A Scalable Energy vs. Latency Trade-Off in Full-Duplex Mobile Edge Computing Systems

In this paper, we investigate the offloading energy and latency trade-off in a multiuser full-duplex (FD) system. We consider a multi-user FD system where a FD base station (BS), equipped with a mobile-edge computing (MEC) server, carries out data transmission in the downlink while at the same time receiving computational tasks from mobile devices in the uplink. Our main aim is to study the trade-off between the offloading energy and latency, which are known to be very important and desirable system objectives for both the system operator and users. In practice, there always exists a trade-off between these two objectives. Toward this aim, we formulate two weighted multi-objective optimization problems (MOOPs), one where the multi-user interference (MUI) is suppressed and the other where MUI is rather exploited. As a result, our proposed MOOPs allow for a scalable trade-off between the two objectives. To tackle the non-convexity of the formulations, we design an iterative algorithm through Lagrangian method. We also address the scenario of imperfect channel state information (CSI) at the FD BS. For the imperfect CSI case, we apply convex relaxations and transformation using the S-procedure to tackle the non-convexity of the formulations. Simulation results show the effectiveness of the proposed FD schemes compared with the existing baseline half duplex schemes and the superiority of MUI exploitation over suppression.