TR2025-165

Constrained Optimization From a Control Perspective via Feedback Linearization

    •  Zhang, R., Raghunathan, A., Shamma, J., Li, N., "Constrained Optimization From a Control Perspective via Feedback Linearization", The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NuerIPS), December 2025.
      BibTeX TR2025-165 PDF
      • @inproceedings{Zhang2025dec,
      • author = {Zhang, Runyu and Raghunathan, Arvind and Shamma, Jeff and Li, Na},
      • title = {{Constrained Optimization From a Control Perspective via Feedback Linearization}},
      • booktitle = {The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NuerIPS)},
      • year = 2025,
      • month = dec,
      • url = {https://www.merl.com/publications/TR2025-165}
      • }
  • MERL Contact:
  • Research Areas:

    Control, Optimization

Abstract:

Tools from control and dynamical systems have proven valuable for analyzing and developing optimization methods. In this paper, we establish rigorous theoretical foundations for using feedback linearization (FL)—a well-established nonlinear control technique—to solve constrained optimization problems. For equality- constrained optimization, we establish global convergence rates to first-order Karush-Kuhn-Tucker (KKT) points and uncover the close connection between the FL method and the Sequential Quadratic Programming (SQP) algorithm. Building on this relationship, we extend the FL approach to handle inequality-constrained problems. Furthermore, we introduce a momentum-accelerated feedback linearization algorithm and provide a rigorous convergence guarantee.

 

  • Related News & Events

    •  NEWS    MERL Researchers at NeurIPS 2025 presented 2 conference papers, 5 workshop papers, and organized a workshop.
      Date: December 2, 2025 - December 7, 2025
      Where: San Diego
      MERL Contacts: Petros T. Boufounos; Anoop Cherian; Radu Corcodel; Stefano Di Cairano; Chiori Hori; Christopher R. Laughman; Suhas Anand Lohit; Pedro Miraldo; Saviz Mowlavi; Kuan-Chuan Peng; Arvind Raghunathan; Diego Romeres; Abraham P. Vinod; Pu (Perry) Wang
      Research Areas: Artificial Intelligence, Computational Sensing, Computer Vision, Control, Data Analytics, Dynamical Systems, Machine Learning, Multi-Physical Modeling, Optimization, Robotics, Signal Processing, Speech & Audio
      Brief
      • MERL researchers presented 2 main-conference papers and 5 workshop papers, as well as organized a workshop, at NeurIPS 2025.

        Main Conference Papers:

        1) Sorachi Kato, Ryoma Yataka, Pu Wang, Pedro Miraldo, Takuya Fujihashi, and Petros Boufounos, "RAPTR: Radar-based 3D Pose Estimation using Transformer", Code available at: https://github.com/merlresearch/radar-pose-transformer

        2) Runyu Zhang, Arvind Raghunathan, Jeff Shamma, and Na Li, "Constrained Optimization From a Control Perspective via Feedback Linearization"

        Workshop Papers:

        1) Yuyou Zhang, Radu Corcodel, Chiori Hori, Anoop Cherian, and Ding Zhao, "SpinBench: Perspective and Rotation as a Lens on Spatial Reasoning in VLMs", NeuriIPS 2025 Workshop on SPACE in Vision, Language, and Embodied AI (SpaVLE) (Best Paper Runner-up)

        2) Xiaoyu Xie, Saviz Mowlavi, and Mouhacine Benosman, "Smooth and Sparse Latent Dynamics in Operator Learning with Jerk Regularization", Workshop on Machine Learning and the Physical Sciences (ML4PS)

        3) Spencer Hutchinson, Abraham Vinod, François Germain, Stefano Di Cairano, Christopher Laughman, and Ankush Chakrabarty, "Quantile-SMPC for Grid-Interactive Buildings with Multivariate Temporal Fusion Transformers", Workshop on UrbanAI: Harnessing Artificial Intelligence for Smart Cities (UrbanAI)

        4) Yuki Shirai, Kei Ota, Devesh Jha, and Diego Romeres, "Sim-to-Real Contact-Rich Pivoting via Optimization-Guided RL with Vision and Touch", Worskhop on Embodied World Models for Decision Making

        5) Mark Van der Merwe and Devesh Jha, "In-Context Policy Iteration for Dynamic Manipulation", Workshop on Embodied World Models for Decision Making

        Workshop Organized:

        MERL members co-organized the Multimodal Algorithmic Reasoning (MAR) Workshop (https://marworkshop.github.io/neurips25/). Organizers: Anoop Cherian (Mitsubishi Electric Research Laboratories), Kuan-Chuan Peng (Mitsubishi Electric Research Laboratories), Suhas Lohit (Mitsubishi Electric Research Laboratories), Honglu Zhou (Salesforce AI Research), Kevin Smith (Massachusetts Institute of Technology), and Joshua B. Tenenbaum (Massachusetts Institute of Technology).
    •