Finite-time analysis of the multiarmed bandit problem

Machine Learning

Volumn 47, Issue 2-3, 2002, Pages 235-256

  Auer, Peter ^a   Cesa Bianchi, Nicolò ^b   Fischer, Paul ^c  

^a GRAZ UNIVERSITY OF TECHNOLOGY   (Austria)

^b UNIVERSITY OF MILAN   (Italy)

^c TU DORTMUND UNIVERSITY   (Germany)

Author keywords

Adaptive allocation rules; Bandit problems; Finite horizon regret

Indexed keywords

COMPUTATION THEORY; PROBLEM SOLVING; STATISTICS; THEOREM PROVING;

FINITE-TIME ANALYSIS;

LEARNING SYSTEMS;

EID: 0036568025     PISSN: 08856125     EISSN: None     Source Type: Journal    
DOI: 10.1023/A:1013689704352     Document Type: Article

References (12)

1. Sample mean based index policies with O(log n) regret for the multi-armed bandit problem
2. Optimal adaptive policies for sequential allocation problems
3. Q-learning for bandit problems
4. Multi-armed bandit allocation indices
5. Multi-armed bandit problem revisited
6. Asymptotically efficient adaptive allocation rules
7. Nonparametric bandit methods