Paper detail

Globally Optimal Distributed Power Control for Nonconcave Utility Maximization

Transmit power control in wireless networks has long been recognized as an effective mechanism to mitigate co-channel interference. Due to the highly non-convex nature, optimal power control is known to be difficult to achieve if a system utility is to be maximized. To date, there does not yet exist a distributed power control algorithm that maximizes any form of system utility, despite the importance of distributed implementation for the wireless infrastructureless networks such as ad hoc and sensor networks. This paper fills this gap by developing a Gibbs Sampling based Asynchronous distributed power control algorithm (referred to as GLAD). The proposed algorithm quickly converges to the global optimal solution regardless of the concavity, continuity, differentiability and monotonicity of the utility function. Same as other existing distributed power control algorithms, GLAD requires extensive message passing among all users in the network, which leads to high signaling overhead and high processing complexity. To address this issue, this paper further proposes a variant of the GLAD algorithm, referred to as I-GLAD, where the prefix "I" stands for infrequent message passing. The convergence of I-GLAD can be proved regardless of the reduction in the message passing rate. To further reduce the processing complexity at each transmitter, we develop an enhanced version of I-GLAD, referred to as NI-GLAD, where only the control messages from the neighboring links are processed. Our simulation results show that I-GLAD approximately converges to the global optimal solution regardless of the type of the system utility function. Meanwhile, the optimality of the solution obtained by NI-GLAD depends on the selection of the neighborhood size.