Dynamic pilot allocation over Markovian fading channels: A restless bandit approach

Maialen Larrañaga (1, 2) , Mohamad Assaad (1) , Apostolos S Destounis (3) , Georgios S Paschos (3)

1 L2S - Laboratoire des signaux et systèmes
2 CentraleSupélec
3 Huawei Technologies France [Boulogne-Billancourt]

—We investigate a pilot allocation problem in wireless networks over Markovian fading channels. In wireless systems, the Channel State Information (CSI) is collected at the Base Station (BS) through either a feedback channel (FDD mode) or a pilot-aided channel estimation method (TDD mode). This paper focuses on the latter. Typically, there are less available pilots than users, hence at each slot the scheduler needs to decide an allocation of pilots to users with the goal of maximizing the long-term average throughput. A trade-off emerges between exploiting users with up-to-date CSI for immediate gains or, exploring users with outdated CSI for a potential larger future gain. As we show, the arising pilot allocation problem is a restless bandit problem and thus its optimal solution is out of reach. In this paper, we propose a Lagrangian relaxation approach to obtain a Whittle index policy, which represents a low-complexity heuristic solution with remarkably good performance.

Domaines: Réseaux et télécommunications [cs.NI]; Optimisation et contrôle [math.OC]

Maialen Larrañaga, Mohamad Assaad, Apostolos S Destounis, Georgios S Paschos. Dynamic pilot allocation over Markovian fading channels: A restless bandit approach. 2016 IEEE Information Theory Workshop (ITW), Sep 2016, Cambridge, United Kingdom. pp.290 - 294, ⟨10.1109/ITW.2016.7606842⟩. ⟨hal-01578943⟩