Optimizing Multicell Scheduling and Beamforming via Fractional Programming and Hungarian Algorithm

Abstract

The problem of optimizing scheduling and beamforming to maximize the network weighted sum rate (WSR) for the downlink of multicell, multi-antenna networks is challenging due to its nonconvexity and NP-hardness. In this paper, we present a novel approach based on the Hungarian algorithm and fractional programming that allows us to converge to an effective solution of the WSR maximization problem. Through extensive simulations, we compare the performance of the proposed algorithm with state-of-the art coordinated and uncoordinated resource allocation schemes in the literature. The proposed algorithm is shown to provide higher sum-log-utility values than previously proposed schemes matched filtering (MF), zero-forcing (ZF) and weighted minimum-mean-squared error (WMMSE), while offering a substantial improvement in cell-edge user rates over WMMSE with proportionally fair scheduling. Furthermore, the proposed scheme has a considerably lower computational complexity than multi-cell WMMSE.