Throughput maximization for full-duplex two-way relay with finite buffers

Abstract

Optimal queueing control of multi-hop networks remains a challenging problem, e specially in two-way relaying systems, even in the most straightforward scenarios. In this paper, we explore two-way relaying having a full-duplex decode-and-forward relay with two fifinite buffers. Principally, we propose a novel concept based on the multi-agent reinforcement learning (that maximizes the cumulative network throughput) based on the combination of the buffer states and the lossy links; a decision is generated as to whether it can transmit, receive or even simultaneously receive and transmit information. Towards this objective, chieflfly, based on the queue state transi tion and the lossy links, an analytic Markov decision process is proposed to analyze this scheme, and the throughput and queueing delay are derived. Our numerical results reveal exciting insights. First, artifificial intelligence based on reinforcement learning is optimal when the length of the buffer is superior to a certain threshold. Second, we demonstrate that reinforcement learning can boost transmission effificiency and prevent buffer overflflow.