mlsanityData to 14 Oct 2026

Latent feedback control of distributed systems in multiple scenarios through deep learning-based reduced order models

arXiv 2412.09942 ↗First posted 13 Dec 2024math.OC cs.LG cs.NAcited by 0

Not on the map. Why?

Abstract (arXiv)

Continuous monitoring and real-time control of high-dimensional distributed systems are often crucial in applications to ensure a desired physical behavior, without degrading stability and system performances. Traditional feedback control design that relies on full-order models, such as high-dimensional state-space representations or partial differential equations, fails to meet these requirements due to the delay in the control computation, which requires multiple expensive simulations of the physical system. The computational bottleneck is even more severe when considering parametrized systems, as new strategies have to be determined for every new scenario. To address these challenges, we propose a real-time closed-loop control strategy enhanced by nonlinear non-intrusive Deep Learning-based Reduced Order Models (DL-ROMs). Specifically, in the offline phase, (i) full-order state-control pairs are generated for different scenarios through the adjoint method, (ii) the essential features relevant for control design are extracted from the snapshots through a combination of Proper Orthogonal Decomposition (POD) and deep autoencoders, and (iii) the low-dimensional policy bridging latent control and state spaces is approximated with a feedforward neural network. After data generation and neural networks training, the optimal control actions are retrieved in real-time for any observed state and scenario. In addition, the dynamics may be approximated through a cheap surrogate model in order to close the loop at the latent level, thus continuously controlling the system in real-time even when full-order state measurements are missing. The effectiveness of the proposed method, in terms of computational speed, accuracy, and robustness against noisy data, is finally assessed on two different high-dimensional optimal transport problems, one of which also involving an underlying fluid flow.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 57 of its 120 references that are in the map. First posted on arXiv 13 Dec 2024.

Builds on

57 of 120 references.

References · 57

  1. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,611
  2. 2019DeepONet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operatorscited by 2,018
  3. 2020Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problemscited by 1,920
  4. 2021Learning the solution operator of parametric partial differential equations with physics-informed DeepOnetscited by 540
  5. 2019Machine Learning for Fluid Mechanicscited by 385
  6. 2021Modern Koopman Theory for Dynamical Systemscited by 276
  7. 2018Optimization of Molecules via Deep Reinforcement Learningcited by 200
  8. 2016Linear predictors for nonlinear dynamical systems: Koopman operator meets model predictive controlcited by 140
  9. 2017Sparse identification of nonlinear dynamics for model predictive control in the low-data limitcited by 140
  10. 2014Dynamic mode decomposition with controlcited by 129
  11. 2020A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEscited by 105
  12. 2018Optimal Control Via Neural Networks: A Convex Approachcited by 85
38 more
  1. 2019Data-driven approximation of the Koopman generator: Model reduction, system identification, and controlcited by 85
  2. 2021POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decompositioncited by 79
  3. 2018Efficient collective swimming by harnessing vortices through deep reinforcement learningcited by 73
  4. 2019A review on Deep Reinforcement Learning for Fluid Mechanicscited by 66
  5. 2023Neural Inverse Operators for Solving PDE Inverse Problemscited by 65
  6. 2021Optimal control of PDEs using physics-informed neural networkscited by 52
  7. 2018Deep Reinforcement Learning for Swarm Systemscited by 45
  8. 2021Solving PDE-constrained Control Problems Using Operator Learningcited by 42
  9. 2017Koopman operator-based model reduction for switched-system control of PDEscited by 31
  10. 2021Physics-Informed Neural Nets for Control of Dynamical Systemscited by 31
  11. 2019Control of chaotic systems by Deep Reinforcement Learningcited by 29
  12. 2016Real-time optimal control via Deep Neural Networks: study on landing problemscited by 28
  13. 2021A Deep Learning approach to Reduced Order Modelling of Parameter Dependent Partial Differential Equationscited by 28
  14. 2023The transformative potential of machine learning for experiments in fluid mechanicscited by 27
  15. 2023Effective control of two-dimensional Rayleigh--B'enard convection: invariant multi-agent reinforcement learning is all you needcited by 27
  16. 2024SINDy-RL: Interpretable and Efficient Model-Based Reinforcement Learningcited by 26
  17. 2021Fast PDE-constrained optimization via self-supervised operator learningcited by 25
  18. 2023Distributed Control of Partial Differential Equations Using Convolutional Reinforcement Learningcited by 24
  19. 2019Zermelo's problem: Optimal point-to-point navigation in 2D turbulent flows using Reinforcement Learningcited by 21
  20. 2022Physics-informed neural networks for PDE-constrained optimization and controlcited by 19
  21. 2023Efficient PDE-Constrained optimization under high-dimensional uncertainty using derivative-informed neural operatorscited by 19
  22. 2022Data-driven control of spatiotemporal chaos with reduced-order neural ODE-based models and reinforcement learningcited by 17
  23. 2021An extended physics informed neural network for preliminary analysis of parametric optimal control problemscited by 16
  24. 2021Learning High-Dimensional Parametric Maps via Reduced Basis Adaptive Residual Networkscited by 16
  25. 2021Hybrid analysis and modeling, eclecticism, and multifidelity computing toward digital twin revolutioncited by 14
  26. 2023AONN: An adjoint-oriented neural network method for all-at-once solutions of parametric optimal control problemscited by 13
  27. 2021TNet: A Model-Constrained Tikhonov Network Approach for Inverse Problemscited by 10
  28. 2021Low Dimensional State Representation Learning with Robotics Priors in Continuous Action Spacescited by 9
  29. 2024Multi-agent reinforcement learning for the control of three-dimensional Rayleigh-B'enard convectioncited by 8
  30. 2020Optimal design of acoustic metamaterial cloaks under uncertaintycited by 7
  31. 2024Parametric PDE Control with Deep Reinforcement Learning and Differentiable L0-Sparse Polynomial Policiescited by 7
  32. 2021Density control of large-scale particles swarm through PDE-constrained optimizationcited by 6
  33. 2021On the Universal Transformation of Data-Driven Models to Control Systemscited by 5
  34. 2024Advanced deep-reinforcement-learning methods for flow control: group-invariant and positional-encoding networks improve learning speed and qualitycited by 5
  35. 2024Navigation in a simplified Urban Flow through Deep Reinforcement Learningcited by 5
  36. 2018Feedback control of nonlinear PDEs using data-efficient reduced order models based on the Koopman operatorcited by 4
  37. 2013Asymptotic Stability of POD based Model Predictive Control for a semilinear parabolic PDEcited by 3
  38. 2014HJB-POD feedback control for Navier-Stokes equationscited by 3

50 most cited of 57 All 57 (JSON)

Built on by

Not cited yet.