{"id":"776b5d59-c501-4787-81d6-dda039c7b67f","arxiv_id":"2412.08104","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a class of nonlinear offset-free MPC designs with quadratic costs and backoff constraints, the authors prove robust asymptotic stability and offset-free tracking despite plant-model mismatch and persistent disturbances.","lead":"This paper proves, under stated assumptions, that a nonlinear offset-free model predictive controller can stay stable and eliminate tracking offset even when the model and plant differ. It is the first general stability guarantee for this widely used control method, which matters for chemical and process control applications.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 6's main hypothesis (Assumption 6, a global quadratic ISS Lyapunov function for the estimator) is not verified for the MHE estimators used in the paper's own examples, leaving the central nonlinear stability claim without a demonstrated instantiation.","rationale":"The reader's weakest-assumption analysis and the stress-test pass converge on the same issue: Assumption 6 is the single most load-bearing hypothesis of the paper's main theorem. The paper is honest about the gap, explicitly stating in Section 8.1 that the MHE estimators used in the examples are not known to satisfy Assumption 6. This is not a hidden flaw but an acknowledged limitation; however, for the central claim, the limitation is decisive. A theorem whose key hypothesis (a global quadratic ISS Lyapunov function for the estimator) is unverified for the demonstrated estimators and unsupported by existing nonlinear estimation theory cannot be presented as establishing the claimed result for nonlinear offset-free MPC. The proof itself appears internally coherent, and the conditional statement 'if Assumption 6 holds, then Theorem 6 holds' is valuable, which is why a verdict of CONDITIONAL is appropriate. No internal inconsistency or more fundamental mathematical error was identified in the proofs of Proposition 7 or the application of Theorem 3. The concern is about the gap between the theorem's hypotheses and the demonstrated cases, not about the logical derivation. Therefore the reader's conditional verdict should stand unchanged. A concrete check — testing whether any practical estimator (including the paper's MHE) satisfies Assumption 6 on the paper's examples — would settle whether the concern is merely a gap in the writeup or a fundamental limitation of the approach.","tokens_in":46249,"tokens_out":9573,"duration_ms":93418,"concrete_test":"Select the MHE estimator (48) used in Section 8.1. First, linearize the pendulum model (52) about the upright setpoint and design a full-order Luenberger observer. Verify whether this observer satisfies (19a)–(19b) for the linearized error dynamics with some positive constants; if not, Assumption 6 already fails for the simplest nontrivial case. Then attempt to extend the same quadratic Lyapunov function (with the same constants) to the full nonlinear error dynamics of (52) on the compact domain used in the simulations, checking (19b) numerically on a dense grid of initial errors and admissible noises. If the inequalities fail at any grid point, Assumption 6 is not satisfied by the demonstrated estimator and Theorem 6 cannot be invoked for the paper's examples.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim, Theorem 6, asserts robust exponential stability of the offset-free MPC closed loop under plant-model mismatch, with Assumptions 1–9. The most load-bearing hypothesis is Assumption 6, which requires the state–disturbance estimator to admit a global quadratic ISS Lyapunov function satisfying (19a)–(19b). This is exceptionally strong: it demands a one-step quadratic Lyapunov inequality for all trajectories of the noisy model (17), with constants c1–c4 uniform over the entire domain. The paper's own Section 8.1 states that the moving-horizon estimators used in the numerical examples 'should be RGES' but 'it is not known if they satisfy Assumption 6' — so the demonstrated estimators do not provably instantiate the theorem. Moreover, the existing stability results for MHE cited by the authors (Allan and Rawlings 2021; Schiller et al. 2023) provide Q-functions or N-step Lyapunov functions, not the one-step quadratic Lyapunov function required here; no conversion to Assumption 6 is given. If no practical nonlinear estimator satisfies Assumption 6, then Theorem 6, while logically valid as a conditional statement, has no nonempty domain of application in the nonlinear setting claimed by the abstract. This is not a matter of consensus but of the theorem's applicability: the central claim of 'first general stability results for nonlinear offset-free MPC' is unsupported unless an estimator is actually constructed or identified that satisfies Assumption 6.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops an offset-free model predictive control architecture for nonlinear discrete-time systems with plant-model mismatch, consisting of a steady-state target problem, a finite-horizon regulator with terminal ingredients constructed by linearization, and a joint state-and-disturbance estimator. It first proves nominal exponential stability with respect to target and setpoint tracking errors (Theorem 4), then robust stability with respect to estimate errors and setpoint/disturbance changes in the absence of mismatch (Theorem 5), and finally robust exponential stability of the joint controller-estimator loop for sufficiently small plant-model mismatch (Theorem 6) under Assumptions 1 to 9. The proof framework introduces an ISS/Lyapunov theory with respect to two measurement functions (Theorems 2 and 3) and is applied to the offset-free MPC loop. Numerical experiments on a pendulum and a CSTR compare the proposed offset-free MPC with a tracking MPC.","tokens_in":46619,"tokens_out":12583,"duration_ms":128907,"significance":"If the results are taken together with a constructive estimator satisfying Assumption 6, this would be a substantial contribution, filling a long-standing gap in the nonlinear offset-free MPC literature. The proof structure is careful and detailed, with complete appendices, and Section 7 provides verifiable local conditions for the steady-state target problem through rank conditions on linearizations. The main weakness is that the paper's key estimator assumption, Assumption 6, is not instantiated by any estimator used in the examples, so the central theorem currently lacks a demonstrated domain of application in the nonlinear setting claimed by the abstract.","major_comments":[{"comment":"Assumption 6, stated in Section 2.2.3 as requiring a global quadratic ISS Lyapunov function satisfying (19a)-(19b), is the most restrictive hypothesis of the central result. The paper itself states in Section 8.1 that the MHE estimators used in the numerical examples 'should be RGES' but 'it is not known if they satisfy Assumption 6'. Consequently, none of the simulations instantiates Theorem 6, and the theorem's domain of application is not demonstrated. The cited stability results for MHE (Allan and Rawlings 2021; Schiller et al. 2023) provide a Q-function and an N-step Lyapunov function, respectively, not the one-step global quadratic Lyapunov function required here, and no conversion argument is supplied. Because the abstract claims 'a nonlinear offset-free MPC design that is robustly stable', the authors need either to construct estimators that provably satisfy Assumption 6 and use them in the examples, or to relax Assumption 6 to a Q-function/N-step Lyapunov condition and adapt Theorems 3 and 6 accordingly. As written, the central nonlinear stability claim is conditional on an assumption that is not shown to be satisfiable by any practical estimator.","section":"§8.1, Assumption 6, Theorem 6"},{"comment":"The abstract and conclusions present the result as an unconditional 'first general stability results' statement, but Theorem 6 is a conditional statement relying on Assumption 6, which the paper's own Section 8.1 admits is unverified for the MHE estimators actually used. Section 9 also lists 'the requirement of a Lyapunov function for the estimator (Assumption 6)' as an open direction for future work. The claims should be rephrased to state explicitly that the stability guarantee holds for estimators satisfying Assumption 6, and that no such estimator is constructed or validated in the paper. This is not merely a wording issue: it determines whether the paper delivers a design or only a conditional theorem.","section":"Abstract, Section 9"}],"minor_comments":[{"comment":"The default simulation parameters in Section 8.1 appear to swap the roles of (wP)4 and (wP)5: the text says 'discretization parameter (wP)4 = 1' and 'no measurement noise (wP)5 = 0', but in equations (50)-(51) (wP)4 is the measurement offset and (wP)5 scales the discretization error; as written, (wP)4 = 1 violates the stated bound (wP)4 ∈ [-0.05, 0.05] and contradicts the claim of no measurement noise.","section":"§8.1"},{"comment":"In the third pendulum experiment, the text says 'we have measurement noise (wP)5 ∼ N(0,10^-4)'; given (50b), the measurement noise is (wP)4, not (wP)5. The same subsection also writes 'where (wP)3 ∼ N(0,10^-2)' for the integrating disturbance, which should presumably be the increment (ΔwP)3.","section":"§8.1"},{"comment":"The phrase 'Runga-Kutta' should be 'Runge-Kutta' in both example sections.","section":"§8.1 and §8.2"},{"comment":"The sentence 'We also consider we softened regulator output constraints' contains an extra 'we' and should read 'We also consider softened regulator output constraints'.","section":"§1"},{"comment":"The statement 'RES w.r.t. (δr, δx̂)' is ambiguous because Definition 7 bounds |(ζ1, ε)| rather than |(ζ1, ζ2)|, and the proof of Theorem 6(c) indeed bounds |(δr, e)|. The authors should state explicitly that the conclusion includes the estimator error e, consistent with Definition 7.","section":"Theorem 6(c), Definition 7"},{"comment":"The CSTR terminal region is reported as cf ≈ 6.5 × 10^-16, which is close to floating-point precision. Since the simulations are intended to illustrate the theory, the authors should report how the terminal constraint Xf is verified numerically and whether the optimizer reliably returns states inside this extremely small terminal region.","section":"§8.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a serious contribution to nonlinear offset-free MPC, with a coherent Lyapunov/ISS proof structure. The decisive issue is Assumption 6: the paper's own examples do not instantiate it, and no estimator is constructed that satisfies it. If the authors can add a section that constructs or identifies estimators satisfying Assumption 6, or extend the theory to the N-step Lyapunov functions available for MHE, the paper could become acceptable. As it stands, the central theorem is conditional on an uninstantiated hypothesis, so I cannot recommend acceptance in the current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper is a serious theory contribution: it gives the first general stability theorem for nonlinear offset-free MPC that does not start by assuming closed-loop stability. The core construction—parameter-varying terminal region, constraint backoffs, and a two-output ISS Lyapunov framework—is coherent, and the proofs in the appendix are detailed and checkable. The linearization lemmas in Section 7 are a useful bridge to existing linear offset-free MPC conditions. I believe the main conditional result, Theorem 6, is sound as a theorem.\n\nThe soft spot is exactly where the stress-test points: Assumption 6. Requiring a global one-step quadratic Lyapunov function for the estimator is very strong, and the paper's own Section 8.1 admits that the MHE estimators used in the numerical examples are not known to satisfy it. That is not a minor caveat—it means the paper offers no demonstrated instantiation of Theorem 6 in the nonlinear setting. The cited MHE stability results give Q-functions or N-step Lyapunov functions, not the one-step quadratic type used here. So the central claim of 'first general stability results' is currently a conditional result whose hypothesis has not been shown to hold for any practical nonlinear estimator. The authors are transparent about this, which I respect, but it is still a load-bearing gap.\n\nThe 'sufficiently small' mismatch constants are left unquantified; the authors flag this as future work. I don't hold that against them much—proving explicit mismatch bounds would be its own paper—but it reinforces the point that the result is existential rather than constructive. The citation pattern looks fine; the self-citation to Kuntz and Rawlings (2024) is relevant and not a red flag.\n\nWho this is for: researchers working on offset-free MPC theory, especially those building on Allan/Rawlings or Schiller et al. MHE stability. They will find the framework useful and the proof strategy transferable. It deserves a serious referee, but I would send it back for major revision, not acceptance as is. The authors should either construct an estimator satisfying Assumption 6 for a nontrivial class, or reframe the headline as 'stability under a robust estimator Lyapunov assumption' and position Assumption 6 as an explicit open problem. With that reframing it could be a strong paper; as it stands, the headline overclaims.","headline":"A solid conditional theorem for nonlinear offset-free MPC, but the key estimator assumption is not demonstrated for any practical estimator; deserves review but needs major revision.","tokens_in":47075,"tokens_out":2419,"would_cite":true,"duration_ms":26029,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["93D09","93D30","93C10","93C55"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that a particular nonlinear offset-free model predictive control design is robustly exponentially stable with respect to setpoint tracking error, despite plant-model mismatch and persistent disturbances, provided the…","keywords":["offset-free model predictive control","nonlinear MPC","robust stability","plant-model mismatch","disturbance estimation","setpoint tracking","Lyapunov stability","moving horizon estimation"],"falsifier":"Take the pendulum or CSTR example with the proposed moving-horizon estimator and a small, asymptotically constant mismatch; Theorem 6 predicts the tracking error converges to zero only if Assumption 6 holds. A single simulation where the mismatch increments vanish but the output offset remains bounded away from zero, with the theorem's other assumptions visibly satisfied, would falsify the claim's applicability, as would an explicit proof that the moving-horizon estimator cannot admit any global quadratic Lyapunov function satisfying (19).","tokens_in":1839,"feed_emoji":"🎛️","tokens_out":2680,"duration_ms":79857,"temperature":0.7,"pith_summary":"The paper supplies the first general stability theory for nonlinear offset-free model predictive control. It shows that a specific design, built from a steady-state target problem, a finite-horizon regulator, and a joint state-and-disturbance estimator, drives the setpoint tracking error to zero when setpoint and disturbance increments vanish, even under plant-model mismatch. The central result, Theorem 6, states that the closed-loop system is regionally robustly exponentially stable with respect to the tracking error $\\delta_r := r - r_{sp}$ under Assumptions 1 to 9. A sympathetic reader cares because offset-free MPC has been used for decades without a stability guarantee that covers mismatch and persistent disturbances.","feed_headline":"Stability proof for offset-free MPC under plant-model mismatch","feed_subtitle":"First general result: tracking error converges to zero when setpoint and disturbance increments vanish.","key_machinery":"The load-bearing structure is the three-part offset-free MPC design: a steady-state target problem (SSTP) that selects a feasible target pair $(x_s,u_s)$ meeting the setpoint, a finite-horizon optimal control regulator driving the state to that target, and a joint state-and-disturbance estimator providing integral action. Stability is carried by a joint Lyapunov theorem (Theorem 3) that combines the regulator value function $V_N^0$ and an estimator Lyapunov function $V_e$ into a single contractive quantity, with plant-model mismatch handled by steady-state correction functions $(\\Delta x_s, d_s)$ that align plant and model steady states. Assumption 6, requiring the estimator to admit a global quadratic Lyapunov function, is what makes the interconnection of controller and estimator tractable.","core_discovery":"The paper's central claim is that offset-free performance can be guaranteed, not assumed, for a nonlinear MPC design. Theorem 6 establishes that there exist constants $\\tau, \\delta_w, \\delta_\\alpha > 0$ such that the closed-loop offset-free MPC system is regionally robustly exponentially stable with respect to the setpoint tracking error $\\delta_r := r - r_{sp}$, and equivalently that as setpoint and disturbance increments vanish, both tracking error and estimator error converge to zero. The proof proceeds in three stages: nominal stability and offset-free performance, robustness to estimate errors and setpoint or disturbance changes, and finally robustness to sufficiently small plant-model mismatch. The result holds under quadratic costs, differentiability of plant and model functions, constraint backoffs at steady state, and a robustly stable state and disturbance estimator.","pith_inferences":["The paper leaves explicit bounds on 'sufficiently small' mismatch unquantified; a natural extension would turn the constants in Proposition 7 into computable margins, as has been done for linear systems.","Assumption 8, requiring known steady-state mismatch corrections, is strong; the theorem does not cover online identification of those corrections, so a practical extension would investigate simultaneous learning and offset-free control.","The proof suggests a modular design philosophy: pair any estimator with a quadratic Lyapunov function with any regulator satisfying the backoff terminal condition to obtain offset-free tracking, decoupling estimation design from controller design.","One testable extension is to seek estimators satisfying Assumption 6, for instance via N-step Lyapunov constructions; success would immediately make the theorem applicable to moving-horizon estimation."],"forward_implications":["Offset-free performance no longer needs to be assumed: under the stated assumptions it follows as a stability property of the closed-loop design.","The design applies to unstable nonlinear plants, as demonstrated on a continuously stirred-tank reactor operating near a Hopf bifurcation.","The theorem implies that tracking error and estimator error converge to zero whenever the setpoint and disturbance increments vanish, not only when signals are asymptotically constant.","For linearized systems, the rank and invertibility conditions on the matrices $M_1$ and $M_2$ recover classical linear offset-free MPC conditions, connecting the nonlinear result to existing practice.","Any estimator that satisfies the global quadratic Lyapunov condition can be plugged into the design and inherit the stability guarantee; the paper notes that moving-horizon estimators are not yet known to satisfy that condition."],"supporting_citations":[{"why":"Established sufficient conditions for offset-free linear MPC under an assumed closed-loop stability, the assumption this paper removes.","marker":"Muske and Badgwell (2002)"},{"why":"Introduced disturbance models for linear offset-free MPC and the detectability conditions that Section 7 recovers in nonlinear form.","marker":"Pannocchia and Rawlings (2003)"},{"why":"First considered offset-free MPC with nonlinear models and tracking costs, but without general stability results.","marker":"Morari and Maeder (2012)"},{"why":"Gave a provably stable state-feedback nonlinear offset-free MPC; the present paper extends this to output feedback and plant-model mismatch.","marker":"Pannocchia et al. (2015)"},{"why":"Supplies the standard MPC value-function properties and recursive-feasibility arguments used throughout the proofs.","marker":"Rawlings et al. (2020)"},{"why":"Proved inherent robustness of optimal and suboptimal nonlinear MPC, which Theorem 5 adapts to estimate errors and setpoint changes.","marker":"Allan et al. (2017)"},{"why":"Established robust stability of full-information and moving-horizon estimators, cited as evidence that example estimators are RGES though not known to satisfy Assumption 6.","marker":"Allan and Rawlings (2021)"},{"why":"Previous work on asymptotic stability of MPC despite plant-model mismatch; supplies the differentiability and quadratic-cost approach this paper builds on.","marker":"Kuntz and Rawlings (2024)"},{"why":"Defined input-to-state stability with respect to two measurement functions, which the paper extends to setpoint-tracking stability.","marker":"Tran et al. (2015)"}],"fun_headline_variants":["First general stability proof for offset-free MPC","Offset-free MPC stable despite model mismatch","Nonlinear MPC offset-free stability proven","MPC offset-free tracking stability guaranteed"],"cache_read_input_tokens":49152,"weakest_assumption_plain":"The whole guarantee rests on the estimator having a global quadratic Lyapunov function (Assumption 6), which the paper notes is not known to hold for the moving-horizon estimators used in its own examples.","fun_headline_variants_meta":{"raw":{"variants":["First general stability proof for offset-free MPC","Offset-free MPC stable despite model mismatch","Nonlinear MPC offset-free stability proven","MPC offset-free tracking stability guaranteed"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000536,"raw_usage":{"total_tokens":2523,"prompt_tokens":843,"completion_tokens":1680,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":459,"completion_tokens_details":{"reasoning_tokens":1628}},"tokens_in":459,"tokens_out":1680,"duration_ms":11551,"temperature":1.0,"reasoning_tokens":1628,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:12:27.037842+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the pendulum or CSTR example with the proposed moving-horizon estimator and a small, asymptotically constant mismatch; Theorem 6 predicts the tracking error converges to zero only if Assumption 6 holds. A single simulation where the mismatch increments vanish but the output offset remains bounded away from zero, with the theorem's other assumptions visibly satisfied, would falsify the claim's applicability, as would an explicit proof that the moving-horizon estimator cannot admit any global quadratic Lyapunov function satisfying (19).","supporting_citations":[],"review_version":1}