Pith. sign in

REVIEW 3 major objections 4 minor 58 references

Uncertainty-Aware Digital Twins: Robust Model Predictive Control using Time-Series Deep Quantile Learning

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read This paper claims that deep-learning quantiles from a one-shot time-series predictor can replace chance constraints in multi-step robust MPC, yielding less-conservative real-time constraint satisfaction for digital twin control.

desk verdict A plausible and well-demonstrated engineering combination of multi-step TiDE prediction with quantile regression for robust MPC, but the closed-loop calibration gap and the unsupported 'guarantee' language need serious work before the core claim holds. read the letter →

arxiv 2501.10337 v3 pith:T5I72LCO submitted 2025-01-17 eess.SY cs.SY

classification eess.SYcs.SY
keywords DigitalTwinRobustModelPredictiveControlQuantileRegressionTime-SeriesDenseEncoderDeepLearningUncertaintyQuantificationDirectedEnergyDepositionReal-Time
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that the quantile outputs of a time-series deep network can serve as the safety tube in multi-step robust model predictive control, replacing chance constraints with deterministic bounds and making uncertainty-aware control fast enough for digital twin use. The predictor, TiDE, generates the entire future horizon in one forward pass, and quantile regression trained on noisy open-loop data supplies lower and upper trajectories that the optimization uses directly as state bounds. On a linear benchmark the method reports a 5.8 percent constraint-violation rate, close to tube-based MPC's 6.2 percent but with a thinner safety margin, and in the directed energy deposition case it holds melt-pool depth within its dilution bounds while solving each step in an average of 0.179 seconds. The motivation is that sampling-based uncertainty quantification and worst-case tube methods are either computationally heavy or overly conservative for real-time decision-making. The paper claims, to the best of its authors' knowledge, the first unification of deep-learning quantiles with multi-step robust MPC.

What carries the argument

The load-bearing mechanism is TiDE, the Time-Series Dense Encoder, trained with quantile loss. TiDE is a residual dense encoder-decoder that ingests past states, past inputs, and future input candidates and outputs, for every horizon step, the median and user-chosen quantiles (here 0.05 and 0.95) of each predicted state in a single forward pass, producing a tensor of shape $[B,N,D,l]$. Because the quantiles are learned directly from noisy trajectories, they form a data-driven tube: the MPC treats $\bar{\hat{x}}_{j,k+i} \le x_{j,\mathrm{ub}}$ and $\underline{\hat{x}}_{j,k+i} \ge x_{j,\mathrm{lb}}$ as the state constraints, and it tightens the input constraint through the ancillary feedback $u_k = v_k + K e_k$. Around this tube, the paper builds a fast solver by differentiating the MPC loss through TiDE with automatic differentiation and folding all constraints into an augmented Lagrangian penalty solved by L-BFGS, which keeps each online optimization cheap.

What would settle it

Run the trained TiDE-based robust MPC in closed loop on the linear example using inputs drawn from the optimized distribution (rather than the uniform $[-5,5]$ training distribution), and measure the maximum over time of the fraction of the 1,000 replicates that leave the $x_1$ and $x_2$ bounds; if that maximum failure rate is materially above the 5 percent target while open-loop quantile coverage on a held-out test set remains near 95 percent, the closed-loop calibration assumption is falsified.

Watch

Extended reading notes

Core claim

The central claim is that trajectory-level quantiles learned by TiDE from a single training pass are valid enough to act as the probabilistic tube in the robust MPC problem: the controller requires the predicted 95th percentile of each state to stay below the upper bound and the 5th percentile to stay above the lower bound, and this is enough to keep the empirical closed-loop failure rate near five percent on the linear example. Because the bounds are learned directly from data, no recursive propagation of disturbances through the dynamics is needed, which is what removes both the conservatism of worst-case tube methods and the computational cost of sampling-based uncertainty quantification. The paper also claims a practical engineering demonstration in directed energy deposition additive manufacturing, where the melt-pool depth constraint is maintained by the quantile tube at the price of reduced temperature tracking accuracy (the reported $r^2$ drops from 0.8261 for constrained nominal MPC to 0.6920 for robust MPC). The authors state explicitly that stability, recursive feasibility, and performance guarantees are not proven.

Load-bearing premise

The load-bearing premise is that the 5th and 95th percentile bounds learned from open-loop training data remain calibrated for the closed-loop state distribution produced by the optimized inputs and the ancillary feedback, so that the quantile tube still contains the true state with roughly 95 percent probability.

Editorial extensions

If this is right

  • A one-shot TiDE pass eliminates recursive state rollout, so the number of model evaluations per MPC step stops growing with the horizon.
  • The learned quantile tube converts probabilistic constraints into simple deterministic interval checks, removing the need for a known disturbance set or distribution at run time.
  • The method is less conservative than worst-case tube-based MPC while keeping a similar violation rate, so the controller can track references more closely without sacrificing constraint satisfaction.
  • The framework applies to nonlinear systems where tube-based MPC requires linear or Lipschitz assumptions, as illustrated by the DED melt-pool case.
  • The reported 0.179-second average solve time in the DED case suggests the optimization is cheap enough for online digital twin control of processes with slower dynamics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves implicit that the quantiles' closed-loop calibration is the load-bearing assumption; a natural extension is to measure the empirical coverage of the 5-95% tube under the optimized closed-loop input distribution and to add an online re-calibration step if coverage degrades.
  • Since stability and recursive-feasibility proofs are absent, a safety-critical deployment would likely pair the quantile tube with a runtime monitor or a barrier-function certificate, which the paper does not provide.
  • The DED case study's uncertainty comes mostly from numerical discretization of the heat source rather than physical sensor noise; testing on real pyrometer readings would clarify whether the tube stays tight under true process variability.
  • A controlled benchmark on nonlinear systems with known ground-truth disturbances would show whether the conservatism advantage over tube-based MPC persists when the quantile network is imperfectly calibrated.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes a robust model predictive control (MPC) framework for digital twins that combines a multi-step-ahead time-series predictor (TiDE) with quantile regression. TiDE is trained on open-loop noisy trajectories to output, in a single forward pass, both the median and quantile bounds of future states. The robust MPC problem is then formulated with deterministic constraints using these quantile bounds, an ancillary feedback controller, and control-input constraint tightening. The method is demonstrated on a linear two-state system, where it is compared with nominal MPC and tube-based robust MPC, and on a directed energy deposition (DED) additive manufacturing simulation, where it is used for melt-pool temperature tracking and melt-pool depth constraint enforcement. The central claims are that the learned quantiles serve as a safety tube that yields approximately the prescribed 0.95 constraint satisfaction probability and that the resulting controller is less conservative than tube-based MPC.

Significance. If the closed-loop calibration claim were established, the framework would be a practical way to provide uncertainty-aware multi-step MPC for digital twins at low computational cost, because TiDE gives trajectory-level predictions and quantiles in one forward pass. The paper has clear strengths: the one-shot multi-step predictor is a sensible way to reduce online optimization cost; quantile regression avoids strong parametric noise assumptions; the linear example provides a concrete comparison with tube-based MPC, showing a 5.8% versus 6.2% maximum failure rate with a smaller safety margin; and the DED case study reports solve times around 0.18 seconds on average. The paper also includes an explicit list of limitations in Section 6, including the absence of stability, recursive-feasibility, and performance proofs, which is commendable. However, the load-bearing statement that the quantile bounds 'guarantee' the 0.95 probability is not derived, and the empirical evaluation does not address the distinction between the open-loop distribution used for training and the closed-loop distribution induced by the feedback controller.

major comments (3)
  1. [Section 3.2.2 and Eqs. (13f), (19i)] The paragraph after Eq. (13) states that 'the value of K does not affect the conservativeness of the predicted error bounds' because TiDE captures the open-loop distribution. This is not justified. The quantiles are trained on open-loop data generated with uniformly sampled inputs (Sec. 4.2.1) or Fourier-designed profiles (Sec. 5.2), but the implemented controller applies u_k = v_k + K e_k. Under this feedback law, the stochastic recursion governing the closed-loop state distribution is different from the open-loop recursion used to generate the training data; for the linear example in Eq. (17), the error dynamics contain an additional feedback term involving K, and the closed-loop distribution is not the distribution TiDE's quantiles were fit to. Unless K = 0 or calibration under closed-loop trajectories is demonstrated, the learned quantiles do not by themselves provide the 0.95 probability required in Eqs. (19d)-(19g). The same paragraph also states that K tightens the design space of u_k in Eq. (13e), so K does affect overall conservativeness, which is internally inconsistent. The reported 5.8% failure rate is an empirical value for one K, one reference trajectory, and one noise model; it is not a consequence of the construction. I recommend either adding a proof or empirical validation of closed-loop calibration, or removing the 'guarantee' language and presenting the result as an empirical evaluation.
  2. [Section 2.3, Table 1] The coverage rates reported for quantile regression contradict the claim that quantile regression 'can effectively capture the skewed distribution' and provides 'accurate uncertainty bounds.' For the 0.05-0.95 interval, the nominal coverage is 90% but the reported coverage is 59.5%; for the 0.001-0.999 interval, nominal coverage is 99.8% but the reported coverage is 73.2%. These numbers indicate substantial under-coverage, and they weaken the motivating argument for using quantile regression. If the low coverage is due to extrapolation outside the training range (training x in [-5,-4]∪[-1,4], testing x in [-7,7]), that should be stated explicitly and the conclusion should be qualified. As written, the table does not support the text.
  3. [Section 5.4, Fig. 10(d)] The paper states that from timestep 2960 to the end of the layer, the lower-bound constraint on melt pool depth is relaxed because no feasible solution exists. This means the robust MPC problem in Eq. (21) is not solved as stated over part of the trajectory, and the reported constraint-violation comparison is not against the original constraints. Since the central claim is constraint satisfaction under uncertainty, the relaxation should be either included as part of the problem formulation from the outset, or the violation statistics should be reported only over the region where all constraints are actually enforced. Without this, the reported results overstate the performance of the method on the original problem.
minor comments (4)
  1. [Eq. (5a)] The cost function in Eq. (5a) uses r_{k+i} while Eq. (2a) uses r_{k+i+1}; this index inconsistency should be corrected.
  2. [Section 1.5] The novelty claim that this is 'the first to unify these two paradigms' should be reconciled with reference [28], which already applies deep learning tubes to tube-based MPC. If the novelty lies specifically in multi-step TiDE quantile predictions for simultaneous multi-step robust MPC, that distinction should be stated explicitly.
  3. [Eqs. (13) and (19)] The notation for quantile bounds is inconsistent: Eq. (13b) uses barred and unbarred x with a superscript f, while Eqs. (13c)-(13d) use ar\hat{x}^{j,k+i}; it should be clarified that f denotes the future trajectory and j denotes the state component, to avoid confusion about indices.
  4. [Section 5.5] The claim that robust MPC is faster than unconstrained MPC because the safety buffer restricts the feasible space is plausible but not a general property; a brief caveat would be helpful, since the observed time difference could also reflect the particular warm-start or initialization used.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the learned quantile bounds are fitted statistical objects, and the MPC formulation applies them as intended; open-loop/closed-loop distribution mismatch is a correctness risk, not a circular reduction.

full rationale

The paper's derivation chain is not circular. TiDE is trained with quantile loss (Eq. 9) on noisy open-loop data, and the resulting median and quantile outputs are then substituted for chance constraints in the robust MPC formulation (Eqs. 13, 19, 20). This is the standard statistical meaning of a fitted conditional quantile: if the model is well calibrated, roughly 95% of states lie below the 0.95 quantile. The paper does not derive these bounds from the controller or from the constraint-satisfaction results; it explicitly calls them learned/deep-learning quantiles throughout. The linear validation is not a hidden reuse of the output: TiDE is evaluated against a separately simulated tube-based MPC benchmark and against nominal MPC, and the UQ method itself is benchmarked against GP, MC dropout, ensembles, and evidential regression in Table 1. The DED case study relies on the same FEA simulator for both training-data generation and closed-loop evaluation, and some DED training details are delegated to prior work by the same authors ([3], [56]); this is a validation-methodology limitation, not a circular derivation, because the MPC decision-making is not constructed from those validation outcomes. The most serious issue in the paper is the assertion in Sec. 3.2.2 that 'the value of K does not affect the conservativeness of the predicted error bounds' despite the closed-loop implementation u_k = v_k + K e_k in Eq. (19i) changing the stochastic recursion relative to the open-loop training distribution. That is a correctness/calibration risk, and the paper itself concedes in Closure that it does not provide 'comprehensive proofs of stability, recursive feasibility, and performance guarantee, convergence, etc.' Unsupported does not mean circular: no fitted parameter is renamed as a prediction, no self-citation is used to force the central claim, and no result is equivalent by construction to its own input.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on fitted quantile outputs rather than a derived bound. The DED validation is in-distribution with respect to the same FEA code used for training, so some of the apparent constraint satisfaction is a fitting artifact. The controller optimization itself is not equivalent to the training objective, which keeps the circularity burden moderate rather than high.

free parameters (5)
  • Quantile levels q = 0.05, 0.95 = 0.05 and 0.95
    User-selected to define the upper and lower probabilistic bounds used as the safety tube in Eq. (13). The choice is part of controller tuning, not derived from the data.
  • Ancillary controller gain K = [-0.0621, -0.2027]
    Assigned by hand in the linear example (Sec. 4.3) without a derivation. It enters the input tightening in Eq. (19h)-(19i) and the closed-loop input u = v + K e, so it influences the actual disturbance response.
  • Augmented Lagrangian constants = lambda0 = 10, mu0 = 1, alpha = 3
    Selected from trial and error (Sec. 3.3.2) to balance constraint satisfaction and convergence. These hyperparameters affect the solver behavior and therefore the reported results.
  • Prediction window and horizon = w = 10, N = 10 (toy); w = 50, N = 50 (DED)
    Chosen hyperparameters for the TiDE model. The DED training details are deferred to previous work [56].
  • Cost weights Q and R for DED = Not reported
    The DED formulation in Eq. (21) includes Q and R, but their values are not given. They directly affect the tradeoff between tracking error and control effort, and hence the reported r^2 values.
assumptions (4)
  • domain assumption The learned TiDE quantiles can replace the chance constraints Pr(state in X) >= alpha in Eq. (12) with deterministic bounds in Eq. (13).
    Invoked in Sec. 3.2.2 after Eq. (13). It requires statistical calibration of the quantile outputs, which is neither proved nor checked on an independent test set.
  • domain assumption Training data disturbances and input distributions are representative of online MPC operating conditions.
    Sec. 3.1 states data must represent behavior under operational disturbances; Sec. 4.2.1 uses uniform random inputs and Sec. 5.2 uses DOE Fourier profiles, while online MPC optimizes inputs. This is a covariate-shift assumption.
  • domain assumption The FEA simulator can serve as the physical plant for validation, and its deterministic numerical fluctuations can be treated as aleatoric uncertainty.
    Sec. 5.1 and 5.2 replace the physical DED system with an in-house FEA code and treat element-size discretization noise as the uncertainty captured by TiDE. No independent experimental validation is provided.
  • ad hoc to paper The ancillary gain K does not alter the learned state distribution.
    Sec. 3.2.2 states "the value of K does not affect the conservativeness of the predicted error bounds". The closed-loop input contains K e_k, while quantiles are trained from open-loop data, so the asserted independence is an unsupported assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Uncertainty-Aware Digital Twins: Robust Model Predictive Control using Time-Series Deep Quantile Learning." pith.science (2026). https://pith.science/paper/T5I72LCO

@misc{pith2026250110337,
  author       = {Pith},
  title        = {Pith review of: Uncertainty-Aware Digital Twins: Robust Model Predictive Control using Time-Series Deep Quantile Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T5I72LCO}},
  note         = {Machine review of arXiv:2501.10337}
}
read the original abstract

Digital Twins, virtual replicas of physical systems that enable real-time monitoring, model updates, predictions, and decision-making, present novel avenues for proactive control strategies for autonomous systems. However, achieving real-time decision-making in Digital Twins considering uncertainty necessitates an efficient uncertainty quantification (UQ) approach and optimization driven by accurate predictions of system behaviors, which remains a challenge for learning-based methods. This paper presents a simultaneous multi-step robust model predictive control (MPC) framework that incorporates real-time decision-making with uncertainty awareness for Digital Twin systems. Leveraging a multistep ahead predictor named Time-Series Dense Encoder (TiDE) as the surrogate model, this framework differs from conventional MPC models that provide only one-step ahead predictions. In contrast, TiDE can predict future states within the prediction horizon in a one-shot, significantly accelerating MPC. Furthermore, quantile regression is employed with the training of TiDE to perform flexible while computationally efficient UQ on data uncertainty. Consequently, with the deep learning quantiles, the robust MPC problem is formulated into a deterministic optimization problem and provides a safety buffer that accommodates disturbances to enhance constraint satisfaction rate. As a result, the proposed method outperforms existing robust MPC methods by providing less-conservative UQ and has demonstrated efficacy in an engineering case study involving Directed Energy Deposition (DED) additive manufacturing. This proactive while uncertainty-aware control capability positions the proposed method as a potent tool for future Digital Twin applications and real-time process control in engineering systems.

Figures

Figures reproduced from arXiv: 2501.10337 by the authors.

Figure 1
Figure 1. Proposed multi-step robust MPC framework. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Illustration of MPC and robust MPC. (a) Illustration of MPC at time = [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Network structure of TiDE, modified from [ [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Data structure of the input and output of TiDE. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Illustration of gradient-based optimization using auto-differentiation. [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Evaluation of TiDE. (a) The training and validation loss of TiDE. (b) Comparison of the state prediction and the [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]
Figure 7
Figure 7. Figure 7: Comparison of simultaneous multi-step MPC with and without robust consideration. (a) Trajectories of the [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]
Figure 8
Figure 8. Figure 8: Distribution of trajectories under 1000 replicates. (a) Trajectories of multi-step (nominal) MPC. (b) Trajectories [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]
Figure 9
Figure 9. Figure 9: Single-track square [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]
Figure 10
Figure 10. Figure 10: DED result comparison. (a) Trajectories of melt pool temperature, melt pool depth, and the corresponding [PITH_FULL_IMAGE:figures/full_fig_p021_10.png]
Figure 11
Figure 11. Figure 11: Histogram of MPC solving time. This case study verifies the suitability of our method for complex engineering applications. It highlights TiDE’s capability as a surrogate model for multi-step predictions with UQ and demonstrates the effectiveness of multi-step robust …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 55 canonical work pages

  1. [1]

    National Academies of Sciences, Engineering, and Medicine and others, 2023.Foundational research gaps and future directions for digital twins

  2. [2]

    Digital twins for the designs of systems: a perspective

    van Beek, A., Nevile Karkaria, V., and Chen, W., 2023. “Digital twins for the designs of systems: a perspective”. Structural and Multidisciplinary Optimization,66(3), p. 49

  3. [3]

    Towards a digital twin framework in additive manufacturing: Machine learning and bayesian optimization for time series process optimization

    Karkaria,V.,Goeckner,A.,Zha,R.,Chen,J.,Zhang,J.,Zhu,Q.,Cao,J.,Gao,R.X.,andChen,W.,2024. “Towards a digital twin framework in additive manufacturing: Machine learning and bayesian optimization for time series process optimization”.Journal of Manufacturing Systems

  4. [4]

    An optimization-centric review for integrating artificial intelligence and digital twin technologies in manufacturing

    Karkaria, V., Tsai, Y.-K., Chen, Y.-P., and Chen, W., 2025. “An optimization-centric review for integrating artificial intelligence and digital twin technologies in manufacturing”.Engineering Optimization, pp. 1–47

  5. [5]

    Security and Privacy of Digital Twins for Advanced Manufacturing: A Survey

    Zemskov,A.D.,Fu,Y.,Li,R.,Wang,X.,Karkaria,V.,Tsai,Y.-K.,Chen,W.,Zhang,J.,Gao,R.,Cao,J.,etal.,2024. “Security and privacy of digital twins for advanced manufacturing: A survey”.arXiv preprint arXiv:2412.13939

  6. [6]

    Overview of predictive maintenance based on digital twin technology

    Zhong, D., Xia, Z., Zhu, Y., and Duan, J., 2023. “Overview of predictive maintenance based on digital twin technology”. Heliyon,9(4)

  7. [7]

    Digital twin: Mitigating unpredictable, undesirable emergent behavior in complex systems

    Grieves, M., and Vickers, J., 2017. “Digital twin: Mitigating unpredictable, undesirable emergent behavior in complex systems”.Transdisciplinary perspectives on complex systems: New findings and approaches, pp. 85–113

  8. [8]

    B., Mayne, D

    Rawlings, J. B., Mayne, D. Q., Diehl, M., et al., 2017.Model predictive control: theory, computation, and design, Vol. 2. Nob Hill Publishing Madison, WI

Show all 58 references
  1. [9]

    Stochastic model predictive control: An overview and perspectives for future research

    Mesbah, A., 2016. “Stochastic model predictive control: An overview and perspectives for future research”.IEEE Control Systems Magazine,36(6), pp. 30–44

  2. [10]

    Robust model predictive control of constrained linear systems with bounded disturbances

    Mayne, D. Q., Seron, M. M., and Raković, S. V., 2005. “Robust model predictive control of constrained linear systems with bounded disturbances”.Automatica,41(2), pp. 219–224. 23

  3. [11]

    Robust learning-based mpc for nonlinear constrained systems

    Manzano, J. M., Limon, D., de la Peña, D. M., and Calliess, J.-P., 2020. “Robust learning-based mpc for nonlinear constrained systems”.Automatica,117, p. 108948

  4. [12]

    Learning-based model predictive control: Towardsafelearningincontrol

    Hewing, L., Wabersich, K. P., Menner, M., and Zeilinger, M. N., 2020. “Learning-based model predictive control: Towardsafelearningincontrol”. AnnualReviewofControl, Robotics,andAutonomousSystems, 3(1), pp.269–296

  5. [13]

    Minimaxapproachestorobustmodelpredictivecontrol ,Vol.812

    Löfberg,J.,2003. Minimaxapproachestorobustmodelpredictivecontrol ,Vol.812. LinköpingUniversityElectronic Press

  6. [14]

    Robust model predictive control: A survey

    Bemporad, A., and Morari, M., 2007. “Robust model predictive control: A survey”. InRobustness in identification and control. Springer, pp. 207–226

  7. [15]

    Online robust tube-based mpc for time-varying systems: A practical approach

    Gonzalez, R., Fiacchini, M., Alamo, T., Guzmán, J. L., and Rodríguez, F., 2011. “Online robust tube-based mpc for time-varying systems: A practical approach”.International Journal of Control,84(6), pp. 1157–1170

  8. [16]

    Robust control co-design using tube-based model predictive control

    Tsai, Y.-K., and Malak, R. J., 2023. “Robust control co-design using tube-based model predictive control”. In 2023 American Control Conference (ACC), IEEE, pp. 769–775

  9. [17]

    Robustmodelpredictivecontrolusingtubes

    Langson,W.,Chryssochoos,I.,Raković,S.,andMayne,D.Q.,2004. “Robustmodelpredictivecontrolusingtubes”. Automatica,40(1), pp. 125–133

  10. [18]

    Constraint-tightening and stability in stochastic model predictive control

    Lorenzen, M., Dabbene, F., Tempo, R., and Allgöwer, F., 2016. “Constraint-tightening and stability in stochastic model predictive control”.IEEE Transactions on Automatic Control,62(7), pp. 3165–3177

  11. [19]

    Control co-design with performance-robustness trade-off using tube-based stochastic model predictive control

    Tsai, Y.-K., and Malak, R., 2025. “Control co-design with performance-robustness trade-off using tube-based stochastic model predictive control”.ASME Letters in Dynamic Systems and Control

  12. [20]

    D., and Todd, M

    Thelen, A., Zhang, X., Fink, O., Lu, Y., Ghosh, S., Youn, B. D., and Todd, M. D., 2022.A comprehensive review of digital twin — part 1 : modeling and twinning enabling technologies, Vol. 65. Springer Berlin Heidelberg

  13. [21]

    Differentiable predictive control: Deep learning alternative to explicit model predictive control for unknown nonlinear systems

    Drgoňa, J., Kiš, K., Tuor, A., Vrabie, D., and Klaučo, M., 2022. “Differentiable predictive control: Deep learning alternative to explicit model predictive control for unknown nonlinear systems”.Journal of Process Control,116, pp. 80–92

  14. [22]

    Modelpredictivecontrolwhenutilizing lstm as dynamic models

    Jung,M.,daCostaMendes,P.R.,Önnheim,M.,andGustavsson,E.,2023. “Modelpredictivecontrolwhenutilizing lstm as dynamic models”.Engineering Applications of Artificial Intelligence,123, p. 106226

  15. [23]

    Real-timeadaptivemachine-learning-basedpredictivecontrolof nonlinear processes

    Wu,Z.,Rincon,D.,andChristofides,P.D.,2019. “Real-timeadaptivemachine-learning-basedpredictivecontrolof nonlinear processes”.Industrial & Engineering Chemistry Research,59(6), pp. 2275–2290

  16. [24]

    Lstm-mpc: A deep learning based predictive control method for multimode process control

    Huang, K., Wei, K., Li, F., Yang, C., and Gui, W., 2022. “Lstm-mpc: A deep learning based predictive control method for multimode process control”.IEEE Transactions on Industrial Electronics,70(11), pp. 11544–11554

  17. [25]

    Neuralnetwork-baseduncertaintyquantification: A survey of methodologies and applications

    Kabir,H.D.,Khosravi,A.,Hosen,M.A.,andNahavandi,S.,2018.“Neuralnetwork-baseduncertaintyquantification: A survey of methodologies and applications”.IEEE access,6, pp. 36218–36234

  18. [26]

    Advancing vehicle trajectory prediction: A probabilistic approach using combined sequential models

    Ren, L., and Xi, Z., 2024. “Advancing vehicle trajectory prediction: A probabilistic approach using combined sequential models”. In International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Vol. 88377, American Society of M...

  19. [27]

    Cautious model predictive control using gaussian process regression

    Hewing, L., Kabzan, J., and Zeilinger, M. N., 2019. “Cautious model predictive control using gaussian process regression”. IEEE Transactions on Control Systems Technology,28(6), pp. 2736–2743

  20. [28]

    Deep learning tubes for tube mpc

    Fan, D. D., Agha-mohammadi, A.-a., and Theodorou, E. A., 2020. “Deep learning tubes for tube mpc”.arXiv preprint arXiv:2002.01587

  21. [29]

    Simultaneous multistep transformer architecture for model predictive control

    Park, J., Babaei, M. R., Munoz, S. A., Venkat, A. N., and Hedengren, J. D., 2023. “Simultaneous multistep transformer architecture for model predictive control”.Computers & Chemical Engineering,178, p. 108396. 24

  22. [30]

    State space models vs. multi-step predictors inpredictivecontrol: Arestatespacemodelscomplicatingsafedata-drivendesigns?

    Köhler, J., Wabersich, K. P., Berberich, J., and Zeilinger, M. N., 2022. “State space models vs. multi-step predictors inpredictivecontrol: Arestatespacemodelscomplicatingsafedata-drivendesigns?”. In2022IEEE61stConference on Decision and Control (CDC), IEEE, pp. 491–498

  23. [31]

    Learning-based predictive control for linear systems: A unitary approach

    Terzi, E., Fagiano, L., Farina, M., and Scattolini, R., 2019. “Learning-based predictive control for linear systems: A unitary approach”.Automatica,108, p. 108473

  24. [32]

    Robust multi-rate predictive control using multi-step prediction models learned from data

    Terzi, E., Farina, M., Fagiano, L., and Scattolini, R., 2022. “Robust multi-rate predictive control using multi-step prediction models learned from data”.Automatica,136, p. 109852

  25. [33]

    Quantile deep learning models for multi-step ahead time series prediction

    Cheung, J., Rangarajan, S., Maddocks, A., Chen, X., and Chandra, R., 2024. “Quantile deep learning models for multi-step ahead time series prediction”.arXiv preprint arXiv:2411.15674

  26. [34]

    A novel time series probabilistic prediction approach based on the monotone quantile regression neural network

    Hu, J., Tang, J., and Liu, Z., 2024. “A novel time series probabilistic prediction approach based on the monotone quantile regression neural network”.Information Sciences,654, p. 119844

  27. [35]

    D., Todd, M

    Thelen, A., Zhang, X., Fink, O., Lu, Y., Ghosh, S., Youn, B. D., Todd, M. D., Mahadevan, S., Hu, C., and Hu, Z.,

  28. [36]

    Design of approximate explicit model predictive controller using parametric optimization

    Tsai, Y.-K., and Malak, R. J., 2022. “Design of approximate explicit model predictive controller using parametric optimization”. Journal of Mechanical Design,144(12)

  29. [37]

    Stochasticmodelpredictivecontrolusingacombination of randomized and robust optimization

    Zhang,X.,Margellos,K.,Goulart,P.,andLygeros,J.,2013.“Stochasticmodelpredictivecontrolusingacombination of randomized and robust optimization”. In 52nd IEEE conference on decision and control, IEEE, pp. 7740–7745

  30. [38]

    Stochastic model predictive control using simplified affine disturbance feedback for chance-constrained systems

    Zhang, J., and Ohtsuka, T., 2021. “Stochastic model predictive control using simplified affine disturbance feedback for chance-constrained systems”.IEEE Control Systems Letters,5(5), pp. 1633–1638

  31. [39]

    A tube-based robust nonlinear predictive control approach to semiautonomous ground vehicles

    Gao, Y., Gray, A., Tseng, H. E., and Borrelli, F., 2014. “A tube-based robust nonlinear predictive control approach to semiautonomous ground vehicles”.Vehicle System Dynamics,52(6), pp. 802–823

  32. [40]

    Long-term forecasting with tide: Time-series dense encoder

    Das, A., Kong, W., Leach, A., Mathur, S., Sen, R., and Yu, R., 2023. “Long-term forecasting with tide: Time-series dense encoder”.arXiv preprint arXiv:2304.08424

  33. [41]

    N-beats: Neural basis expansion analysis for interpretable time series forecasting

    Oreshkin, B. N., Carpov, D., Chapados, N., and Bengio, Y., 2019. “N-beats: Neural basis expansion analysis for interpretable time series forecasting”.arXiv preprint arXiv:1905.10437

  34. [42]

    Quantileregression

    Koenker,R.,andHallock,K.F.,2001. “Quantileregression”. Journalofeconomicperspectives, 15(4),pp.143–156

  35. [43]

    Quantile regression: applications and current research areas

    Yu, K., Lu, Z., and Stander, J., 2003. “Quantile regression: applications and current research areas”.Journal of the Royal Statistical Society Series D: The Statistician,52(3), pp. 331–350

  36. [44]

    Deep evidential regression

    Amini, A., Schwarting, W., Soleimany, A., and Rus, D., 2020. “Deep evidential regression”.Advances in Neural Information Processing Systems,33, pp. 14927–14937

  37. [45]

    Stochastic model predictive control—how does it work?

    Heirung, T. A. N., Paulson, J. A., O’Leary, J., and Mesbah, A., 2018. “Stochastic model predictive control—how does it work?”.Computers & Chemical Engineering,114, pp. 158–170

  38. [46]

    Probabilistic tubes in linear stochastic model predictive control

    Cannon, M., Kouvaritakis, B., and Ng, D., 2009. “Probabilistic tubes in linear stochastic model predictive control”. Systems & Control Letters,58(10-11), pp. 747–753

  39. [47]

    Alatentvariableapproachfornon-hierarchicalmulti-fidelity adaptive sampling

    Chen,Y.-P.,Wang,L.,Comlek,Y.,andChen,W.,2024.“Alatentvariableapproachfornon-hierarchicalmulti-fidelity adaptive sampling”.Computer Methods in Applied Mechanics and Engineering,421, p. 116773

  40. [48]

    Nonlinearmodelpredictivecontrolalgorithmwithiterativenonlinear prediction and linearization for long short-term memory network models

    Schwedersky,B.B.,andFlesch,R.C.,2022. “Nonlinearmodelpredictivecontrolalgorithmwithiterativenonlinear prediction and linearization for long short-term memory network models”.Engineering Applications of Artificial Intelligence,115, p. 105247. 25

  41. [49]

    Acomparativestudyofstochasticmodelpredictive controllers

    González,E.,Sanchis,J.,García-Nieto,S.,andSalcedo,J.,2020.“Acomparativestudyofstochasticmodelpredictive controllers”. Electronics,9(12), p. 2078

  42. [50]

    Dynamic tube mpc for nonlinear systems

    Lopez, B. T., Slotine, J.-J. E., and How, J. P., 2019. “Dynamic tube mpc for nonlinear systems”. In 2019 American Control Conference (ACC), IEEE, pp. 1655–1662

  43. [51]

    Pytorch: Animperativestyle,high-performancedeeplearninglibrary

    Paszke,A.,Gross,S.,Massa,F.,Lerer,A.,Bradbury,J.,Chanan,G.,Killeen,T.,Lin,Z.,Gimelshein,N.,Antiga,L., etal.,2019. “Pytorch: Animperativestyle,high-performancedeeplearninglibrary”. Advancesinneuralinformation processing systems,32

  44. [52]

    J., 1999.Numerical optimization

    Nocedal, J., and Wright, S. J., 1999.Numerical optimization. Springer

  45. [53]

    Pytorch-minimize: a library for numerical optimization with autograd

    Feinman, R., 2021. Pytorch-minimize: a library for numerical optimization with autograd

  46. [54]

    Machine learning-assistedin-situadaptivestrategiesforthecontrolofdefectsandanomaliesinmetaladditivemanufacturing

    Gunasegaram,D.,Barnard,A.,Matthews,M.,Jared,B.,Andreaco,A.,Bartsch,K.,andMurphy,A.,2024. “Machine learning-assistedin-situadaptivestrategiesforthecontrolofdefectsandanomaliesinmetaladditivemanufacturing”. Additive Manufacturing, p. 104013

  47. [55]

    Efficient gpu-accelerated thermomechanical solver for residual stress prediction in additive manufacturing

    Liao, S., Golgoon, A., Mozaffar, M., and Cao, J., 2023. “Efficient gpu-accelerated thermomechanical solver for residual stress prediction in additive manufacturing”.Computational Mechanics,71(5), pp. 879–893

  48. [56]

    Real-time decision-making for Digital Twin in additive manufacturing with Model Predictive Control using time-series deep neural networks

    Chen, Y.-P., Karkaria, V., Tsai, Y.-K., Rolark, F., Quispe, D., Gao, R. X., Cao, J., and Chen, W., 2025. “Real-time decision-making for Digital Twin in additive manufacturing with Model Predictive Control using time-series deep neural networks”.Journal of Manufacturing Systems...

  49. [57]

    State of the art in directed energy deposition: From additive manufacturing to materials design

    Dass, A., and Moridi, A., 2019. “State of the art in directed energy deposition: From additive manufacturing to materials design”.Coatings,9(7), p. 418. 26

  50. [2022]

    A comprehensive review of digital twin—part 1: modeling and twinning enabling technologies

    “A comprehensive review of digital twin—part 1: modeling and twinning enabling technologies”.Structural and Multidisciplinary Optimization,65(12), p. 354

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.