Wirelessly Powered Mobile Computation Offloading: Energy Savings Maximization

Achieving long battery lives or even self sustainability has been a long standing challenge for designing mobile devices. This paper seamlessly integrates two promising energy-saving technologies, namely mobile computation offloading (MCO) and microwave power transfer (MPT), and proposes a novel design framework of wirelessly powered MCO. Consider a single-user system where a base station (BS) either transfers power to or offloads computation from a mobile. Two mobile operation modes, namely local computation and offloading, are optimized separately for maximizing the mobile energy savings. For local computation, the non-convex problem of optimizing the CPU-cycle frequencies under the deadline and energy causality constraints is solved via convex relaxation. The optimal CPU-cycle frequencies are shown to have different forms depending on the BS transmission power. For offloading, the time duration before the deadline is divided for separate MPT and offloading and the optimal division is derived in a closed form. By combining above results, the optimal offloading decision is analyzed with respect to the deadline, data-input size and BS transmission power and validated by simulation.