Anil Vullikanti Wins CAREER Award

Anil Vullikanti of the Virginia Bioinformatics Institute recently received a CAREER award from the National Science Foundation. The Faculty Early Career Development (CAREER) Program is a Foundation-wide activity that offers the National Science Foundation's most prestigious awards in support of junior faculty who exemplify the role of teacher-scholars through outstanding research, excellent education and the integration of education and research within the context of the mission of their organizations. A description of Anil’s research appears below.

Title:
Cross-Layer Optimization in Cognitive Radio Networks in the Physical Interference Model Based on SINR Constraints: Algorithmic Foundations

Abstract:
The concept of Cognitive Radio Networks (CRN) represents one of the most significant recent advances in wireless networks. CRNs allow unlicensed (or secondary) users to access spectrum bands allocated to licensed (primary) users, without disrupting their performance. Since many licensed spectrum bands have been found to be greatly under-utilized, CRNs can potentially enhance the spectrum usage significantly. The basic principle underlying CRNs is to first sense the spectrum usage by primary users, and then allocate power levels and channels opportunistically to the secondary users, so that the interference levels at primary users are within an acceptable threshold. Most theoretical analyses of protocols in such networks use disk/graph based approximations (in which "close-by" links cannot transmit simultaneously) to model wireless interference; however, these are inadequate and can lead to infeasible solutions with unacceptable interference levels at the primary users. The goal of this proposal is to examine the theoretical foundations of cross-layer optimization in Cognitive Radio Networks in the physical interference model, which is considered a much better approximation of interference than disk based models. The results of this proposal will contribute to the theoretical underpinnings of the broader area of wireless networks, not just the application of CRN, because of the central role interference plays.

The proposal will focus on distributed algorithms for these problems, and in addition to traditional metrics of running time and number of messages, we study the work complexity, which is the total energy used by the distributed algorithm. In light of the large body of work involving disk based models, we study formal reductions between these models, and propose to develop techniques to allow results in disk based models to be translated to those in the physical model. Our focus will be on efficient algorithms (i.e., with polynomial running time) with provable performance guarantees relative to the optimum, that hold for every instance of arbitrarily distributed nodes and connections (i.e., these are worst case approximation guarantees).