXiv21 Sep 2026PreprintCC BY 4.0Full textAnalysis & PDEs ProbabilityarXiv
Euler Scheme for Stochastic Functional Differential Equations Driven by Fractional Brownian Motion via Fractional Calculus Techniques
Alexander Abreu, Héctor Araya, Lisandro Fermin +2
HAL21 Sep 2026PreprintCC BY 4.0Full textNumerical analysis & scientific computing ProbabilityHALarXiv

Cite

BibTeX RIS

BibTeX

@misc{pang2026datadriven,
  title = {{Data-Driven Brownian Reflection Control}},
  author = {Guodong Pang and Dacheng Yao and Hao Yin},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.03870},
  eprint = {2609.03870},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.03870},
}

RIS

TY  - UNPB
TI  - Data-Driven Brownian Reflection Control
AU  - Guodong Pang
AU  - Dacheng Yao
AU  - Hao Yin
PY  - 2026
DA  - 2026/09/03
PB  - arXiv
DO  - 10.48550/arxiv.2609.03870
UR  - https://arxiv.org/abs/2609.03870
AB  - We study a data-driven reflection control problem for a Brownian model with unknown drift and volatility. We first propose a learn-then-optimize (LTO) algorithm: it estimates the policy-relevant parameter during exploration, plugs the estimate into the optimality equation, and exploits the resulting policy---achieving an O(√T) finite-time expected regret bound. We further propose two algorithms, adaptive-updating (AU) and full-history adaptive-updating (AU-FH), which continuously update the estimator and reflecting level, attaining an improved O(log T) regret bound. Notably, AU-FH algorithms leverages all historical data, yielding better performance in numerical simulations. Our analysis decomposes regret into exploration, transient, and learning components. Transient regret from nonstationarity is bounded by the time-integrated deviation of the transition semigroup from stationarity evaluated on the holding cost, which can be further bounded via a Foster-Lyapunov inequality for exponential convergence of the controlled reflected Brownian motion (RBM). For learning regret, we establish local regularity properties together with consistency and mean-squared error bounds for the estimator, which control the stationary cost gap between the learned and optimal reflection policies. In addition, we leverage the monotonicity of the moving boundary Skorokhod map to derive moment bounds for the AU algorithms' switching states, via pathwise comparison with fixed-boundary RBMs.
ER  -

Abstract

Related

Cite