Budget-Constrained Multi-Armed Bandits with Multiple Plays
Многорукие бандиты с несколькими выборами и ограничением бюджета
2017-11-16
SCID: 54.1/m96fs3ww
Discuss with AI
Exp3 algorithmadversarial banditsbudget-constrained multi-armed banditsmultiple playsstochastic bandits
Figures from the paper
Abstract (AI)
We study the multi-armed bandit problem with multiple plays and a budget constraint for both the stochastic and the adversarial setting. At each round, exactly $K$ out of $N$ possible arms have to be played (with $1\leq K \leq N$). In addition to observing the individual rewards for each arm played, the player also learns a vector of costs which has to be covered with an a-priori defined budget $B$. The game ends when the sum of current costs associated with the played arms exceeds the remaining budget. Firstly, we analyze this setting for the stochastic case, for which we assume each arm to have an underlying cost and reward distribution with support $[c_{\min}, 1]$ and $[0, 1]$, respectively. We derive an Upper Confidence Bound (UCB) algorithm which achieves $O(NK^4 \log B)$ regret. Secondly, for the adversarial case in which the entire sequence of rewards and costs is fixed in advance, we derive an upper bound on the regret of order $O(\sqrt{NB\log(N/K)})$ utilizing an extension of the well-known $\texttt{Exp3}$ algorithm. We also provide upper bounds that hold with high probability and a lower bound of order $Ω((1 - K/N)^2 \sqrt{NB/K})$.
Key Findings
1
For adversarially fixed reward and cost sequences, an extension of Exp3 achieves O(√(NB log(N/K))) regret.
2
For stochastic rewards and costs, a UCB-based algorithm achieves O(NK^4 log B) regret when arm costs and rewards lie in the specified bounded supports.
3
The adversarial setting also admits high-probability regret guarantees and a lower bound of Ω((1−K/N)^2√(NB/K)).
4
The paper formulates a multiple-play multi-armed bandit problem where exactly K of N arms are selected each round under a shared budget constraint.
Research Object
Budget-constrained multi-armed bandit systems with multiple plays, involving N arms, K simultaneously selected arms, and reward and cost sequences or distributions
Research Subject
Regret bounds and learning performance in stochastic and adversarial settings under budget exhaustion, including UCB- and Exp3-based guarantees
Publication Details
Publication Date
2017-11-16
Journal
Publisher
ISSN
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest