Pith. sign in

REVIEW 5 major objections 8 minor 56 references

QI-MPC: A Hybrid Quantum-Inspired Model Predictive Control for Learning Optimal Policies

T0 review · 5 major / 8 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read QI-MPC claims that a variational-quantum-circuit controller can in theory beat classical MPC on smooth high-dimensional control problems, while failing globally on nonlinear oscillatory systems.

desk verdict New VQC-in-the-loop MPC recipe with honest experimental reporting, but the regret and stability claims do not hold up under scrutiny. read the letter →

arxiv 2504.13041 v1 pith:UOCQIWGZ submitted 2025-04-17 quant-ph math.OC

Muhammad Al-Zafar Khan , Jamal Al-Karaki This is my paper · ORCID
classification quant-phmath.OC
keywords modelpredictivecontrolvariationalquantumcircuitsquantum-inspiredhybridquantum-classicalregretboundsoscillatorysystemsnonlinearstabilityguarantees
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

QI-MPC wraps a variational quantum circuit inside the standard model-predictive-control loop: the circuit encodes the current state, a parameterized ansatz produces candidate control actions, and only the first action is applied before the horizon shifts. The paper reports five experiments — target tracking, building climate control, autonomous vehicle dynamics, a simple pendulum, and a compound pendulum — and finds the approach works cleanly for the smooth, low-frequency cases and poorly for the oscillatory ones. From those observations it derives two results: hybrid quantum-classical control cannot achieve globally stable MPC for nonlinear oscillatory systems with finite-dimensional quantum ansätze, and for systems with high-dimensional state spaces and sufficiently smooth dynamics, QI-MPC can in theory attain a lower regret bound than classical MPC. A sympathetic reader would take the paper's aim to be identifying, with experiments and supporting arguments, the class of control problems where a VQC-based MPC loop is viable and the class where it is not.

What carries the argument

The central object is Algorithm 1's hybrid loop: encode the current state into a quantum state $|\psi(x_k)\rangle$, apply the parameterized ansatz $U(\theta)$ built from single-qubit rotations $\mathrm{Rot}(\theta)=R_Z(\theta_1)R_Y(\theta_2)R_X(\theta_3)$ and CNOT entanglers, measure Pauli-Z expectations to obtain controls $u_k$, clip to the admissible range, evolve the system dynamics $x_{k+1}=f(x_k,\tilde{u}_k)$, compute the loss, and update $\theta$ with a gradient-based rule under a receding horizon. The two theorems are carried by four lemmas for the oscillatory negative result (Nyquist feedback-delay violation, energy-scale mismatch, non-contractible reachable sets in phase space, and decoherence timescales) and three lemmas for the positive result (quantum parallelism in cost evaluation, exponential state compression, and quantum gradient regret). Proposition 2 also states explicit conditions: polynomial qubit scaling, polylogarithmic circuit depth in the horizon, smooth dynamics, and no ultra-fast feedback requirement.

What would settle it

Take a smooth high-dimensional linear-quadratic tracking problem, run Algorithm 1 with a fixed VQC ansatz over many steps, and compare cumulative regret against a standard interior-point MPC solver; if the classical solver's regret grows at or below $O(\sqrt{T})$ while the quantum loop's regret does not, Proposition 2's advantage claim is contradicted.

Watch

Extended reading notes

Core claim

The paper's central claim is that wrapping a variational quantum circuit around the standard receding-horizon MPC loop is a viable way to learn control policies, and that its viability has a sharp boundary. On one side, Proposition 1 asserts that for nonlinear oscillatory systems, any hybrid quantum-classical control method using a finite-dimensional quantum ansatz fails to achieve globally stable MPC because the classical-optimizer feedback delay violates the Nyquist rate for the oscillation, qubit-level energies are orders of magnitude below mechanical energies, a finite-dimensional Hilbert space cannot topologically cover the non-contractible reachable sets such as tori that oscillatory phase spaces contain, and decoherence acts before mechanical damping. On the other side, Proposition 2 asserts that for high-dimensional state spaces with sufficiently smooth dynamics, QI-MPC with polynomially many qubits can in theory achieve lower regret than classical MPC, relying on quantum parallelism for $O(1/\epsilon)$ cost evaluation, exponential state compression, and a claimed $O(\sqrt{T})$ quantum-gradient regret. The empirical section shows clean success only in the smooth low-frequency cases, with marginal results for the vehicle model and clear divergence for the two pendula.

Load-bearing premise

The positive advantage claim rests on an unproven comparison: quantum gradient updates are assumed to accumulate regret like the square root of the time horizon while classical MPC accumulates regret linearly, and the paper supplies no circuit construction and no online-learning proof for that comparison, so if it fails, Proposition 2 has no basis.

Editorial extensions

If this is right

  • For classical oscillatory control tasks, such as pendula or vibrating mechanical systems, classical MPC remains the recommended tool because QI-MPC cannot be globally stable under the paper's assumptions.
  • For nanoscale or quantum-mechanical systems, the timescale and energy scales match quantum hardware capabilities, so the hybrid loop is identified as a plausible control approach.
  • For smooth, high-dimensional, low-frequency control problems, QI-MPC can in theory achieve a lower regret bound than classical MPC while requiring exponentially less memory to discretize the state space.
  • The three safety guarantees imply that, under Lipschitz dynamics and positive-definite stage costs, the closed loop is stable, constraints are satisfied by clipping, and measurement noise can be exponentially suppressed by additional circuit readouts.
  • The experiments indicate the approach performs best on target-tracking and building climate control, marginally on autonomous vehicle dynamics, and poorly on the two pendulum systems.

Reading between the lines

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

  • The negative result for oscillatory systems may reflect the finite ansatz and measurement noise rather than a fundamental quantum limit; continuous-variable quantum control or adaptive ansätze could evade the topological obstruction, as the paper itself leaves open.
  • The claimed $O(\sqrt{T})$ versus $O(T)$ regret comparison is not the standard regret baseline for convex MPC, where strongly convex costs often yield logarithmic regret; replacing the baseline with standard bounds would likely shrink or eliminate the stated quantum advantage.
  • If Lemmas 5 through 7 were made constructive with explicit circuits for the MPC cost evaluation, QI-MPC could be benchmarked directly against an interior-point MPC solver on a smooth high-dimensional problem, giving a practical test of the regret claim.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 8 minor

Summary. This paper proposes QI-MPC, a hybrid quantum-classical framework in which a Variational Quantum Circuit (VQC) is trained online to generate control inputs that are clipped and applied to a dynamical system under a receding-horizon update. The approach is illustrated on five systems (target tracking, building climate control, autonomous vehicle dynamics, a simple pendulum, and a compound pendulum), and three 'safety guarantees' are stated in Sec. 3.2. On the basis of the five experiments, the paper then derives two propositions: (P1) hybrid quantum-classical MPC cannot achieve globally stable control for nonlinear oscillatory systems with finite-dimensional quantum ansatze, and (P2) QI-MPC can, in theory, achieve a lower regret bound than classical MPC for high-dimensional systems with sufficiently smooth dynamics. The positive claim rests on Lemmas 5-7 in Sec. 5.2 (quantum parallelism, state compression, and regret), and the negative claim rests on Lemmas 1-4 in Sec. 5.1 (sampling rates, energy scales, topology, and decoherence).

Significance. If the two propositions were established, the paper would provide a rare theoretical characterization of when hybrid quantum-classical control can beat classical MPC, and its identification of favorable versus unfavorable application domains (smooth low-bandwidth systems versus oscillatory, chaotic, and ultrafast systems) would interest both the quantum-control and MPC communities. The paper deserves credit for transparency: Experiments 4 and 5 are explicitly reported as suboptimal, the authors list possible reasons for the vacuous temperature trajectories in Experiment 2, Algorithm 1 is given in full pseudocode, and the Conclusion candidly states that only experimentation can determine where the hybrid approach wins or loses. However, the theoretical content consists of lemmas asserted without proof or construction, the stability guarantee is delegated to an unnamed theorem, and the experimental section contains no classical MPC baseline; the significance of the claims therefore cannot be assessed from the present manuscript.

major comments (5)
  1. [Sec. 3.1, Algorithm 1; Eq. (1)] Algorithm 1 does not implement the MPC problem stated in Eq. (1): the update at step 11 minimizes the one-step loss ||x_{k+1} - x_target||^2 + lambda||u_k||^2, not the finite-horizon cost J(x_t, U), and no N-step prediction or open-loop solution u* = arg min_U J is ever formed. The quantities 'prediction horizon N' and 'receding-horizon update' play no role in the optimization, so the procedure is better described as an online-trained feedback policy than as an MPC controller; this mismatch propagates into the safety guarantees and the regret claims, which presuppose that the algorithm solves Eq. (3).
  2. [Sec. 3.2, Formal Verification II] The stability guarantee is asserted by invoking 'the stability theorem' without stating or proving it, and no argument shows that the VQC parameter update decreases a control Lyapunov function. Standard MPC stability results require conditions (terminal cost or terminal constraint sets, and exact solution of the open-loop problem) that Algorithm 1 neither assumes nor satisfies; positive-definite stage cost and Lipschitz dynamics do not suffice even in the classical setting. The claimed convergence lim_{k to infinity} x_k = x_target is also contradicted by the paper's own Experiments 4 and 5, which fail to converge under this algorithm.
  3. [Sec. 5.1, Lemmas 1-4] The four lemmas underpinning Proposition 1 do not support it. Lemma 1 confuses clock speeds with the control sampling rate and misapplies the Nyquist theorem to closed-loop feedback: the sampling rate is a design choice and the relevant constraint is feedback computation latency, not the processor clock; moreover the inequality direction is wrong for the stated numbers, since for a macroscopic oscillator with omega_sys of order 1-10^3 rad/s the required update rate f_control >= omega_sys/pi is orders of magnitude below the GHz clock rate, so the stated condition is satisfied trivially. Lemma 2 is a category error: the VQC output is a classical number that is clipped and mapped to a torque or force applied through classical actuators, so the single-excitation energy hbar*omega of a qubit imposes no bound on control authority. Lemma 3 states that the n-qubit pure-state space is homeomorphic to R^{2n}; the relevant space is the complex projective space CP^{2^n - 1}, and the control output <psi|Z_i|psi> is an element of R^m, not a state vector in the Hilbert space, so no topological obstruction to covering the physical reachable set is established. Lemma 4 is contradicted by the paper's own cited coherence time of up to one year for nuclear spins and in any case does not imply global instability of the closed loop. Proposition 1 therefore has no valid proof.
  4. [Sec. 5.2, Lemmas 5-7; Proposition 2] Proposition 2's proof rests on Lemmas 5-7, none of which is established. Lemma 5 asserts without construction that quantum amplitude estimation evaluates the MPC cost in O(1/epsilon) steps; no circuit computing J(x_k, U) over a superposition of control trajectories is given, and the stated classical baseline O(1/epsilon^2) is the Monte Carlo estimation rate, not the cost of evaluating the deterministic objective in Eq. (1). Lemma 6 invokes superposition as exponential state compression with no measurement or retrieval scheme, which is the well-known gap between Hilbert-space dimension and algorithmic speedup. Lemma 7 asserts that classical MPC has regret O(T) on convex costs while quantum gradient updates achieve O(sqrt(T)); the standard online-convex-optimization result for convex Lipschitz costs is O(sqrt(T)) (and O(log T) for strongly convex costs), so the asserted classical baseline is not the standard one, and no regret benchmark for Algorithm 1 is even defined since the algorithm minimizes a one-step loss and no cumulative-regret analysis is supplied. With the correct baseline, the strict advantage claimed in Proposition 2 disappears.
  5. [Sec. 4 (Experiments); Sec. 6 (Conclusion)] The experiments do not provide the empirical support claimed for the propositions. No classical MPC baseline is run in any of the five experiments, so the section title 'Outperformance of Classical MPC by QI-MPC' (Sec. 5.2) is unsupported numerically. The two experiments invoked as evidence for Proposition 2 are unpersuasive: Experiment 1 (Sec. 4.1) uses the trivially stabilizable dynamics x(t+1) = x(t) + alpha(u(t) - x(t)) and yet reports a non-vanishing final loss of about 0.75, and Experiment 2 (Sec. 4.2) reports all temperature trajectories constant at 20 degrees C although the declared set point is 22 degrees C, a non-result that the paper itself attributes to possibly over-simplified dynamics. The validation is also circular: Propositions 1 and 2 are motivated by the outcomes of Experiments 4-5 and 1-2 (Secs. 5.1-5.2), and the Conclusion then invokes those propositions to explain the experimental outcomes, which provides neither independent evidence for the propositions nor a genuine test of them.
minor comments (8)
  1. [Eq. (1), Sec. 1] The minimization in Eq. (1) is written as min over u_k in U, but the decision variable is the whole sequence U = (u_0, ..., u_{N-1}); the text 'x_N contained in X_f contained in X' also states set inclusions for what is a single terminal state.
  2. [Sec. 1] The phrase 'the horizon is sifted forward' should read 'shifted forward'.
  3. [Sec. 5.1, Lemma 2] The energy hbar*f_qubit ~ 10^-24 J is described as 'Planck scale energies', which is incorrect; the Planck energy is of order 10^9 J.
  4. [Sec. 4.2, Fig. 7] Fig. 7 shows all temperature trajectories constant at 20 degrees C while the text declares the set point to be 22 degrees C; the paper should reconcile the set point with the observed trajectories or correct the experimental description.
  5. [Sec. 3.2, Formal Verification III] Measurement outcomes are modeled as Gaussian noise, eta ~ N(0, sigma^2), and then bounded with Hoeffding's inequality, which requires bounded random variables; the relationship between the two noise models should be clarified.
  6. [Figs. 2-4, 6-8, 10-12, 15-17, 20-22] The shaded regions are described as confidence intervals, but the paper does not state how many independent runs or seeds they are computed from, so the uncertainty bands cannot be interpreted.
  7. [Title and Abstract] The title uses 'Quantum-Inspired' while the body describes 'hybrid quantum-classical' control and the experiments appear to be classical simulations of quantum circuits; the distinction between these terms should be clarified.
  8. [Sec. 5.1, Lemma 1] The derivation f_control >= 2 f_sys = 2 omega_sys/(2 pi) = omega_sys/pi mixes angular and cyclic frequencies; with f_sys = omega_sys/(2 pi) the bound is f_control >= omega_sys/pi in Hz, and the units should be stated consistently throughout the lemma.

Circularity Check

2 steps flagged · score 6.0 of 10

Proposition 2's regret advantage is assumed in Lemma 7, and the theory's scope is fit to the experiments it then explains.

  1. other [Section 5.2, Lemma 7 and Proposition 2 (proof bullet 3).]
    "Lemma 7. Hybrid MPC quantum gradient updates have a lower regret cost than classical MPC gradient updates. Proof. For a finite time horizon T <∞ and a convex cost function, classical MPC methods have a regret cost of O(T ). Quantum gradient update rules have a regret cost ofO(√ T )."

    The proof of Proposition 2 depends on Lemma 7 for the decisive comparative claim that quantum updates attain sublinear regret while classical MPC incurs linear regret. Lemma 7's entire proof is the assertion of exactly those two regret bounds, without derivation of either bound, without identification of the quantum update rule, and without a defined regret benchmark. The lemma is therefore the proposition's conclusion restated as a premise: Proposition 2 reduces to Lemma 7, and Lemma 7 reduces to the claim of Proposition 2. The central advantage is assumed rather than derived.

  2. fitted input called prediction [Abstract; Section 5.2 (p. 35); Section 6 (p. 36).]
    "the experiments gave the motivation for two important theoretical results that, in essence, identify systems for which the approach works best. ... We note that the type of systems that proposition 2 holds for are systems analogous to experiments 4.1 and 4.2. ... The suboptimal outcomes of the latter two experiments are because they are oscillating systems and therefore invalidate the conditions of proposition 2."

    The paper explicitly says the propositions came from the experimental observations. Proposition 2's domain is then identified with the successful experiments (4.1 and 4.2), while the failed oscillatory experiments (4.4 and 4.5) are excluded from the domain. In the conclusion, the same proposition is invoked to explain why those experiments succeeded or failed. The theory's applicability conditions are thus selected post hoc from the very outcomes they are used to explain; the 'identification' of systems where QI-MPC works best is a restatement of the experimental selection, not an independent prediction. This makes the explanatory loop self-supporting rather than a testable theoretical result.

full rationale

The paper contains no load-bearing self-citation chain: the self-citations [41]-[47] support only background ML/QC applications, and the safety guarantees in Sec. 3.2 are elementary consequences of clipping and Lyapunov-style reasoning, not circular. The main circularity is in Proposition 2. Lemma 7, which Proposition 2's proof uses for the decisive 'sublinear vs linear regret' claim, is not derived: its proof is simply the statement that classical MPC regret is O(T) and quantum gradient regret is O(sqrt T). Thus the claimed advantage is the lemma, and the lemma is the claim; the proof chain is Proposition 2 -> Lemma 7 -> assertion of Proposition 2. Separately, the paper states explicitly that the experiments motivated the two propositions, and Proposition 2's domain is identified with the successful experiments (4.1, 4.2) while oscillatory unsuccessful experiments are excluded; the conclusion then uses Proposition 2 to explain those same successes and failures. This is a post-hoc fit: the 'identification' of systems where QI-MPC works best restates the experimental selection rather than predicting new cases. These two issues make the central positive claim partially circular. Proposition 1 has independent (though debatable) lemmas on timescale, energy, topology, and decoherence, so it is not charged with circularity here; concerns about unsupported assumptions or nonstandard regret baselines are correctness issues, not circularity.

Assumptions & free parameters 7 free parameters · 7 assumptions · 0 invented entities

The central claims rest on hand-picked hyperparameters per experiment and on unproved lemmas in Section 5.2. No code or data is provided, so the experiments cannot be independently checked. The theoretical results are post hoc rationalizations of observed behavior, not derivations from first principles.

free parameters (7)
  • Building control penalty weights lambda1, lambda2 = lambda1=0.01, lambda2=0.005
    Chosen by hand in Experiment 2; no sensitivity analysis or selection rule.
  • Vehicle control penalty weights lambda1-lambda4 = lambda1=lambda2=0.1, lambda3=lambda4=0.01
    Chosen by hand in Experiment 3; no ablation.
  • Pendulum control penalty weight lambda = lambda=0.05
    Chosen by hand in Experiment 4.
  • Compound pendulum loss weights lambda1, lambda2, lambda3 = lambda1=1, lambda2=0.1, lambda3=0.01
    lambda3 was added post hoc in Experiment 5 because the loss repeatedly increased without it.
  • Experiment 4 optimizer hyperparameters = initial lr=0.3, min lr=0.01, decay=0.95, momentum=0.85, clip=[-0.5,0.5]
    Hand-picked to reduce oscillatory training loss; no justification.
  • Circuit sizes and ansatz depth = 10 qubits (Exp1), 6 qubits (Exp2, Exp3), 4 qubits (Exp4, Exp5)
    Chosen without a stated rule; number of variational layers is not specified.
  • State encoding rotation angles
    Custom encoding choices such as RY(pi*x_i), RX(pi*(x_i+1/2)), RZ(pi*x_i/2) in Exp1 and RY(theta) in Exp4; no comparison of encodings.
assumptions (7)
  • standard math Nyquist-Shannon sampling theorem
    Invoked in Lemma 1 to assert that control updates must be at least twice the system frequency.
  • standard math Brouwer invariance of domain
    Invoked in Lemma 3 to argue that a finite-dimensional Hilbert space cannot densely cover a torus.
  • standard math KAM theory: phase space of integrable Hamiltonian systems is foliated by invariant tori
    Remark 3 takes this result as true and uses it to infer that oscillatory systems contain tori in phase space.
  • ad hoc to paper Finite-dimensional Hilbert space is homeomorphic to R^(2n) and VQC outputs live in that space
    Lemma 3 relies on this to compare dimensions of controls and reachable sets, conflating classical measurement outputs with the quantum state space.
  • domain assumption Positive definite stage cost and Lipschitz dynamics imply closed-loop stability via control Lyapunov function and recursive feasibility
    Formal Verification II invokes this without proving that VQC parameter updates decrease the stage cost.
  • ad hoc to paper Quantum amplitude estimation evaluates the MPC cost in O(1/epsilon) steps and quantum gradient updates have O(sqrt T) regret
    Lemmas 5 and 7 assert these claims without a construction; the classical O(T) regret baseline is not generally correct for convex costs.
  • domain assumption Measurement noise is zero-mean Gaussian and Hoeffding's inequality applies to readouts
    Formal Verification III assumes normal noise and independent readouts without empirical support.

how reviews work

0 comments
Cite this review

Pith. "Pith review of QI-MPC: A Hybrid Quantum-Inspired Model Predictive Control for Learning Optimal Policies." pith.science (2026). https://pith.science/paper/UOCQIWGZ

@misc{pith2026250413041,
  author       = {Pith},
  title        = {Pith review of: QI-MPC: A Hybrid Quantum-Inspired Model Predictive Control for Learning Optimal Policies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UOCQIWGZ}},
  note         = {Machine review of arXiv:2504.13041}
}
read the original abstract

In this paper, we present Quantum-Inspired Model Predictive Control (QIMPC), an approach that uses Variational Quantum Circuits (VQCs) to learn control polices in MPC problems. The viability of the approach is tested in five experiments: A target-tracking control strategy, energy-efficient building climate control, autonomous vehicular dynamics, the simple pendulum, and the compound pendulum. Three safety guarantees were established for the approach, and the experiments gave the motivation for two important theoretical results that, in essence, identify systems for which the approach works best.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 54 canonical work pages

  1. [1]

    Automatica 29(5), 1251–1274 (1993)

    Richalet, J.: Industrial applications of model based predictive control. Automatica 29(5), 1251–1274 (1993)

    work page 1993

  2. [2]

    Control Engineering Practice 10(8), 819–831 (2002)

    Zanin, A.C., De Gouvea, M.T., Odloak, D.: Integrating real-time optimiza- tion into the model predictive controller of the fcc system. Control Engineering Practice 10(8), 819–831 (2002)

    work page 2002

  3. [3]

    Control engineering practice 11(7), 733–764 (2003)

    Qin, S.J., Badgwell, T.A.: A survey of industrial model predictive control technology. Control engineering practice 11(7), 733–764 (2003)

    work page 2003

  4. [4]

    Computers & Chemical Engineering 34(10), 1671–1686 (2010)

    Y¨ uzge¸ c, U., Palazoglu, A., Romagnoli, J.A.: Refinery scheduling of crude oil unloading, storage and processing using a model predictive control strategy. Computers & Chemical Engineering 34(10), 1671–1686 (2010)

    work page 2010

  5. [5]

    Industrial & Engineering Chemistry Research 53(45), 17696–17704 (2014)

    Sildir, H., Arkun, Y., Ari, A., Dogan, I., Harmankaya, M.: Economic model pre- dictive control of an industrial fluid catalytic cracker. Industrial & Engineering Chemistry Research 53(45), 17696–17704 (2014)

    work page 2014

  6. [6]

    Industrial & Engineering Chemistry Research 56(31), 8961–8971 (2017)

    Wei, M., Yang, M., Qian, F., Du, W., He, W., Zhong, W.: Dynamic modeling and economic model predictive control with production mode switching for an indus- trial catalytic naphtha reforming process. Industrial & Engineering Chemistry Research 56(31), 8961–8971 (2017)

    work page 2017

  7. [7]

    In: 2012 15th International IEEE 37 Conference on Intelligent Transportation Systems, pp

    Wang, M., Daamen, W., Hoogendoorn, S.P., Arem, B.: Driver assistance sys- tems modeling by model predictive control. In: 2012 15th International IEEE 37 Conference on Intelligent Transportation Systems, pp. 1543–1548 (2012). IEEE

    work page 2012

  8. [8]

    Vehicle System Dynamics 53(12), 1705–1720 (2015)

    Lef` evre, S., Carvalho, A., Gao, Y., Tseng, H.E., Borrelli, F.: Driver models for personalised driving assistance. Vehicle System Dynamics 53(12), 1705–1720 (2015)

    work page 2015

Show all 56 references
  1. [9]

    IEEE Intelligent transportation systems magazine 9(3), 23–35 (2017)

    Cesari, G., Schildbach, G., Carvalho, A., Borrelli, F.: Scenario model predictive control for lane change assistance and autonomous driving on highways. IEEE Intelligent transportation systems magazine 9(3), 23–35 (2017)

    2017

  2. [10]

    IEEE Transactions on Human-Machine Systems 47(5), 700–710 (2017)

    Ercan, Z., Carvalho, A., Gokasan, M., Borrelli, F.: Modeling, identification, and predictive control of a driver steering assistance system. IEEE Transactions on Human-Machine Systems 47(5), 700–710 (2017)

    2017

  3. [11]

    Robotics and Autonomous Systems 112, 291–303 (2019)

    Sajadi-Alamdari, S.A., Voos, H., Darouach, M.: Nonlinear model predictive con- trol for ecological driver assistance systems in electric vehicles. Robotics and Autonomous Systems 112, 291–303 (2019)

    2019

  4. [12]

    Energies 14(23), 7974 (2021)

    Musa, A., Pipicelli, M., Spano, M., Tufano, F., De Nola, F., Di Blasio, G., Gimelli, A., Misul, D.A., Toscano, G.: A review of model predictive controls applied to advanced driver-assistance systems. Energies 14(23), 7974 (2021)

    2021

  5. [13]

    IEEE transactions on control systems technology 15(3), 449–456 (2007)

    Ortner, P., Del Re, L.: Predictive control of a diesel engine air path. IEEE transactions on control systems technology 15(3), 449–456 (2007)

    2007

  6. [14]

    In: 2008 47th IEEE Conference on Decision and Control, pp

    Stewart, G., Borrelli, F.: A model predictive control framework for industrial tur- bodiesel engine control. In: 2008 47th IEEE Conference on Decision and Control, pp. 5704–5711 (2008). IEEE

    2008

  7. [15]

    Applied energy 169, 555–566 (2016)

    Pizzonia, F., Castiglione, T., Bova, S.: A robust model predictive control for efficient thermal management of internal combustion engines. Applied energy 169, 555–566 (2016)

    2016

  8. [16]

    In: 2012 IEEE International Conference on Control Applications, pp

    Hrovat, D., Di Cairano, S., Tseng, H.E., Kolmanovsky, I.V.: The development of model predictive control in automotive industry: A survey. In: 2012 IEEE International Conference on Control Applications, pp. 295–302 (2012). IEEE

    2012

  9. [17]

    IEEE Transactions on industrial electronics 56(6), 1826–1838 (2008)

    Kouro, S., Cort´ es, P., Vargas, R., Ammann, U., Rodr´ ıguez, J.: Model predic- tive control—a simple and powerful method to control power converters. IEEE Transactions on industrial electronics 56(6), 1826–1838 (2008)

    2008

  10. [18]

    IEEE Transactions on Power Systems 25(2), 988–998 (2009)

    Jin, L., Kumar, R., Elia, N.: Model predictive control-based real-time power sys- tem protection schemes. IEEE Transactions on Power Systems 25(2), 988–998 (2009)

    2009

  11. [19]

    IEEE Transactions on Power Systems 32(2), 1651–1662 (2016) 38

    Martin, J.A., Hiskens, I.A.: Corrective model-predictive control in large electric power systems. IEEE Transactions on Power Systems 32(2), 1651–1662 (2016) 38

    2016

  12. [20]

    Energy 172, 555–565 (2019)

    Liu, J., Yao, Q., Hu, Y.: Model predictive control for load frequency of hybrid power system with wind power and thermal power. Energy 172, 555–565 (2019)

    2019

  13. [21]

    Journal of guidance, control, and dynamics 29(2), 427–434 (2006)

    Van Soest, W., Chu, Q., Mulder, J.: Combined feedback linearization and con- strained model predictive control for entry flight. Journal of guidance, control, and dynamics 29(2), 427–434 (2006)

    2006

  14. [22]

    Journal of Guidance, Control, and Dynamics 40(7), 1541–1566 (2017)

    Eren, U., Prach, A., Ko¸ cer, B.B., Rakovi´ c, S.V., Kayacan, E., A¸ cıkme¸ se, B.: Model predictive control in aerospace systems: Current state and opportunities. Journal of Guidance, Control, and Dynamics 40(7), 1541–1566 (2017)

    2017

  15. [23]

    In: 2018 Annual American Control Conference (ACC), pp

    Di Cairano, S., Kolmanovsky, I.V.: Real-time optimization and model predictive control for aerospace and automotive applications. In: 2018 Annual American Control Conference (ACC), pp. 2392–2409 (2018). IEEE

    2018

  16. [24]

    IEEE Transactions on Control Systems Technology 25(4), 1505–1511 (2016)

    Faulwasser, T., Weber, T., Zometa, P., Findeisen, R.: Implementation of non- linear model predictive path-following control for an industrial robot. IEEE Transactions on Control Systems Technology 25(4), 1505–1511 (2016)

    2016

  17. [25]

    IEEE Robotics and Automation Letters 4(4), 3758–3765 (2019)

    Carron, A., Arcari, E., Wermelinger, M., Hewing, L., Hutter, M., Zeilinger, M.N.: Data-driven model predictive control for trajectory tracking with a robotic arm. IEEE Robotics and Automation Letters 4(4), 3758–3765 (2019)

    2019

  18. [26]

    IFAC-PapersOnLine 53(2), 9891–9898 (2020)

    Gold, T., V¨ olz, A., Graichen, K.: Model predictive interaction control for industrial robots. IFAC-PapersOnLine 53(2), 9891–9898 (2020)

    2020

  19. [27]

    Energy and buildings 45, 15–27 (2012)

    Oldewurtel, F., Parisio, A., Jones, C.N., Gyalistras, D., Gwerder, M., Stauch, V., Lehmann, B., Morari, M.: Use of model predictive control and weather fore- casts for energy efficient building climate control. Energy and buildings 45, 15–27 (2012)

    2012

  20. [28]

    IEEE Transactions on Control Systems Technology 22(3), 1198–1205 (2013)

    Oldewurtel, F., Jones, C.N., Parisio, A., Morari, M.: Stochastic model predic- tive control for building climate control. IEEE Transactions on Control Systems Technology 22(3), 1198–1205 (2013)

    2013

  21. [29]

    Building and Environment 160, 106204 (2019)

    Wang, J., Li, S., Chen, H., Yuan, Y., Huang, Y.: Data-driven model predictive control for building climate control: Three case studies on different buildings. Building and Environment 160, 106204 (2019)

    2019

  22. [30]

    Computer Methods and Programs in Biomedicine 116(3), 193–204 (2014)

    Sopasakis, P., Patrinos, P., Sarimveis, H.: Robust model predictive control for optimal continuous drug administration. Computer Methods and Programs in Biomedicine 116(3), 193–204 (2014)

    2014

  23. [31]

    Computer Methods and Programs in Biomedicine 144, 13–19 (2017) 39

    Sharifi, N., Ozgoli, S., Ramezani, A.: Multiple model predictive control for opti- mal drug administration of mixed immunotherapy and chemotherapy of tumours. Computer Methods and Programs in Biomedicine 144, 13–19 (2017) 39

    2017

  24. [32]

    IEEE Transactions on NanoBioscience 16(3), 149–156 (2017)

    Moussa, H.G., Husseini, G.A., Abel-Jabbar, N., Ahmad, S.E.: Use of model pre- dictive control and artificial neural networks to optimize the ultrasonic release of a model drug from liposomes. IEEE Transactions on NanoBioscience 16(3), 149–156 (2017)

    2017

  25. [33]

    Automatica 47(8), 1638–1645 (2011)

    Kouramas, K.I., Fa´ ısca, N.P., Panos, C., Pistikopoulos, E.N.: Explicit/multi- parametric model predictive control (mpc) of linear discrete-time systems by dynamic and multi-parametric programming. Automatica 47(8), 1638–1645 (2011)

    2011

  26. [34]

    Automatica 61, 27–34 (2015)

    Shekhar, R.C., Manzie, C.: Optimal move blocking strategies for model predictive control. Automatica 61, 27–34 (2015)

    2015

  27. [35]

    In: 1999 European Control Conference (ECC), pp

    Donkelaar, E.T., Bosgra, O.H., Hof, P.M.J.: Model predictive control with gener- alized input parametrization. In: 1999 European Control Conference (ECC), pp. 1693–1698 (1999)

    1999

  28. [36]

    part i: Theory

    Wu, Z., Tran, A., Rincon, D., Christofides, P.D.: Machine learning-based predic- tive control of nonlinear processes. part i: Theory. AIChE Journal 65(11), 16729 (2019)

    2019

  29. [37]

    IEEE Access 12, 23265–23278 (2024)

    Hassan, A., Ruiz-Moreno, S., Frejo, J.R.D., Maestre, J.M., Camacho, E.F.: Neural-network based mpc for enhanced lateral stability in electric vehicles. IEEE Access 12, 23265–23278 (2024)

    2024

  30. [38]

    IEEE Control Systems Magazine 36(6), 30–44 (2016)

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

    2016

  31. [39]

    Journal of Process Control 22(3), 540–550 (2012)

    Kirches, C., Wirsching, L., Bock, H.G., Schl¨ oder, J.P.: Efficient direct multi- ple shooting for nonlinear model predictive control on long horizons. Journal of Process Control 22(3), 540–550 (2012)

    2012

  32. [40]

    IFAC-PapersOnLine 52(23), 73–80 (2019)

    Baayen, J., Becker, B., Heeringen, K.-J., Miltenburg, I., Piovesan, T., Rauw, J., Toom, M., VanderWees, J.: An overview of continuation methods for non-linear model predictive control of water systems. IFAC-PapersOnLine 52(23), 73–80 (2019). 1st IFAC Workshop on Control Method...

    2019

  33. [41]

    Quantum Information Processing 22(10), 374 (2023)

    Innan, N., Khan, M.A.-Z., Panda, B., Bennai, M.: Enhancing quantum sup- port vector machines through variational kernel training. Quantum Information Processing 22(10), 374 (2023)

    2023

  34. [42]

    arXiv preprint arXiv:2304.10786 (2023)

    Innan, N., Khan, M.A.-Z.: Classical-to-quantum sequence encoding in genomics. arXiv preprint arXiv:2304.10786 (2023)

    2023 arXiv

  35. [43]

    40 Materials Today Communications 38, 107760 (2024)

    Innan, N., Khan, M.A.-Z., Bennai, M.: Quantum computing for electronic structure analysis: Ground state energy and molecular properties calculations. 40 Materials Today Communications 38, 107760 (2024)

    2024

  36. [44]

    arXiv preprint arXiv:2401.15804 (2024)

    Khan, M.A.-Z., Innan, N., Galib, A.A.O., Bennai, M.: Brain tumor diagnosis using quantum convolutional neural networks. arXiv preprint arXiv:2401.15804 (2024)

    2024 arXiv

  37. [45]

    arXiv preprint arXiv:2411.18141 (2024)

    Khan, M.A.-Z., Al-Karaki, J., Omar, M.: Predicting water quality using quantum machine learning: The case of the umgeni catchment (u20a) study region. arXiv preprint arXiv:2411.18141 (2024)

    2024 arXiv

  38. [46]

    International Journal of Quantum Information 22(02), 2350044 (2024)

    Innan, N., Khan, M.A.-Z., Bennai, M.: Financial fraud detection: a comparative study of quantum machine learning models. International Journal of Quantum Information 22(02), 2350044 (2024)

    2024

  39. [47]

    Quantum Machine Intelligence 6(1), 28 (2024)

    Innan, N., Siddiqui, O.I., Arora, S., Ghosh, T., Ko¸ cak, Y.P., Paragas, D., Galib, A.A.O., Khan, M.A.-Z., Bennai, M.: Quantum state tomography using quantum machine learning. Quantum Machine Intelligence 6(1), 28 (2024)

    2024

  40. [48]

    Acta Astro- nautica 205, 68–79 (2023)

    De Grossi, F., Circi, C.: Nonlinear model predictive control leveraging quantum- inspired optimization in the three body problem with uncertainty. Acta Astro- nautica 205, 68–79 (2023)

    2023

  41. [49]

    Scientific Reports 10(1), 1591 (2020)

    Inoue, D., Yoshida, H.: Model predictive control for finite input systems using the d-wave quantum annealer. Scientific Reports 10(1), 1591 (2020)

    2020

  42. [50]

    Quantum 6, 837 (2022)

    Goldschmidt, A.J., DuBois, J.L., Brunton, S.L., Kutz, J.N.: Model predictive control for robust quantum state preparation. Quantum 6, 837 (2022)

    2022

  43. [51]

    IEEE Transactions on Quantum Engineering 3, 1–23 (2022)

    Clouˆ atr´ e, M., Khojasteh, M.J., Win, M.Z.: Model-predictive quantum control via hamiltonian learning. IEEE Transactions on Quantum Engineering 3, 1–23 (2022)

    2022

  44. [52]

    arXiv preprint arXiv:2410.19467 (2024)

    Novara, C., Boggio, M., Volpe, D.: Quantum optimization for nonlinear model predictive control. arXiv preprint arXiv:2410.19467 (2024)

    2024 arXiv

  45. [53]

    In: 2015 IEEE Intelligent Vehicles Symposium (IV), pp

    Kong, J., Pfeiffer, M., Schildbach, G., Borrelli, F.: Kinematic and dynamic vehicle models for autonomous driving control design. In: 2015 IEEE Intelligent Vehicles Symposium (IV), pp. 1094–1099 (2015). IEEE

    2015

  46. [54]

    IEEE Transactions on Intelligent Transportation Systems 21(1), 48–58 (2019)

    Xu, S., Peng, H.: Design, analysis, and experiments of preview path tracking control for autonomous vehicles. IEEE Transactions on Intelligent Transportation Systems 21(1), 48–58 (2019)

    2019

  47. [55]

    TMH (1991)

    Taub, H., Schilling, D.L.: Principles of communication systems. TMH (1991)

    1991

  48. [56]

    Advances in Computers, vol

    Black, P.E., Kuhn, D.R., Williams, C.J.: Quantum computing and communica- tion. Advances in Computers, vol. 56, pp. 189–244. Elsevier (2002) 41

    2002

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.