Distributed optimization in wireless networks using broadcast advantage

摘要

In this paper, we consider cross layer optimizationudin wireless networks with wireless broadcast advantage,udfocusing on the problem of distributed scheduling of broadcastudlinks. The wireless broadcast advantage is most usefuludin multicast scenarios. For a multicast scenario, we give audsubgradient algorithm for distributed joint congestion control,udnetwork coding and session scheduling, which however requiresudcentralized link scheduling. Under the primary interferenceudmodel, link scheduling problem is equivalent to a maximumudweighted hypergraph matching problem that is NP-complete.udTo solve the scheduling problem distributedly, locally greedyudand randomized approximation algorithms are proposed andudshown to have bounded worst-case performance. With randomudnetwork coding, we obtain a fully distributed cross-layer design.udNumerical results show promising throughput gain using theudproposed algorithms, and surprisingly, in some cases evenudwith less complexity than cross-layer design without broadcastudadvantage.