Cross-layer utility maximization subject to stability constraints for multi-channel wireless networks

• Authors: M. Codreanu;C. Weeraddana;M. Latva-Aho
• Venue: Asilomar'09 Proceedings of the 43rd Asilomar conference on Signals, systems and computers
• Year: 2009

We consider the problem of cross-layer utility maximization subject to stability constraints for a multicommodity wireless network where all links are sharing a number of orthogonal channels. We assume a time slotted network, where the channel gains are changing randomly from slot to slot. The optimal cross-layer network control policy can be decomposed into three subproblems: 1) flow control at the transport and network layers, 2) routing and scheduling at the network layer, and 3) resource allocation (RA) at the medium access control and physical layers. Every time slot, a network controller decides the amount of each commodity data admitted to the network layer, schedules different commodities over networks' links and controls the power and rate allocated to every link in every channel. To fully exploit the available multichannel diversity, we consider the general case, where multiple links can be activated in the same channel during the same time slot, and the interference is controlled solely via power and rate control. The links' achievable rates became coupled due to interference, and this makes the RA subproblem a difficult to solve nonconvex optimization problem. The main contribution of this paper is a computationally efficient power and rate control algorithm, derived via signomial programming. In contrast to other previously proposed solutions, our method is applicable to the whole range of signal-to-interference-and-noise-ratio (SINR) values of practical interest. Even though the global optimality of the solution can not be guaranteed due to the nonconvexity of the problem, the numerical results show that the proposed algorithm is able to exploit efficiently the available multichannel diversity and it can provide significant gains at the network layer in terms of end-to-end rates and network congestion.