Optimal distributed flow control over multiple parallel wireless channels

This paper addresses a distributed flow and channel allocation method for point-to-multipoint (PMP) networks operating over multiple parallel channels, where a central node shares its rate capacity with multiple subscribers. The problem formulation takes into account fairness and quality of service (QoS) considerations by means of generalized utility functions. Assuming that a local area network (LAN), wired or wireless, is attached to each of the subscribers, a centralized allocation scheme becomes prohibitive in terms of required network signalling when the number of connections within the LANs become high. In order to overcome this and to keep practicality in terms of iterations needed to converge, we apply a combination of a mean value cross (MVC) decomposition strategy with the novel coupled-decompositions method that achieves the optimal solution with a reduced number of iterations when compared to classical approaches.