Computation offloading leveraging computing resources from edge cloud and mobile peers

In this paper, we study the joint computation offloading and resource allocation problem exploiting computing resources from both mobile edge cloud and mobile peers. Our design aims to optimize the computation load assignments to local processors in the mobile users, mobile peers and the edge cloud jointly with the resource allocation to achieve the minimum weighted energy consumption subject to practical constraints on the bandwidth and computing resources and allowable latency. To tackle this non-convex optimization problem, we employ the successive convex approximation (SCA) method where we transform the underlying problem and iteratively solve a sequence of approximated convex problems. Moreover, the geometric programming (GP) method is applied to find the optimal solution of the approximated problem. The proposed SCA-based approach employs the arithmetic-geometric mean (AGM) approximation and the proposed algorithm is proved to converge to a local optimal solution. Finally, numerical studies confirm that the proposed scheme achieves energy saving gains about 60% and 10% in comparison with the local computation strategy and cloud offloading strategy under the strict required latency of 0.25s, respectively.