Pith. sign in

REVIEW 2 major objections 4 minor 42 references

One gradient step per time update stabilizes unknown linear time-varying systems while tracking frozen-time LQR policies.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

One-step policy-gradient LQR updates with normalized sliding-window least-squares stabilize unknown slowly varying and piecewise-constant linear systems and track frozen-time optima on average.

T0 review reviewed 2026-07-12 challenge →

load-bearing objection Solid Automatica-style LTV extension of PGAC: one-step gradient updates plus normalized window LS, with real PES and average optimality-gap theorems for slow drift and switched modes. the 2 major comments →

arxiv 2607.03251 v1 pith:HDB4UVAS submitted 2026-07-03 math.OC cs.SYeess.SY

Adaptive Linear Quadratic Control of Unknown Linear Time-Varying Systems via Policy Gradient Methods

Feiran Zhao , Florian D\"orfler This is my paper
classification math.OC cs.SYeess.SY MSC 93C4093B5249N1093C05
keywords adaptive controllinear quadratic regulatorlinear time-varying systemspolicy gradientdata-driven controlsequential stabilitynormalized least-squares
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Unknown linear plants whose parameters drift or jump need a feedback gain that keeps changing from live closed-loop data. Most existing data-driven schemes recompute a full optimizer (Riccati equation or SDP) at every step; that is expensive and can swing wildly when the model estimate is noisy. This paper instead inserts a single policy-gradient step of the linear-quadratic cost into the feedback loop, using a local model estimated by normalized sliding-window least-squares. The resulting algorithm is cheap, its step-size directly limits how much the gain can jump, and the paper proves that the closed loop remains practically exponentially stable for two standard classes of time-varying systems. For slow continuous drift the proof needs no dwell-time condition; for abrupt switches a dwell-time contraction across modes is enough. Average frozen-time optimality gaps are also bounded, so the adapted gain tracks the ideal LQR of the current frozen model.

Core claim

Under small enough variation (or long enough dwell), small enough normalized identification error, and a sufficiently small gradient step-size, one-step policy-gradient adaptive control produces a sequentially stable gain sequence for slowly varying unknown LTV systems and therefore practical exponential stability without dwell time; for piecewise-constant LTV systems the same controller yields within-mode decay plus interval-wise practical exponential stability under a dwell-time condition, together with explicit average frozen-time LQR optimality-gap bounds in both cases.

What carries the argument

Policy-gradient adaptive control (PGAC): at each instant form a normalized sliding-window least-squares model of the recent closed-loop trajectory, evaluate the LQR policy gradient of that frozen surrogate, and take exactly one descent step; sequential stability of the resulting gain sequence then supplies the closed-loop certificates.

Load-bearing premise

The closed-loop data must stay persistently exciting under the adapting gain itself; if the normalized excitation level collapses, the identification-error bounds fail and the stability thresholds become empty.

What would settle it

Run the algorithm on a slowly varying LTV plant whose continuous drift exceeds the paper's (non-constructive) variation threshold while keeping excitation, noise and step-size fixed; if the state still decays practically exponentially, the slow-variation claim is false. Alternatively, drop the probing signal so that the normalized Gramian falls below the assumed floor and check whether the identification error and subsequent stability guarantees remain valid.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper proposes policy gradient adaptive control (PGAC) for unknown LTV systems: at each time the state-feedback gain is updated by one gradient step of a frozen-time LQR cost built from a normalized sliding-window least-squares model estimate. For slowly varying systems (Assumptions 2–4, 11), Theorem 16 shows sequential stability of the policy sequence and practical exponential stability of the closed loop without a dwell-time condition, plus an average frozen-time optimality-gap bound. For piecewise-constant systems (Assumptions 2, 3, 7, 11), Theorem 18 gives within-mode decay and interval-wise PES under a dwell-time condition, with a corresponding optimality-gap bound that depends on switching frequency. Identification-error lemmas (14, 17), cost boundedness by induction, sequential-stability arguments, and telescoping gap estimates are developed in the appendix; three numerical examples (continuous LTV, switched LTV, nonlinear planar quadrotor) illustrate the method.

Significance. The contribution is a useful middle ground between one-shot certainty-equivalence LQR/SDP updates and purely direct data-driven schemes: the first-order update is computationally light and the stepsize directly limits policy chatter under noisy estimates. Extending sequential-stability certificates to continuous slow variation without dwell time, and obtaining uniform constants for infinitely many switches in the piecewise-constant case, are genuine advances over the authors’ preliminary switched result and over several recent LTV data-driven methods. The normalized sliding-window estimator cleanly separates variation, noise, and excitation in the error bounds. The analysis is fully spelled out in the appendix and the numerical examples (including a nonlinear plant) support the claims. The PE and smallness conditions are standard for this literature and are disclosed.

major comments (2)
  1. Assumption 11 requires a uniform lower bound γ on the normalized data Gramian for every window while the gain is adapting. Lemmas 14 and 17 and all subsequent thresholds in Theorems 16 and 18 collapse if closed-loop excitation fails. The manuscript should either (i) give a concrete, checkable construction of e_t that preserves γ under the sequential-stability bounds already proved, or (ii) state explicitly that PE is an external hypothesis and discuss how γ can be monitored online. Without one of these, the certificates remain conditional on an assumption that is not automatically inherited from the closed-loop dynamics.
  2. The admissible ranges for δ, ε (or ε_sw) and η in Theorems 16 and 18 are expressed via existential constants ν_i that depend on (a_m, b_m, Q, R, K_t0) but are never made explicit or estimated. While this is common in sequential-stability arguments, a short remark (or a numerical illustration of how large η may be taken for the examples of Section 5) would make the results more usable and would clarify that the conditions are not vacuous for the simulated systems.
minor comments (4)
  1. In Algorithm 1 the data window used for identification is written (X_{t+1}, U_{t+1}, X_{t+2}) while the surrounding text uses the indexing of (12); a one-line clarification of the time shift would avoid confusion.
  2. Figures 2–5 would benefit from a short caption note on the magnitude of the residual set relative to e_m and w_m, so that the practical-stability claim is visually linked to the theory.
  3. A few typos remain (e.g., “D¨ orfler”, “Z¨ urich” in the author block; “the pol-icy update” line break in Section 4.2). A light copy-edit pass is enough.
  4. The discussion of DeePO and discounted LQR in Section 4.4 is interesting but could be shortened; the main theorems already stand without it.

Circularity Check

0 steps flagged

No significant circularity: LTV stability and optimality-gap certificates are derived from PE, small-variation/dwell-time, and stepsize assumptions via standard Lyapunov/gradient-dominance arguments, not by construction from fitted quantities or load-bearing self-citation.

full rationale

The derivation chain is self-contained against its stated assumptions. Identification errors (Lemmas 14, 17) follow from the normalized data equation plus uniform PE (Assumption 11) and variation bounds; they are not defined in terms of the stability claim. Policy updates use certainty-equivalence PG of the frozen-time LQR cost of the windowed estimate; sequential stability (Definition 15, from the literature) is then proved under small δ, ε (or ε_sw), and η (Lemmas 25–27, Theorems 16 and 18), and PES / interval-wise PES and average gaps C_t(K_t)−C_t^* follow as consequences. Gaps are measured against true frozen-time LQR costs of (A_t,B_t), not against a quantity defined by the fitted gain. Self-citations to the authors’ LTI PGAC work supply standard gradient formulas and perturbation lemmas; they do not force the LTV certificates by uniqueness or ansatz. Remark 12 explicitly breaks the ordinary-LS state-bound circularity by using normalization. No fitted-input-as-prediction, self-definitional, or renaming pattern appears in the load-bearing theorems.

Axiom & Free-Parameter Ledger

3 free parameters · 7 axioms · 0 invented entities

The central claims rest on standard adaptive-control domain assumptions (frozen-time controllability, bounded matrices/noise, PE probing, initial stabilizing gain) plus smallness of drift/identification error/stepsize relative to non-constructive constants. Free design knobs are the step size, window length, and probing amplitude. No new physical entities are postulated; sequential stability and PES are imported notions used as proof tools.

free parameters (3)
  • policy gradient stepsize η
    Hand-chosen design parameter; theorems require η small enough relative to identification error and smoothness constants; simulations use η=0.05.
  • sliding window length L
    Design parameter with L ≥ m+n for identifiability; enters identification bias as Lδ/γ (slow case) or N=⌈L/τ⌉ (switched case); simulations use L=20.
  • probing signal bound e_m and excitation level γ
    Excitation amplitude and normalized PE lower bound are design choices that trade residual state size against identification accuracy; assumed to exist for all t.
axioms (7)
  • domain assumption Frozen-time controllability of (A_t, B_t) for every t (Assumption 2)
    Needed so each frozen LQR problem is well-posed and stabilizing gains exist.
  • domain assumption Uniform bounds on ||A_t||, ||B_t|| and process noise ||w_t|| (Assumption 3)
    Used to get uniform cost bounds and polynomial constants in the perturbation lemmas.
  • domain assumption Slow variation ||Δ_t|| ≤ δ (Assumption 4) or bounded switching jumps (Assumption 7)
    Load-bearing for sequential stability (slow case) and identification bias (both cases).
  • domain assumption Bounded normalized PE of closed-loop data for all t (Assumption 11)
    Gives uniform identification error bounds independent of state magnitude; must hold while the gain adapts.
  • domain assumption Initial policy K_t0 stabilizes the initial frozen-time pair
    Standard adaptive LQR starting assumption; induction of cost boundedness begins from it.
  • standard math Policy gradient formula and gradient dominance for LQR (Lemma 1 / Fazel et al.)
    Background RL/control result used to define the one-step update and descent inequalities.
  • standard math Sequential stability implies practical exponential stability / ISS-type bounds (Cohen et al.)
    Bridge from policy-sequence properties to state bounds in Theorem 16.

reviewed 2026-07-12 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Linear Quadratic Control of Unknown Linear Time-Varying Systems via Policy Gradient Methods." pith.science (2026). https://pith.science/paper/HDB4UVAS

@misc{pith2026260703251,
  author       = {Pith},
  title        = {Pith review of: Adaptive Linear Quadratic Control of Unknown Linear Time-Varying Systems via Policy Gradient Methods},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HDB4UVAS}},
  note         = {Machine review of arXiv:2607.03251}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Unknown linear time-varying (LTV) systems require the control policy to adapt from online closed-loop data as dynamics evolve. Existing methods usually update the policy by solving a one-shot optimization problem, which can be computationally demanding and sensitive to noisy model estimates. In this paper, we propose a policy gradient adaptive control (PGAC) method for LTV system control with unknown model parameters. Specifically, PGAC integrates online policy optimization into feedback by updating the state-feedback policy with one-step gradient descent of the linear quadratic regulator cost at each time instant. This incremental update is computationally light and naturally limits policy variations caused by noisy data. To explicitly compute the policy gradient online, we estimate local models from recent closed-loop trajectories using normalized sliding-window least-squares. We provide stability and convergence certificates of PGAC for two classes of LTV systems. For slowly time-varying systems, we prove that the closed-loop system achieves practical exponential stability without a dwell-time condition. For piecewise-constant LTV systems, we establish practical stability through a dwell-time contraction argument. We also provide average frozen-time optimality-gap bounds of the policy sequence for both classes. Finally, we validate the effectiveness of our method via numerical case studies of both LTV and nonlinear systems.

Figures

Figures reproduced from arXiv: 2607.03251 by Feiran Zhao, Florian D\"orfler.

Figure 1
Figure 1. Figure 1: Illustration of the policy gradient adaptive control [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: illustrates the state norm ∥xt∥ and the spectral ra￾dius of the open-loop matrix ρ(At). As indicated by The￾orem 16, the PGAC algorithm successfully stabilizes the system, and the state rapidly converges to a small neigh￾borhood of the origin induced by the probing signal. This happens despite the fact that the open-loop matrix At is time-varying and unstable for parts of the trajectory, as reflected by ρ(… view at source ↗
Figure 3
Figure 3. Figure 3: Frozen-time optimality gap of PGAC and the fixed [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Spectral radius ρ(At) and state norm ∥xt∥ for the piecewise-constant time-varying system under PGAC with bounded uniform noise. Vertical dashed lines indicate switch￾ing instants. 0 50 100 150 200 Time step t 10 10 10 7 10 4 10 1 10 2 optimality gap PGAC fixed nominal LQR [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Frozen-time optimality gap of PGAC and the fixed [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references · 3 linked inside Pith

  1. [1]

    Derivative-free data-enabled optimal tracking (DF-DeeOT) controller for a PV grid-connected inverter.IEEE Access, 14:16406–16420, 2026

    Said Al-Abri, Myada Shadoul, and Hassan Yousef. Derivative-free data-enabled optimal tracking (DF-DeeOT) controller for a PV grid-connected inverter.IEEE Access, 14:16406–16420, 2026

    2026

  2. [2]

    Courier Corporation, 2007

    Brian DO Anderson and John B Moore.Optimal control: linear quadratic methods. Courier Corporation, 2007

    2007

  3. [3]

    Stability of certainty-equivalent adaptive LQR for linear systems with unknown time-varying parameters.arXiv preprint arXiv:2511.08236, 2025

    Marcell Bartos, Johannes K¨ ohler, Florian D¨ orfler, and Melanie N Zeilinger. Stability of certainty-equivalent adaptive LQR for linear systems with unknown time-varying parameters.arXiv preprint arXiv:2511.08236, 2025

    Pith/arXiv arXiv 2025

  4. [4]

    Data-driven model predictive control with stability and robustness guarantees.IEEE Transactions on Automatic Control, 66(4):1702–1717, 2020

    Julian Berberich, Johannes K¨ ohler, Matthias A M¨ uller, and Frank Allg¨ ower. Data-driven model predictive control with stability and robustness guarantees.IEEE Transactions on Automatic Control, 66(4):1702–1717, 2020

    2020

  5. [5]

    Data-driven predictive control in a stochastic setting: a unified framework.Automatica, 152:110961, 2023

    Valentina Breschi, Alessandro Chiuso, and Simone Formentin. Data-driven predictive control in a stochastic setting: a unified framework.Automatica, 152:110961, 2023

    2023

  6. [6]

    Online linear quadratic control

    Alon Cohen, Avinatan Hasidim, Tomer Koren, Nevena Lazic, Yishay Mansour, and Kunal Talwar. Online linear quadratic control. InInternational Conference on Machine Learning, pages 1029–1038. PMLR, 2018

    2018

  7. [7]

    Learning linear-quadratic regulators efficiently with only √ Tregret

    Alon Cohen, Tomer Koren, and Yishay Mansour. Learning linear-quadratic regulators efficiently with only √ Tregret. In International Conference on Machine Learning, pages 1300–

  8. [8]

    Data- enabled predictive control: In the shallows of the DeePC

    Jeremy Coulson, John Lygeros, and Florian D¨ orfler. Data- enabled predictive control: In the shallows of the DeePC. In 18th European Control Conference (ECC), pages 307–312, 2019

    2019

  9. [9]

    Online adaptation of data-driven controllers for unknown nonlinear systems.International Journal of Robust and Nonlinear Control, 36(3):1086–1095, 2026

    Xiaoyan Dai, Claudio De Persis, and Nima Monshizadeh. Online adaptation of data-driven controllers for unknown nonlinear systems.International Journal of Robust and Nonlinear Control, 36(3):1086–1095, 2026

    2026

  10. [10]

    Formulas for data-driven control: Stabilization, optimality, and robustness.IEEE Transactions on Automatic Control, 65(3):909–924, 2019

    Claudio De Persis and Pietro Tesi. Formulas for data-driven control: Stabilization, optimality, and robustness.IEEE Transactions on Automatic Control, 65(3):909–924, 2019

    2019

  11. [11]

    On the sample complexity of the linear quadratic regulator.Foundations of Computational Mathematics, 20(4):633–679, 2020

    Sarah Dean, Horia Mania, Nikolai Matni, Benjamin Recht, and Stephen Tu. On the sample complexity of the linear quadratic regulator.Foundations of Computational Mathematics, 20(4):633–679, 2020

    2020

  12. [12]

    Data-driven control: Part two of two: Hot take: Why not go with models?IEEE Control Systems Magazine, 43(6):27–31, 2023

    Florian D¨ orfler. Data-driven control: Part two of two: Hot take: Why not go with models?IEEE Control Systems Magazine, 43(6):27–31, 2023

    2023

  13. [13]

    On the certainty-equivalence approach to direct data-driven LQR design.IEEE Transactions on Automatic Control, 68(12):7989–7996, 2023

    Florian D¨ orfler, Pietro Tesi, and Claudio De Persis. On the certainty-equivalence approach to direct data-driven LQR design.IEEE Transactions on Automatic Control, 68(12):7989–7996, 2023

    2023

  14. [14]

    Predictive active steering control for autonomous vehicle systems.IEEE Transactions on Control Systems Technology, 15(3):566–580, 2007

    Paolo Falcone, Francesco Borrelli, Jahan Asgari, Hongtei Eric Tseng, and Davor Hrovat. Predictive active steering control for autonomous vehicle systems.IEEE Transactions on Control Systems Technology, 15(3):566–580, 2007

    2007

  15. [15]

    Global convergence of policy gradient methods for the linear quadratic regulator

    Maryam Fazel, Rong Ge, Sham Kakade, and Mehran Mesbahi. Global convergence of policy gradient methods for the linear quadratic regulator. InInternational Conference on Machine Learning, pages 1467–1476, 2018

    2018

  16. [16]

    Toward a theoretical foundation of policy optimization for learning control policies.Annual Review of Control, Robotics, and Autonomous Systems, 6:123–158, 2023

    Bin Hu, Kaiqing Zhang, Na Li, Mehran Mesbahi, Maryam Fazel, and Tamer Ba¸ sar. Toward a theoretical foundation of policy optimization for learning control policies.Annual Review of Control, Robotics, and Autonomous Systems, 6:123–158, 2023

    2023

  17. [17]

    A hybrid systems framework for data-based adaptive control of linear time- varying systems.IEEE Transactions on Automatic Control, 2025

    Andrea Iannelli and Romain Postoyan. A hybrid systems framework for data-based adaptive control of linear time- varying systems.IEEE Transactions on Automatic Control, 2025

    2025

  18. [18]

    Khalil.Nonlinear Systems

    Hassan K. Khalil.Nonlinear Systems. Prentice Hall, 3 edition, 2002

    2002

  19. [19]

    Lakshmikantham, S

    V. Lakshmikantham, S. Leela, and A. A. Martynyuk. Practical Stability of Nonlinear Systems. World Scientific, 1990

    1990

  20. [20]

    Adaptive control of unknown linear switched systems via policy gradient methods.European Control Conference (ECC), 2026, accepted

    Felix Laurent, Feiran Zhao, Jaap Eising, and Florian D¨ orfler. Adaptive control of unknown linear switched systems via policy gradient methods.European Control Conference (ECC), 2026, accepted

    2026

  21. [21]

    Online data- driven adaptive control for unknown linear time-varying systems

    Shenyu Liu, Kaiwen Chen, and Jaap Eising. Online data- driven adaptive control for unknown linear time-varying systems. In2023 62nd IEEE Conference on Decision and Control (CDC), pages 8775–8780. IEEE, 2023

    2023

  22. [22]

    Almost surely √ Tregret for adaptive lqr.IEEE Transactions on Automatic Control, 70(8):5145– 5159, 2025

    Yiwen Lu and Yilin Mo. Almost surely √ Tregret for adaptive lqr.IEEE Transactions on Automatic Control, 70(8):5145– 5159, 2025

    2025

  23. [23]

    Certainty equivalence is efficient for linear quadratic control

    Horia Mania, Stephen Tu, and Benjamin Recht. Certainty equivalence is efficient for linear quadratic control. In Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019

    2019

  24. [24]

    Persistency of excitation criteria for linear, multivariable, time-varying systems.Mathematics of Control, Signals and Systems, 1(3):203–226, 1988

    Iven MY Mareels and Michel Gevers. Persistency of excitation criteria for linear, multivariable, time-varying systems.Mathematics of Control, Signals and Systems, 1(3):203–226, 1988

    1988

  25. [25]

    Online control of unknown time-varying dynamical systems.Advances in Neural Information Processing Systems, 34:15934–15945, 2021

    Edgar Minasyan, Paula Gradu, Max Simchowitz, and Elad Hazan. Online control of unknown time-varying dynamical systems.Advances in Neural Information Processing Systems, 34:15934–15945, 2021

    2021

  26. [26]

    Direct data- driven control of linear time-varying systems.IEEE Transactions on Automatic Control, 68(8):4888–4895, 2023

    Benita Nortmann and Thulasi Mylvaganam. Direct data- driven control of linear time-varying systems.IEEE Transactions on Automatic Control, 68(8):4888–4895, 2023

    2023

  27. [27]

    An adaptive data- enabled policy optimization approach for autonomous bicycle control.IEEE Transactions on Control Systems Technology, 2026 (early access)

    Niklas Persson, Feiran Zhao, Mojtaba Kaheni, Florian D¨ orfler, and Alessandro V Papadopoulos. An adaptive data- enabled policy optimization approach for autonomous bicycle control.IEEE Transactions on Control Systems Technology, 2026 (early access)

    2026

  28. [28]

    Stable online control of linear time-varying systems

    Guannan Qu, Yuanyuan Shi, Sahin Lale, Anima Anandkumar, and Adam Wierman. Stable online control of linear time-varying systems. InLearning for Dynamics and Control, pages 742–753. PMLR, 2021

    2021

  29. [29]

    Online learning of data-driven controllers for unknown switched linear systems.Automatica, 145:110519, 2022

    Monica Rotulo, Claudio De Persis, and Pietro Tesi. Online learning of data-driven controllers for unknown switched linear systems.Automatica, 145:110519, 2022

    2022

  30. [30]

    Rugh and Jeff S

    Wilson J. Rugh and Jeff S. Shamma. Research on gain scheduling.Automatica, 36(10):1401–1425, 2000

    2000

  31. [31]

    Gain and phase margin for multiloop LQG regulators.IEEE Transactions on Automatic Control, 22(2):173–179, 1977

    Michael Safonov and Michael Athans. Gain and phase margin for multiloop LQG regulators.IEEE Transactions on Automatic Control, 22(2):173–179, 1977

    1977

  32. [32]

    Naive exploration is optimal for online LQR

    Max Simchowitz and Dylan Foster. Naive exploration is optimal for online LQR. InInternational Conference on Machine Learning, pages 8937–8948. PMLR, 2020

    2020

  33. [33]

    Adaptive model predictive control for linear time varying mimo systems.Automatica, 105:237–245, 2019

    Marko Tanaskovic, Lorenzo Fagiano, and Vojislav Gligorovski. Adaptive model predictive control for linear time varying mimo systems.Automatica, 105:237–245, 2019

    2019

  34. [34]

    Teel and Laurent Praly

    Andrew R. Teel and Laurent Praly. Tools for semiglobal stabilization by partial state and output feedback.SIAM Journal on Control and Optimization, 33(5):1443–1488, 1995

    1995

  35. [35]

    Statistical learning theory for control: A finite-sample perspective.IEEE Control Systems Magazine, 43(6):67–97, 2023

    Anastasios Tsiamis, Ingvar Ziemann, Nikolai Matni, and George J Pappas. Statistical learning theory for control: A finite-sample perspective.IEEE Control Systems Magazine, 43(6):67–97, 2023

    2023

  36. [36]

    Data informativity: a new perspective on data-driven analysis and control.IEEE Transactions on Automatic Control, 65(11):4753–4768, 2020

    Henk J Van Waarde, Jaap Eising, Harry L Trentelman, and M Kanat Camlibel. Data informativity: a new perspective on data-driven analysis and control.IEEE Transactions on Automatic Control, 65(11):4753–4768, 2020

    2020

  37. [37]

    Xuerui Wang, Feiran Zhao, Andres J¨ urisson, Florian D¨ orfler, and Roy S. Smith. Unified aeroelastic flutter and loads control via data-enabled policy optimization. IEEE Transactions on Aerospace and Electronic Systems, 61(5):11437–11449, 2025

    2025

  38. [38]

    Data-driven online control for real-time optimal economic dispatch and temperature regulation in district heating systems.arXiv preprint arXiv:2603.23748, 2026

    Xinyi Yi and Ioannis Lestas. Data-driven online control for real-time optimal economic dispatch and temperature regulation in district heating systems.arXiv preprint arXiv:2603.23748, 2026

    Pith/arXiv arXiv 2026

  39. [39]

    Policy gradient adaptive control for the LQR: Indirect and direct approaches.arXiv preprint arXiv:2505.03706, 2025

    Feiran Zhao, Alessandro Chiuso, and Florian D¨ orfler. Policy gradient adaptive control for the LQR: Indirect and direct approaches.arXiv preprint arXiv:2505.03706, 2025

    Pith/arXiv arXiv 2025

  40. [40]

    Data-enabled policy optimization for direct adaptive learning of the LQR.IEEE Transactions on Automatic Control, 70(11):7217–7232, 2025

    Feiran Zhao, Florian D¨ orfler, Alessandro Chiuso, and Keyou You. Data-enabled policy optimization for direct adaptive learning of the LQR.IEEE Transactions on Automatic Control, 70(11):7217–7232, 2025

    2025

  41. [41]

    Convergence and sample complexity of policy gradient methods for stabilizing linear systems.IEEE Transactions on Automatic Control, 70(3):1455–1466, 2025

    Feiran Zhao, Xingyun Fu, and Keyou You. Convergence and sample complexity of policy gradient methods for stabilizing linear systems.IEEE Transactions on Automatic Control, 70(3):1455–1466, 2025. 17

    2025

  42. [42]

    Direct adaptive control of grid-connected power converters via output-feedback data- enabled policy optimization

    Feiran Zhao, Ruohan Leng, Linbin Huang, Huanhai Xin, Keyou You, and Florian D¨ orfler. Direct adaptive control of grid-connected power converters via output-feedback data- enabled policy optimization. In2025 European Control Conference (ECC), pages 2563–2568. IEEE, 2025. 18

    2025

This paper was first reviewed by grok-4.5 on July 12, 2026.