Information Theory, Queueing, and Resource Allocation in Multi-user Fading Communications

摘要

We establish a unified cross-layer analytical frameworkrnfor optimal resource allocation in multiaccess (uplink)rnand broadcast (downlink) wireless communication systems wherernpackets arrive randomly to buffered transmitters. Viewing thernmultiaccess and broadcast fading channels as controlled queueingrnsystems, we demonstrate adaptive joint power control (subject tornboth long-term average and short-term peak power constraints)rnand rate allocation policies which maximize system throughput,rnwhile providing guarantees on delay performance. Since wernassume information-theoretic optimal coding at the physicalrnlayer, the resulting optimal resource allocation policies providernfundamental performance bounds.