Throughput and Delay Optimal Resource Allocation in Multiple Access Fading Channels

Edmund Yeh, Yale Unversity:

Thursday, April 10, 2003, 4:30 PM, Princeton University, Friend Center 006

The central problem in multiple-access (many-to-one) communications is the design of mechanisms for resource sharing. The development of these mechanisms is influenced by the stochastic nature of data traffic as well as the choice of coding and modulation schemes. Traditionally, these two sets of issues have been analyzed in virtual isolation from each other: some bodies of literature focus mainly on network-layer throughput and delay, while others concentrate on physical-layer channel modeling, coding, and detection. We present an integrated, cross-layer view of multiple-access communications over fading channels. This approach combines basic communication limits with network quality-of-service issues such as throughput and delay. We consider a multiple-access model with random packet arrivals and queuing at the transmitters. Optimal coding is assumed at the physical layer, so that all rates in the information-theoretic capacity region are achievable. We then consider resource allocation policies which assign power and rate dynamically as a function of the joint fading state and the queue state. Policies which optimize network throughput and delay are characterized. Joint work with Aaron Cohen, Brown University.