WorksMathematics
Real-Time Nonlinear Mixed-Integer Model Predictive Control With Horizon-Tail Relaxations
Abstract
Model predictive control (MPC) of nonlinear systems with discrete decisions leads to mixed-integer nonlinear programs (MINLPs) whose combinatorial complexity often prevents real-time deployment. In this paper, we propose a decomposition-based MPC method that reorders and reduces computations to shorten the delay between receiving a state measurement and applying the control input. Before the measurement arrives, we solve a continuous relaxation of the MINLP at a predicted state and construct a mixed-integer quadratic program (MIQP). Once the measured state is available, we solve the MIQP that reconstructs the discrete controls and simultaneously performs one sequential quadratic programming step toward the corresponding nonlinear solution. To reduce the combinatorial complexity in the horizon length, we either fix the integer controls or relax their integrality over the prediction horizon tail, while retaining an integer-feasible first control input. We derive error bounds relating the nonlinear objective values obtained with the horizon-tail approximations to the optimal value of the original mixed-integer problem. In a data center cooling example, the proposed method reduces mean feedback computation time by 94% relative to a state-of-the-art MINLP method with a total closed-loop cost difference of less than 1%.
Introduction
Model predictive control (MPC) is a feedback control strategy that repeatedly solves optimal control problems. Its appeal lies in its ability to directly specify objectives and constraints while using predictive models to anticipate the effects of control decisions. MPC algorithms are well established for linear and nonlinear systems with continuous control inputs [1]. Yet, many real-world systems also require discrete controls.
Combined with nonlinear dynamics, these discrete decisions give rise to nonlinear mixed-integer optimal control problems (MIOCPs), whose time discretization yields mixed-integer nonlinear programs (MINLPs). Generic MINLP solvers [2], even when terminated early, are usually too slow for real-time MPC deployment. Particularly for nonlinear systems, this has motivated methods tailored to MIOCPs that rapidly compute high-quality approximate solutions, often with well-understood theoretical properties. These methods include, for example, switching-time optimization [3, 4], decomposition methods [5, 6] or machine-learning-based approaches [7, 8, 9].
Decomposition methods for MIOCPs solve continuous nonlinear programs (NLPs) related to the original MINLP, combined with a discrete reconstruction step [5, 6, 10], yielding high-quality approximations with well-understood error bounds [5, 6]. Discrete reconstruction recovers the integer controls using different methods ranging from OCP-tailored rounding [10] and variants accommodating dwell-time constraints [11, 12] to integer programming [5, 13] and tailored shortest-path methods incorporating combinatorial constraints [14]. However, recovering integer controls without explicitly accounting for their effects on the dynamics can cause performance losses or infeasibility on coarse time-discretization grids [15]. Reconstruction through a mixed-integer quadratic program (MIQP) incorporates local approximations of the dynamics and objective, improving solution quality at greater computational cost [16]. For real-time MPC, this trade-off motivates further approximations that reduce computation while preserving solution quality, a direction pursued in this paper.
Optimization problem solution approximations are already standard in real-time MPC with continuous controls, where their theoretical properties are well understood [17]. Rawlings and Risbeck Rawlings and Risbeck [18] show how these properties extend conceptually to approximate mixed-integer MPC. This motivates us to adapt algorithmic ideas from real-time continuous MPC to the mixed-integer setting [1].
A shared idea in many of these algorithms is to reorder optimization computations to reduce the feedback delay [1]. The computations are split into a preparation phase, which performs operations before the next state measurement is known, such as evaluating functions and derivatives or presolving related problems, and a feedback phase, which computes a suitable control approximation with minimal effort once is known. Advanced-step methods compute in the preparation phase an (approximate) solution of the MPC problem at a predicted state [19, 20, 21].
In this paper, we adapt these ideas to mixed-integer MPC by reordering and reducing the computations of decomposition algorithms. In the preparation phase, we solve the continuous relaxation of the MINLP, and optionally an MIQP at a predicted advanced state . In the feedback phase, we update the prepared MIQP with the measured state and solve it using the available warm start. To reduce the MIQP’s combinatorial complexity, we either fix the integer controls or relax their integrality over the prediction horizon tail. This
The author thanks Nathalie Janes and Moritz Diehl for interesting discussions on mixed-integer optimal control. OpenAI’s Code was used to assist with drafting and language editing, as well as implementing the accompanying code. The author reviewed and verified all AI-assisted content and takes full responsibility for the final manuscript and results. reduces the feedback delay while providing bounds on the resulting performance losses.
Contributions. The main contributions of this paper are: (i) a real-time mixed-integer nonlinear MPC method that separates preparation and feedback computations, combining integer reconstruction and one SQP step in a single MIQP; (ii) two horizon-tail approximations that fix integer controls or relax their integrality beyond a selected horizon, reducing the combinatorial search while retaining an integer-feasible first control input; (iii) bounds relating the nonlinear objective values obtained with these approximations to the globally optimal value of the original MINLP; (iv) a numerical comparison with a state-of-the-art MINLP method and a rule-based controller on a data center cooling example; and (v) an open-source implementation supporting several algorithmic variants and several MPC examples from the literature.2
Decomposition methods for MIOCP
In this section, we state the MIOCP of interest, and recall a standard decomposition method for solving such problems.
Problem formulation
We consider a discrete-time nonlinear system with states , continuous controls , and binary controls . General bounded integer controls can be encoded by binary variables, so this restriction is without loss of generality. For an initial state and a horizon of intervals, define , , and , to be the discrete-time states and controls. The MIOCP is stated as follows:
Here, is the stage cost and is the terminal cost. The map denotes the discrete-time state-transition map. The functions collect the path inequality constraints. The polyhedron contains additional combinatorial constraints coupling several control intervals, e.g., dwell-time constraints [12].
Next, we define a few shorthands. We aggregate the binary controls as , where , and collect the continuous primal variables into . Including the bounds in its linear representation, we write . The feasible binary set is . The vector denotes the parameter, in our case, without loss of generality, this is the initial state . Let collect the equality constraints in (1b) and (1c), and let collect the inequality constraints in (1d). A function is a generic nonlinear function collecting the stage and terminal costs in (1). In the remainder of this and the next section, we assume that the parameter is fixed. Using this notation, problem (1) can be written compactly as
It will be later useful to define a standard NLP where is a parameter and not a decision variable:
Here, denotes the optimal value function of this NLP. Note that we can compactly express the integer part of the MINLP (2) solution as [16]:
A relaxed integer solution , and thus a lower bound for is obtained by solving the relaxed MINLP (2), compactly written as: s.t. .
A three-phase decomposition method
For optimal control and MPC applications, decomposition methods typically compute high-quality approximate solutions of (2) more efficiently than generic MINLP methods [5, 6, 10]. The decomposition method introduced by Sager [10] consists of three steps: solve a relaxation of the MINLP, integer reconstruction, and nonlinear polishing. Following the exposition in [16], these steps are:
(S1) The integrality condition in (2) is relaxed by replacing with , yielding a nonlinear program (NLP). A local minimizer of this NLP is denoted by , where but not necessarily . If , then steps two and three are omitted.
(S2) An integer solution is reconstructed by minimizing a distance function:
(S3) The integer variables are fixed to , and the resulting NLP (3), referred to as the polishing NLP, is solved. Whenever this NLP admits a solution , the pair is feasible for (2).
This procedure typically yields a pair that is locally optimal for the fixed-integer NLP (3) and feasible, but generally suboptimal, for the MINLP (2). Let denote a global minimizer of (2). If the continuous relaxation is solved globally, the optimal values satisfy [16]. Any local minimizer of the relaxed NLP can also be used for integer reconstruction in (S2), although the first inequality does not necessarily hold anymore. The practical performance and computational complexity of decomposition methods depend strongly on the choice of the distance function . In the remainder of the section, we review existing choices for . In Sec. III we propose new ones tailored to MPC, and establish their relations to .
Ideal distance function
The ideal choice for the distance function is . If is a global solution of the continuous relaxation, this function is nonnegative for all , since . Because is constant, solving (5) with this choice is equivalent to solving (4), and hence the original MINLP (2). This choice is therefore usually intractable and we look for more tractable approximations of .
Combinatorial integral approximation (CIA) distance
A simpler choice is , which yields a mixed-integer linear programming (MILP) formulation of (5) for or . The frequently used CIA norm , with an invertible matrix defined in [5], also admits an MILP formulation. If , sum-up rounding provides a polynomial-time solution to the CIA problem under suitable assumptions [5 Theorem 3]. Norm-based distance functions are attractive for their computational efficiency, as they admit tailored solution methods, such as those implemented in pycombina [13], which we use in this work. On the downside, such approximations reconstruct based solely on , disregarding or the nonlinear dynamics and constraints, which can lead to performance losses [15].
Quadratic-model distance
Since minimizing over is computationally intractable, [16] proposes a quadratic approximation of , turning (5) into an MIQP. We recall this approximation here and further adapt it in Sec. III.
Define the Lagrangian of the fixed- NLP (3) as
The Lagrangian of the continuous relaxation of (2) is . Since the additional term is affine in , both Lagrangians have the same Hessian with respect to . At a Karush–Kuhn–Tucker (KKT) point of the relaxed NLP in (S1), we approximate this Hessian by a symmetric positive definite matrix . With and , define the linearization of compactly as
The linearizations and are defined analogously. Furthermore, a quadratic approximation of the objective of (2) using is defined as
For the linearization point , the MIQP approximation of (2) reads
In analogy to (3), for fixed , the quadratic value function is defined as
For , is convex, continuous, and piecewise quadratic in on its feasible domain [16]. Moreover, approximates , with an error bound given in Lemma 1 below. For fixed , the quadratic-model distance is defined as
The relaxed NLP’s KKT point also satisfies the KKT conditions of the MIQP’s (6) continuous relaxation, which is convex because and its constraints are affine. Hence, globally minimizes this QP, and fixing any feasible cannot decrease its optimal value, so , hence . Minimizing this distance over is equivalent to solving (6) constructed at . Note that we could use also , then (6) reduces to an MILP.
Relaxations in the integer reconstruction
Solving the MIQP (6) is usually the most expensive step of the decomposition approach with a quadratic-model distance. Since MPC applies only the first control input, only must be integer feasible. This motivates reducing the combinatorial complexity through two horizon-tail approximations inspired by [22]. While [22] uses these approximations in a tailored solver for fixed-matrix MIQPs, we study them within the decomposition approach and apply them to real-time MPC in Sec. IV.
For either approximation, partition at a horizon , where
The corresponding block is empty when or . The objective and linearized constraints retain the full horizon.
Relaxation of optimality
Suppose we have an integer-feasible vector . In practice, it may be obtained via sum-up rounding or by solving the CIA-MILP with pycombina [13]. For , the first block remains a binary decision and the second block is fixed to the corresponding part of . Thus, in (S2) we solve an MIQP with fewer free binary variables:
Every entry of is binary, so (10) supplies an integer-feasible first input. For , the complete integer sequence is fixed to and only one continuous QP is solved. For , problem (10) is the full MIQP (6). For any , subsequent nonlinear polishing yields a feasible, potentially suboptimal solution of the full MINLP (2), provided the polishing NLP in (S3) is feasible.
Relaxation of feasibility
For , only the first block is required to be binary, while integrality is relaxed in the remaining block. We solve the following MIQP:
Since , the first control input is integer feasible, while the second block can be fractional. For , this problem is the full MIQP (6). With this relaxation, the decomposition no longer necessarily provides a feasible solution of the full MINLP (2). Nevertheless, the relaxed problem provides a lower bound on the full MIQP value, as discussed below, and retains an integer-feasible first input for MPC. From a dynamic programming perspective, both approximations replace the full MIQP’s tail value function by one obtained with either prescribed binary controls or relaxed integrality. Figure 1 illustrates both approximations and compares them with the full MIQP.
Figure 1. Horizon-tail approximations for a binary control in the parallel-chiller example of Sec. V. The full MIQP optimizes all binary decisions. The feasibility relaxation () allows a continuous tail, whereas the optimality relaxation () fixes the tail to prescribed binary values. Shading indicates the respective horizon tails.
Relation to the nonlinear value function
Throughout this section, is fixed and we abbreviate . We now relate the optimal value functions of the relaxed MIQPs (10) and (11) to , the optimal value function of (3). We begin by relating the optimal value functions of the three MIQP variants: the full MIQP and its optimality and feasibility relaxations.
Theorem 1 (Horizon-tail gap bound). Suppose that the continuous relaxation of (6) has a compact feasible set and fix with . Fix a binary sequence feasible for (6) and assume that, for every partition (9), replacing the tail of any feasible sequence with a binary prefix by preserves feasibility after reoptimizing . Let , , and denote the globally optimal values of (6), (10), and (11), respectively, with and . Then there exists , independent of , such that
Moreover, is nondecreasing, is nonincreasing, and .
Proof. Every feasible point of the full MIQP (6) is feasible for the feasibility-relaxed problem (11), hence . Conversely, every optimizer of (10) is feasible for the full MIQP (6), hence . Together, these inequalities establish (12a). Increasing shrinks the feasible set of (11) and enlarges that of (10), proving monotonicity of and . At , both problems coincide with (6).
By Theorem 6.1, is continuous and piecewise quadratic on its feasible domain. Its finitely many quadratic pieces and compactness of the feasible set of the relaxation of (6) imply Lipschitz continuity of with a constant . Let be the -component of a global optimizer of (11), so that . Similarly, let be the corresponding component of a global optimizer of (10), so that . Then
In the first inequality, we used , which follows because the tail-replacement assumption makes feasible in (10), while is optimal. The second inequality follows from Lipschitz continuity of . The last uses the fact that has components, each differing by at most one. Thus, , which establishes (12b). □
We proceed towards relating to . Recall that is interpreted here as a fixed parameter in the NLP (3), for which we will assume some standard regularity conditions.
Assumption 1 (Local NLP regularity). Let be a KKT point of (3) at , and define and . It holds that:
(a) The functions , , and are twice continuously differentiable near .
(b) Linear independence constraint qualification (LICQ) holds with respect to , i.e., the vectors
are linearly independent.
(c) Strong second order sufficient condition (SSOSC) holds with respect to , i.e.,
for every nonzero in the strong critical cone
Next, we establish an intermediate result similar to [16 Corollary 1], but more directly using only elementary NLP perturbation theory. Throughout the following analysis, denotes the locally unique NLP solution map in the polishing step used in Lemma 1, with and .
Lemma 1. Suppose Assumption 1 holds for the NLP (3) at the fixed parameter and the corresponding solution , and that in (7). Then there exist and such that for all :
Proof. Fix the parameter also in the QP (7). By construction, the NLP and QP have the same KKT conditions at . Since , the unique solution of this QP is therefore , with . Hence, , and it can be verified by direct computation that , where is the Lagrangian of the QP (7).
By Assumption 1 and [24 Theorem 2.4.5 (iv)], the local primal–dual solution of the NLP (3) is Lipschitz continuous in . Because ensures SSOSC and Assumption 1 implies LICQ also for the QP (7) at , we can apply [24 Theorem 2.4.5 (iv)] to the QP’s solution map as well. Thus, let , , denote these primal–dual solution maps, with .
By twice continuous differentiability of , , and (Ass. 1(a)) and Lipschitz continuity of the solution maps, and , evaluated along the respective solution maps, are locally Lipschitz in .
Furthermore, by applying [24 Theorem 2.4.5 (v)] to the NLP and QP, we have that and . The same identities hold locally along the respective solution maps, so both value-function gradients are locally Lipschitz.
Finally, define . It follows from the arguments above that there exists such that is Lipschitz with a constant for . Let . Using and and applying the fundamental theorem of calculus for , with we obtain
Thus, (13) holds with .
Finally, we relate the value obtained by the decomposition approach to the globally optimal MINLP value in (4). Lemma 1 requires , which need not hold for arbitrary binary vectors and a relaxed solution . Hence, we make an assumption specific to mixed-integer OCPs, analogous to [16 Assumption 1], that allows us to apply the preceding result.
Assumption 2. Let be as in Lemma 1, and fix . Let be the -component of a global optimizer of either (10) with or (11) with . Assume that and that there exists a global minimizer of (2) satisfying , .
Recall that is the relaxed solution obtained in (S1), while is the value of the polished solution obtained in (S3) by fixing in (3), with . Next, our main theoretical result follows.
Theorem 2 (Nonlinear value bounds). Fix and let be a local solution of the continuous relaxation of (2). Suppose Assumption 1 holds for the fixed- NLP (3) at . Construct the MIQP (7) at with , and let and be given by Lemma 1. Suppose the Assumption of Theorem 1 hold for this MIQP, and let be its gap-bound constant.
Fix and let be the -component of a global optimizer of either (10) with or (11) with . Let be a global minimizer of (2), and suppose Assumption 2 holds for and with the radius . For the polishing solution , define ; by the same assumption, .
(a) If is obtained from the optimality-MIQP (10), then
\leq K\sqrt{(N-M)n_v}+C(\lVert\hat y-y^\rVert^2+\lVert y^\circ-y^\rVert^2).
J_{\mathrm{NLP}}(\hat y)-J_{\mathrm{NLP}}(y^\circ)\leq C(\lVert\hat y-y^\rVert^2+\lVert y^\circ-y^\rVert^2)-(J^*_{\mathrm{QP}}-L_M).
Proof. By Assumption 2, Lemma 1 applies at and . Adding and subtracting the corresponding QP values gives
Since , QP (7) with admits a solution. Let denote the globally optimal value of the full MIQP (6). Together with Theorem 1 and , this yields
In case (a), optimality of the reconstruction gives . Comparison with the full MIQP and Theorem 1 yields
Substitution into (16) proves the upper bound in (14). Its lower bound follows from feasibility of the polished pair and global optimality of , hence .
In case (b), optimality of the reconstruction gives . Using (17), we obtain
Substitution into (16) proves (15).
For the optimality relaxation, the nonlinear loss is bounded by the horizon-tail gap and a local quadratic modeling error. The feasibility relaxation does not enforce integrality over the full horizon, which tightens the upper bound by the relaxation gap at the expense of full integer feasibility. At , the upper bounds of (14) and (15) are identical and reduce to the form of [16 Theorem 2].
Advanced-step mixed-integer NMPC
In real-time MPC, the delay between receiving a state measurement and applying the control input should be as short as possible. To reduce this feedback delay, we reorder and reduce the computations of the three-step decomposition into a preparation phase, which solves the continuous relaxation with a predicted state, and a feedback phase, which solves a horizon-tail relaxed MIQP with the measured state.
Preparation and feedback
Following the advanced-step ideas [19, 20, 21], we separate computations performed before the next measurement from those needed to compute the feedback. At sampling time , denote the true initial state by and its prediction by .
During the preparation phase, while the previously computed input is applied, we solve the continuous relaxation at to obtain and its multipliers. This completes (S1) and provides the linearization point for the MIQP. For the optimality relaxation, we also compute the binary sequence used to fix the tail. In principle, this can be obtained with any feasible method, e.g., heuristics, rule based controllers or learned policies [8]. In our implementation, we solve the CIA-MILP via to obtain . Optionally, the MIQP can be solved at to provide a warm start for the feedback phase MIQP with .
Once is available, the feedback phase updates the initial state to and solves the MIQP. All other data in the MIQP remain unchanged. Algorithm 1 summarizes the procedure. All three MIQP variants enforce integrality of the first control input, including the case , where the entire binary sequence is prescribed, and only a QP is solved in the feedback phase.
Algorithm 1 Advanced-step mixed-integer nonlinear MPC (AS-MINMPC)
Require: Horizons , and a choice of (6), (10), or (11)
1: for each sample do
Preparation before
2: Predict the next initial state .
3: Solve the relaxed NLP of (S1) with to obtain and its multipliers.
4: Set up the MIQP of (S2) at the lin. point .
5: For (10), compute and fix its tail.
6: Optional: Solve the selected MIQP at to obtain a warm start.
Feedback after measuring
7: Set the initial state in the MIQP to .
8: Solve the selected MIQP of (S2) using the available warm start, to obtain .
9: Optional: Solve the polishing NLP in (S3) with .
10: Apply the first continuous and integer control inputs.
11: Shift the solution for the next preparation phase.
12: end for
MIQP step as a full SQP step
To reduce feedback delay, the proposed algorithm does not in general solve the polishing NLP in (S3) to convergence. We observe that the feedback MIQP simultaneously reconstructs the integer decisions and performs one SQP step toward the corresponding polishing solution, yielding .
For the full and optimality-relaxed MIQPs, fixing the reconstructed integers yields the SQP subproblem obtained by linearizing the NLP in at and imposing . For the feasibility relaxation, the same interpretation holds after fixing only the binary prefix and retaining the tail as continuous variables.
This interpretation connects the method to real-time iteration (RTI) and advanced-step RTI [1, 20, 21] for continuous NLPs. If the integer decisions are the same for the measured and predicted states, i.e., , we expect a local predictor–corrector error analysis for the proposed AS-MINMPC method to yield bounds similar to [21]. If , the solution map may be discontinuous, similarly to nonsmooth MPC [25], which would need to be accounted for in the error analysis. Due to page restrictions, we leave a detailed error analysis for future work.
Numerical example
We compare several variants of the proposed method with the state-of-the-art MINLP solver CAMINO [26], and a rule-based controller (RBC) from [8] on a parallel-chiller example [8].
Implementation
We implement several variants of Algorithm 1 (including the optional steps, and solving MILPs instead of MIQPs) in the accompanying repository.3 The implementation uses IPOPT [27] for NLPs, BnB-DAQP [28] for MIQPs via CasADi 3.8.0 [29], and pycombina [13] for integer-feasible reconstruction. Thus, the proposed method uses only open-source software packages.
Benchmark example
To evaluate the proposed algorithms, we consider a data center parallel-chiller plant example from [8 Sec. 2]. The objective is to meet the cooling demand while reducing power consumption and chiller switching frequency. For chillers, the state vector is , comprising the shared return and individual supply temperatures. The continuous controls are , comprising the evaporator temperatures and mass flows, while contains the binary on/off decisions for each chiller operating in parallel. The cooling powers and thermal dynamics are:
The chiller and pump power consumption are:
All model parameters, bounds, and load profiles are taken from [8 Table 1]. Using the compact notation of the OCP (1), we have
Here, denotes the previously applied binary control input. The stage costs terms consist of
the power consumption and load demand tracking error: . We use the cost weights and . The discrete-time map incorporates the prescribed effective demand and one fourth-order Runge–Kutta (RK4) step per control interval of 180 s. Closed-loop simulation uses four RK4 substeps per control interval to obtain the next measured state.
All controllers use the same load profile (Figure 3) and initial states over 240 MPC steps (12 h of sim. time). Preparation times are never the computational bottleneck, therefore, we focus on and report feedback times, which are dominated by the MIQP time. The realized closed-loop cost sums the switching and stage costs incurred by the applied inputs over the simulation and is normalized by the lowest cost among all tested algorithms for each experiment. We compare several variants of the proposed AS-MINMPC method with different and values, the three-phase algorithm [16] of Sec. II, and a state-of-the-art MINLP method: SB-MIQP, implemented in CAMINO [26]. The latter two perform all computations in the feedback phase. We use the default settings of SB-MIQP, except for a time limit of 30 s per MPC step, with Gurobi as MIQP solver. A single full SB-MIQP iteration is equivalent to the three phase algorithm [26].
Figure 3. Representative MPC trajectories for delivered cooling and demand (top), and number of active chillers (bottom), for .
Figure 2 compares mean feedback costs for with chillers. For , the variant computes feedback 16.68 times faster than CAMINO, with a slightly lower closed-loop cost. For , it reduces the mean and maximum observed feedback times by factors of 15.56 and 28.30, respectively, relative to full AS-MINMPC, with only a 0.27% increase in closed-loop cost. RBC is computationally cheaper, but its cost is approximately 2.5 times that of the best optimization-based controller. Figure 3 shows that closely follows the cooling and chiller activation trajectories of the three-phase method, with better demand tracking than RBC.
Figure 2. Mean feedback time and normalized realized closed-loop cost for three chillers with and . Horizon-tail relaxations yield greater AS-MINMPC speedups for longer prediction horizons.
In this example, feasibility-relaxed MIQPs yielded substantial speedups, often with minor losses of optimality below 1%. Optimality-relaxed MIQPs resulted in higher costs in these experiments due to the less favorable integer-sequences obtained with pycombina. Our experiments on other examples available in the public repository showed that this relative performance is problem dependent. But in general, horizon-tail approximations substantially reduced feedback delay while preserving near-reference control performance across all implemented examples.
Conclusions
We presented a real-time mixed-integer nonlinear MPC method that reorders the computations of a decomposition algorithm [10, 16] into preparation and feedback phases. The feedback MIQP reconstructs the integer controls and simultaneously performs one SQP step toward the solution of the problem with the latest state estimate. Horizon-tail approximations in the MIQP reduce the combinatorial complexity associated with the horizon length. Numerical results show reductions of more than 95% in the maximum observed feedback time, with only minor changes in closed-loop cost. The computational advantage of the proposed real-time algorithms becomes increasingly pronounced as the prediction horizon grows. The feedback phase still requires solving an MIQP, which can be expensive when many integer control variables are present. This motivates development of more efficient open-source MIQP solvers.
References
- [1]M. Diehl, H. J. Ferreau, and N. Haverbeke, “Efficient numerical methods for nonlinear MPC and moving horizon estimation,” in Nonlinear model predictive control, ser. Lecture Notes in Control and Information Sciences, L. Magni, M. Raimondo, and F. Allgöwer, Eds. Springer, 2009, vol. 384, pp. 391–417.DOI
- [2]P. Belotti, C. Kirches, S. Leyffer, J. Linderoth, J. Luedtke, and A. Mahajan, “Mixed-integer nonlinear optimization,” Acta Numerica, vol. 22, pp. 1–131, 2013.DOI
- [3]M. Gerdts, “A variable time transformation method for mixed-integer optimal control problems,” Optimal Control Applications and Methods, vol. 27, no. 3, pp. 169–182, 2006.DOI
- [4]R. Abbasi-Esfeden, A. Nurkanović, M. Diehl, P. Patrinos, and J. Sw-evers, “An efficient mixed-integer formulation and an iterative method for optimal control of switched systems under dwell time constraints,” Optimal Control Applications and Methods, 2026.
- [5]S. Sager, M. Jung, and C. Kirches, “Combinatorial integral approximation,” Mathematical Methods of Operations Research, vol. 73, no. 3, pp. 363–380, 2011.DOI
- [6]C. Kirches, F. Lenders, and P. Manns, “Approximation properties and tight bounds for constrained mixed-integer optimal control,” SIAM Journal on Control and Optimization, vol. 58, no. 3, pp. 1371–1402, 2020.DOI
- [7]C. A. Orrico, W. Heemels, and D. Krishnamoorthy, “Learning myopic mixed-integer nonlinear model predictive control from expert demonstrations,” arXiv preprint arXiv:2605.07401, 2026.arxiv.org/abs/2605.07401
- [8]J. Boldkovský, C. Faulker, E. Michael, M. Gulan, A. Tuor, and J. Drgoňa, “Data center chiller plant optimization via mixed-integer nonlinear differentiable predictive control,” Control Engineering Practice, vol. 174, p. 107063, 2026.
- [9]J. Adamek, L. Lüken, and S. Lucia, “Enabling robust mixed-integer nonlinear model predictive control via self-supervised learning and combinatorial integral approximation,” Journal of Process Control, vol. 159, p. 103636, 2026.DOI
- [10]S. Sager, “Reformulations and algorithms for the optimization of switching decisions in nonlinear optimal control,” J. Proc. Contr., vol. 19, pp. 1238–1247, 2009.DOI
- [11]C. Zeile, N. Robuschi, and S. Sager, “Mixed-integer optimal control under minimum dwell time constraints,” Mathematical Programming, vol. 188, no. 2, pp. 653–694, 2021.
- [12]Y. Chen and M. Lazar, “An efficient mpc algorithm for switched systems with minimum dwell time constraints,” Automatica, vol. 143, p. 110453, 2022.
- [13]A. Bürger, C. Zeile, M. Hahn, A. Altmann-Dieses, S. Sager, and M. Diehl, “pycombina: An open-source tool for solving combinatorial approximation problems arising in mixed-integer optimal control,” in Proceedings of the IFAC World Congress, vol. 53, no. 2, 2020, pp. 6502–6508.DOI
- [14]F. Bestehorn, C. Hansknecht, C. Kirches, and P. Manns, “Mixed-integer optimal control problems with switching costs: a shortest path approach,” Mathematical Programming, vol. 188, no. 2, pp. 621–652, 2021.DOI
- [15]A. Bürger, “Nonlinear mixed-integer model predictive control of renewable energy systems,” Ph.D. dissertation, University of Freiburg, 2020.DOI
- [16]A. Bürger, C. Zeile, A. Altmann-Dieses, S. Sager, and M. Diehl, “A Gauss–Newton-based decomposition algorithm for nonlinear mixed-integer optimal control problems,” Automatica, vol. 152, p. 110967, 2023.DOI
- [17]J. B. Rawlings, D. Q. Mayne, and M. M. Diehl, Model Predictive Control: Theory, Computation, and Design, 2nd ed. Nob Hill, 2017.
- [18]J. B. Rawlings and M. J. Risbeck, “Model predictive control with discrete actuators: Theory and application,” Automatica, vol. 78, 2017.DOI
- [19]V. M. Zavala and L. T. Biegler, “The advanced step NMPC controller: Optimality, stability and robustness,” Automatica, vol. 45, pp. 86–93, 2009.DOI
- [20]A. Nurkanović, A. Zanelli, S. Albrecht, and M. Diehl, “The Advanced Step Real Time Iteration for NMPC,” in Proceedings of the IEEE Conference on Decision and Control (CDC), 2019.DOI
- [21]J. Frey, A. Nurkanović, and M. Diehl, “Advanced-step real-time iterations with four levels – new error bounds and fast implementation in acados,” IEEE Control Systems Letters, 2024.DOI
- [22]D. Frick, A. Domahidi, and M. Morari, “Embedded optimization for mixed logical dynamical systems,” Computers & Chemical Engineering, vol. 72, no. Supplement C, pp. 21–33, 2015, a Tribute to Ignacio E. Grossmann.DOI
- [24]A. Fiacco, Introduction to sensitivity and stability analysis in nonlinear programming. New York: Academic Press, 1983.
- [25]A. Nurkanović, P. Pozharskiy, and M. Diehl, “Real-time algorithms for model predictive control of hybrid dynamical systems,” arXiv preprint arXiv:2604.18432, 2026.arxiv.org/abs/2604.18432
- [26]A. Ghezzi, W. V. Roy, S. Sager, and M. Diehl, “A sequential Benders-based mixed-integer quadratic programming algorithm and its implementation in the CAMINO toolbox,” Mathematical Programming Computation, vol. 1–70, 2026.arxiv.org/abs/2404.11786
- [27]A. Wächter and L. T. Biegler, “On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming,” Mathematical Programming, vol. 106, no. 1, pp. 25–57, 2006.DOI
- [28]D. Arnström and D. Axehill, “Bnb-daqp: A mixed-integer qp solver for embedded applications,” IFAC-PapersOnLine, vol. 56, no. 2, pp. 7420–7427, 2023.arxiv.org/abs/2211.12902
- [29]J. A. E. Andersson, J. Gillis, G. Horn, J. B. Rawlings, and M. Diehl, “CasADi – a software framework for nonlinear optimization and optimal control,” Mathematical Programming Computation, vol. 11, no. 1, pp. 1–36, 2019.


