Regret bounds for Narendra-Shapiro bandit algorithms

DOI10.1080/17442508.2018.1457675zbMATH Open1498.60303arXiv1502.04874OpenAlexW1693596837MaRDI QIDQ5086451FDOQ5086451

Authors: Sébastien Gadat, Fabien Panloup, Sofiane Saadane 

Publication date: 5 July 2022

Published in: Stochastics (Search for Journal in Brave)

Abstract: Narendra-Shapiro (NS) algorithms are bandit-type algorithms that have been introduced in the sixties (with a view to applications in Psychology or learning automata), whose convergence has been intensively studied in the stochastic algorithm literature. In this paper, we adress the following question: are the Narendra-Shapiro (NS) bandit algorithms competitive from a extit{regret} point of view? In our main result, we show that some competitive bounds can be obtained for such algorithms in their penalized version (introduced in cite{Lamberton_Pages}). More precisely, up to an over-penalization modification, the pseudo-regret $$ related to the penalized two-armed bandit algorithm is uniformly bounded by $Csqrtn$ (where $C$ is made explicit in the paper). oindent We also generalize existing convergence and rates of convergence results to the multi-armed case of the over-penalized bandit algorithm, including the convergence toward the invariant measure of a Piecewise Deterministic Markov Process (PDMP) after a suitable renormalization. Finally, ergodic properties of this PDMP are given in the multi-armed case.

Full work available at URL: https://arxiv.org/abs/1502.04874

Recommendations

• Bounded Regret for Finitely Parameterized Multi-Armed Bandits
• Regret and Convergence Bounds for a Class of Continuum-Armed Bandit Problems
• Regret lower bound and optimal algorithm for high-dimensional contextual linear bandit
• Regret bounds for restless Markov bandits
• Regret Bounds for Restless Markov Bandits
• Regret analysis of stochastic and nonstochastic multi-armed bandit problems
• Near-optimal regret bounds for reinforcement learning
• Near-optimal regret bounds for Thompson sampling
• Deviations of stochastic bandit regret
• Constrained regret minimization for multi-criterion multi-armed bandits

zbMATH Keywords

regretpiecewise deterministic Markov processesstochastic bandit algorithms

Mathematics Subject Classification ID

Sequential statistical analysis (62L10) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Continuous-time Markov processes on general state spaces (60J25)

Cites Work

• Stochastic approximation methods for constrained and unconstrained systems
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• The Nonstochastic Multiarmed Bandit Problem
• Some aspects of the sequential design of experiments
• 10.1162/153244303321897663
• Kullback-Leibler upper confidence bounds for optimal sequential allocation
• Total variation estimates for the TCP process
• When can the two-armed bandit algorithm be trusted?
• On the linear model with two absorbing barriers
• Use of Stochastic Automata for Parameter Self-Optimization with Multimodal Performance Criteria
• A penalized bandit algorithm
• How Fast Is the Bandit?
• Long time behavior of Markov processes and beyond

Cited In (3)

• Regret and Convergence Bounds for a Class of Continuum-Armed Bandit Problems
• Regret Bounds for Restless Markov Bandits
• Bounded Regret for Finitely Parameterized Multi-Armed Bandits

This page was built for publication: Regret bounds for Narendra-Shapiro bandit algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5086451)