Joint Resource Allocation, Computation Offloading, and Path Planning for UAV Based Hierarchical Fog-Cloud Mobile Systems

Abstract

In this paper, the computation offloading problem for the hierarchical fog-cloud computing (FCC) system with unmanned aerial vehicles (UAVs) is studied. The hierarchical FCC, which exploits both centralized and distributed computing architectures, is very promising to support computation offloading in emerging computation-demanding mobile applications. In our design, UAVs integrating computing platforms act as small distributed clouds while the macro base station (BS) integrates a more powerful central cloud server. Furthermore, the multiple input multiple output (MIMO) technology is employed for data communication. We assume that mobile users (UEs) and (UAVs) can change their locations over time and we consider the joint task offloading, user-cloud/cloudlet association, transmit power allocation, and path planning to minimize the total weighted consumed power of the system. To tackle the underlying non-convex mixed integer non-linear program (MINLP), we propose an iterative two-phase algorithm. Specifically, we iteratively solve the user-cloud/cloudlet association problem in the first phase and address the joint resource allocation, path planning problem in the second phase. Furthermore, we employ the difference of convex (DC) optimization method in the second phase to approximate the non-convex bilinear functions and propose to transform the non-convex INLP to the integer linear program (ILP) in the first phase. Numerical studies confirm that the proposed design for the FCC architecture achieves great performance benefits for executing mobile computation tasks.