Throughput Optimal Control of Cooperative Relay Networks

In cooperative relaying, multiple nodes cooperate to forward a packet within a network. To date, such schemes have been primarily investigated at the physical layer with the focus on communication of a single end-to-end flow. This paper considers cooperative relay networks with multiple stochastically varying flows, which may be queued within the network. Throughput optimal network control policies are studied that take into account queue dynamics to jointly optimize routing, scheduling and resource allocation. To this end, a generalization of the maximum differential backlog algorithm is given, which takes into account the cooperative gains in the network. Several structural characteristics of this policy are discussed for the special case of parallel relay networks.