Computation offloading for mobile edge computing with accuracy guarantee

In this paper, we investigate the problem of energy expenditure minimization under latency and accuracy constraints in mobile edge computing (MEC)-based computation offloading. Given the non-convexity of the formulated problem, we first propose an energy-efficient computation resource allocation scheme inspired by the recent successive convex approximation (SCA) advances. After carefully exploring the problem structure, we fortunately derive the optimal solution, whose optimality is theoretically proved. Numerical results show that, compared with the SCA-based scheme and two other benchmark schemes, the optimal computation resource allocation scheme achieves the lowest energy consumption while satisfying the latency and accuracy requirements.