arxiv: 2604.12912 · v1 · submitted 2026-04-14 · 📡 eess.SY · cs.SY

Recognition: unknown

Nonlinear Stochastic Model Predictive Control with Generative Uncertainty in Homogeneous Charge Compression Ignition

Xu Chen , Kevin Kluge , Maximilian Basler , Lorenz D\"orschel , Heike Vallery

Authors on Pith no claims yet

Pith reviewed 2026-05-10 14:39 UTC · model grok-4.3

classification 📡 eess.SY cs.SY

keywords stochastic model predictive controlhomogeneous charge compression ignitionmaximum mean discrepancypolynomial chaos expansionuncertainty modelingcombustion controlengine control

0 comments

The pith

Nonlinear stochastic MPC with learned uncertainty distributions reduces HCCI combustion variation by over 28 percent.

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

The paper develops a control strategy for homogeneous charge compression ignition engines that must handle uncertain combustion timing and load from complex dynamics and external disturbances. It learns an uncertainty model directly from empirical residual data and integrates it into a nonlinear model predictive controller. Polynomial chaos expansion propagates the nonlinear uncertainties forward in time, while a cost function based on maximum mean discrepancy penalizes differences between the full predicted distribution of combustion outcomes and the desired distribution. Simulations show this yields more than 28 percent less variation in combustion phasing and over 26 percent better load tracking accuracy than traditional nonlinear or Gaussian-based controllers. The work demonstrates the value of operating explicitly at the level of probability distributions rather than means or variances alone.

Core claim

The authors present a nonlinear stochastic model predictive control framework that explicitly incorporates distributional information of uncertainties by learning an uncertainty model from empirical residual data, handling nonlinear additive uncertainty propagation via polynomial chaos expansion, and employing a maximum mean discrepancy-based cost function to penalize discrepancies between predicted and desired distributions of combustion indicators, resulting in over 28 percent reduction on combustion phasing variation and more than 26 percent improvement in load tracking accuracy.

What carries the argument

The maximum mean discrepancy cost function that directly penalizes the discrepancy between predicted and desired distributions of combustion indicators, combined with polynomial chaos expansion to propagate nonlinear additive uncertainty within the MPC prediction horizon.

Load-bearing premise

The uncertainty model learned from empirical residual data accurately captures the realistic probabilistic characteristics of combustion dynamics and external disturbances, and the simulation environment sufficiently represents real engine behavior for the claimed improvements to transfer.

What would settle it

Running the controller on a physical HCCI engine testbed under measured real-world disturbances and checking whether combustion phasing variation drops by at least 28 percent and load tracking error improves by at least 26 percent relative to the baseline controllers.

Figures

Figures reproduced from arXiv: 2604.12912 by Heike Vallery, Kevin Kluge, Lorenz D\"orschel, Maximilian Basler, Xu Chen.

**Figure 2.** Figure 2: Comparison of the marginal uncertainty distributio [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗

**Figure 5.** Figure 5: Comparison of mean CA50 (top) and mean IMEP (bottom) tr [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗

**Figure 4.** Figure 4: Comparison of CA50 trajectories over 120 engine cycles under different control strategies: nonlinear MPC, Gaussian-based SMPC, PC-based SMPC, and GEM-SMPC. Each plot shows results from 50 simulation runs. The bold solid line represents the mean CA50, while the dashed line indicates the CA50 set point. The shaded area reflects the variability across runs, highlighting both control performance and combustion… view at source ↗

read the original abstract

This work addresses the challenge of ignition timing and load control in homogeneous charge compression ignition engines operating subject to uncertainty from complex combustion dynamics and external disturbances. To handle this issue, we propose a nonlinear stochastic model predictive control approach explicitly incorporating distributional information of uncertainties. Specifically, we integrate an uncertainty model learned from empirical residual data to capture realistic probabilistic characteristics and handle the nonlinear additive uncertainty propagation within the prediction horizon based on polynomial chaos expansion. Additionally, we introduce a novel cost function based on maximum mean discrepancy, enabling direct penalization of the discrepancy between predicted and desired distributions of combustion indicators. The simulation results demonstrate that our proposed method achieves over a 28 \% reduction on combustion phasing variation and more than a 26 \% improvement in load tracking accuracy compared to traditional nonlinear and Gaussian-based predictive control strategies. These findings indicate the effectiveness of explicitly modeling uncertainty distributions and highlight the advantages of distribution-level performance index in robust combustion control.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper applies learned residual uncertainty, PCE propagation, and an MMD cost inside stochastic MPC for HCCI control and reports simulation gains, but all claims rest on an unvalidated model with no hardware results.

read the letter

Here's the core of it: this is a simulation-based application of nonlinear stochastic MPC to HCCI ignition and load control. They learn an uncertainty model from empirical residuals, propagate the nonlinear uncertainty with polynomial chaos expansion, and add a maximum mean discrepancy term to penalize distribution mismatch in the cost. That combination is new for this specific engine problem. The paper does a solid job laying out how to handle non-Gaussian uncertainties in the prediction without assuming normality, and the simulation comparisons to nonlinear MPC and Gaussian stochastic MPC show clear improvements in the metrics they track. The soft spot is the complete reliance on simulation. The performance numbers come from a closed loop that uses the same learned model for both the plant and the controller. Without any real-engine data or even a separate validation set for the uncertainty model against physical disturbances, it's hard to know if the 28 percent and 26 percent figures would survive outside the model. The abstract doesn't detail statistical validation of the uncertainty model either. The math itself looks standard and consistent for this style of work. This paper is aimed at control engineers working on combustion engines or similar uncertain systems where distribution-aware control could matter. A reader looking for an example of MMD in MPC or PCE in engine applications would find it useful. It deserves peer review because the method is coherent and the problem is practical, even with the simulation limitation. I would recommend sending it to referees.

Referee Report

3 major / 3 minor

Summary. The paper proposes a nonlinear stochastic MPC framework for HCCI engine ignition timing and load control. It learns a generative uncertainty model from empirical residual data to capture non-Gaussian combustion disturbances, propagates uncertainties over the horizon via polynomial chaos expansion, and introduces an MMD-based cost that directly penalizes distributional mismatch between predicted and target combustion indicators. Closed-loop simulations report >28% reduction in combustion phasing variation and >26% improvement in load tracking versus standard nonlinear MPC and Gaussian-assumption baselines.

Significance. If the learned uncertainty model is shown to match real residual statistics and the simulation plant is representative, the work would demonstrate a practical route to distribution-aware stochastic control for highly nonlinear, uncertain combustion systems. The combination of data-driven generative uncertainty, PCE propagation, and MMD cost is a technical step beyond moment-based or Gaussian stochastic MPC and could influence robust control design in other domains with complex additive disturbances.

major comments (3)

[Abstract / Simulation Results] Abstract and Simulation Results section: the headline claims of 28% reduction in phasing variation and 26% load-tracking improvement are obtained exclusively inside a simulation loop that uses the fitted uncertainty model as both disturbance source and plant; no quantitative validation (e.g., MMD or Wasserstein distance between learned and held-out residual distributions, cross-validation error, or sensitivity plots) is reported, making the performance deltas directly dependent on an unverified modeling assumption.
[Methodology] Methodology section on uncertainty propagation: while PCE is invoked to handle nonlinear additive uncertainty, the manuscript does not specify the polynomial order, the number of samples used to estimate the MMD cost inside the optimizer, or how the generative model parameters are kept fixed versus re-estimated online; these choices are load-bearing for both computational tractability and the claimed advantage over Gaussian MPC.
[Simulation Results / Conclusions] No experimental or HIL validation is presented. All quantitative results rest on a closed simulation whose fidelity to physical HCCI combustion (including sensor noise, actuator dynamics, and cycle-to-cycle variability not captured by the residual model) is asserted but not quantified, directly undermining transferability of the reported gains.

minor comments (3)

[Methodology] Notation for the MMD kernel and the PCE basis functions should be introduced with explicit definitions and reference to standard implementations to avoid ambiguity.
[Abstract / Introduction] The abstract states 'generative uncertainty' without clarifying whether the model is a VAE, GAN, or normalizing flow; a short sentence in the introduction or methods would improve readability.
[Figures] Figure captions for the closed-loop trajectories should include the number of Monte-Carlo realizations or confidence bands used to generate the plotted statistics.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which help improve the clarity and rigor of the manuscript. We address each major point below, indicating planned revisions where appropriate.

read point-by-point responses

Referee: [Abstract / Simulation Results] Abstract and Simulation Results section: the headline claims of 28% reduction in phasing variation and 26% load-tracking improvement are obtained exclusively inside a simulation loop that uses the fitted uncertainty model as both disturbance source and plant; no quantitative validation (e.g., MMD or Wasserstein distance between learned and held-out residual distributions, cross-validation error, or sensitivity plots) is reported, making the performance deltas directly dependent on an unverified modeling assumption.

Authors: We agree that explicit quantitative validation of the learned generative uncertainty model strengthens the interpretation of the simulation results. In the revised manuscript we will add MMD and Wasserstein distance metrics comparing the model output to held-out residual data, cross-validation error statistics, and sensitivity plots with respect to training-set size. These additions will demonstrate that the learned distribution faithfully reproduces the empirical non-Gaussian statistics before the closed-loop results are presented. revision: yes
Referee: [Methodology] Methodology section on uncertainty propagation: while PCE is invoked to handle nonlinear additive uncertainty, the manuscript does not specify the polynomial order, the number of samples used to estimate the MMD cost inside the optimizer, or how the generative model parameters are kept fixed versus re-estimated online; these choices are load-bearing for both computational tractability and the claimed advantage over Gaussian MPC.

Authors: We will insert the missing implementation details into the revised Methodology section. The polynomial order for the PCE expansion, the number of Monte-Carlo samples used to estimate the MMD cost at each optimizer iteration, and the fact that the generative model is trained offline and held fixed during online operation will be stated explicitly, together with a short discussion of the resulting computational cost relative to the Gaussian baseline. revision: yes
Referee: [Simulation Results / Conclusions] No experimental or HIL validation is presented. All quantitative results rest on a closed simulation whose fidelity to physical HCCI combustion (including sensor noise, actuator dynamics, and cycle-to-cycle variability not captured by the residual model) is asserted but not quantified, directly undermining transferability of the reported gains.

Authors: The work is a simulation study that employs a high-fidelity HCCI engine model previously validated against experimental data in the literature. In the revised manuscript we will expand the discussion of simulation fidelity, cite the relevant validation studies, and explicitly acknowledge that sensor noise, actuator dynamics, and any residual variability not captured by the learned model remain untested in hardware. We agree that hardware-in-the-loop or experimental validation would further strengthen transferability claims, but the current simulation results still isolate the benefit of the distribution-aware cost under realistic disturbance statistics. revision: partial

Circularity Check

0 steps flagged

No circularity in derivation chain; results are simulation outputs from independent components

full rationale

The paper learns an uncertainty model from external empirical residual data, propagates it via standard polynomial chaos expansion, and optimizes a maximum mean discrepancy cost within a nonlinear stochastic MPC formulation. Simulation comparisons to baseline controllers produce the reported performance deltas as independent closed-loop outcomes rather than quantities defined by the fitted model itself. No self-citations, uniqueness theorems, or ansatzes are invoked to force the central claims, and the derivation remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on a data-driven uncertainty model whose parameters are fitted to residuals and on standard assumptions of polynomial chaos expansion for uncertainty propagation; no new entities are postulated.

free parameters (1)

Parameters of the learned uncertainty model
Fitted from empirical residual data to capture probabilistic characteristics; specific values and fitting procedure not detailed in abstract.

axioms (1)

domain assumption Polynomial chaos expansion accurately approximates nonlinear additive uncertainty propagation over the prediction horizon
Invoked to handle uncertainty spreading in the nonlinear system dynamics.

pith-pipeline@v0.9.0 · 5468 in / 1489 out tokens · 41359 ms · 2026-05-10T14:39:10.185355+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

49 extracted references · 4 canonical work pages

[1]

Zhao, HCCI and CAI engines for the automotive industry

H. Zhao, HCCI and CAI engines for the automotive industry . Elsevier, 2007

2007
[2]

Progress and recent trends i n homoge- neous charge compression ignition (HCCI) engines,

M. Y ao, Z. Zheng, and H. Liu, “Progress and recent trends i n homoge- neous charge compression ignition (HCCI) engines,” Progress in energy and combustion science , vol. 35, no. 5, pp. 398–437, 2009

2009
[3]

L ¨ u, W

X.-C. L ¨ u, W. Chen, and Z. Huang, “A fundamental study on t he control of the HCCI combustion and emissions by fuel design concept c ombined with controllable EGR. part 2. effect of operating conditio ns and EGR on HCCI combustion,” Fuel, vol. 84, no. 9, pp. 1084–1092, 2005

2005
[4]

Fundamental phenomena affec ting low temperature combustion and HCCI engines, high load limits a nd strate- gies for extending these limits,

S. Saxena and I. D. Bedoya, “Fundamental phenomena affec ting low temperature combustion and HCCI engines, high load limits a nd strate- gies for extending these limits,” Progress in Energy and Combustion Science, vol. 39, no. 5, pp. 457–488, 2013

2013
[5]

An overview of industrial mo del predictive control technology,

S. J. Qin and T. A. Badgwell, “An overview of industrial mo del predictive control technology,” in AIche symposium series , vol. 93, no
[6]

New Y ork, NY: American Institute of Chemical Engineers , 1971- c2002., 1997, pp. 232–256

1971
[7]

Model predictive control of HCCI using variable valve actu ation and fuel injection,

N. Ravi, H.-H. Liao, A. F. Jungkunz, A. Widd, and J. C. Gerd es, “Model predictive control of HCCI using variable valve actu ation and fuel injection,” Control Engineering Practice , vol. 20, no. 4, pp. 421– 430, 2012

2012
[8]

Multi-cyl inder HCCI control with cam phaser variable valve actuation using mode l predictive control,

S. M. Erlien, A. F. Jungkunz, and J. C. Gerdes, “Multi-cyl inder HCCI control with cam phaser variable valve actuation using mode l predictive control,” in Dynamic Systems and Control Conference , vol. 45301. American Society of Mechanical Engineers, 2012, pp. 823–83 1

2012
[9]

Multicylinder HCCI control with coupled valve actu ation using model predictive control,

——, “Multicylinder HCCI control with coupled valve actu ation using model predictive control,” Journal of Dynamic Systems, Measurement, and Control, vol. 135, no. 5, p. 051018, 2013

2013
[10]

Control of exhaust recompression HCCI using hybrid model p redictive control,

A. Widd, H.-H. Liao, J. C. Gerdes, P . Tunest˚ al, and R. Joh ansson, “Control of exhaust recompression HCCI using hybrid model p redictive control,” in Proceedings of the 2011 American control conference . IEEE, 2011, pp. 420–425

2011
[11]

Phys ics-based model predictive control of HCCI combustion phasing using f ast thermal management and vva,

A. Widd, K. Ekholm, P . Tunestal, and R. Johansson, “Phys ics-based model predictive control of HCCI combustion phasing using f ast thermal management and vva,” IEEE Transactions on Control Systems Technol- ogy, vol. 20, no. 3, pp. 688–699, 2011

2011
[12]

Model predictive control for c ombustion timing and load control in HCCI engines,

K. Ebrahimi and C. Koch, “Model predictive control for c ombustion timing and load control in HCCI engines,” SAE Technical Pape r, Tech. Rep., 2015

2015
[13]

Real-time control of HCCI en gine using model predictive control,

K. Ebrahimi and C. B. Koch, “Real-time control of HCCI en gine using model predictive control,” in 2018 Annual American Control Conference (ACC). IEEE, 2018, pp. 1622–1628. 13

2018
[14]

Introducing a deep neural network-based mod el predictive control framework for rapid controller impleme ntation,

D. C. Gordon, A. Winkler, J. Bedei, P . Schaber, S. Pischi nger, J. Andert, and C. R. Koch, “Introducing a deep neural network-based mod el predictive control framework for rapid controller impleme ntation,” in 2024 American Control Conference (ACC) . IEEE, 2024, pp. 5232– 5237

2024
[15]

Nonlinea r model predictive control of a gasoline HCCI engine using extreme l earning machines,

V . M. Janakiraman, X. Nguyen, and D. Assanis, “Nonlinea r model predictive control of a gasoline HCCI engine using extreme l earning machines,” arXiv preprint arXiv:1501.03969 , 2015

work page arXiv 2015
[16]

Nonlinear model predictive control of a discr ete-cycle gasoline-controlled auto ignition engine model: Simulati ve analysis,

E. Nuss, M. Wick, J. Andert, J. De Schutter, M. Diehl, D. A bel, and T. Albin, “Nonlinear model predictive control of a discr ete-cycle gasoline-controlled auto ignition engine model: Simulati ve analysis,” International Journal of Engine Research , vol. 20, no. 10, pp. 1025– 1036, 2019

2019
[17]

Nonlinear MPC and MHE of a gasoline controlled auto-igniti on engine based on reduced-order differential equation models,

J. De Schutter, E. Nuss, K. Baumg¨ artner, D. Abel, and M. Diehl, “Nonlinear MPC and MHE of a gasoline controlled auto-igniti on engine based on reduced-order differential equation models,” in 2022 IEEE 61st Conference on Decision and Control (CDC) . IEEE, 2022, pp. 2411– 2416

2022
[18]

Robust model predictive con trol: A survey,

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

2007
[19]

Stochastic model predictive control—how does it work?

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

2018
[20]

Stochastic model predictive control: An ov erview and perspectives for future research,

A. Mesbah, “Stochastic model predictive control: An ov erview and perspectives for future research,” IEEE Control Systems Magazine , vol. 36, no. 6, pp. 30–44, 2016

2016
[21]

S tochastic model predictive control for building climate control,

F. Oldewurtel, C. N. Jones, A. Parisio, and M. Morari, “S tochastic model predictive control for building climate control,” IEEE Transactions on Control Systems Technology, vol. 22, no. 3, pp. 1198–1205, 2013

2013
[22]

Scenario-based model predictive operation contr ol of islanded microgrids,

C. A. Hans, P . Sopasakis, A. Bemporad, J. Raisch, and C. R eincke- Collon, “Scenario-based model predictive operation contr ol of islanded microgrids,” in 2015 54th IEEE conference on decision and control (CDC). IEEE, 2015, pp. 3272–3277

2015
[23]

A stoc hastic MPC based approach to integrated energy management in micro grids,

Y . Zhang, F. Meng, R. Wang, W. Zhu, and X.-J. Zeng, “A stoc hastic MPC based approach to integrated energy management in micro grids,” Sustainable cities and society , vol. 41, pp. 349–362, 2018

2018
[24]

Sequential monte carlo for model predictive control,

N. Kantas, J. Maciejowski, and A. Lecchini-Visintini, “Sequential monte carlo for model predictive control,” in Nonlinear model predictive control: Towards new challenging applications . Springer, 2009, pp. 263–273

2009
[25]

NMP C for complex stochastic systems using a markov chain monte carlo approac h,

J. M. Maciejowski, A. L. Visintini, and J. Lygeros, “NMP C for complex stochastic systems using a markov chain monte carlo approac h,” in As- sessment and Future Directions of Nonlinear Model Predicti ve Control. Springer, 2007, pp. 269–281

2007
[26]

Stochastic nonlinear model predictive control based on gaussian mixture approxi mations,

F. Weissel, M. F. Huber, and U. D. Hanebeck, “Stochastic nonlinear model predictive control based on gaussian mixture approxi mations,” in Informatics in Control, Automation and Robotics: Selected Papers from the International Conference on Informatics in Control, Au tomation and Robotics 2007 . Springer, 2009, pp. 239–252

2007
[27]

Stochastic nonlinear mode l predictive control using gaussian processes,

E. Bradford and L. Imsland, “Stochastic nonlinear mode l predictive control using gaussian processes,” in 2018 european control conference (ECC). IEEE, 2018, pp. 1027–1034

2018
[28]

Successive linearizati on based stochastic model predictive control for batch processes de scribed by daes,

L. L. Atmaram and H. Kodamana, “Successive linearizati on based stochastic model predictive control for batch processes de scribed by daes,” IF AC-PapersOnLine, vol. 53, no. 1, pp. 380–385, 2020

2020
[29]

Re inforcement learning with ensemble model predictive safety certiﬁcati on,

S. Gronauer, T. Haider, F. S. da Roza, and K. Diepold, “Re inforcement learning with ensemble model predictive safety certiﬁcati on,” arXiv preprint arXiv:2402.04182, 2024

work page arXiv 2024
[30]

Soft-constrained model pre dictive control based on data-driven distributionally robust optimization,

S. Lu, J. H. Lee, and F. Y ou, “Soft-constrained model pre dictive control based on data-driven distributionally robust optimization,” AIChE Journal, vol. 66, no. 10, p. e16546, 2020

2020
[31]

An introduction to deep gener ative model- ing,

L. Ruthotto and E. Haber, “An introduction to deep gener ative model- ing,” GAMM-Mitteilungen, vol. 44, no. 2, p. e202100008, 2021

2021
[32]

An efﬁcient method for stoc hastic optimal control with joint chance constraints for nonlinea r systems,

J. A. Paulson and A. Mesbah, “An efﬁcient method for stoc hastic optimal control with joint chance constraints for nonlinea r systems,” International Journal of Robust and Nonlinear Control , vol. 29, no. 15, pp. 5017–5037, 2019

2019
[33]

Polynomial chaos expa nsion for parametric problems in engineering systems: A review,

D. Shen, H. Wu, B. Xia, and D. Gan, “Polynomial chaos expa nsion for parametric problems in engineering systems: A review,” IEEE Systems Journal, vol. 14, no. 3, pp. 4500–4514, 2020

2020
[34]

Recursive feasibility and stability for stochastic MPC based on polyn omial chaos,

Z. Ma, H. Schl¨ uter, F. Berkel, T. Specker, and F. Allg¨ o wer, “Recursive feasibility and stability for stochastic MPC based on polyn omial chaos,” IF AC-PapersOnLine, vol. 56, no. 1, pp. 204–209, 2023

2023
[35]

Stochas tic model predictive control of discrete-time cooperative unmanned system,

R. Bhusal, D. Bhattacharjee, and K. Subbarao, “Stochas tic model predictive control of discrete-time cooperative unmanned system,” in AIAA SCITECH 2022 F orum , 2022, p. 0961

2022
[36]

St ochastic nonlinear model predictive control with probabilistic con straints,

A. Mesbah, S. Streif, R. Findeisen, and R. D. Braatz, “St ochastic nonlinear model predictive control with probabilistic con straints,” in 2014 American control conference . IEEE, 2014, pp. 2413–2419

2014
[37]

Nonlinear stochastic mode l predictive control via regularized polynomial chaos expansions,

L. Fagiano and M. Khammash, “Nonlinear stochastic mode l predictive control via regularized polynomial chaos expansions,” in 2012 IEEE 51st IEEE conference on decision and control (cdc) . IEEE, 2012, pp. 142–147

2012
[38]

Nonlinear model predictiv e control with explicit backoffs for stochastic systems under arbitrary u ncertainty,

J. A. Paulson and A. Mesbah, “Nonlinear model predictiv e control with explicit backoffs for stochastic systems under arbitrary u ncertainty,” IF AC-PapersOnLine, vol. 51, no. 20, pp. 523–534, 2018

2018
[39]

Statistical aspects of wa sserstein distances,

V . M. Panaretos and Y . Zemel, “Statistical aspects of wa sserstein distances,” Annual review of statistics and its application , vol. 6, no. 1, pp. 405–431, 2019

2019
[40]

Wasserstein auto-encoders.arXiv preprint arXiv:1711.01558,

I. Tolstikhin, O. Bousquet, S. Gelly, and B. Schoelkopf , “Wasserstein auto-encoders,” arXiv preprint arXiv:1711.01558 , 2017

work page arXiv 2017
[41]

Generative adversar ial nets,

I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. W arde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversar ial nets,” Advances in neural information processing systems , vol. 27, 2014

2014
[42]

Training generative neural networks via Maximum Mean Discrepancy optimization

G. K. Dziugaite, D. M. Roy, and Z. Ghahramani, “Training generative neural networks via maximum mean discrepancy optimization ,” arXiv preprint arXiv:1505.03906, 2015

work page Pith review arXiv 2015
[43]

Luo, Wiener chaos expansion and numerical solutions of stochast ic partial differential equations

W. Luo, Wiener chaos expansion and numerical solutions of stochast ic partial differential equations . California Institute of Technology, 2006

2006
[44]

Comparison of non-intrusiv e polynomial chaos and stochastic collocation methods for uncertainty q uantiﬁcation,

M. Eldred and J. Burkardt, “Comparison of non-intrusiv e polynomial chaos and stochastic collocation methods for uncertainty q uantiﬁcation,” in 47th AIAA aerospace sciences meeting including the new hori zons forum and aerospace exposition , 2009, p. 976

2009
[45]

Evaluation of un certainty in dynamic simulations of power system models: The probabil istic collocation method,

J. R. Hockenberry and B. C. Lesieutre, “Evaluation of un certainty in dynamic simulations of power system models: The probabil istic collocation method,” IEEE Transactions on Power Systems , vol. 19, no. 3, pp. 1483–1491, 2004

2004
[46]

On distributionally ro bust chance- constrained linear programs,

G. C. Calaﬁore and L. E. Ghaoui, “On distributionally ro bust chance- constrained linear programs,” Journal of Optimization Theory and Ap- plications, vol. 130, pp. 1–22, 2006

2006
[47]

Gasoline c ontrolled auto-ignition with learning-based uncertainty using stoc hastic model predictive control,

X. Chen, M. Basler, M. Stemmler, and D. Abel, “Gasoline c ontrolled auto-ignition with learning-based uncertainty using stoc hastic model predictive control,” in 2023 62nd IEEE Conference on Decision and Control (CDC). IEEE, 2023, pp. 6328–6335

2023
[48]

Dynamic measurement with in-cycle process exci tation of HCCI combustion: The key to handle complexity of data-dri ven control?

J. Bedei, M. Oberlies, P . Schaber, D. Gordon, E. Nuss, L. Li, and J. Andert, “Dynamic measurement with in-cycle process exci tation of HCCI combustion: The key to handle complexity of data-dri ven control?” International Journal of Engine Research , vol. 24, no. 3, pp. 1155–1174, 2023

2023
[49]

Gr ammar vari- ational autoencoder,

M. J. Kusner, B. Paige, and J. M. Hern´ andez-Lobato, “Gr ammar vari- ational autoencoder,” in International conference on machine learning . PMLR, 2017, pp. 1945–1954

2017