Optimal Distributed Routing and Power Control Decomposition for Wireless TDMA/CDMA Networks

Efficiently transmitting data in wireless time-division multiple access (TDMA) networks requires an integrated routing, time- scheduling, and power control strategy. In this paper we present a method to solve this comprehensive optimization problem by decomposing it into practical sub-problems, but still meeting main requirements such as optimality, distributed implementation, multiple path routing, and per-hop error performance. Due to the highly complex nature of designing scheduler, we purposely manage scheduling separately, but include it in the constraint set of a joint routing and power control problem. We present a framework in which routing on the network and power allocation on the link layer is jointly solved in a cross-layer optimal fashion by using a decomposition method that fully decouples the joint problem into two convex sub-problems. Aside the framework we give a Routing and Power Control Decomposition (RPCD) algorithm that leverages distributed standard methods to calculate optimum routes and power variables while requiring minor inter-node communications. We prove its convergence to a KKT-point for typical network utilization functions. This convergence is ensured albeit the links suffer multiple access interference as it is the case for TDMA/CDMA networks. For illustration, we apply the RPCD algorithm to a single frequency purely TDMA based wireless mesh backhaul network with the objective to minimize total power consumption while meeting delay, buffer and power constraints. Impressive convergence results indicate to reach the optimum solution in one iteration step only.