Distributed Bayesian pricing for sum-rate maximization in small-cell networks

In this paper we address the problem of distributed sum-rate maximization in small cell networks in the special case where users have incomplete information about the interfering cross-channels. The sum-rate maximization problem admits a decentralized solution based on pricing mechanism. This requires the exchange of some control parameters (prices) over a common signaling channel. In femtocell networks, this exchange of control data may occur through the wired backhaul (e.g., an ADSL channel) connecting the femto-access points (FAP´s). The problem with the implementation of pricing is that it requires knowledge of the cross channels between the interfering nodes and this information typically is not available. In this paper, we address this problem by using only statistical information about the cross-channels. This leads to a Bayesian pricing mechanism. We start modeling the interference mechanism through a directed random graph. Then, we model the interference incorporating the fading statistics and the spatial distribution of the nodes. We provide a closed form expression for the expected value of the pricing term and derive distributed schemes able to converge to a stationary point of the sum-utility function. Finally, we compare the proposed approach to a purely competitive mechanism and to a clairvoyant (gene-aided) scheme.