Distributed Learning and Optimal Assignment in Multiplayer Heterogeneous Networks

Abstract--We consider an ad hoc network where multiple users access the same set of channels. The channel characteristics are unknown and could be different for each user (heterogeneous). No controller is available to coordinate channel selections by the users, and if multiple users select the same channel, they collide and none of them receive any rate (or reward). For such a completely decentralized network we develop algorithms that aim to achieve optimal network throughput. Due to lack of any direct communication between the users, we allow each user to exchange information by transmitting in a specific pattern and sense such transmissions from others. However, such transmissions and sensing for information exchange do not add to network throughput. For the wideband sensing and narrowband sensing scenarios, we first develop explore-and-commit algorithms that converge to near-optimal allocation with high probability in a small number of rounds. Building on this, we develop an algorithm that gives logarithmic regret, even when the number of users changes with time. I. INTRODUCTION Cognitive cellular networks are one of the key components of the next generation wireless networks. It promises seamless and high-speed connectivity by combining the features of both cellular and ad hoc networks [1, 2], which is a much-desired requirement in all mission-critical and Internet of Things (IoT) applications. In the ad hoc component of such networks, a central controller may not always exist which makes the coordination amongthe users challenging. Further, users may not know characteristics (mean rewards) of the available channels, and these characteristics could be statistically different across users due to their geographical separations. Thus for effective utilization of network resources, users not only need to learn the channel characteristics experienced by them but also that experienced by the others. This work develops distributed learning algorithms for such networks that achieve optimal network performance using signaling schemes. A. Sensing and Signaling Ad hoc networks are usually dynamic in nature and users may not know how many others are present in the network.