Constrained stochastic games in wireless networks

We consider the situation where N nodes share a common access point. With each node i there is an associated buffer and channel state that change in time. Node i dynamically chooses both the power and the admission control to be adopted so as to maximize the expected throughput, which depends on the actions and states of all the players, given its power and delay constraints. The information structure that we consider is such that each player knows the state of its own buffer and channel and its own actions. It does not know the states of, and the actions taken by other players. Using Markov Decision Processes we analyze the single player optimal policies under different model parameters. In the context of the stochastic games we study the equilibria of the N player scenario.