UC Riverside

We present optimization algorithms for source and relay power allocations in a multicarrier relay system with direct link, where the source power is allowed to transmit in both phases in a two-phase relay scheme. We show that there is a significant benefit to the system capacity by allowing the source power to be distributed over both phases. Specifically, we consider the joint optimization of source and relay power to minimize a general cost function. The joint optimization problem is non-convex and the complexity of finding the optimal solution is extremely high. Using the alternating optimization (AO) method, the joint problem is decomposed into a convex source power allocation problem and a non-convex relay power allocation problem. By exploiting the specific structure of the problem, we present efficient algorithms that yield the exact optimal solutions for both source and (non-convex) relay power allocation problems. Then we show that the overall AO algorithm converges to a stationary point of the joint problem. Moreover, the proposed AO algorithm is asymptotically optimal for large relay transmit power or large source-relay channel gain. Finally, simulations show that the proposed AO algorithm achieves significant gain over various baselines.