Exact Resource Allocation for Weighted Proportional Fair Wireless Relay Networks

Abstract

In this paper, we consider a relay-enabled wireless network and optimize the weighted proportional fair (wPF) bandwidth allocation by means of exact algorithms with lin- ear complexity on the number of users and relays. Complex architectures typical of relay-enabled systems pose the need of developing efficient techniques that account for the intertwined nature of all the network agents, as resources must be split not only between users, but also between relays and relay- served users, altogether constrained by backhaul capacities and the traffic bottleneck present at the wired base stations. Here, traditional schemes for bandwidth allocation cannot be applied, as resources from one point of the network cannot be allocated regardless the allocation performed at other entities of the same network. Hence, we develop a compact weighted proportional fair resource management with very lightweight complexity, able to jointly allocate access and backhaul resources optimally in real time. We benchmark the results on network capacity and fairness with state-of-art proposals and show that the wPF exact algorithms proposed yield the best trade-off between capacity and fairness in relay networks.