Abstract
In multihop relay networks with a large number of hops, it is challenging to allocate the available resources optimally among the relay nodes due to the large number of required control signals between the controller and the relay nodes. In addition, the latency caused by this signaling can result in using outdated resource-allocation information as the channels are time varying. Hence, a distributed resource-allocation scheme is proposed for multihop decode-and-forward (DF) relay systems, with low complexity in analytical computation and practical implementation. Based on the type of division multiplexing in time or frequency, this scheme can be used to obtain the optimal per-hop time slot length or bandwidth, respectively. This distributed iterative scheme needs no additional resource for optimization. The required communication to fulfill this optimization is only with the adjacent nodes of each relay and can be performed through the forwarding and acknowledgment packets (ACKs). It is shown that this scheme converges to the global optimal solution with an exponential convergence rate. The performance of this scheme is also investigated for the time-variant transmission channels by computer simulations.