Weighted sum rate maximization for non-regenerative multi-way relay channels with multi-user decoding

This paper studies the maximization of the weighted sum rate in multi-way relay channels with simultaneous non-unique decoding at the receivers. We state the resource allocation problem as a global optimization problem of the transmit powers and achievable rates, and transform it into a monotonic optimization problem. The computational complexity of monotonic optimization problems is exponential in the number of variables. We observe that for fixed powers the problem is a linear program with much lower complexity and exploit this structural property by decomposing the optimization problem into an inner linear and an outer monotonic program. This reduces the computational complexity significantly and allows computing the global solution. We compare the achievable throughput with multi-user decoding and optimal power allocation numerically to state-of-the-art single-user decoding and to simply transmitting at maximum power. We observe that multi-user decoding performs much better than single-user decoding in terms of throughput and fairness for medium to high SNRs.