Here is a description of my research organized by topic area. In addition to papers, you'll find links to several talk videos.

The design of search and matching platforms: In recent years there has been a proliferation of online search and matching platforms, including those for dating, labor and lodging. A core challenge for such platforms is that match formation requires both parties to be mutually interested in each other. Under typical platform designs, participants must spend effort interacting with many potential partners before they can match. Design levers available to online platforms include marketplace rules (for example, a man cannot send the first message on the dating platform Bumble), match recommendations and interest-signaling mechanisms. In a series of papers, we characterize how matching platforms should deploy these design levers so that they produce high quality match outcomes while minimizing demands on participants.

Y. Kanoria and D. Saban, “Facilitating the Search for Partners on Matching Platforms,” Management Science (Featured Article), 67(10), 2021 (also in EC 2017). (talk video, MS Blog post)

I. Ashlagi, M. Braverman, Y. Kanoria and P. Shi, “Clearing matching markets efficiently: informative signals and match recommendations,” Management Science, 66(5), 2020 (appeared in EC 2017). (video of short talk by Peng)

N. Arnosti, R. Johari and Y. Kanoria, “Managing Congestion in Matching Markets,” Manufacturing and Service Operations Management, 23(3):620-636, 2021, also in EC 2014.

Centralized college and school admissions: School and college admissions constitute an important class of matching markets. I designed and helped implement a new joint seat allocation process for undergraduate admissions to over 500 programs spread across 80 technical universities in India, including the prestigious Indian Institutes of Technology (IITs). The process has been in successful use since 2015 handling over 1 million candidates a year, and has produced a significant reduction in vacancies.

S. Baswana, P. P. Chakrabarti, S. Chandran, Y. Kanoria and U. Patange, “Centralized admissions for engineering colleges in India,” INFORMS Journal on Applied Analytics, 49(5), special issue for finalists for the Wagner Prize for Excellence in OR Practice 2018 (appeared in EC 2019). (Wagner finalist talk video)
Covered by The Times of India, IFORS News, Ideas at Work and Al Roth's blog post.
The continuing non-IIT vacancies have a simple fix. See my Medium article.

In another paper (Irene Lo's job market paper), we design mechanisms to address the practically important challenge of systematically reassigning seats that are vacated by students at a late stage in the admissions process.

I. Feigenbaum, Y. Kanoria, I. Lo and J. Sethuraman, “Dynamic Matching in School Choice: Efficient Seat Reallocation After Late Cancellations,” Management Science, 66(11), 2020. (appendix, video of talk by Irene)

Equilibria in matching markets: In two papers, we characterize equilibria (stable matchings) in matching markets without and with transfers, and show that equilibria are approximately unique. The JPE paper which has received widespread attention. A new working paper revisits random matching markets and explains why real world markets do not exhibit a stark effect of competition.

I. Ashlagi, Y. Kanoria and J. Leshno, “Unbalanced random matching markets: the stark effect of competition” (online appendix), Journal of Political Economy, 125(1), 2017 (appeared in EC 2013). (talk video)
Al Roth's blog post, Gil Kalai's blog post.

Y. Kanoria, D. Saban and J. Sethuraman, “Convergence of the core in assignment markets” (electronic companion), Operations Research, 66(3):597-892, 2018, also in SODA 2015.

Y. Kanoria, S. Min and P. Qian, “Which random matching markets exhibit a stark effect of competition?” to appear in SODA 2021.

Algorithms for dynamic matching: Matching platforms must dynamically manage and balance the "inventory" of agents on the two sides of the market, while grappling with agent heterogeneity and noisy predictions regarding future arrivals.

My latest work: Y. Kanoria, “Dynamic Spatial Matching,” working paper. (talk video)

Shared transportation (e.g., ridehailing) systems: This is a "simple" setting where agents and supply units are (primarily) distinguished by geographical location. Even so, prior state-of-the-art policies required perfect knowledge of future demand arrival rates and were impractical. We introduce simple and practical state-dependent control policies that make decisions based on the current geographical distribution of supply, and show them to be near optimal, even without demand predictions. Our first paper studies a setting with "small" geographical imbalance and obtains a sharp large deviations characterization of the (vanishing) optimality gap.

S. Banerjee, Y. Kanoria and Pengyu Qian, “Dynamic Assignment Control of a Closed Queueing Network under Complete Resource Pooling,” major revision at Operations Research (appeared in ACM SIGMETRICS 2018).

In our second paper, we consider a very general setting and provide a highly practical and near optimal "mirror backpressure" control methodology for joint entry/pricing/dispatch control in the absence of demand predictions. Mirror backpressure autocorrects supply imbalances by aggressively protecting and replenishing cars where they are scarce, while maximally deploying cars from regions where they are plentiful. Notably, mirror backpressure provides a powerful generalization of the celebrated backpressure policy for control of queuing networks.

Y. Kanoria and Pengyu Qian, “Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure,” appeared in EC 2020.

In a prior paper (Vijay Kamble's job market paper), we introduce and analyze a novel model combining matching and learning.

R. Johari, V. Kamble and Y. Kanoria, “Matching while learning,” Operations Research, 69(2):655-681, 2021, also in EC 2017. (video of short talk by Vijay)

In an early paper, we study dynamic matching with i.i.d. random agent compatibilities and show near optimality of greedy matching for (i) two-way cycles (matching), (ii) three-way cycles and (iii) chains. Cases (ii) and (iii) are important in kidney exchange, and are also the technically challenging cases. The matching setting (case (i)), is related to the concurrent paper of Akbarpour, Li & Oveis Gharan (for their setting where agent departures occur without warning): in comparison to that paper, our model avoids a nuisance parameter and our result is sharper.

R. Anderson, I. Ashlagi, D. Gamarnik and Y. Kanoria, “Efficient dynamic barter exchange” (online appendix), Operations Research, 65(6):1446-1459, 2017, also in SODA 2015.

Other recent research: Internet publishers earn revenue by selling the opportunity to advertise to their viewers via auctions. Publishers may try to use past advertiser bids to learn their willingness to pay and set auction reserve prices accordingly, but this approach gives advertisers the incentive to deflate their bids, which hurts revenues. We provide a novel approach for publishers that leverages competition between advertisers to learn their willingness to pay without compromising long-run advertiser incentives.

Y. Kanoria and H. Nazerzadeh, “Incentive-compatible Learning of Reserve Prices for Repeated Auctions,” Operations Research, 69(2):509-524, 2021.

Customer churn is a significant challenge for service platforms, and is typically managed using discounts and promotions. We propose a novel alternative control lever, namely, how a platform should choose between risky and safe customer service modes to maximize customer lifetime value. Notably, our (optimal) policy results in vastly larger customer lifetime value than a service policy which ignores customer churn (the latter reflects current practice).

Y. Kanoria, I. Lobel and Jiaqi Lu, “Managing Customer Churn Via Service Mode Control,” submitted.

Some exciting research directions have emerged from my Fall 2019 PhD class on Statistical Physics, Markets and Algorithms:

Y. Kanoria, J. Gan, and X. Zhang, “Local algorithms for dynamic optimization in networks,” work in progress.

N. Immorlica, Y. Kanoria, and J. Lu, “When does competition and costly information acquisition lead to a deadlock?” work in progress.

Earlier journal papers:

M. Bayati, C. Borgs, J. Chayes, Y. Kanoria and A. Montanari, “Bargaining dynamics in exchange networks,” Journal of Economic Theory, 156, 2015 (also in SODA 2011).

M. Ibrahimi, Y. Kanoria, M. Kraning and A. Montanari, “The Set of Solutions of Random XORSAT Formulae,” Annals of Applied Probability, Vol. 25, No. 5, 2015 (also in SODA 2012).

Y. Kanoria and A. Montanari, “Majority dynamics on trees and the dynamic cavity method,” Annals of Applied Probability, Vol. 21, No. 5, 2011.

Y. Kanoria and O. Tamuz, “Tractable Bayesian Social Learning on Trees,” IEEE Journal on Selected Areas in Communications, Vol. 31, No. 4, 2013. Also in ISIT 2012.

Y. Kanoria and A. Montanari, “Optimal Coding for the Binary Deletion Channel With Small Deletion Probability,” IEEE Transactions on Information Theory, Vol. 59, No. 10, 2013. Also in ISIT 2010, Student Paper award.

Other Refereed Conference Papers:

Y. Kanoria and H. Nazerzadeh, “Dynamic Reserve Prices for Repeated Auctions: Learning from Bids,” Web and Internet Economics (WINE), 2014.

Y. Kanoria, A. Montanari, D. Tse and B. Zhang, “Distributed Storage for Intermittent Energy Sources: Control Design and Performance Limits,” [slides pptx, pdf] Annual Allerton Conference on Communication, Control, and Computing, 2011.

M. Ibrahimi, A. Javanmard, Y. Kanoria and A. Montanari, “Robust max-product belief propagation,” Asilomar Conference on Signals, Systems and Computers, 2011 (Invited).

Y. Kanoria and A. Montanari, “Subexponential convergence for information aggregation on regular trees,” (long version) IEEE Conf. on Decision and Control and European Control Conference (CDC-ECC), 2011.

Y. Kanoria, “An FPTAS for Bargaining Networks with Unequal Bargaining Powers,” Workshop on Internet & Network Economics (WINE), 2010.

Y. Kanoria, S. Mitra and A. Montanari, “Statistical Static Timing Analysis using Markov Chain Monte Carlo,” Design, Automation & Test in Europe (DATE), 2010.

C. Dutta, Y. Kanoria, D. Manjunath and J. Radhakrishnan, “A Tight Lower Bound for Parity in Noisy Communication Networks,” ACM-SIAM Symp. on Discrete Alg. (SODA), 2008.

Y. Kanoria and D. Manjunath, “On Distributed Computation in Noisy Random Planar Networks,” IEEE Intl. Symp. on Information Theory (ISIT), 2007.

Awards: IIT Bombay YAA award, NSF CAREER Award, Sigecom ToT award blog and listing, ISIT student paper

External profiles: Medium, Simons, IEEE Xplore, Semantic Scholar, Campion