pith. sign in

arxiv: 1907.04213 · v1 · pith:LLHU7DE2new · submitted 2019-07-06 · 📡 eess.SY · cs.SY· math.OC

Dynamic Real-time Optimization of Batch Processes using Pontryagin's Minimum Principle and Set-membership Adaptation

Radoslav Paulen , Miroslav Fikar This is my paper

Pith reviewed 2026-05-25 01:53 UTC · model grok-4.3

classification 📡 eess.SY cs.SYmath.OC
keywords real-time optimizationbatch processesPontryagin's minimum principleset-membership estimationreachable setsmodel predictive controlparametric uncertaintyadaptive control
0
0 comments X

The pith

Batch processes can be optimized in real time under model mismatch by projecting reachable sets into parameterized optimality conditions from Pontryagin's principle.

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

This paper tackles the high computational cost of repeated full re-optimizations in receding-horizon control for batch processes that have parametric model errors. It proposes to parameterize the optimality conditions and treat prediction uncertainty explicitly by computing reachable sets from the current parameter bounds and projecting those sets into the conditions. Model parameters are then adapted online through set-membership estimation inside an adaptive predictive-control loop. The method is illustrated on batch membrane separation processes, where it is shown to lower the re-optimization burden while still accounting for uncertainty.

Core claim

Dynamic real-time optimization of time-optimal batch operation under parametric mismatch can be performed by using parameterized conditions of optimality in an adaptive predictive-control fashion, with model uncertainty handled explicitly via reachable sets that are projected into the optimality conditions and with parameters adapted online by set-membership estimation, thereby reducing the computational load of frequent full re-optimizations.

What carries the argument

Parameterized conditions of optimality (from Pontryagin's Minimum Principle) into which reachable sets computed from the current parameter uncertainty set are projected, combined with set-membership estimation for online adaptation.

If this is right

  • The computational burden of frequent re-optimization is reduced while uncertainty is accounted for explicitly through the projection step.
  • Online set-membership adaptation progressively tightens the reachable sets and reduces conservatism.
  • The optimality conditions remain satisfied despite ongoing model mismatch.
  • The scheme applies directly to batch membrane separation and similar processes with parametric uncertainty.

Where Pith is reading between the lines

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

  • This projection technique could be combined with other uncertainty representations beyond reachable sets for different classes of processes.
  • The same parameterization might reduce computation in continuous processes where receding-horizon updates are also expensive.
  • Explicit handling of reachable sets inside optimality conditions offers a route to certify robustness of the computed control trajectories.
  • Industrial deployment would require checking whether the projection step itself can be computed fast enough on embedded hardware.

Load-bearing premise

Reachable sets computed from the current parameter uncertainty set can be projected into the parameterized optimality conditions without destroying the optimality guarantee or introducing excessive conservatism.

What would settle it

Apply the method to a concrete batch process, compare the achieved batch completion time and total computation time against a standard receding-horizon solver that fully re-optimizes at every step; if the proposed method either takes more total time or produces a longer batch duration, the efficiency claim is falsified.

Figures

Figures reproduced from arXiv: 1907.04213 by Miroslav Fikar, Radoslav Paulen.

Figure 1
Figure 1. Figure 1: Illustration of the switching function evolution under uncertainty using (set [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the parameterization of the optimal control policy under uncer [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Schematic representation of a generalized diafiltration process. [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Results of the set-membership estimation over time (top and middle plots) with [PITH_FULL_IMAGE:figures/full_fig_p020_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: A box plot with the statistical information (the median, the 25 [PITH_FULL_IMAGE:figures/full_fig_p021_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Results of the set-membership estimation over time (top three plots) with pro [PITH_FULL_IMAGE:figures/full_fig_p024_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: A box plot with the statistical information (the median, the 25 [PITH_FULL_IMAGE:figures/full_fig_p025_7.png] view at source ↗
read the original abstract

This paper studies a dynamic real-time optimization in the context of model-based time-optimal operation of batch processes under parametric model mismatch. In order to tackle the model-mismatch issue, a receding-horizon policy is usually followed with frequent re-optimization. The main problem addressed in this study is the high computational burden that is usually required by such schemes. We propose an approach that uses parameterized conditions of optimality in the adaptive predictive-control fashion. The uncertainty in the model predictions is treated explicitly using reachable sets that are projected into the optimality conditions. Adaptation of model parameters is performed online using set-membership estimation. A class of batch membrane separation processes is in the scope of the presented applications, where the benefits of the presented approach are outlined.

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.

Referee Report

2 major / 2 minor

Summary. The paper proposes a dynamic real-time optimization method for batch processes subject to parametric model mismatch. It parameterizes the optimality conditions from Pontryagin's Minimum Principle and embeds them in a receding-horizon adaptive predictive-control scheme. Model uncertainty is handled by computing reachable sets from the current parameter uncertainty set and projecting those sets into the parameterized optimality conditions; parameters are adapted online via set-membership estimation. The approach is illustrated on a class of batch membrane separation processes, with the central claim being that the method achieves substantial computational savings relative to repeated full re-optimization while preserving acceptable performance.

Significance. If the projection step can be shown to retain a quantifiable optimality guarantee or at least a bounded performance degradation, the work would offer a practical route to real-time optimal control of uncertain batch systems by combining PMP necessary conditions with set-membership techniques. This could be particularly relevant for membrane separation and similar processes where frequent re-optimization is computationally prohibitive.

major comments (2)
  1. [Method description (reachable-set projection step)] The projection of reachable sets into the parameterized PMP optimality conditions is the load-bearing step for both the claimed computational savings and the preservation of performance. No explicit mathematical definition of the projection operator is supplied, nor is any sub-optimality bound or guarantee that the projected conditions remain sufficient for the time-optimal objective derived.
  2. [Numerical examples / case studies] The numerical validation on membrane separation processes reports computational-time reductions but does not include a side-by-side comparison of closed-loop performance (e.g., batch time or yield) against a baseline of repeated full re-optimization under identical uncertainty realizations, leaving the performance-conservatism trade-off unquantified.
minor comments (2)
  1. Notation for the reachable-set projection and the parameterized Hamiltonian could be introduced earlier and used consistently to improve readability.
  2. [Introduction] The introduction would benefit from a short paragraph contrasting the proposed projection approach with existing robust or tube-based MPC formulations that also employ reachable sets.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We respond to each major comment below, indicating planned revisions where the manuscript can be strengthened without misrepresenting the original contributions.

read point-by-point responses

  1. Referee: [Method description (reachable-set projection step)] The projection of reachable sets into the parameterized PMP optimality conditions is the load-bearing step for both the claimed computational savings and the preservation of performance. No explicit mathematical definition of the projection operator is supplied, nor is any sub-optimality bound or guarantee that the projected conditions remain sufficient for the time-optimal objective derived.

    Authors: We agree that the projection operator requires an explicit mathematical definition, which was only described at a high level in the original text. The revised manuscript will add a formal definition of the projection as the operator that maps the reachable set of state trajectories (under the current parameter uncertainty) onto the admissible set of parameterized PMP conditions. However, no sub-optimality bound or sufficiency guarantee is derived in the work; the emphasis is on computational tractability for the membrane-separation class. We will revise the text to state this limitation explicitly and identify it as future work. revision: partial

  2. Referee: [Numerical examples / case studies] The numerical validation on membrane separation processes reports computational-time reductions but does not include a side-by-side comparison of closed-loop performance (e.g., batch time or yield) against a baseline of repeated full re-optimization under identical uncertainty realizations, leaving the performance-conservatism trade-off unquantified.

    Authors: The original manuscript reports computation-time savings and states that performance remains acceptable, but does not provide the requested side-by-side closed-loop metrics under matched uncertainty realizations. We will add this comparison in the revised version by including batch-time and yield results from full re-optimization runs on the same uncertainty sequences already used in the case studies. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation combines standard PMP and set-membership tools without self-referential reduction

full rationale

The provided abstract and context contain no equations or explicit derivation steps that reduce a claimed prediction or optimality condition to a fitted input by construction, nor any load-bearing self-citation chain. The approach parameterizes PMP conditions and projects reachable sets from set-membership uncertainty, but without quoted text exhibiting a self-definitional loop (e.g., adaptation rule tuned directly to performance metric) or renaming of known results, the chain remains independent of its own outputs. This is the expected honest non-finding for a methods paper whose central construction is not shown to collapse into its inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no explicit free parameters, axioms, or invented entities; the method implicitly assumes that model mismatch is purely parametric and that reachable sets remain tractable.

pith-pipeline@v0.9.0 · 5661 in / 1205 out tokens · 20334 ms · 2026-05-25T01:53:01.356793+00:00 · methodology

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

32 extracted references · 32 canonical work pages

  1. [1]

    Adaptive model predictive control for constrained nonlinear systems

    Adetola, V., DeHaan, D., Guay, M., 2009. Adaptive model predictive control for constrained nonlinear systems. Systems & Control Letters 58 (5), 320 – 326

    work page 2009

  2. [2]

    Extents of reaction and flow for homogeneous reaction systems with inlet and outlet streams

    Amrhein, M., Bhatt, N., Srinivasan, B., Bonvin, D., 11 2010. Extents of reaction and flow for homogeneous reaction systems with inlet and outlet streams. AIChE Journal 56

    work page 2010

  3. [3]

    Toward fast dynamic optimiza- tion: An indirect algorithm that uses parsimonious input parameterization

    Aydin, E., Bonvin, D., Sundmacher, K., 2018. Toward fast dynamic optimiza- tion: An indirect algorithm that uses parsimonious input parameterization. Industrial & Engineering Chemistry Research 57 (30), 10038–10048

    work page 2018

  4. [4]

    Adjoint sensitivity analysis for differential-algebraic equations: The adjoint dae system and its numerical solution

    Cao, Y., Li, S., Petzold, L., Serban, R., 2003. Adjoint sensitivity analysis for differential-algebraic equations: The adjoint dae system and its numerical solution. SIAM Journal on Scientific Computing 24 (3), 1076–1089. 29

    work page 2003

  5. [5]

    D., Rajyaguru, J., Vil- lanueva, M

    Chachuat, B., Houska, B., Paulen, R., Peri´ c, N. D., Rajyaguru, J., Vil- lanueva, M. E., 2015. Set-theoretic approaches in analysis, estimation and control of nonlinear systems. IFAC-PapersOnLine 48 (8), 981–995

    work page 2015

  6. [6]

    Ultrafiltration and microfiltration handbook

    Cheryan, M., 1998. Ultrafiltration and microfiltration handbook. CRC press,

    work page 1998

  7. [7]

    de Prada, C., Rodriguez, M., Sarabia, D., 2011

    Florida, USA. de Prada, C., Rodriguez, M., Sarabia, D., 2011. On-line scheduling and con- trol of a mixed continuous-batch plant. Industrial & Engineering Chemistry Research 50 (9), 5041–5049

    work page 2011

  8. [8]

    On the value of information in system identifica- tion – bounded noise case

    Fogel, E., Huang, Y., 1982. On the value of information in system identifica- tion – bounded noise case. Automatica 18 (2), 229 – 238. Fran¸ cois, G., Bonvin, D., 2013. Chapter one - measurement-based real-time optimization of chemical processes. In: Pushpavanam, S. (Ed.), Control and Optimisation of Process Systems. Vol. 43 of Advances in Chemical Enginee...

    work page 1982

  9. [9]

    M., Bokor, J., Szederk´ enyi, G., 2006

    Hangos, K. M., Bokor, J., Szederk´ enyi, G., 2006. Analysis and control of nonlinear process systems. Springer

    work page 2006

  10. [10]

    Control of particle size distribu- tion in emulsion polymerization using mid-course correction under struc- tural plant-model mismatch*

    Hosseini, A., Oshaghi, M., Engell, S., 2013. Control of particle size distribu- tion in emulsion polymerization using mid-course correction under struc- tural plant-model mismatch*. IFAC Proceedings Volumes 46 (32), 529 – 534, 10th IFAC International Symposium on Dynamics and Control of Process Systems

    work page 2013

  11. [11]

    H., Biegler, L

    Jang, H., Lee, J. H., Biegler, L. T., 2016. A robust nmpc scheme for semi- batch polymerization reactors. IFAC-PapersOnLine 49 (7), 37 – 42, 11th IFAC Symposium on Dynamics and Control of Process Systems Including Biosystems DYCOPS-CAB 2016

    work page 2016

  12. [12]

    Set inversion via interval analysis for nonlinear bounded-error estimation

    Jaulin, L., Walter, E., 1993. Set inversion via interval analysis for nonlinear bounded-error estimation. Automatica 29 (4), 1053–1064

    work page 1993

  13. [13]

    D., Gibson, J

    Johnson, C. D., Gibson, J. E., 1963. Singular solutions in problems of optimal control 8 (1), 4–15

    work page 1963

  14. [14]

    Exact linearization and control of a continuous stirred tank reactor

    Liou, C., Hsiue, T., 1995. Exact linearization and control of a continuous stirred tank reactor. Journal of the Chinese Institute of Engineers 18 (6), 825–833

    work page 1995

  15. [15]

    Multi-stage nonlinear model predictive control applied to a semi-batch polymerization reactor under uncertainty

    Lucia, S., Finkler, T., Engell, S., 2013. Multi-stage nonlinear model predictive control applied to a semi-batch polymerization reactor under uncertainty. J Process Contr 23 (9), 1306 – 1319. Mart´ ı, R., Lucia, S., Sarabia, D., Paulen, R., Engell, S., de Prada, C., 2015. 31 Improving scenario decomposition algorithms for robust nonlinear model predictive...

    work page 2013

  16. [16]

    K., Braatz, R

    Nagy, Z. K., Braatz, R. D., 2003. Robust nonlinear model predictive control of batch processes. AIChE Journal 49 (7), 1776–1786

    work page 2003

  17. [17]

    Optimal Operation of Batch Membrane Pro- cesses

    Paulen, R., Fikar, M., 2016. Optimal Operation of Batch Membrane Pro- cesses. Springer

    work page 2016

  18. [18]

    Optimal feeding strategy of diafiltration buffer in batch membrane processes

    Paulen, R., Fikar, M., Foley, G., Kov´ acs, Z., Czermak, P., 2012. Optimal feeding strategy of diafiltration buffer in batch membrane processes. Jour- nal of Membrane Science 411-412, 160–172

    work page 2012

  19. [19]

    Economically opti- mal batch diafiltration via analytical multi-objective optimal control

    Paulen, R., Jelemensk´ y, M., Kov´ acs, Z., Fikar, M., 2015. Economically opti- mal batch diafiltration via analytical multi-objective optimal control. Jour- nal of Process Control 28, 73 – 82

    work page 2015

  20. [20]

    E., Chachuat, B., 2016

    Paulen, R., Villanueva, M. E., Chachuat, B., 2016. Guaranteed parameter estimation of non-linear dynamic systems using high-order bounding tech- niques with domain and cpu-time reduction strategies. IMA Journal of Mathematical Control and Information 33 (3), 563–587

    work page 2016

  21. [21]

    S., Boltyanskii, V

    Pontryagin, L. S., Boltyanskii, V. G., Gamkrelidze, R. V., Mishchenko, E. F., 1962. The Mathematical Theory of Optimal Processes. John Wiley & Sons, Inc., New York

    work page 1962

  22. [22]

    Process Optimization in Ultrafiltration: Flux-Time Considerations in the Purification of Macromolecules 106 (1), 57–69

    Rajagopalan, N., Cheryan, M., 1991. Process Optimization in Ultrafiltration: Flux-Time Considerations in the Purification of Macromolecules 106 (1), 57–69. 32

    work page 1991

  23. [23]

    Dynamic optimization of reaction systems via exact parsimonious input parameterization

    Rodrigues, D., Bonvin, D., 2019. Dynamic optimization of reaction systems via exact parsimonious input parameterization. Industrial & Engineering Chemistry ResearchIn Press

    work page 2019

  24. [24]

    Detection and exploitation of the con- trol switching structure in the solution of dynamic optimization problems

    Schlegel, M., Marquardt, W., 2006. Detection and exploitation of the con- trol switching structure in the solution of dynamic optimization problems. Journal of Process Control 16 (3), 275 – 290, selected Papers from Dycops 7 (2004), Cambridge, Massachusetts Seventh {IFAC} Symposium on the Dymanics and Control of Process Systems (Dycops-7)

    work page 2006

  25. [25]

    Dynamic optimization using adaptive control vector parameterization

    Schlegel, M., Stockmann, K., Binder, T., Marquardt, W., 2005. Dynamic optimization using adaptive control vector parameterization. Computers & Chemical Engineering 29 (8), 1731 – 1751

    work page 2005

  26. [26]

    Recursive state estimation: Unknown but bounded errors and system inputs

    Schweppe, F., 1968. Recursive state estimation: Unknown but bounded errors and system inputs. IEEE Transactions on Automatic Control 13 (1), 22–28

    work page 1968

  27. [27]

    Implementa- tion of optimal strategy to economically improve batch membrane separa- tion

    Sharma, A., Valo, R., Kal ˜Aˇ zz, M., Paulen, R., Fikar, M., 2019. Implementa- tion of optimal strategy to economically improve batch membrane separa- tion. Journal of Process Control 76, 155 – 164

    work page 2019

  28. [28]

    D., 1998

    Sontag, E. D., 1998. Mathematical Control Theory: Deterministic Finite Dimensional Systems (2nd Ed.). Springer-Verlag, Berlin, Heidelberg

    work page 1998

  29. [29]

    E., Houska, B., Chachuat, B., 2015

    Villanueva, M. E., Houska, B., Chachuat, B., 2015. Unified framework for the propagation of continuous-time enclosures for parametric nonlinear odes. Journal of Global Optimization 62 (3), 575–613

    work page 2015

  30. [30]

    Bounded-error opti- mal experimental design via global solution of constrained minˆ a ˘A¸ Smax program

    Walz, O., Djelassi, H., Caspari, A., Mitsos, A., 2018. Bounded-error opti- mal experimental design via global solution of constrained minˆ a ˘A¸ Smax program. Computers & Chemical Engineering 111, 92 – 101

    work page 2018

  31. [31]

    F., 1997

    Yabuki, Y., MacGregor, J. F., 1997. Product quality control in semibatch re- actors using midcourse correction policies. Industrial & Engineering Chem- istry Research 36 (4), 1268–1275

    work page 1997

  32. [32]

    F., 2002

    Yabuki, Y., Nagasawa, T., MacGregor, J. F., 2002. Industrial experiences with product quality control in semi-batch processes. Computers & Chem- ical Engineering 26 (2), 205 – 212. 34

    work page 2002