UC San Diego

As the demand for high-speed communication networks grows, it becomes increasingly necessary to re-examine system design with respect to resource allocation. Particularly in wireless networks where the spectrum is a shared and limited resource, it is important for networks to respond to randomly fluctuating conditions by adapting the allocation of resources in an efficient, fair, and scalable manner. The difficulty in doing this lies in the inherently distributed structure of large-scale networks with many autonomous and possibly selfish users. Recently, a large body of work in which resource allocation is modeled as a constrained optimization problem has emerged. In such work, the parallel computation of the solution to an optimization problem is mimicked in order to design practical, distributed algorithms for resource allocation. This paradigm of design philosophy can be applied to a wide-range of networks and applications. The contributions of this research lie in four main areas: i) distributed and adaptive rate assignments for CDMA 1xEVDO-type systems, ii) cross-layer optimal rate assignments for hybrid wired/ wireless CDMA networks, iii) leveraging downlink for socially optimal and incentive compatible uplink rate allocation in cellular networks, and iv) capacity- achieving rate assignments in unicast networks that employ network coding. In each of these problems, I address issues related to modeling, analysis, implementation, and validation