SWIPT-based Mobile Edge Computing Systems: A Stochastic Geometry Perspective

In this paper, we have considered a simultaneous wireless information and power transfer (SWIPT)-enabled mobile edge computing (MEC) system where a low power device (LPD) has to decide either to perform local computation or offloading. A decision strategy has been proposed and the impact of this strategy on a triple trade-off between energy, local computation and offloading has been evaluated. Moreover, in the local computation case, due to the existence of an uploading phase in which results are sent to the MEC server, another communication-computation trade-off has been analysed. In order to provide a system-level perspective on the performance of such a SWIPT-MEC network, we have employed stochastic geometry to characterise these communication-computation trade-offs of the system and we have investigated joint success probability of local computation and offloading.