Pith. sign in

REVIEW 3 major objections 6 minor 45 references

A regularized projection of the fundamental-lemma weights cuts prediction error under process noise and keeps adaptive data-driven control recursively feasible with high probability.

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 ↗

T0 review · grok-4.5

2026-07-31 17:09 UTC pith:AJVCCTIB

load-bearing objection Clean incremental package: regularized-g collapsed to fixed covariances, modest EIV gains, and a standard SNM tube for adaptive SPC—solid systems work, not a foundational reset. the 3 major comments →

arxiv 2607.28123 v1 pith:AJVCCTIB submitted 2026-07-30 eess.SY cs.SY

Projection-Regularized Indirect Data-Driven Predictive Control

Mahmood Mazare , Hossein Ramezani This is my paper
classification eess.SY cs.SY
keywords data-driven controlpredictive controlregularizationadaptive controllinear time-varying systemssubspace methodsmartingale concentration
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.

Indirect data-driven predictive control is usually brittle when process noise corrupts the regressors or when sliding data windows lose rank. This paper keeps the latent weight vector from the fundamental lemma, applies a Tikhonov-regularized projection that matches past data softly and future inputs strictly, and shows the whole construction collapses analytically into fixed-dimension covariance matrices. The resulting Projection-Regularized Predictive Control (PRPC) strictly lowers mean-squared prediction error relative to unregularized subspace methods under errors-in-variables noise and remains well-defined when the sample size falls below the regressor dimension. The same covariances support a sliding-window adaptive controller for linear time-varying plants; closed-loop correlations are handled by a uniform-in-time finite-sample radius from vector-valued self-normalized martingales. Embedding that radius into dynamically tightened constraints yields robust recursive feasibility and input-to-state practical stability at a user-chosen confidence level, while the online problem stays a low-dimensional quadratic program independent of data length.

Core claim

Under process noise and rank-deficient windows, the regularized fundamental-lemma projection, condensed into empirical cross-covariances, produces a multi-step predictor whose mean-squared error is strictly smaller than unregularized subspace predictive control for some positive regularization, and whose uniform statistical uncertainty radius, once used to tighten constraints, guarantees recursive feasibility and input-to-state practical stability with probability at least 1−δ.

What carries the argument

Analytical covariance collapse of the regularized projection: the high-dimensional weight vector is eliminated offline into fixed-size matrices (Σpp, Σup, Σyp, …), so the online controller is a low-dimensional QP whose latency does not grow with data length; the same objects carry recursive blending and the self-normalized martingale radius that tightens the constraints.

Load-bearing premise

Safety needs a known uniform bound on how far the true multi-step map can drift from the nominal predictor at every time, plus a known noise size; neither is learned from data in the paper.

What would settle it

On a process-noise plant with scarce data, check whether cross-validated PRPC yields lower predictor MSE than unregularized SPC; under slow LTV drift, check whether the mismatch-augmented radius achieves the claimed coverage while the naive noise-only radius does not. If the MSE ratio never falls below one, or coverage fails the target, the central claims fail.

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

If this is right

  • Online controller latency is decoupled from the length of the collected trajectory.
  • Sliding-window LTV adaptation stays mathematically well-posed even when feedback kills persistent excitation.
  • Constraint tightening can be driven by a finite-sample, uniform-in-time statistical radius rather than purely worst-case tubes.
  • Under pure measurement noise and abundant data the method recovers ordinary subspace predictive control, so no extra conservatism is forced.
  • Rank-deficient windows no longer make the predictor undefined.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same covariance collapse could fix the online dimension of other regularized Hankel or instrumental-variable predictors.
  • If the drift bound were estimated from residual growth instead of supplied a priori, the guarantees would become fully data-driven for unknown time variation.
  • The bias–variance analysis implies the largest practical gains appear exactly in early commissioning and sudden plant-change regimes where data are scarce and process noise dominates.

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

3 major / 6 minor

Summary. The paper proposes Projection-Regularized Predictive Control (PRPC), an indirect data-driven predictive controller that retains a Tikhonov-regularized fundamental-lemma weight g subject to a hard future-input constraint and a soft past-data match. The KKT system is analytically collapsed onto fixed-dimension sample covariances (Proposition 1), recovering unregularized SPC as λ→0 (Theorem 1) while remaining well-posed under rank deficiency. A bias–variance analysis argues that under process noise (errors-in-variables) there exists λ*>0 strictly reducing predictor MSE relative to SPC (Propositions 3–4), and that PRPC stays defined when M<Th (Proposition 5). These properties support an adaptive sliding-window scheme via convex blending of offline and exponentially forgotten online covariances. Closed-loop prediction error is bounded uniformly in time by a self-normalized martingale radius plus a deterministic mismatch term (Theorem 2); embedding the radius into dynamically tightened constraints yields high-probability recursive feasibility and ISpS (Theorems 3–4). LTI (Boeing 747) and polytopic LTV simulations illustrate conditioning gains, modest MSE reductions, tracking, and empirical coverage of the bound.

Significance. If the claims hold, the work cleanly unifies computational advantages of subspace predictors with regularization-based robustness and supplies finite-sample, closed-loop-valid safety certificates for adaptive LTV data-driven MPC—addressing a recognized gap between DeePC-style methods and classical robust tube MPC under correlated online data. Strengths that deserve explicit credit: the covariance collapse (Proposition 1) is algebraically exact and numerically verified to machine precision against SPC (Fig. 1); the PRPC–SPC limit (Theorem 1) and rank-deficient well-posedness (Proposition 5) are sharp; the multi-output SNM argument and standard robust-MPC feasibility/ISpS scaffolding are carefully adapted to the blended-covariance predictor. The free design knobs (λ, γ, ρ_f) are standard and the circularity risk is low. The main practical caveat is that the safety radius inherits a known uniform mismatch bound L_Θ and noise proxy c_w, which limits turnkey use on fully unknown LTV plants; the EIV “strict reduction” claim is real but modest in the reported numerics.

major comments (3)
  1. [§4.2, Proposition 4, Appendix D; Abstract; §8] Proposition 4 and Appendix D assert existence of λ*>0 with MSE(PRPC)<MSE(SPC) under EIV by showing ∂MSE/∂λ|λ=0<0. The argument is only qualitative: it requires the residual-variance term to dominate a bounded bias derivative “whenever σ_ζ² > θ_i² a_i(1−a_i)/s_i⁰ for the significant directions,” without an explicit, checkable condition in terms of (σ_w/σ_v, M, spectrum of Σ_Φ0). The Monte-Carlo evidence (Figs. 2–4) shows only modest gains (~7% at M/Th≈1.3 under mild process noise; ~20% at σ_w/σ_v=10 and M/Th≈2.6). Please either (i) supply a sufficient condition that can be verified from the offline covariances and noise proxies, or (ii) temper the abstract/introduction language (“strictly reduces,” “transformative advantages,” “up to a 7% reduction”) to match the derivative-level guarantee and the reported effect sizes.
  2. [§6.1 Assumption 1; Theorem 2; Theorems 3–4; §7.2] Assumption 1 requires a known uniform operator-norm bound L_Θ on the entire parametric mismatch ΔΘ(k) for all k, together with a known conditional sub-Gaussian proxy c_w. Both enter the radius r_k (Theorem 2) and therefore the tightened sets Y_tight(k), the terminal set inclusion, recursive feasibility (Theorem 3), and the ISpS ultimate ball O(r_∞) (Theorem 4). For a truly unknown LTV plant this bound is not identified from data in the paper. The LTV experiments use a known polytopic hull, so L_Θ is available by construction; the manuscript should state clearly how L_Θ (and c_w) are obtained in practice, discuss sensitivity of the closed-loop certificates when L_Θ is over-/under-estimated, or provide a data-driven outer bound. Without this, the high-probability safety claim for adaptive LTV operation rests on an a-priori quantity that the method does not produce.
  3. [§5, Lemma 2; §7.2 ablation; Lemma 3] Lemma 2 and the adaptive scheme rely on a strictly positive offline anchor weight γ∈(0,1) to guarantee Σ_act_pp(k) ⪰ γ Σ_off_pp ≻ 0 and hence uniform invertibility of W_p and S. The ablation (Fig. 9) shows that un-anchored recursion (γ=0) matches average-case cost under persistent excitation, so the anchor’s value is purely structural. Please quantify how large γ must be relative to the offline PE margin and the forgetting factor ρ_f so that the spectral floor remains useful (i.e., r_∞ does not become vacuous), and discuss the bias introduced into the predictor when the offline model is far from the current LTV vertex. This is load-bearing for the claim that the adaptive controller remains well-posed and non-conservative under loss of online excitation.
minor comments (6)
  1. [§7.1, Fig. 6] Fig. 6 caption and text refer to “γ-DPC” and “GDPC” somewhat interchangeably; please align nomenclature with the cited GDPC reference and define every baseline acronym at first use in §7.1.
  2. [§5, Eq. (13)] In (13) the factor M that “normalizes the update to the scale of the offline data” is not defined in the notation section; clarify whether it is the offline column count or a free gain.
  3. [Theorem 2; Appendix E] Theorem 2 writes β_k with √(1+λ ϱ(V(k)⁻¹)); a one-line justification that this factor absorbs the ridge initialization bias (beyond the sketch in Appendix E) would help readers reproduce the constant.
  4. [Throughout; §2–§4 headings] Typos / style: “F undamental Lemma”, “V ariance”, “T uning”, “F easibility” (spurious spaces after capitals in several headings); “HV AC”; “minλ MSE” formatting in Fig. 3; “plim” vs. “p-lim” consistency in Appendix C.
  5. [§8] The conclusion claims “reduces the dominant offline computational load by over 50% compared to unregularized SPC”; SPC’s classical cost is also O(T_h² M) or better with covariance form—please cite the exact baseline complexity being compared or soften the percentage claim.
  6. [§7.1–7.2] Table 1 / Table 2 matrices are hard to parse in text form; if the journal allows, provide them as supplementary files or increase column formatting for readability.

Circularity Check

0 steps flagged

No significant circularity: derivations are algebraic (KKT/covariance collapse), classical EIV bias-variance, and standard SNM concentration, not self-defining or fit-then-predict.

full rationale

The load-bearing chain is self-contained and non-circular. PRPC is defined by a Tikhonov-regularized projection (Eq. 6); the predictor matrices follow from KKT elimination and the matrix push-through identity (Prop. 1, App. A), recovering SPC as λ→0 by minimum-norm selection (Thm. 1)—an algebraic limit, not a definition of the target in terms of itself. The bias–variance claims (Props. 2–4, Apps. B–D) are classical EIV/ridge arguments: unregularized SPC is inconsistent under process-noise contamination of the regressor, and ∂MSE/∂λ|λ=0 can be negative when residual variance dominates, so some λ*>0 exists; λ is then chosen by open-loop cross-validation, a design knob, not a fitted quantity re-labeled as a theorem prediction. Adaptive well-posedness (Lemma 2) is a convex blend with offline PE data. Finite-sample safety (Thm. 2) applies vector-valued self-normalized martingale concentration to the closed-loop estimation error under stated sub-Gaussian and mismatch assumptions; recursive feasibility and ISpS (Thms. 3–4) embed that radius into tightened sets with standard robust-MPC terminal ingredients. No self-citations by the present authors carry uniqueness or forbid alternatives; external citations (Willems, SPC, SNM literature) are used as tools, not as smuggled ansätze that force the result. Simulations validate coverage and tracking; they do not close a definitional loop. Score 0 is appropriate.

Axiom & Free-Parameter Ledger

5 free parameters · 6 axioms · 2 invented entities

The central guarantees rest on behavioral PE/fundamental lemma, standard robust-MPC terminal ingredients, known noise and mismatch bounds for concentration, and design hyperparameters (λ, γ, ρ_f, δ). No new physical entities are postulated; the ‘invented’ object is the PRPC predictor construction itself. Free parameters are the usual control knobs plus the a-priori mismatch radius that makes the probabilistic tube non-vacuous.

free parameters (5)
  • λ (Tikhonov regularization) = tuned per dataset via CV; limit λ→0 recovers SPC
    Balances initial-condition fit vs weight norm; chosen by k-fold CV on open-loop errors; existence of λ*>0 is claimed under EIV but value is data-dependent.
  • γ (offline covariance anchor weight) = 0.01 in §7.2
    Convex blend Σ_act = γ Σ_off + (1−γ) Σ_on; enforces spectral floor for well-posedness; set to 0.01 in LTV experiments.
  • ρ_f (exponential forgetting factor) = 0.95 in §7.2
    Controls online covariance adaptation speed; hand-chosen for LTV runs.
  • L_Θ (uniform parametric mismatch bound) = treated as known constant; used in mismatch-augmented radius
    Enters r_k additively; required known for all k under Assumption 1; not estimated in the paper.
  • c_w (sub-Gaussian variance proxy) and δ (failure probability) = δ=0.05; c_w from σ²=0.02 in LTV
    Set the SNM radius β_k and confidence level 1−δ; c_w linked to noise variance in sims (e.g. 2√σ²).
axioms (6)
  • domain assumption Willems’ Fundamental Lemma: PE inputs of sufficient order imply all trajectories lie in the column space of the Hankel data matrix (Lemma 1).
    Foundation for representing future I/O via g; invoked throughout §§2–3.
  • domain assumption Noise is a martingale-difference sequence, conditionally sub-Gaussian with known proxy c_w (Assumption 1).
    Required for vector SNM concentration in Theorem 2 / Appendix E.
  • ad hoc to paper Parametric LTV mismatch is uniformly bounded in operator norm by a known L_Θ for all time (Assumption 1).
    Makes the deterministic part of r_k finite; strong for unknown plants and not learned online.
  • domain assumption Robust control-invariant terminal set Y_f and local law κ_f exist inside Y⊖B_{r_∞} (Assumption 2).
    Standard robust MPC ingredient used for recursive feasibility and ISpS (Theorems 3–4).
  • domain assumption Offline data are persistently exciting so Σ_off_pp ≻ 0 and U_off_f has full row rank (Lemma 2).
    Anchors active covariances away from singularity under loss of online PE.
  • standard math Self-normalized martingale concentration inequalities for vector-valued processes apply to the closed-loop regressor-noise pair.
    Cited tool behind the uniform-in-time β_k bound; proof sketch in Appendix E.
invented entities (2)
  • Projection-Regularized Predictive Control (PRPC) predictor (regularized g with hard U_f g = u_N, soft Z_p match) no independent evidence
    purpose: Define a noise-robust indirect multi-step map that collapses to fixed covariances and recovers SPC as λ→0.
    Methodological construct, not a physical entity; independent evidence is the algebraic equivalence proofs and MSE simulations, not an external measurement.
  • Active covariance blend Σ_act = γ Σ_off + (1−γ) Σ_on no independent evidence
    purpose: Keep W_p and S uniformly well-posed during closed-loop loss of excitation for adaptive LTV control.
    Design mechanism introduced for Lemma 2; validated only inside the paper’s simulations.

pith-pipeline@v1.2.0-daily-grok45 · 25193 in / 4085 out tokens · 82555 ms · 2026-07-31T17:09:31.051486+00:00 · methodology

0 comments
read the original abstract

Indirect data-driven predictive control methods often suffer under process noise and data scarcity. This paper introduces Projection-Regularized Predictive Control (PRPC), retaining the fundamental-lemma weight vector via a regularized projection analytically condensed into an efficient, fixed-dimension covariance update. A rigorous bias--variance analysis proves PRPC strictly reduces prediction error under process noise (errors-in-variables) and structural rank deficiencies compared to unregularized subspace methods. We leverage these properties to develop an adaptive sliding-window controller for linear time-varying (LTV) systems. To guarantee safety despite closed-loop data correlations, we derive a uniform-in-time, finite-sample confidence bound on the empirical predictor using vector-valued martingale concentration inequalities. Embedding this statistical uncertainty radius into a dynamically tightened constraint set rigorously ensures robust recursive feasibility and Input-to-State practical Stability (ISpS) with high probability. Simulations on LTI and LTV benchmarks demonstrate real-time tractability and strict constraint satisfaction.

Figures

Figures reproduced from arXiv: 2607.28123 by Hossein Ramezani, Mahmood Mazare.

Figure 1
Figure 1. Figure 1: PRPC → SPC as λ → 0 +. The covariance￾collapse predictor matches SPC to machine precision, verifying Theorem 1. LTI benchmark to isolate the fundamental bias-variance trade-offs and benchmark closed-loop performance; and second, a multi-variable polytopic LTV system to ex￾plicitly evaluate the adaptive recursive mechanisms and the probabilistic finite-sample confidence bounds. 7.1 LTI Benchmark: Boeing 747… view at source ↗
Figure 3
Figure 3. Figure 3: Best-case MSE ratio vs. data length. The Regularization benefit concentrates in data-scarce regimes (M/Th → 1) under process noise. 10 5 10 3 10 1 10 1 10 3 ¸ 10 2 10 3 10 4 10 5 co n d(© © > + ¸ I) K = 200 K = 500 K = 2500 10 5 10 3 10 1 10 1 10 3 10 5 ¸ 0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 M S E P R P C=M S ESP C ¾w=¾v = 2 ¾w=¾v = 5 ¾w=¾v = 10 [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 6
Figure 6. Figure 6: Statistical comparison of the Input Energy [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 4
Figure 4. Figure 4: Top: Design matrix conditioning vs. λ for sev￾eral data lengths. Bottom: Predictor MSE ratio vs. λ at M/Th ≈ 2.6 for varying process-to-measurement noise ratios; both the benefit and the optimal λ ⋆ increase monotonically with σw/σv, and σw/σv = 10 yields the largest reduction. At very large λ the curves turn back upward as the predictor over-regularizes. dominates. For very large λ every curve turns back … view at source ↗
Figure 7
Figure 7. Figure 7: Adaptive PRPC regulation under i.i.d. poly [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Adaptive PRPC regulation under slow sinu [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Ablation under slow drift (σ 2 = 10−3 , 50 Monte-Carlo runs): closed-loop regulation cost for the fixed offline predictor, the un-anchored recursion, and the full Adaptive PRPC (log scale). Adaptation is es￾sential, as the fixed predictor is markedly worse with a heavy failure tail; the anchor matches the un-anchored performance while additionally guaranteeing uniform well-posedness (Lemma 2). 0.0 0.5 1.0 … view at source ↗
Figure 10
Figure 10. Figure 10: Empirical coverage of the prediction bound [PITH_FULL_IMAGE:figures/full_fig_p010_10.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

45 extracted references · 2 linked inside Pith

  1. [1]

    Handbook of linear data-driven predictive con- trol: Theory, implementation and design,

    P. C. N. Verheijen, V. Breschi, and M. Lazar, “Handbook of linear data-driven predictive con- trol: Theory, implementation and design,”Annual Reviews in Control, vol. 56, p. 100914, 2023

    2023

  2. [2]

    Learning-Based Efficient Approximation of Data- Enabled Predictive Control,

    Y. Zhou, Y. Lu, Z. Li, J. Yan, and Y. Mo, “Learning-Based Efficient Approximation of Data- Enabled Predictive Control,”IEEE Conference on Decision and Control, 2024

    2024

  3. [3]

    Formulas for Data- Driven Control: Stabilization, Optimality, and Ro- bustness,

    C. De Persis and P. Tesi, “Formulas for Data- Driven Control: Stabilization, Optimality, and Ro- bustness,”IEEE Transactions on Automatic Con- trol, 2019

    2019

  4. [4]

    Learning-Based Model Predictive Con- trol: Toward Safe Learning in Control,

    L. Hewing, K. P. Wabersich, M. Menner, and M. Zeilinger, “Learning-Based Model Predictive Con- trol: Toward Safe Learning in Control,”Annu. Rev. Control. Robotics Auton. Syst., 2020

    2020

  5. [5]

    Gain- Scheduling Data-Enabled Predictive Control for Nonlinear Systems with Linearized Operating Re- gions,

    S. Zieglmeier, M. Hudoba de Badyn, N. Warak- agoda, T. R. Krogstad, and P. Engelstad, “Gain- Scheduling Data-Enabled Predictive Control for Nonlinear Systems with Linearized Operating Re- gions,”arXiv.org, 2025

    2025

  6. [6]

    Behavioral systems theory in data-driven analysis, signal processing, and control,

    I. Markovsky and F. D¨ orfler, “Behavioral systems theory in data-driven analysis, signal processing, and control,”Annual Reviews in Control, 2021

    2021

  7. [7]

    Maximum Likelihood Estimation in Data-Driven Modeling and Control,

    M. Yin, A. Iannelli, and R. S. Smith, “Maximum Likelihood Estimation in Data-Driven Modeling and Control,”IEEE Transactions on Automatic Control, 2020

    2020

  8. [8]

    A note on persistency of excita- tion,

    J. C. Willems, P. Rapisarda, I. Markovsky, and B. L. M. De Moor, “A note on persistency of excita- tion,”Systems & Control Letters, vol. 54, no. 4, pp. 325–329, 2005

    2005

  9. [9]

    Data Informativity: A New Perspec- tive on Data-Driven Analysis and Control,

    H. J. van Waarde, J. Eising, H. Trentelman, and M. C ¸ amlibel, “Data Informativity: A New Perspec- tive on Data-Driven Analysis and Control,”IEEE Transactions on Automatic Control, 2019

    2019

  10. [10]

    Data-Driven Model Predictive Control With Stability and Robustness Guarantees,

    J. Berberich, J. K¨ ohler, M. M¨ uller, and F. Allg¨ ower, “Data-Driven Model Predictive Control With Stability and Robustness Guarantees,”IEEE Transactions on Automatic Control, 2019

    2019

  11. [11]

    Data-Enabled Predictive Control for Fast Charging of Lithium-Ion Batteries with Constraint Handling,

    K. Chen, K. Zhang, X. Lin, Y. Zheng, X. Yin, X. Hu, Z. Song, and Z. Li, “Data-Enabled Predictive Control for Fast Charging of Lithium-Ion Batteries with Constraint Handling,”arXiv.org, 2022

    2022

  12. [12]

    Data- Enabled Predictive Control: In the Shallows of the DeePC,

    J. Coulson, J. Lygeros, and F. D¨ orfler, “Data- Enabled Predictive Control: In the Shallows of the DeePC,”European Control Conference, 2018

    2018

  13. [13]

    Data-Enabled Predictive Control for Building HV AC Systems,

    V. Chinde, Y. Lin, and M. J. Ellis, “Data-Enabled Predictive Control for Building HV AC Systems,” Journal of Dynamic Systems Measurement, and Control, 2022

    2022

  14. [14]

    Decentralized Data-Enabled Predictive Control for Power System Oscillation Damping,

    L. Huang, J. Coulson, J. Lygeros, and F. D¨ orfler, “Decentralized Data-Enabled Predictive Control for Power System Oscillation Damping,” IEEE Transactions on Control Systems Technol- ogy, 2019

    2019

  15. [15]

    Data-enabled predictive control for quadcopters,

    E. Elokda, J. Coulson, P. Beuchat, J. Lygeros, and F. D¨ orfler, “Data-enabled predictive control for quadcopters,”International Journal of Robust and Nonlinear Control, 2021

    2021

  16. [16]

    DeeP-LCC: Data-EnablEd Predictive Leading Cruise Control in Mixed Traffic Flow,

    J. Wang, Y. Zheng, K. Li, and Q. Xu, “DeeP-LCC: Data-EnablEd Predictive Leading Cruise Control in Mixed Traffic Flow,”IEEE Transactions on Control Systems Technology, 2022

    2022

  17. [17]

    Data- driven predictive control in a stochastic setting: a unified framework,

    V. Breschi, A. Chiuso, and S. Formentin, “Data- driven predictive control in a stochastic setting: a unified framework,”Automatica, 2022

    2022

  18. [18]

    On the impact of regularization in data-driven predictive control,

    V. Breschi, A. Chiuso, M. Fabris, and S. For- mentin, “On the impact of regularization in data-driven predictive control,”arXiv preprint arXiv:2304.00263, 2023

    Pith/arXiv arXiv 2023

  19. [19]

    Quadratic Regularization of Data-Enabled Pre- dictive Control: Theory and Application to Power Converter Experiments,

    L. Huang, J. Zhen, J. Lygeros, and F. D¨ orfler, “Quadratic Regularization of Data-Enabled Pre- dictive Control: Theory and Application to Power Converter Experiments,”IF AC-PapersOnLine, 2020

    2020

  20. [20]

    Data-enabled pre- dictive control with instrumental variables: the di- rect equivalence with subspace predictive control,

    J.-W. van Wingerden, S. P. Mulders, R. Dinkla, T. Oomen, and M. Verhaegen, “Data-enabled pre- dictive control with instrumental variables: the di- rect equivalence with subspace predictive control,” IEEE 61st Conference on Decision and Control (CDC), pp. 2111–2116, 2022

    2022

  21. [21]

    Distribu- tionally Robust Chance Constrained Data-Enabled Predictive Control,

    J. Coulson, J. Lygeros, and F. D¨ orfler, “Distribu- tionally Robust Chance Constrained Data-Enabled Predictive Control,”IEEE Transactions on Auto- matic Control, 2020

    2020

  22. [22]

    Dis- tributionally Robust Stochastic Data-Driven Pre- dictive Control with Optimized Feedback Gain,

    R. Li, J. Simpson-Porco, and S. L. Smith, “Dis- tributionally Robust Stochastic Data-Driven Pre- dictive Control with Optimized Feedback Gain,” IEEE Conference on Decision and Control, 2024

    2024

  23. [23]

    Robust Data-Driven Pre- dictive Control of Unknown Nonlinear Systems us- ing Reachability Analysis,

    M. Farjadnia, A. Alanwar, M. U. B. Niazi, M. Moli- nari, and K. Johansson, “Robust Data-Driven Pre- dictive Control of Unknown Nonlinear Systems us- ing Reachability Analysis,”European Journal of Control, 2022

    2022

  24. [24]

    Cloud-Based Computational Data-Enabled Pre- dictive Control,

    L. Dai, T. Huang, R. Gao, Y. Zhang, and Y. Xia, “Cloud-Based Computational Data-Enabled Pre- dictive Control,”IEEE Internet of Things Journal, 2022. 13

    2022

  25. [25]

    Dimension Re- duction for Efficient Data-Enabled Predictive Con- trol,

    K. Zhang, Y. Zheng, and Z. Li, “Dimension Re- duction for Efficient Data-Enabled Predictive Con- trol,”IEEE Control Systems Letters, 2022

    2022

  26. [26]

    Online Reduced-Order Data- Enabled Predictive Control,

    A. Vahidi-Moghaddam, K. Zhang, X. Yin, V. Sri- vastava, and Z. Li, “Online Reduced-Order Data- Enabled Predictive Control,”IEEE Transactions on Automation Science and Engineering, 2024

    2024

  27. [27]

    Deep DeePC: Data-enabled predictive control with low or no online optimization using deep learning,

    X. Zhang, K. Zhang, Z. Li, and X. Yin, “Deep DeePC: Data-enabled predictive control with low or no online optimization using deep learning,” AIChE Journal, 2024

    2024

  28. [28]

    SPC: Subspace predictive control,

    W. Favoreel, B. De Moor, and M. Gevers, “SPC: Subspace predictive control,”Proceedings of the 14th IF AC World Congress, pp. 4004–4009, 1999

    1999

  29. [29]

    On the relationship be- tween data-enabled predictive control and sub- space predictive control,

    F. Fiedler and S. Lucia, “On the relationship be- tween data-enabled predictive control and sub- space predictive control,”European Control Con- ference, 2021

    2021

  30. [30]

    Bridg- ing Direct and Indirect Data-Driven Control For- mulations via Regularizations and Relaxations,

    F. D¨ orfler, J. Coulson, and I. Markovsky, “Bridg- ing Direct and Indirect Data-Driven Control For- mulations via Regularizations and Relaxations,” IEEE Transactions on Automatic Control, 2021

    2021

  31. [31]

    Generalized Data–Driven Predictive Control: Merging Subspace and Hankel Predictors,

    M. Lazar and P. Verheijen, “Generalized Data–Driven Predictive Control: Merging Subspace and Hankel Predictors,”Mathematics, 2023

    2023

  32. [32]

    Op- timal Data-Driven Prediction and Predictive Con- trol using Signal Matrix Models,

    R. S. Smith, M. Abdalmoaty, and M. Yin, “Op- timal Data-Driven Prediction and Predictive Con- trol using Signal Matrix Models,”arXiv preprint arXiv:2403.15329, 2024

    Pith/arXiv arXiv 2024

  33. [33]

    Trajectory-based data-driven predictive control and the state-space predictor,

    L. D. Premer, A. Khabbazi, and K. J. Kircher, “Trajectory-based data-driven predictive control and the state-space predictor,” 2026

    2026

  34. [34]

    Data-Driven Predictive Control Using Closed-Loop Data: An Instrumental Variable Ap- proach,

    Y. Wang, Y. Qiu, M. Sader, D. Huang, and C. Shang, “Data-Driven Predictive Control Using Closed-Loop Data: An Instrumental Variable Ap- proach,”IEEE Control Systems Letters, 2023

    2023

  35. [35]

    Adap- tive Robust Data-Driven Building Control via Bilevel Reformulation: An Experimental Result,

    Y. Lian, J. Shi, M. Koch, and C. Jones, “Adap- tive Robust Data-Driven Building Control via Bilevel Reformulation: An Experimental Result,” IEEE Transactions on Control Systems Technol- ogy, 2021

    2021

  36. [36]

    Data-Driven Predictive Control for Linear Parameter-Varying Systems,

    C. Verhoek, H. S. Abbas, R. T’oth, and S. Haesaert, “Data-Driven Predictive Control for Linear Parameter-Varying Systems,”IF AC- PapersOnLine, 2021

    2021

  37. [37]

    Robust and Kernelized Data-Enabled Predictive Control for Nonlinear Systems,

    L. Huang, J. Lygeros, and F. D¨ orfler, “Robust and Kernelized Data-Enabled Predictive Control for Nonlinear Systems,”IEEE Transactions on Con- trol Systems Technology, 2022

    2022

  38. [38]

    Neural Parameter-varying Data-enabled Predictive Control of Cold Atmospheric Pressure Plasma Jets,

    P. GhafGhanbari, M. Lazar, and J. Mohammad- pour, “Neural Parameter-varying Data-enabled Predictive Control of Cold Atmospheric Pressure Plasma Jets,”arXiv.org, 2025

    2025

  39. [39]

    Linear Tracking MPC for Nonlinear Systems—Part II: The Data-Driven Case,

    J. Berberich, J. K¨ ohler, M. M¨ uller, and F. Allg¨ ower, “Linear Tracking MPC for Nonlinear Systems—Part II: The Data-Driven Case,”IEEE Transactions on Automatic Control, 2021

    2021

  40. [40]

    Efficient Recur- sive Data-enabled Predictive Control,

    J. Shi, Y. Lian, and C. N. Jones, “Efficient Recur- sive Data-enabled Predictive Control,”European Control Conference, 2023

    2023

  41. [41]

    Data-Enabled Neighboring Extremal Optimal Control: A Computationally Efficient DeePC,

    A. Vahidi-Moghaddam, K. Zhang, Z. Li, and Y. Wang, “Data-Enabled Neighboring Extremal Optimal Control: A Computationally Efficient DeePC,”IEEE Conference on Decision and Con- trol, 2023

    2023

  42. [42]

    Gain-Scheduled Data-Enabled Predictive Control: A DeePC Approach for Nonlinear Sys- tems,

    M. A. Guerrero, B. Lakshminarayanan, and C. R. Rojas, “Gain-Scheduled Data-Enabled Predictive Control: A DeePC Approach for Nonlinear Sys- tems,”IEEE Control Systems Letters, 2025

    2025

  43. [43]

    On a Stochas- tic Fundamental Lemma and Its Use for Data- Driven Optimal Control,

    G. Pan, R. Ou, and T. Faulwasser, “On a Stochas- tic Fundamental Lemma and Its Use for Data- Driven Optimal Control,”IEEE Transactions on Automatic Control, 2021

    2021

  44. [44]

    Towards data- driven stochastic predictive control,

    G. Pan, R. Ou, and T. Faulwasser, “Towards data- driven stochastic predictive control,”International Journal of Robust and Nonlinear Control, 2022

    2022

  45. [45]

    Learning-Based Data-Enable Predictive Control for Trajectory Tracking of Unknown Systems,

    T. Huang, L. Dai, S. Li, M. Lin, and Y. Xia, “Learning-Based Data-Enable Predictive Control for Trajectory Tracking of Unknown Systems,”In- ternational Journal of Robust and Nonlinear Con- trol, 2025. 14

    2025