Hello, I'm Siddhartha. Since October 2024, I've been a Postdoctoral researcher at the Applied Mathematics and Physics Department, Kyoto University, Japan. I'm working with the Control Systems Theory Group, led by Prof. Kenji Kashima.
In September 2024, I completed my Ph.D. at The Centre for Systems & Control Engineering (SysCon), IIT Bombay. I had the privilege of being supervised by Prof. Debasish Chatterjee and Prof. Ravi N. Banavar.
My Ph.D. thesis received the prestigious Naik and Rastogi Award for Excellence in Ph.D. Research for the year 2023-2025.
Google Scholar, LinkedIn, math genealogy (academic tree), My Erdős number is 4.
See research for more details about my previous and current work.
Contact details:
Electronic: ganguly.siddhartha.7p@kyoto-u.ac.jp, sganguly@iitb.ac.in.
Coordinates: Dept. of Applied Mathematics and Physics, Graduate School of Informatics, Faculty of Engineering, Integrated Research Bldg, 4th floor, Yoshida Main Campus, Room 413.
Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-8501, Japan.
I'm broadly interested in the mathematical theory of control and dynamical systems. My current research focus is on:
Optimal transport, especially discrete optimal transport and the numerical methods associated with such problems.
Density transportation/steering of noisy partial differential equations.
The interplay between bilevel and min-max optimization and their applications in machine learning and control.
Sparsity in optimization and optimal control.
Previously, I have worked on topics related to:
Optimal control (see the excellent introductory articles -- OptCon1, OptCon2); more specifically computational/numerical optimal control
Approximation theory (a nice introductory text -- CheLig:AppTheo) and its applications to optimization and control problems
Model predictive control (a nice introductory text -- GruPan:MPC); more specifically robust, stochastic, and distributionally robust MPC algorithms with emphasis on explicit and fast implementation
Probability and learning theory (see Francis Bach's notes on learning theory) and their application to control problems
Optimization algorithms, their applications in the area of machine learning and control theory, and numerical software development.
News and activities:
Our article, titled: QuITO v.2: Numerical Solutions with Uniform Error Guarantees to Optimal Control Problems under Path Constraints, with R.A. D'Silva and D. Chatterjee, was accepted to the IEEE Transactions on Automatic Control.
Long version and software: arXiv version, Software information page, GitHub Repo.
Our article, titled: Data-driven Gromov-Wasserstein density steering, with Haruto Nakashima and Kenji Kashima was accepted to the CDC 2025, Rio De Janeiro, Brazil.
Our article, titled: Formation Shape Control using the Gromov-Wasserstein Metric, with Haruto Nakashima, Kohei Morimoto and Kenji Kashima, was accepted to the Learning for Dynamics and Control (L4DC) 2025, arXiv version.
Submitted an article titled: Approximation-aware constrained feedback policy learning via neural networks and KANs: a detailed case study of predictive control, with Gakul Rajaraman, Vipul Mishra, Ashwin Aravind, and Debasish Chatterjee.
Our article, titled: Explicit Feedback Synthesis Driven by Quasi-Interpolation for Nonlinear Robust Model Predictive Control, with Debasish Chatterjee, was accepted to the IEEE Transactions on Automatic Control (IEEE TACON).
Our article, titled Exact solutions to minmax optimal control problems for constrained linear systems, with Debasish Chatterjee, was accepted to the IEEE Control Systems Letters (IEEE LCSS).
Our article, titled Discrete-time Pontryagin maximum principle under rate constraints: necessary conditions for optimality, with Souvik Das, Debasish Chatterjee, and Ravi Banavar, was accepted by the Asian Journal of Control.
Our article, titled Data-driven distributionally robust MPC via semi-infinite semidefinite programming: an application to finance, with Souvik Das, Ashwin Aravind, and Debasish Chatterjee, was accepted to the MTNS (mathematical theory of networks and systems); Souvik presented it at Cambridge, UK in August, 2024.
Version 2 of QuITO --- QuITO v.2 is now out; see the Software information page and the GitHub Repo. It is more optimized, and faster, solves a wide range of problems (especially singular optimal control problems), and comes with an automatic change point localization and mesh refinement module and a GuI. Give it a try! The associated article that documents the theoretical developments can be found here: https://arxiv.org/abs/2404.13681v3.
Presented our article Towards continuous-time MPC: a novel trajectory optimization algorithm (With Souvik Das, Debasish Chatterjee, and Muthyala Anjali), at the IEEE CDC23, Singapore.
Our article, QuITO: Numerical software for constrained nonlinear optimal control problems, with Nakul Randad, RihanAaron D'Silva, Mukesh Raj, and Debasish Chatterjee, was accepted to the SoftwareX.
Gave a talk on robust explicit model predictive control for low- through moderate-dimensional linear systems in Students' workshop SysCon Workshop for Graduate Students, 15-16 September 2023, IIT Bombay.
10 Jul 2023: Organized an invited session on computationally tractable constrained control synthesis, with Souvik Das and Debasish Chatterjee, at the IFAC world congress 2023, Yokohama, Japan.
Presented our article, An illustration of a quasi-interpolation driven technique for feedback synthesis (with Manav Doshi, Debasish Chatterjee, and Ravi Banavar), at the IFAC World Congress, Yokohama, Japan, 2023.
Gave a talk on approximate constrained trajectory synthesis via quasi-interpolation, IIT Gandhinagar, India, March 30, 2023.
28 Apr 2023: Patent number 430622 corresponding to application 202221040362 on Method and trajectory management controller for constrained trajectory optimization granted.
Give our new solver QuITO (comes with a Graphical User Interface) a try to solve your optimal control problems directly.
Workshop on Innovations in Nonlinear Control, details can be found on the workshop website.
Gave a talk on constrained optimal control and approximation theory, organized by Control and Dynamical System Reading Group, IIT Bombay, India, January 31, 2022.
Nakul presented our paper, constrained trajectory synthesis via quasi-interpolation (with Debasish Chatterjee and Ravi Banavar) at the IEEE CDC, Cancun, Mexico, 2022.
Our article, Discrete-time rate constrained maximum principle, with Souvik Das, Debasish Chatterjee and Ravi Banavar, was accepted at the IFAC conference on Lagrangian and Hamiltonian Methods for Nonlinear Control (LHMNC), Berlin, Germany, 2021.
Lecture series on brain-inspired robotics by Hiroaki Wagatsuma, Nov 24 to Dec 2, 2021.
Seminars and talks on the broad area of systems and control: SysConTalks.
Received the Prime Minister's Research Fellows (PMRF) scholarship.