We consider location-dependent opportunistic bandwidth sharing between staticand mobile downlink users in a cellular network. Each cell has some fixednumber of static users. Mobile users enter the cell, move inside the cell forsome time and then leave the cell. In order to provide higher data rate tomobile users, we propose to provide higher bandwidth to the mobile users atfavourable times and locations, and provide higher bandwidth to the staticusers in other times. We formulate the problem as a long run average rewardMarkov decision process (MDP) where the per-step reward is a linear combinationof instantaneous data volumes received by static and mobile users, and find theoptimal policy. The transition structure of this MDP is not known in general.To alleviate this issue, we propose a learning algorithm based on singletimescale stochastic approximation. Also, noting that the unconstrained MDP canbe used to solve a constrained problem, we provide a learning algorithm basedon multi-timescale stochastic approximation. The results are extended toaddress the issue of fair bandwidth sharing between the two classes of users.Numerical results demonstrate performance improvement by our scheme, and alsothe trade-off between performance gain and fairness.
展开▼