Location Aware Opportunistic Bandwidth Sharing between Static and Mobile Users with Stochastic Learning in Cellular Networks

In this paper, we consider the problem of location-dependent opportunistic bandwidth sharing between static and mobile (i.e., moving) downlink users in a cellular network. Each cell of the network has some fixed number of static users. Mobile users enter the cell, move inside the cell for some time and then leave the cell. In order to provide higher data rate to the mobile users, we propose location dependent bandwidth sharing between the two classes of static and mobile users; the idea is to provide higher bandwidth to the mobile users at favourable times and locations, and provide higher bandwidth to the static users in other times. Our approach is agnostic to the way the bandwidth is further shared within the same class of users; it can be combined with any particular bandwidth allocation policy employed for one of these two classes of users. We formulate the problem as a long run average reward Markov decision process (MDP) where the per-step reward is a linear combination of instantaneous data volumes received by static and mobile users, and find the optimal policy. The transition structure of this MDP is not known in general, and it may change with time. To alleviate these issues, we propose a learning algorithm based on single timescale stochastic approximation. Also, noting that the MDP problem can be used to maximize the long run average data rate for mobile users subject to a constraint on the long run average data rate of static users, we provide a learning algorithm based on multi-timescale stochastic approximation. We prove asymptotic convergence of the bandwidth sharing policies under these learning algorithms to the optimal policy. The results are extended to address the issue of fair bandwidth sharing between the two classes of static and mobile users, where the notion of fairness is motivated by the popular notion of α-fairness in the literature. Numerical results exhibit significant performance improvement by our scheme, and also demonstrate the trade-off between performance gain and fairness requirement.