Deep learning method for delay minimization in MANET

Abstract

A transmission delay is a critical metric when dealing with ad hoc networks in 5G, particularly for real-time applications and multimedia. In this paper, we describe the challenge of managing mobile ad-hoc networks (MANET) based on multi-hop time-slotted time-division multiple access (TDMA) under routing delay minimization with heterogeneous traffic flows. In particular, we consider the challenge of request scheduling and power control in TDMA, for minimizing the overall weighted end-to-end packet delay when the weights are determined according to the priorities of the requests. A delay minimization network that uses deep learning is also introduced (DMNet). Simulations show that DMNet outperforms other state-of-art methods. Our approach is one of the first to utilize a DNN to solve end-to-end delay minimization through scheduling and power control.

Keywords

• 5G and beyond
• Delay minimization
• MANET
• Deep learning
• TDMA