Modeling Buffer-Aided Relay Selection in Networks With Direct Transmission Capability

TLDR

We consider a wireless relay network with a source, half‑duplex decode‑and‑forward buffer‑aided relays, and a destination, and model its evolution as a Markov chain to associate outage probability with the steady state. The study lifts the common assumption of no direct source‑destination link and proposes a link‑selection policy that exploits both buffering and direct reception opportunities. The scheme selects the relay based on the instantaneous strength of the wireless links, choosing the strongest available link for transmission. The analysis shows that, despite the unreliability of the direct link, it is always advantageous for the source to multicast a packet to both the relay with the strongest link and the destination.

Abstract

We consider a wireless relay network that consists of a source, half-duplex decode-and-forward buffer-aided relays and a destination. While the majority of previous works on relay selection assume no direct transmission between source and destination in such a setting, we lift this assumption and propose a link selection policy that exploits both the buffering ability and the opportunity for successful reception of a packet directly from the source. The proposed relay selection scheme incorporates the instantaneous strength of the wireless links and adapts the relay selection decision based on the strongest available link. The evolution of the network as a whole is modeled by means of a Markov chain and thus, the outage probability is associated with the steady state of the Markov chain. It is deduced that even if the link between the source and the destination is in principle a very unreliable link, it is always beneficial for the source to multicast a packet to both the relay with the strongest available link and the destination.