I am an Assistant Professor of Operations Research in the Grado Department of Industrial and Systems Engineering at Virginia Tech. Previously, I was a postdoc in the Center for Nonlinear Studies and Applied Mathematics and Plasma Physics groups at Los Alamos National Laboratory, and a postdoc in the Wisconsin Institute for Discovery at UW-Madison. I obtained my Bachelor's in Chemical Engineering from IIT Madras and my Master's and Ph.D. in Chemical Engineering from MIT.
My research interests lie at the interface of optimization and applied machine learning. I am interested in exploring new paradigms that integrate machine learning and optimization under uncertainty, designing algorithms for their solution with theoretical guarantees, and developing scalable implementations for energy systems applications.
I am looking for highly-motivated Ph.D. students with a background in mathematics, operations research, and/or machine learning. Feel free to reach out to me with your resume and unofficial transcripts.
Ph.D. in Chemical Engineering, 2018
Massachussets Institute of Technology
Master of Science, 2014
Massachussets Institute of Technology
Bachelor of Technology, 2012
Indian Institute of Technology, Madras
Sept. 2023: Joint work with Avinash on Optimization under uncertainty of a hybrid waste tire and natural gas feedstock flexible polygeneration system has been accepted by Energy! [Paper]
Aug. 2023: Excited to start as an Assistant Professor of Operations Research in the Grado Department of Industrial and Systems Engineering at Virginia Tech!
Aug. 2023: Joint work with Guzin and Jim on Residuals-based DRO with covariate information has been accepted by Mathematical Programming! [Paper]
July 2023: Invited talk on Data-Driven Multistage Stochastic Optimization on Time Series at the 2023 XVI International Conference on Stochastic Programming (ICSP).
June 2023: Excited to begin working with Erin George on accelerating the global solution of quadratically-constrained quadratic programs using graph-based machine learning as part of the 2023 LANL Applied ML Summer Research Fellowship Program! Welcome, Erin!
May 2023: Invited talk on Learning to Accelerate the Global Optimization of QCQPs at the 2023 SIAM Conference on Optimization (OP23).
Feb. 2023: Check out my new preprint with Evren and Johannes on Optimality-Based Discretization Methods for Nonconvex Semi-Infinite Programs! [Preprint]
Dec. 2022: New preprint with Harsha and Deep on Learning to Accelerate the Global Optimization of Nonconvex QCQPs is up on arXiv! [Preprint]
Dec. 2022: Happy to announce that my proposal with Harsha and Deep on Using Graph Neural Networks to Accelerate Solutions to Nonconvex Optimization Problems will be funded by the 2023 LANL Applied ML Summer Research Fellowship Program.
Nov. 2022: I gave two talks at the 2022 AIChE Annual Meeting:
Oct. 2022: I chaired two sessions at the 2022 INFORMS Annual Meeting:
Oct. 2022: Excited to begin working as a Co-Investigator on the LANL LDRD Exploratory Research project Learning to Accelerate Global Solutions for Non-Convex Optimization.
July 2022: Invited talk on Data-Driven Multistage Stochastic Optimization on Time Series at the 2022 International Conference on Continuous Optimization (ICCOPT).
May 2022: Excited to begin working with Mithun Goutham on using multistage stochastic programming to model the resilience of the power grid to hurricanes! Welcome, Mithun!
This project explores the use of ML to predict optimal instance-specific heuristic parameters within global optimization algorithms without sacrificing global optimality guarantees. We have devised algorithms to optimally partition variable domains for piecewise convex relaxation of nonconvex problems. Our ML algorithms are able to accelerate the solution of challenging problems by up to three orders of magnitude for specific instances and families of nonconvex problems by an order of magnitude on average! |
This project aims to leverage ensemble forecasts of extreme weather events (such as hurricanes and winter storms) to make proactive grid operation decisions and mitigate power outages. We model the operational resilience of the grid using multistage stochastic programming and use stochastic dual dynamic programming to joinly optimize anticipative and restorative actions. Our preliminary results show that using hurricane forecasts to make proactive decisions can reduce load shed by up to 5% on average! |
Section Under Construction! |
* = Graduate Student
R. Kannan and J. R. Luedtke (2021). A stochastic approximation method for approximating the efficient frontier of chance-constrained nonlinear programs, Mathematical Programming Computation, 13, 705–751. [Journal] [Preprint] [Slides] [Poster] [Code]
R. Kannan, J. R. Luedtke, and L. A. Roald (2020). Stochastic DC optimal power flow with reserve saturation, Electric Power Systems Research (special issue for the XXI Power Systems Computation Conference), pp. 1-9. [Journal] [Preprint] [Slides] [Code]
R. Kannan and P. I. Barton (2018). Convergence-order analysis of branch-and-bound algorithms for constrained problems, Journal of Global Optimization, 71(4), pp. 753-813. [Journal] [PDF] [Slides]
R. Kannan and P. I. Barton (2017). The cluster problem in constrained global optimization, Journal of Global Optimization, 69(3), pp. 629-676. [Journal] [PDF] [Slides]
R. Kannan and A. K. Tangirala (2014). Correntropy-based partial directed coherence for testing multivariate Granger causality in nonlinear processes, Physical Review E, 89(6), 062144, pp. 1-15. [Journal] [PDF]
E. M. Turan*, R. Kannan, and J. Jäschke (2022). Design of PID controllers using semi-infinite programming, Proceedings of the 14th International Symposium on Process Systems Engineering, pp. 439-444. [Conference] [Preprint]
R. Kannan and P. I. Barton (2016). The cluster problem in constrained global optimization, Proceedings of the XIII Global Optimization Workshop (GOW’16), pp. 9-12. [Conference]
R. Kannan, G. Bayraksan, and J. R. Luedtke. Heteroscedasticity-aware residuals-based contextual stochastic optimization, pp. 1-15. [Preprint]
Ph.D. (2018): Algorithms, analysis and software for the global optimization of two-stage stochastic programs, MIT. [DSpace@MIT] [PDF]
B.Tech. (2012): Partial directed coherence for nonlinear Granger causality: A generalized correlation function-based approach, IIT Madras.
I have been fortunate to work with several excellent collaborators over the years.
Students
|
Faculty
|
Instructor, Dept. of Industrial and Systems Engineering, Virginia Tech, 2023-present
Teaching Assistant, Dept. of Chemical Engineering, MIT, 2015
Math Lecturer for the IIT Joint Entrance Exam (IIT JEE), MIT, 2016
Math Olympiad Trainer, Science & Math Academy for Real Talents, India, 2008-2011
Volunteer, National Services Scheme, IIT Madras, 2008-2009
Videos recorded as a Math Lecturer for the IIT JEE
Check out my CV for other talks I have given.
Invited External Examiner for the following students:
Invited Peer-Reviewer for the following journals/conferences (Reviewer Profile):
|
|
Served in the following committees at Virginia Tech:
Session Chair at the following conferences:
Computational Stochastic Programming 2018 INFORMS Annual Meeting
Organized monthly teleconference meetings on Optimization Under Uncertainty as part of the DOE MACSER Project (Nov. 2018 to Nov. 2020)