Fast Change Identification in Multi-Play Bandits and its Applications in Wireless Networks

원문 초록

Next-generation wireless services are characterized by a diverse set of requirements, to sustain which, the wireless access points need to probe the users in the network periodically. In this regard, we study a novel multi-armed bandit (MAB) setting that mandates probing all the arms periodically while keeping track of the best current arm in a non-stationary environment. In particular, we develop \texttt{TS-GE} that balances the regret guarantees of classical Thompson sampling (TS) with the broadcast probing (BP) of all the arms simultaneously in order to actively detect a change in the reward distributions. The main innovation in the algorithm is in identifying the changed arm by an optional subroutine called group exploration (GE) that scales as $\log_2(K)$ for a $K-$armed bandit setting. We characterize the probability of missed detection and the probability of false-alarm in terms of the environment parameters. We highlight the conditions in which the regret guarantee of \texttt{TS-GE} outperforms that of the state-of-the-art algorithms, in particular, \texttt{ADSWITCH} and \texttt{M-UCB}. We demonstrate the efficacy of \texttt{TS-GE} by employing it in two wireless system application - task offloading in mobile-edge computing (MEC) and an industrial internet-of-things (IIoT) network designed for simultaneous wireless information and power transfer (SWIPT).