pith. sign in

arxiv: 2002.08556 · v2 · submitted 2020-02-20 · 🧮 math.OC

Diffusing-Horizon Model Predictive Control

Sungho Shin , Victor M. Zavala This is my paper

Pith reviewed 2026-05-24 14:14 UTC · model grok-4.3

classification 🧮 math.OC
keywords model predictive controltime coarseningexponential decay of sensitivityparametric perturbationlinear dynamicsHVAC controlcomputational efficiencyoptimal control
0
0 comments X

The pith

A time-coarsening strategy for linear model predictive control preserves exponential decay of sensitivity by treating the grid as a parametric perturbation, enabling two-order-of-magnitude speedups with a 3 percent cost increase.

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

The paper establishes conditions under which the exponential decay of sensitivity holds in constrained model predictive control problems that have linear dynamics and costs. This decay means that the effect of a change far in the future on decisions made now shrinks exponentially as one looks backward in time. The authors then show that their proposed diffusing-horizon coarsening, in which the time grid becomes exponentially sparser into the future, is mathematically equivalent to a parametric perturbation of the original problem and therefore inherits the decay property. In a heating, ventilation, and air conditioning example that uses real operating data, the approach reduces computation time by roughly a hundred times while increasing the realized closed-loop cost by only three percent.

Core claim

The central claim is that the diffusing-horizon MPC formulation, whose time discretization grid becomes exponentially more sparse as one moves forward, satisfies the exponential decay of sensitivity condition for constrained problems with linear dynamics and costs because the coarsening can be rewritten as a parametric perturbation of the original MPC problem.

What carries the argument

The diffusing-horizon time discretization grid that sparsens exponentially forward in time, reformulated as a parametric perturbation of the MPC optimization problem.

If this is right

  • The exponential decay of sensitivity holds for constrained MPC with linear dynamics and costs under the conditions derived in the paper.
  • The proposed coarsening scheme inherits the decay property because it is a parametric perturbation.
  • Computational solution times are reduced by two orders of magnitude on the HVAC plant example.
  • Closed-loop cost rises by only three percent despite the drastic reduction in grid points.

Where Pith is reading between the lines

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

  • The same coarsening idea could be applied to any linear MPC problem whose sensitivity decay rate can be bounded a priori, even if the explicit conditions of the paper are not met.
  • Real-time implementation on embedded hardware becomes feasible for prediction horizons that would otherwise be computationally prohibitive.
  • The approach may combine naturally with move-blocking or other horizon-reduction heuristics to produce still larger speed gains.

Load-bearing premise

The exponential decay of sensitivity property must hold for the constrained MPC problem with linear dynamics and costs.

What would settle it

A concrete linear constrained MPC instance in which the sensitivity of the first-stage decisions to a parametric perturbation at a distant future time step fails to decay exponentially with the number of steps between them.

Figures

Figures reproduced from arXiv: 2002.08556 by Sungho Shin, Victor M. Zavala.

Figure 1
Figure 1. Figure 1: Schematic representation of different coarsening s [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of perturbation analysis setting. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of algebraic coarsening strategy (diff [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Schematic representation of central HVAC plant unde [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 7
Figure 7. Figure 7: Solution trajectories for different coarsening sch [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 5
Figure 5. Figure 5: Time-varying data for HVAC plant problem. Electrici [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Effect of perturbations at different time locations [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Closed-loop economic performance for different coar [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Total solution times for different coarsening schem [PITH_FULL_IMAGE:figures/full_fig_p014_9.png] view at source ↗
read the original abstract

We analyze a time-coarsening strategy for model predictive control (MPC) that we call diffusing-horizon MPC. This strategy seeks to overcome the computational challenges associated with optimal control problems that span multiple timescales. The coarsening approach uses a time discretization grid that becomes exponentially more sparse as one moves forward in time. This design is motivated by a recently established property of optimal control problems that is known as exponential decay of sensitivity. This property states that the impact of a parametric perturbation at a future time decays exponentially as one moves backward in time. We establish conditions under which this property holds for a constrained MPC formulation with linear dynamics and costs. Moreover, we show that the proposed coarsening scheme can be cast as a parametric perturbation of the MPC problem and thus the exponential decay condition holds. We use a heating, ventilation, and air conditioning plant case study with real data to demonstrate the proposed approach. Specifically, we show that computational times can be reduced by two orders of magnitude while increasing the closed-loop cost by only 3%.

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 / 1 minor

Summary. The paper proposes 'diffusing-horizon MPC,' a time-coarsening strategy for model predictive control that employs an exponentially sparse discretization grid forward in time. It is motivated by the exponential decay of sensitivity property in optimal control problems. The authors establish conditions under which this property holds for constrained MPC with linear dynamics and costs, cast the coarsening scheme as a parametric perturbation of the original MPC problem (so that the decay applies), and demonstrate via an HVAC case study with real data that computational times can be reduced by two orders of magnitude while increasing closed-loop cost by only 3%.

Significance. If the conditions for exponential decay and the parametric-perturbation casting hold rigorously, the approach would provide a principled way to reduce the computational cost of long-horizon MPC on multi-timescale systems while preserving near-optimal closed-loop performance; the reported 100x speedup with 3% cost penalty in a realistic HVAC example would be a practically relevant result.

major comments (2)
  1. [Abstract claim on parametric perturbation casting] The central claim that the coarsening scheme equals a parametric perturbation of the original constrained linear-quadratic MPC (so that the established exponential sensitivity decay applies directly) is load-bearing. When the time grids differ and inequality constraints are active, altering the decision-variable dimension and the integrated dynamics/constraint matrices is not automatically a bounded perturbation of the original QP data; an explicit interpolation or zero-order-hold map must be shown to preserve the exact sensitivity structure used in the decay proof. This equivalence is asserted in the abstract but requires a concrete verification that the perturbation norm decays exponentially backward in time under active constraints.
  2. [Section establishing conditions for exponential decay] The conditions under which exponential decay of sensitivity is established for the constrained MPC formulation (linear dynamics and costs) are not yet shown to survive the grid coarsening when the active-set changes; the proof must explicitly address whether the sensitivity decay bound remains uniform across the sequence of perturbed QPs induced by the diffusing horizon.
minor comments (1)
  1. [Case study section] The abstract reports specific numerical outcomes (two orders of magnitude reduction, 3% cost increase) from the HVAC case study; the main text should include the exact horizon lengths, constraint active-set statistics, and solver tolerances used so that the speedup claim can be reproduced.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. The comments highlight important points regarding the rigor of the parametric perturbation argument and the uniformity of the sensitivity decay under grid coarsening. We address each major comment below and indicate where revisions will strengthen the presentation.

read point-by-point responses

  1. Referee: [Abstract claim on parametric perturbation casting] The central claim that the coarsening scheme equals a parametric perturbation of the original constrained linear-quadratic MPC (so that the established exponential sensitivity decay applies directly) is load-bearing. When the time grids differ and inequality constraints are active, altering the decision-variable dimension and the integrated dynamics/constraint matrices is not automatically a bounded perturbation of the original QP data; an explicit interpolation or zero-order-hold map must be shown to preserve the exact sensitivity structure used in the decay proof. This equivalence is asserted in the abstract but requires a concrete verification that the perturbation norm decays exponentially backward in time under active constraints.

    Authors: We agree that the casting requires an explicit map and verification under active constraints. Section 3 defines the coarsening via a zero-order-hold interpolation from the fine to the coarse grid, which produces a structured perturbation to the QP data matrices. However, the current proof does not explicitly bound the perturbation norm when the active set is nonempty. We will add a new lemma in the revised manuscript that shows the induced perturbation in the KKT system remains exponentially small in the backward direction by exploiting the exponential sparsity pattern, even when the active-set projection is applied. This will be supported by a short numerical verification on the HVAC example. revision: yes

  2. Referee: [Section establishing conditions for exponential decay] The conditions under which exponential decay of sensitivity is established for the constrained MPC formulation (linear dynamics and costs) are not yet shown to survive the grid coarsening when the active-set changes; the proof must explicitly address whether the sensitivity decay bound remains uniform across the sequence of perturbed QPs induced by the diffusing horizon.

    Authors: The decay result in Section 2 is derived for a fixed active set. For the sequence of perturbed QPs arising from coarsening, the active sets are not guaranteed to be identical. We will revise the manuscript to include an additional argument (a short corollary) showing that the decay constant remains uniform across sufficiently small perturbations, which holds by the exponential grid construction. When active-set changes occur, the bound is understood to apply locally to each QP; we will clarify this scope in the text rather than claiming global uniformity without qualification. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation chain is self-contained

full rationale

The paper establishes conditions for the exponential decay of sensitivity property directly for constrained linear MPC (a mathematical result presented as new), then shows the coarsening scheme can be cast as a parametric perturbation so the property applies. No quoted reduction of any central claim to a self-citation chain, fitted parameter renamed as prediction, or self-definitional loop appears in the provided text. The case-study performance numbers are empirical outcomes, not forced by construction. This is the normal finding for a paper whose core proofs and casting step stand independently of the inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the existence of exponential sensitivity decay for the linear constrained MPC problem and on the ability to represent the coarsening as a parametric perturbation; both are stated as established in the paper but their precise assumptions are not visible in the abstract.

axioms (1)
  • domain assumption Exponential decay of sensitivity holds for constrained MPC with linear dynamics and costs under the conditions established in the paper
    This property is invoked to justify the coarsening and is claimed to be preserved when the scheme is viewed as a parametric perturbation.

pith-pipeline@v0.9.0 · 5701 in / 1327 out tokens · 21801 ms · 2026-05-24T14:14:29.790110+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Controllability and Observability Imply Exponential Decay of Sensitivity in Dynamic Optimization

    math.OC 2021-01 unverdicted novelty 6.0

    Uniform controllability and observability imply exponential decay of sensitivity under uniform Hessian boundedness, uSOSC, and uLICQ in dynamic optimization.

  2. Exponential Decay of Sensitivity in Graph-Structured Nonlinear Programs

    math.OC 2021-01 conditional novelty 6.0

    Sensitivity of primal-dual solutions in graph-induced NLPs decays exponentially with graph distance under SOSC and LICQ.

  3. On the Convergence of Overlapping Schwarz Decomposition for Nonlinear Optimal Control

    math.OC 2020-05 unverdicted novelty 5.0

    Overlapping Schwarz decomposition for nonlinear OCPs achieves local linear convergence with rate improving exponentially with overlap size, based on exponential decay of sensitivity for primal and dual solutions.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · cited by 3 Pith papers · 1 internal anchor

  1. [1]

    Control of integrated proce ss networks- a multi-time scale perspective,

    M. Baldea and P . Daoutidis, “Control of integrated proce ss networks- a multi-time scale perspective,” Computers & chemical engineering , vol. 31, no. 5-6, pp. 426–444, 2007

    work page 2007

  2. [2]

    Time domain parti- tioning of electricity production cost simulations,

    C. Barrows, M. Hummon, W. Jones, and E. Hale, “Time domain parti- tioning of electricity production cost simulations,” Nati onal Renewable Energy Lab.(NREL), Golden, CO (United States), Tech. Rep., 2014

    work page 2014

  3. [3]

    Pre- dictive active steering control for autonomous vehicle sys tems,

    P . Falcone, F. Borrelli, J. Asgari, H. E. Tseng, and D. Hro vat, “Pre- dictive active steering control for autonomous vehicle sys tems,” IEEE Transactions on control systems technology , vol. 15, no. 3, pp. 566–580, 2007

    work page 2007

  4. [4]

    Temporal decompositi on scheme for nonlinear multisite production planning and distribut ion models,

    J. R. Jackson and I. E. Grossmann, “Temporal decompositi on scheme for nonlinear multisite production planning and distribut ion models,” Industrial & engineering chemistry research , vol. 42, no. 13, pp. 3045– 3055, 2003

    work page 2003

  5. [5]

    A multi-scale o ptimization framework for electricity market participation,

    A. W. Dowling, R. Kumar, and V . M. Zavala, “A multi-scale o ptimization framework for electricity market participation,” Applied energy, vol. 190, pp. 147–164, 2017

    work page 2017

  6. [6]

    Architectures for distributed and hier archical model predictive control–a review,

    R. Scattolini, “Architectures for distributed and hier archical model predictive control–a review,” Journal of process control , vol. 19, no. 5, pp. 723–731, 2009

    work page 2009

  7. [7]

    A n MPC approach to the design of two-layer hierarchical control sy stems,

    B. Picasso, D. De Vito, R. Scattolini, and P . Colaneri, “A n MPC approach to the design of two-layer hierarchical control sy stems,” Automatica, vol. 46, no. 5, pp. 823–831, 2010

    work page 2010

  8. [8]

    Hierarchical MPC Schemes for Periodic Systems using Stochastic Programming

    R. Kumar, M. J. Wenzel, M. J. Ellis, M. N. ElBsat, K. H. Dree s, and V . M. Zavala, “Hierarchical MPC schemes for periodic system s using stochastic programming,” arXiv preprint arXiv:1804.10866 , 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  9. [9]

    Distributed optimization and statistical learning via the alternating direction method of multipliers,

    S. Boyd, N. Parikh, E. Chu, B. Peleato, J. Eckstein et al. , “Distributed optimization and statistical learning via the alternating direction method of multipliers,” F oundations and Trends® in Machine learning , vol. 3, no. 1, pp. 1–122, 2011

    work page 2011

  10. [10]

    Generalized benders decomposition,

    A. M. Geoffrion, “Generalized benders decomposition, ” Journal of optimization theory and applications , vol. 10, no. 4, pp. 237–260, 1972

    work page 1972

  11. [11]

    A par allel decom- position scheme for solving long-horizon optimal control p roblems,

    S. Shin, T. Faulwasser, M. Zanon, and V . M. Zavala, “A par allel decom- position scheme for solving long-horizon optimal control p roblems,” arXiv preprint arXiv:1903.01055 , 2019

    work page arXiv 1903

  12. [12]

    Benchmarking large-s cale dis- tributed convex quadratic programming algorithms,

    A. Kozma, C. Conte, and M. Diehl, “Benchmarking large-s cale dis- tributed convex quadratic programming algorithms,” Optimization Meth- ods and Software , vol. 30, no. 1, pp. 191–214, 2015

    work page 2015

  13. [13]

    Accelerated gradient methods and dual decomposition in di stributed model predictive control,

    P . Giselsson, M. D. Doan, T. Keviczky, B. De Schutter, an d A. Rantzer, “Accelerated gradient methods and dual decomposition in di stributed model predictive control,” Automatica, vol. 49, no. 3, pp. 829–833, 2013

    work page 2013

  14. [14]

    Real-time optimization and nonlinear model predictive control of processes governed by differential-algebraic e quations,

    M. Diehl, H. G. Bock, J. P . Schl¨ oder, R. Findeisen, Z. Na gy, and F. Allg¨ ower, “Real-time optimization and nonlinear model predictive control of processes governed by differential-algebraic e quations,” Jour- nal of Process Control , vol. 12, no. 4, pp. 577–585, 2002

    work page 2002

  15. [15]

    A continuation/gmres method for fast comp utation of nonlinear receding horizon control,

    T. Ohtsuka, “A continuation/gmres method for fast comp utation of nonlinear receding horizon control,” Automatica, vol. 40, no. 4, pp. 563– 574, 2004

    work page 2004

  16. [16]

    The advanced-step nmpc c ontroller: Optimality, stability and robustness,

    V . M. Zavala and L. T. Biegler, “The advanced-step nmpc c ontroller: Optimality, stability and robustness,” Automatica, vol. 45, no. 1, pp. 86–93, 2009

    work page 2009

  17. [17]

    Efficient num erical methods for nonlinear mpc and moving horizon estimation,

    M. Diehl, H. J. Ferreau, and N. Haverbeke, “Efficient num erical methods for nonlinear mpc and moving horizon estimation,” in Nonlinear model predictive control. Springer, 2009, pp. 391–417

    work page 2009

  18. [18]

    Real-time nonlinear opti mization as a generalized equation,

    V . M. Zavala and M. Anitescu, “Real-time nonlinear opti mization as a generalized equation,” SIAM Journal on Control and Optimization , vol. 48, no. 8, pp. 5444–5467, 2010

    work page 2010

  19. [19]

    New architectures for hierarchical pred ictive control,

    V . M. Zavala, “New architectures for hierarchical pred ictive control,” IF AC-PapersOnLine, vol. 49, no. 7, pp. 43–48, 2016

    work page 2016

  20. [20]

    Multigrid methods for pde opti mization,

    A. Borz` ı and V . Schulz, “Multigrid methods for pde opti mization,” SIAM review, vol. 51, no. 2, pp. 361–395, 2009

    work page 2009

  21. [21]

    General highly accurate algebraic coarsen ing,

    A. Brandt, “General highly accurate algebraic coarsen ing,” Electronic Transactions on Numerical Analysis , vol. 10, no. 1, p. 21, 2000

    work page 2000

  22. [22]

    An aggregation-based algebraic multigrid m ethod,

    Y . Notay, “An aggregation-based algebraic multigrid m ethod,” Electronic transactions on numerical analysis , vol. 37, no. 6, pp. 123–146, 2010

    work page 2010

  23. [23]

    Multi-grid schemes for multi- scale coordi- nation of energy systems,

    S. Shin and V . M. Zavala, “Multi-grid schemes for multi- scale coordi- nation of energy systems,” in Energy Markets and Responsive Grids . Springer, 2018, pp. 195–222

    work page 2018

  24. [24]

    A hierarchi cal optimization architecture for large-scale power networks,

    S. Shin, P . Hart, T. Jahns, and V . M. Zavala, “A hierarchi cal optimization architecture for large-scale power networks,” IEEE Transactions on Control of Network Systems , 2019

    work page 2019

  25. [25]

    Move blocking strategies in receding horizon control,

    R. Cagienard, P . Grieder, E. C. Kerrigan, and M. Morari, “Move blocking strategies in receding horizon control,” Journal of Process Control , vol. 17, no. 6, pp. 563–570, 2007

    work page 2007

  26. [26]

    Least-restrictive mo ve-blocking model predictive control,

    R. Gondhalekar and J.-i. Imura, “Least-restrictive mo ve-blocking model predictive control,” Automatica, vol. 46, no. 7, pp. 1234–1240, 2010

    work page 2010

  27. [27]

    Exponentially accurate tempora l decompo- sition for long-horizon linear-quadratic dynamic optimiz ation,

    W. Xu and M. Anitescu, “Exponentially accurate tempora l decompo- sition for long-horizon linear-quadratic dynamic optimiz ation,” SIAM Journal on Optimization , vol. 28, no. 3, pp. 2541–2573, 2018

    work page 2018

  28. [28]

    Exponential decay in the sensiti vity analysis of nonlinear dynamic programming,

    S. Na and M. Anitescu, “Exponential decay in the sensiti vity analysis of nonlinear dynamic programming,” SIAM Journal on Optimization , vol. 30, no. 2, pp. 1527–1554, 2020

    work page 2020

  29. [29]

    Exponential se nsitivity and turnpike analysis for linear quadratic optimal control of g eneral evolution equations,

    L. Gr¨ une, M. Schaller, and A. Schiela, “Exponential se nsitivity and turnpike analysis for linear quadratic optimal control of g eneral evolution equations,” Journal of Differential Equations, vol. 268, no. 12, pp. 7311– 7341, 2020

    work page 2020

  30. [30]

    Gr¨une, M

    ——, “Abstract nonlinear sensitivity and turnpike anal ysis and an appli- cation to semilinear parabolic pdes,” arXiv preprint arXiv:2008.13001 , 2020

    work page arXiv 2008

  31. [31]

    Model predictive co ntrol with non- uniformly spaced optimization horizon for multi-timescal e processes,

    C. K. Tan, M. J. Tippett, and J. Bao, “Model predictive co ntrol with non- uniformly spaced optimization horizon for multi-timescal e processes,” Computers & Chemical Engineering , vol. 84, pp. 162–170, 2016. 15

    work page 2016

  32. [32]

    Stochastic model predictive control for central hvac plants,

    R. Kumar, M. J. Wenzel, M. N. ElBsat, M. Risbeck, M. J. Ell is, K. H. Drees, and V . M. Zavala, “Stochastic model predictive control for central hvac plants,” 2019, under Review

    work page 2019

  33. [33]

    Economic model predic tive control for time-varying cost and peak demand charge optimization,

    M. J. Risbeck and J. B. Rawlings, “Economic model predic tive control for time-varying cost and peak demand charge optimization, ” IEEE Transactions on Automatic Control , 2019

    work page 2019

  34. [34]

    Bounds for error in the solution set of a perturbed linear program,

    S. M. Robinson, “Bounds for error in the solution set of a perturbed linear program,” Linear Algebra and its applications , vol. 6, pp. 69–81, 1973

    work page 1973

  35. [35]

    On approximate solutions of systems of l inear inequali- ties,

    A. J. Hoffman, “On approximate solutions of systems of l inear inequali- ties,” in Selected Papers Of Alan J Hoffman: With Commentary . World Scientific, 2003, pp. 174–176

    work page 2003

  36. [36]

    Schrijver, Theory of linear and integer programming

    A. Schrijver, Theory of linear and integer programming . John Wiley & Sons, 1998

    work page 1998

  37. [37]

    Bertsimas and J

    D. Bertsimas and J. N. Tsitsiklis, Introduction to linear optimization . Athena Scientific Belmont, MA, 1997, vol. 6

    work page 1997

  38. [38]

    R. A. Horn and C. R. Johnson, Matrix analysis. Cambridge university press, 2012

    work page 2012

  39. [39]

    Borrelli, A

    F. Borrelli, A. Bemporad, and M. Morari, Predictive control for linear and hybrid systems . Cambridge University Press, 2017

    work page 2017

  40. [40]

    Condit ions under which suboptimal nonlinear mpc is inherently robust,

    G. Pannocchia, J. B. Rawlings, and S. J. Wright, “Condit ions under which suboptimal nonlinear mpc is inherently robust,” Systems & Control Letters, vol. 60, no. 9, pp. 747–755, 2011

    work page 2011

  41. [41]

    On the inherent robustness of optimal and suboptimal nonlinear mp c,

    D. A. Allan, C. N. Bates, M. J. Risbeck, and J. B. Rawlings , “On the inherent robustness of optimal and suboptimal nonlinear mp c,” Systems & Control Letters , vol. 106, pp. 68–78, 2017

    work page 2017

  42. [42]

    Gurobi optimizer reference m anual,

    L. Gurobi Optimization, “Gurobi optimizer reference m anual,” 2019. [Online]. Available: http://www.gurobi.com

    work page 2019

  43. [43]

    Jump: A modeling language for mathematical optimization,

    I. Dunning, J. Huchette, and M. Lubin, “Jump: A modeling language for mathematical optimization,” SIAM Review, vol. 59, no. 2, pp. 295–320, 2017. Sungho Shin received the B.S. degree in chemical engineering and mathe- matics from Seoul National University, Seoul, South Korea, in 2016. He is currently working toward the Ph.D. degree with the Departme nt of...

    work page 2017