Title:Version Age of Information Minimization over Fading Broadcast Channels

Abstract:We consider a base station (BS) that receives version update packets from multiple exogenous streams and broadcasts them to corresponding destination users over a fading broadcast channel using a non-orthogonal multiple access (NOMA) scheme. In each stream, packets arrive randomly, with each new packet's index being one higher than that of the immediate previous packet. Moreover, the arrival of a new version update renders previous versions obsolete. In this case, we consider the version age of information (VAoI) at a user, defined as the difference in the version index of the latest available packet at the BS and that at the user, as a metric of freshness of information. Our objective is to minimize a weighted sum of the long-term expected average VAoI across users and average power by optimally scheduling the update packets from different streams for transmission and transmitting them with appropriate powers to guarantee their successful delivery. We obtain the optimal policy within the class of channel-only stationary randomized policies (CO-SRP), which make transmission decisions based only on the channel power gain realizations at each decision time, and the optimal Markov decision process (MDP) based solution, along with its structural properties. Via numerical simulations, we show that the proposed optimal CO-SRP performs close to the MDP-based solution. Additionally, a time division multiple access (TDMA) scheme that allows transmission to at most one user at any given time instant exhibits comparable performance to NOMA when the average power consumed is low. However, as the average power consumed increases, NOMA outperforms TDMA.