Computation Offloading and Resource Allocation for the Internet of Things in Energy-Constrained MEC-Enabled HetNets

Abstract

In this paper, we present our investigation of a latency-minimization offloading problem for internet of things (IoT) terminals in multiple access edge computing (MEC)-enabled heterogeneous networks (HetNets), which jointly optimizes computation and communication resource allocation. Different from related works, the inter-user interferences caused by computation offloading demonstrate effective management in this paper. We also consider the limited battery capacity for IoT terminals for an energy-limited network. Then, we formulate a joint computation offloading and resource allocation optimization problem to minimize the weight-sum delay of users under the constraint of inter-user interference and energy consumption. Since the problem we formulated is a mixed integer non-linear programming (MINLP) problem, the optimal solution can't be easily obtained. Thus, we decompose the problem into multiple sub-problems. First, we obtain the optimal close solution for local CPU frequencies for each user. Then we propose a low complexity algorithm by using the CVX tool and the successive convex approximation approach (SCA). Finally, we propose a distributed computation offloading algorithm. The simulation results compare the performance of the proposed offloading scheme with different algorithms. We also analyze the influence of network parameters on the network latency and obtain some interesting conclusions.