|
|
|
Research Statement
Teaching Statement
|
|
|
|
|
Journal papers:
|
|
with Demosthenis Teneketzis
Under review in Games and Economic Behavior
|
|
|
with Volodymyr Babich, Demosthenis Teneketzis and Mark Van Oyen
Under review in Operations Research
|
|
|
with Demosthenis Teneketzis
Telecommunication Systems, 2010
|
|
|
with Asser Tantawi, Michael Spreitzer and Malgorzata Steinder
Performance Evaluation, 2010
|
|
|
with Demosthenis Teneketzis
IEEE/ACM Transactions on Networking, 2009
|
|
|
with Gaurav Yadav and Ajit K. Chaturvedi
IEEE Transactions on Wireless Communications, 2007
|
Conference papers:
|
|
with Demosthenis Teneketzis
Proceedings of GameComm (Performance Evaluation Methodologies and Tools), 2008
|
|
|
with Demosthenis Teneketzis
Proceedings of IEEE SECON, 2007
|
Working papers:
|
|
A supermodular Nash implementation mechanism for local public good provision in networks
|
|
|
Resource allocation in local public good networks: A realization perspective
with Demosthenis Teneketzis
|
|
|
An analysis of efficiency-user satisfaction tradeoff for sponsored search
with Aditya Gupta and Dinesh Garg
We study the tradeoff between efficiency and user satisfaction for a family of sponsored search ranking rules under the cascade model of user behavior. We define efficiency as the expected value of an ad slate to all the advertisers and user satisfaction as the Expected Reciprocal Rank (ERR) of the ad slate. The family of ranking rules we investigate includes the efficiency maximizing rule, the ERR maximizing rule, as well as the max eCPM rule - the ranking rule used in the Generalized Second Price (GSP) auction. For this family we characterize the tradeoff between efficiency and ERR, and use that to propose guidelines for choosing an optimum (in terms of both efficiency and user satisfaction) ad ranking rule.
|
|
|
An algorithm for online allocation of display ads in the presence of user fatigue
with Koyel Mukherjee and Dinesh Garg
We study the problem of online display ad allocation to multiple streams of users. Each advertiser targets a subset of user streams and has an aggregate advertising budget (maximum number of ad impressions the advertiser can pay for) over all these streams. Because of user fatigue, the bid of an advertiser for each additional ad impression to a given user stream decreases with the number of impressions. We assume that the advertisers' bids for ad impressions are revealed online with the arrival of requests from user streams. For this system we present an algorithm for online ad allocation that achieves a competitive ratio of 1-1/e with respect to the revenue maximizing offline ad allocation when advertising budgets are large.
|
Invited talks:
|
|
with Volodymyr Babich, Demosthenis Teneketzis and Mark Van Oyen
INFORMS annual meeting, October 2008
|
|
|
with Demosthenis Teneketzis, Asser Tantawi, Michael Spreitzer and Malgorzata Steinder
Applied probability seminar, IBM T. J. Watson research center, August 2008
|
|
|
with Demosthenis Teneketzis
Department of Electrical and Computer Engineering, University of California, San Diego, June 2007
Department of Electrical Engineering, University of California, Los Angeles, June 2007
|
|
|
Department of Electrical Engineering, Indian Institute of Technology, Kanpur, August 2006
|
Conference presentations:
|
|
with Demosthenis Teneketzis
Brazilian Workshop of the Game Theory Society, University of Sao Paulo, Brazil, August 2010
STIET research workshop, University of Michigan, Ann Arbor, May 2010
|
|
|
with Demosthenis Teneketzis
GameComm, Athens, Greece, October 2008
|
|
|
Service management middleware group, IBM T. J. Watson research center, August 2008
|
|
|
Society of Women Engineers (SWE) national conference, Nashville, October 2007
|
|
|
with Demosthenis Teneketzis
IEEE SECON, San Diego, June 2007
NSF/CEME Decentralization conference, University of Michigan, Ann Arbor, April 2007
STIET research workshop, University of Michigan, Ann Arbor, October 2006
|
|
