From Theory to Practice with RAVEN-UCB: Addressing Non-Stationarity in Multi-Armed Bandits through Variance Adaptation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Fang, Junyi, Chen, Yuxun, Chen, Yuxin, Zhang, Chen
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913872397467648
author Fang, Junyi
Chen, Yuxun
Chen, Yuxin
Zhang, Chen
author_facet Fang, Junyi
Chen, Yuxun
Chen, Yuxin
Zhang, Chen
contents The Multi-Armed Bandit (MAB) problem is challenging in non-stationary environments where reward distributions evolve dynamically. We introduce RAVEN-UCB, a novel algorithm that combines theoretical rigor with practical efficiency via variance-aware adaptation. It achieves tighter regret bounds than UCB1 and UCB-V, with gap-dependent regret of order $K σ_{\max}^2 \log T / Δ$ and gap-independent regret of order $\sqrt{K T \log T}$. RAVEN-UCB incorporates three innovations: (1) variance-driven exploration using $\sqrt{\hatσ_k^2 / (N_k + 1)}$ in confidence bounds, (2) adaptive control via $α_t = α_0 / \log(t + ε)$, and (3) constant-time recursive updates for efficiency. Experiments across non-stationary patterns - distributional changes, periodic shifts, and temporary fluctuations - in synthetic and logistics scenarios demonstrate its superiority over state-of-the-art baselines, confirming theoretical and practical robustness.
format Preprint
id arxiv_https___arxiv_org_abs_2506_02933
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle From Theory to Practice with RAVEN-UCB: Addressing Non-Stationarity in Multi-Armed Bandits through Variance Adaptation
Fang, Junyi
Chen, Yuxun
Chen, Yuxin
Zhang, Chen
Machine Learning
I.2.6; I.2.8; G.3
The Multi-Armed Bandit (MAB) problem is challenging in non-stationary environments where reward distributions evolve dynamically. We introduce RAVEN-UCB, a novel algorithm that combines theoretical rigor with practical efficiency via variance-aware adaptation. It achieves tighter regret bounds than UCB1 and UCB-V, with gap-dependent regret of order $K σ_{\max}^2 \log T / Δ$ and gap-independent regret of order $\sqrt{K T \log T}$. RAVEN-UCB incorporates three innovations: (1) variance-driven exploration using $\sqrt{\hatσ_k^2 / (N_k + 1)}$ in confidence bounds, (2) adaptive control via $α_t = α_0 / \log(t + ε)$, and (3) constant-time recursive updates for efficiency. Experiments across non-stationary patterns - distributional changes, periodic shifts, and temporary fluctuations - in synthetic and logistics scenarios demonstrate its superiority over state-of-the-art baselines, confirming theoretical and practical robustness.
title From Theory to Practice with RAVEN-UCB: Addressing Non-Stationarity in Multi-Armed Bandits through Variance Adaptation
topic Machine Learning
I.2.6; I.2.8; G.3
url https://arxiv.org/abs/2506.02933