Time-Slotted Resource Allocation in a Two-User Computationally-Constrained Offloading System

Abstract

Computation offloading is a promising technique that enables mobile devices to utilize additional computation resources provided by the Mobile Edge Computing (MEC) paradigm to reduce the energy consumption and the latency required to complete a computational task. In order to exploit the offloading opportunity efficiently, the available communication and computation resources must optimally be allocated to the devices. In this paper, we seek to address the joint optimization problem of minimizing the total energy consumption of a two-user offloading system over limited communication and computation resources, under various multiple access schemes. We will propose an optimal time-slotted resource allocation strategy in which the resources are either shared or assigned to a single user in each slot. We will determine the optimal arrangement of the time slots and then obtain closed-form expressions for the jointly optimal slot lengths and resource allocations over all time slots. Our numerical results illustrate that the combination of the time-slotted structure and a capacity-approaching multiple access scheme enables the proposed resource allocation approach to significantly reduce the total energy consumption of an offloading system as compared to the existing approaches.