Asynchronous distributed optimization via dual decomposition for delay-constrained flying ad hoc networks

Abstract

The rapidly changing network topologies bring some challenges to optimize congestion control, routing and power control in delay-constrained flying ad hoc networks (FANETs). Traditional methods designed for mobile ad hoc networks (MANETs) underperform due to their poor adaptability to the changing network topologies. In this work, we propose an optimization framework with end-to-end (E2E) delay constraint, and formulate the problems as a utility maximization problem with respect to link capacity, E2E delay and flow conservation by coupling data flows and decomposing the E2E delay constraint. Further, to optimize such a non-linear maximization problem, we employ primal–dual decomposition and the delay scale factors to cast the maximization problem into several sub-problems with lower complexity. Subsequently, we propose a delay-constrained path selection algorithm and an asynchronous distributed optimization method which requires only the local channel information and outdated messages to optimize different parameters. Finally, to highlight the advantages of the proposed method, we further give the related analysis from the viewpoint of the different solutions of the maximization problem. Simulation results demonstrate that the proposed framework obtains significant network performance in terms of energy efficiency and packet timeout rate, compared with traditional methods.