Routing, Scheduling and Power Allocation in Generic OFDMA Wireless Networks: Optimal Design and Efficiently Computable Bounds

The goal of this paper is to determine the data routes, subchannel schedules, and power allocations that maximize a weighted-sum rate of the data communicated over a generic OFDMA wireless network in which the nodes are capable of simultaneously transmitting, receiving and relaying data. Two instances are considered. In the first instance, subchannels are allowed to be time-shared by multiple links, whereas in the second instance, each subchannel is exclusively used by one of the links. Using a change of variables, the first problem is transformed into a convex form. In contrast, the second problem is not amenable to such a transformation and results in a complex mixed integer optimization problem. To develop insight into this problem, we utilize the first instance to obtain efficiently computable lower and upper bounds on the weighted-sum rate that can be achieved in the absence of time-sharing. Another lower bound is obtained by enforcing the scheduling constraints through additional power constraints and a monomial approximation technique to formulate the design problem as a geometric program. Numerical investigations show that the obtained rates are higher when time-sharing is allowed, and that the lower bounds on rates in the absence of time-sharing are relatively tight.