Modelling energy-aware task allocation in mobile workflows

Abstract

Mobile devices are becoming the platform of choice for both business and personal computing needs. For a group of users to efficiently collaborate over the execution of a set workflow using their mobile devices, the question then arises as to which device should run which task of the workflow and when? In order to answer this question, we study two common energy requirements: in the minimum group energy cost problem (MGECP) we build the model as a quadratic 0-1 program and solve the optimisation problem with the objective to minimise the total energy cost of the devices as a group. In the minimum max-utilisation problem (MMUP) we aim to improve the fairness of the energy cost within the group of devices and present two adjustment algorithms to achieve this goal. We demonstrate the use of a Mixed Integer Quadratic Programming (MIQP) solver in both problem’s solutions. Simulation result shows that both problems are solved to good standards. Data generated by different workload pattern also give us a good indication of the type of workflow that benefit the most from MMUP. The model used in this work can also be adapted for other energy critical scenarios.