Joint Hybrid Beamforming and Offloading for mmWave Mobile Edge Computing Systems

In this paper, we investigate the joint design of hybrid beamforming and demanding computation tasks offloading in mmWave-based mobile edge computing (MEC) systems, in order to minimize the maximum latency. The resulting optimization problem is challenging, mainly due to the highly nonlinear objective function and the unit modulus constraints on the analog beamformers. By seeking the special structure of the problem, we divide it into two separate problems. An iterative weighted mean-square error minimization (WMMSE) approach is adopted to address the first optimization problem, and the second problem is solved in closed-form. We also investigate a more practical scheme when only finite resolution phase shifters are implemented. Simulation results are provided to confirm that the proposed strategy achieves significant better performance than recent reported beamforming algorithm with fixed offloading ratio, and it is effective when low-resolution phase shifters are used.