Energy-Efficient Dynamic Task Offloading for Energy Harvesting Mobile Cloud Computing

Mobile-edge cloud computing (MEC) as an emerging and prospective computing paradigm, can significantly enhance computation capability and prolong the lifetime of mobile devices (MDs) by offloading computation-intensive tasks to the cloud. This paper considers applying simultaneous wireless information and power transfer (SWIPT) technique to a multi-user computation offloading problem for mobile-edge cloud computing, where energy-limited mobile devices (MDs) harvest energy form the ambient radio-frequency (RF) signal. We investigate partial computation offloading by jointly optimizing MDs' clock frequency, transmit power and offloading ratio with the system design objective of minimizing energy cost of mobile devices. To this end, we first formulate an energy cost minimization problem constrained by task completion time and finite mobile-edge cloud computation capacity. Then, by exploiting alternative optimization (AO) based on difference of convex function (DC) programming and linear programming, we design an iterative algorithm for clock frequency control, transmission power allocation, offloading ratio and power splitting ratio to solve the non-convex optimization problem. Our simulation results reveal that the proposed algorithm can converge within a few iterations and yield minimum system energy cost.