WorksComputer Science
Large-Scale Hierarchical Decision-Making in Robot Control via ADMM
Abstract
Sparse hierarchical nonlinear programming (SH-NLP) based on the -norm provides a powerful framework for integrated, efficient hierarchical decision-making in robot kinematic control, for example in the presence of a large set of possible end-effector candidate locations. While Sequential Sparse Hierarchical Quadratic Programming (S-SHQP) with interior-point method based SHQP sub-problem solvers exploit the hierarchical, sparse structure of such problems, they are less amenable to effective warm-starting between subsequent, coherent control steps. In this work, we formulate an Alternating Direction Method of Multipliers (ADMM) to efficiently solve related SHQP sub-problems. We introduce efficient substitutions for auxiliary variables representing -norm surrogates. The solver efficiently handles separable decision functions by evaluating linear algebra expressions in the appropriate order. Results in large-scale kinematic and dynamic control show reduced computation times of up to 23.5% compared with a structure-exploiting interior-point method in low- to moderate-accuracy regimes. Furthermore, hierarchical decision-making is extended to whole-body dynamic control for humanoid walking, where the need for high-accuracy solutions is highlighted.
Introduction
Hierarchical decision-making in robot control frequently requires selecting among a large number of candidate actions or locations while simultaneously satisfying whole-body kinematic and physical constraints of different priorities. Examples include selecting footholds [1], or activating only a subset of joints in highly redundant robotic systems [2], all while respecting higher or lower priority constraints [3]. Sparse hierarchical nonlinear programming (SH-NLP) [4] provides an attractive formulation for such problems, as discrete decisions can be embedded directly into continuous programs without explicitly introducing combinatorial decision variables.
SH-NLPs can be addressed by iteratively solving sparse hierarchical quadratic programs (SHQP), for example by the interior-point method (IPM) [4]. By exploiting the sparse hierarchical structure, a high degree of efficiency with linear dependency in the number of sparse candidate constraints can be achieved, unlike standard QP solvers with cubic dependency.
However, exploitable structure exists also between individual SHQP subproblems. Robot control typically consists of high-frequency updates with gradual changes between subsequent instances, for example similar sets of active and
inactive inequality constraints. The IPM is less amenable to exploiting this additional sequential coherence through warm-starting.
The Alternating Direction Method of Multipliers (ADMM) [5] is a suitable candidate for such parametric programs [6], as its primal and dual can be effectively initialized from previous optimization instances. However, specialized solver formulations for SHQPs do not exist. This work aims to close this gap with the following main contributions:
We derive a structure-exploiting ADMM solver for SHQPs which eliminates sparsity-related auxiliary variables for linear dependence on the number of decision candidates, while enabling warm-starting, unlike existing IPM based methods.
For separable decision functions, we exploit their structure and implement computationally favorable linear algebra evaluations, which improves efficiency for large candidate sets.
We evaluate the proposed solver on large-scale sparse programs and robot-control problems, isolating the effects of sparse structure, warm-starting, and separable decision-function evaluation. The results demonstrate reduced computation times over the structure-exploiting IPM [4] in low- to moderate-accuracy regimes.
We investigate the extension of hierarchical decision-making from kinematic to dynamic robot control. The experiments demonstrate efficient drone control, while a humanoid walking experiment exposes the need for high-accuracy solutions, which is less suitable for first-order methods like ADMM.
The remainder of this paper is organized as follows. Sec. II reviews related work on optimization methods for hierarchical decision-making in robot control. Sec. III recalls the SHQP formulation arising within S-SHQP. Sec. IV derives the proposed efficient ADMM solver, and compares it with other SHQP solvers. Sec. V discusses equilibration and warm-starting, while Sec. VI outlines how structure in separable decision functions can be exploited. Sec. VII evaluates the computational performance and scalability of the proposed solver on numerical and robot control problems.
Related Work
Hierarchical Decision-Making in Robotics
Mixed-integer formulations provide a powerful tool for globally optimal decision-making in robot planning [7], but their combinatorial complexity can limit their applicability to large candidate sets and real-time control. Instead, sparsity-promoting continuous -norm formulations have been efficiently applied to parsimonious kinematic control [8] and footstep selection [1]. However, these methods either rely on specialized algorithms like the simplex method or additional brute-force layers to achieve accurate sparsity.
The approach in [9] enhances the sparsity-promoting qualities of the -norm by sequentially reweighting the linear objective according to a logarithm-based -norm surrogate. The -norm directly measures cardinality and therefore provides a natural representation of sparse decision-making, but needs to be smoothly approximated for continuous optimization. While [9] is limited to linear unconstrained problems, the approach has been extended to hierarchical nonlinear decision-making by incorporation within a SQP-type framework (Sequential Quadratic Programming [10]; from hereon referred to as Sequential Sparse Hierarchical Quadratic Programming (S-SHQP) [4]). The corresponding dedicated IPM exploits the sparse hierarchical structure of the SHQP sub-problems, resulting in linear computational complexity with the number of candidate constraints. However, warm-starting is not exploited, and the solver has only been applied to purely kinematic control scenarios.
This work addresses this gap and proposes an ADMM solver which can be effectively warm-started for the efficient repeated solution of subsequent SHQP subproblems. Furthermore, hierarchical decision-making is integrated within a humanoid whole-body controller under consideration of instantaneous robot dynamics.
Parametric Quadratic Programming
Repeatedly solving closely related parametric optimization problems is fundamental to real-time robot control. Active-set methods like qpOASES [11] are well suited to exploit this similarity to accelerate subsequent optimizations, because the active set of a preceding QP often provides a good estimate for that of the subsequent problem in high-frequency control. However, active-set methods can suffer from issues like repeated constraint activation and deactivation (cycling) when constraints are linearly dependent, leading to long computation times.
Interior-point methods are numerically robust alternatives to the active-set method, for example by integrating proximal regularization which enables stable convergence even in the case of linearly dependent constraints [12]. An extension to SHQPs, which are a member of the broad class of QPs, has been proposed in [4]. Linear dependence on the number of decision candidates as in the unconstrained case [13] is maintained by efficient exploitation of the underlying sparse hierarchical structure introduced by -norm surrogates.
Warm-starting interior-point methods is less direct than for active-set methods. Some techniques modify previous solutions to meet central feasibility conditions before restarting the method [14], which limits the direct use of a preceding solution.
Alternating Direction Method of Multipliers (ADMM)
First-order methods such as ADMM update primal and dual variables through gradient and projection operations with factorization reuse at the cost of reduced accuracy, rather than repeatedly forming and factorizing KKT systems as required by second-order methods such as the IPM. Their computational efficiency can be further improved through effective warm-starting and by caching factorizations and other quantities which remain invariant across inner iterations [15] and consecutive control steps [6]. ADMM has been successfully deployed as an SQP subproblem solver for complex whole-body trajectory optimization, while efficiently exploiting temporal sparsity of optimal control problems [16].
In this work, we adapt the ADMM to the specific sparse and hierarchical structure of SHQPs in order to maintain linear dependency of the solver in the number of candidates in hierarchical decision-making.
Background
Sparse Hierarchical Nonlinear Programming for Hierarchical Decision-Making
This work formulates a continuous optimization problem for hierarchical decision-making with a more aggressive sparsity promoting approximation of the -norm with strong gradient decay [17] (see Figure 2). This can be advantageous in the presence of a large number of candidates, when perpetual influence as seen for the logarithmic penalty [4, 9] is undesired. For completeness, and unlike the work in [4], we also consider -norm objectives:

Figure 2. Comparison between continuous log [9] and fraction function surrogates [17] of the -norm on squared decision functions (see (1)).
s.t.
is the optimization variable, for example representing the robot joint state. The functions denote nonlinear robot constraints associated with the index sets , , , or . Each scalar entry in represents a decision candidate, for example a squared expression of kinematic translational reaching with a constant desired value (see Sec. VI on how this separable structure can be exploited computationally)
The set contains least-squares tasks, while inactive or active tasks from higher priority levels are represented in , respectively. These sets may include both equality and inequality constraints, collectively denoted by the relation .
The constraint sets and are associated with a slack variable , which permits relaxation when exact satisfaction is not possible, for example when a task specifies an unreachable target. On each priority level , the slack is optimized until the optimal one is found, in which case the associated constraints enter the active set of the next level with fixed optimal slack . The auxiliary variable , together with the appropriate bound constraint set on , provides a smooth representation of the elementwise magnitude . For compactness, parentheses denote the stacked vector of quantities indexed by a given set, for example .
In the following, the explicit set memberships , , , and are omitted whenever they are clear from the context, for example .
Sequential Sparse Hierarchical Quadratic Programming
S-HNLP is solved by iteratively solving SHQP subproblems [4] for the approximate step variables , , and . The step is subject to a trust region of radius for globalization. In this work focusing on instantaneous robot control, the trust region is kept constant while a filter framework as in [18, 19] is omitted. SHQP is a constrained version of the reweighted majorization-minimization -optimization problem of [9]:
The weights are defined as follows, with a small constant which improves numerical robustness:
The positive-definite hierarchical Lagrangian Hessian [20] captures second-order terms of the constraints and objectives, while and represent linearizations of S-HNLP at the current iterate . Following the approaches in [21, 4], we introduced the change of variables: with the overall primal solution: . Here, is the optimal primal of levels to , and is a nullspace basis of the vectors (; ). A vector or matrix marked by represents the projected value (or ). The value denotes the nullspace projection offset from previous priority levels: . is the projected primal of level , with for efficiency.
Below are the primal and dual residuals and as the derivative of the Lagrangian of a level of SHQP [4]. For convenience, we define the placeholder
such that denotes the Lagrange multiplier associated with the corresponding constraint class. Specifically, represents , , , or for , respectively. Additionally, the constraint vectors , , and are stacked row-wise into the matrices , , and , according to their respective index sets , , and . The -norm auxiliary constraints , the inequality constraints in and , and are relaxed by positive slack variables (marked by checkered boxes):
An ADMM for SHQP
ADMM solves problems of the following form [15]:
s.t.
Applying it to SHQP, we get the following variable and split variable vectors and (excluding ):
The slack variables relax inequality constraints, while the operator projects them onto the set of positive numbers . The matrices and vectors are composed as:
The augmented Lagrangian of level of (6) is:
The parameters and are the individual step parameters for each individual constraint and . The component is variable and its update rule is described in Sec. V-B, while are fixed hand-tuned values for each different constraint set in ( for equalities, for inequalities).
The ADMM computes the primal (see Sec. IV-A), the split variables (see Sec. IV-B), and the Lagrange multipliers (see Sec. IV-C) in alternating fashion. Once the alternating steps of level have converged with , the active constraint set and corresponding to activity in , , and are composed. The algorithm then continues with the ADMM of the next level projected into the nullspace of the new active set . The remaining inactive constraints are contained in the updated inactive set . More details on convex hierarchical programming can be found in [22].
Primal update
The parenthesized superscript (abbreviation for ) denotes values of the next ADMM iteration based on the values of the current one (omitted for readability). The primal update follows from:
This can be rewritten into the linear system:
where:
and:
The matrices are positive-definite and invertible. The overall matrix is then positive definite as well for and positive-definite . Note that computing is the most computationally demanding operation of the algorithm, but only needs to be done once, even in case of updates due to the separable nature of .
The right hand side writes as:
The vector can be directly computed by:
Furthermore, the following substitutions of and in dependence of can be computed by:
Finally, is efficiently computed from:
for example using a Cholesky decomposition of the normal-matrix :
Split variable update
The split variable update results from the condition :
This leads to the updates and:
Finally, the operator is realized by:
Dual update
The Lagrange multipliers are updated by dual ascent [5], which writes as :
Specifically, this leads to the updates:
Computational comparison
Each iteration of our solver (referred to as ADM), IPM [4], PIQP [12], and OSQP [15] involves the sets of operations summarized in Table 1. Due to the lack of elimination of -norm auxiliaries, the complexities of both PIQP and OSQP are strongly influenced by the number of constraints (short of cubic due to sparse linear algebra implementations), which is typically large in decision-making. On the other hand, both ADM and IPM are linearly dependent on , while ADM needs to conduct the expensive normal-matrix assembly (23) of order only once, even if there are step-size parameter updates.
Table 1. Computational cost over iterations for L0-equality constraints . Solvers marked with * are implemented using sparse linear algebra, such that the computational cost is typically governed by sparsity rather than the raw matrix dimensions. For qpOASES, denotes the number of active-set changes, and is the number of linearly independent active constraints.
Further ADMM components
Equilibration
We apply Ruiz equilibration [23] for preconditioning in order to reduce the number of iterations . Here, matrix rows and columns are iteratively scaled by their inverse square root maximum absolute value until they equal to 1. For QPs, it is suggested to scale the Lagrangian Hessian [15]. In the specific case of SHQPs, this leads to the scaled matrix (symmetric indicated by ): is a diagonal matrix containing the scaling factors:
Gray entries indicate the individual Hessian rows and columns as derivatives with respect to the given vector. Overlined matrices and vectors are the equilibrated equivalents of their original counterparts. The matrices , , , , and are scaled diagonal matrices.
In the following, we make an observation which leads to simplifications in the scaled version of (11), as off-diagonal elements of the form are canceled out. We use the operator which returns a vector containing the maximum value of each row of a matrix, or of a row / column block of .
Proposition 1. Throughout the Ruiz equilibration of , it holds that and therefore .
Proof. Initially, , so follow immediately. At the same time, since for any maximum row value , we have that . We then clearly have . In the next Ruiz iteration, still holds due to the unchanged and . By induction over the Ruiz iterations, the claims follow.
Step-size parameter updates
We update the step-size parameter (initially, ) by the update rule presented in [15], where the operator returns the absolute maximum value of all additive components in and :
It balances reduction in the normalized primal and dual residuals and , respectively. If is updated, (22) needs to be updated and re-factorized.
Over-relaxation
In over-relaxed ADMM, the relaxed quantity is used in the split-variable and dual-variable updates (Sec. IV-B and Sec. IV-C) in order to reduce the number of iterations [24]. The parameter is the over-relaxation parameter (typically [15]).
Warm-starting
Between problem instances, for each priority level the previously found step-size parameter , and the unscaled primal, split, and dual variables , , and , respectively, are stored. A level is only warm-started if the current active and inactive constraint sets and are identical with the ones of the previous problem instance at that level and , while standard initialization is used otherwise.
Separable decision functions
The separable structure of decision functions of the form as in (1) can be exploited for computational gain. One row of the Jacobian of writes as:
We use . Then, a matrix product evaluated in (11) can be rewritten to:
Evaluating this matrix naively by with is of order , so the large factor is amplified by . Instead, by first computing , the overall evaluation is of order , where is amplified by only . The factor is typically small in robotics, for example for translational reaching.
Evaluation
We compare the proposed solver ADM (cold and warm) with IPM [4] and PIQP [12] (both no warm-start), and OSQP [15] and qpOASES [11] (both cold and warm). PIQP, OSQP, and qpOASES solve SHQP directly without eliminating the auxiliary variables and the related constraints. Except for the walking scenario (Sec. VII-D), all solvers terminate at . Timings purely reflect the solve times, while problem setup / update times, ... are excluded. The robot SHQPs are composed by Pinocchio [25], with PD controllers expressed in the velocity domain [20] (Sec. VII-C and Sec. VII-D). Evaluations are run on an Intel Core Ultra 9 185H CPU with 64 GB RAM.
Large-scale LPs
We first isolate the benefit of exploiting sparse structure over general-purpose solvers based on sparse linear algebra [12, 15], independently of hierarchical effects. For this, we solve large-scale linear programs (LPs) corresponding to the case of and in SHQP. Equality constraints are composed as randomized, dense matrices with and (, increment: 1000). For each , we repeatedly () solve slightly disturbed problems in order to simulate warm-starting, where is a randomized dense matrix with elements in the range of .
Figure 3 shows that the proposed warm-started solver ADM achieves a mean speedup of around with respect to IPM (123 29 and 16 1 iterations, s and s solve time over all runs for ADM and IPM, respectively). On the contrary, the reference solvers PIQP and OSQP are between one to two orders of magnitude slower. qpOASES is significantly slower due to its dense linear algebra implementation, but exhibits a significant drop in active-set changes (typically 0 to 3 iterations) after the initial cold solve. While PIQP does not possess explicit warm-starting functionality, its iteration count is low at iterations, which emphasizes the need to exploit the structure of sparse programs. PIQP, OSQP, and qpOASES could only be evaluated up to constraints due to their increased memory requirements ( variables and more constraints) as no elimination of the sparse auxiliaries takes place.

Figure 3. Speedup of different solvers with reference to IPM on large-scale LPs with , , and equality constraints.
Humanoid inverse kinematic control
| 1 | Vel. lim. / trust region | |
| 2 | Angle lim. | |
| 3 | Stance | |
| 4 | Coll. av. | |
| 5 | Goal DM |
Table 2. Inverse kinematic control of humanoid robot, task hierarchy.

Figure 4. Inverse kinematic control of humanoid robot, solve times per control iteration.
Here, we evaluate warm-starting and the linear algebra implementation for separable decision functions as outlined in Sec. VI. A simulation is designed where the velocity controlled humanoid robot Unitree G1 needs to touch as many as possible passing objects out of 500.
Figure 4 shows that after 200 iterations of warm-starting (maximum solve time: 4.1 ms, : 3.0 ms), the solver reaches a steady level of around ms (: ), with being 23.5% faster than over all control iterations. (Sec. VI) and (Sec. VI), which exploit the structure of the decision functions according to Sec. VI, improve the solve time per control iteration by 10% and 12% over their baselines, respectively.
Drone control
| 1 | Vel. lim. / trust region | |
| 2 | Force lim. | |
| 3 | Coll. av. | |
| 4 | Dynamics | |
| 5 | Goal DM | |
| 6 | Goal tracking |
Table 3. Drone control, task hierarchy.
In this example, we create an SH-NLP with and for navigation of a drone with linear dynamics through an obstacle course to choose one of 100 targets (see Figure 5). The problem is solved with an instantaneous controller with ms. The drone is described by , where is the current goal location, is the current drone translation [4], and is the control force. Since the drone is controlled via a proportional-derivative (PD) controller expressed in

Figure 5. Instantaneous hierarchical control of a drone to navigate through an obstacle course, while deciding on one of 100 possible goal locations.

Figure 6. Drone control, number of inner iterations per control iteration.
the velocity domain [20], imposing the decision-making directly on acceleration commands is less effective and does not accurately reflect the true decision-making error to the physical target. Instead, the decision-making is imposed on a separate goal position which chooses one of the goal locations . The goal location is then realized by a PD controller on the last priority level. Further constraints on velocities (max 0.5 m/s), control forces (6 N and 18 N max. horizontal and vertical), and collision avoidance complete the hierarchy.
Figure 6 and Table 4 show how is warm-started very effectively, where in many iterations the controller applies the same control in subsequent iterations. In contrast, warm-started OSQP still requires around 200 iterations, while qpOASES reaches steady-state after around 24 control iterations.
Table 4. Drone control, comparison of number of inner iterations and solve time (ms) per control iteration.
Humanoid whole-body control for walking over stepping stones
In this experiment, we demonstrate the integration of hierarchical decision-making within a dynamic humanoid walking controller see Figure 1. Both the left and right feet have to make a decision on a candidate set of stepping stones to step on next, while the other foot is fixed on the previously chosen stepping stone. Whole-body dynamics, joint torque limits, 6D contact wrench conditions, collision avoidance, and DCM stability constraints [26] complete the hierarchy given in Table 5. The variable vector is composed of the goal location , joint velocities , joint torques , and left and right foot contact wrenches . The problem can only be faithfully solved under high-accuracy settings (), which is less suitable for first-order methods like NADM and OSQP. NADM’s maximum iteration count is set to , making its computation time of ms per control iteration not competitive with NIPM’s and PIQP’s runtime of ms and ms, respectively. This highlights the existence of problem classes which require high-accuracy solution regimes.

Figure 1. Unitree G1 whole-body control with integrated hierarchical decision-making for walking over 200 stepping stones (current decision: yellow sphere; Sec. VII-D).
| 1 | Angle lim. / Vel. lim. | |
| 2 | Torque lim. | |
| 3 | Dynamics | |
| 4 | Wrench cone | |
| 5 | Stance-foot | |
| 6 | Coll. av. | |
| 7 | DCM / upright torso | |
| 8 | Goal DM | |
| 9 | Swing-foot | |
| 10 | Regularization |
Table 5. Humanoid whole-body control for walking over stepping stones, task hierarchy.
Conclusion
In this work, we have presented an ADMM solver for hierarchical decision-making control, which effectively exploits temporal similarities between nonlinear solver subproblems. Evaluation shows that in low- to moderate-accuracy regimes, the solver achieves speedups of up to 23.5% over the structure-exploiting IPM [4]. At the same time, complex scenarios like humanoid walking expose the limitations of ADMM based solvers when high-accuracy solutions are required.
Acknowledgments
This work was supported by the RDF-24-02-060 XJTLU Research Development Fund.
References
- [1]D. Song, P. Fernbach, T. Flayols, A. D. Prete, N. Mansard, S. Tonneau, and Y. J. Kim, “Solving footstep planning as a feasibility problem using ℓ1-norm minimization,” IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 5961–5968, 2021.DOI
- [2]T. Dinev, W. Merkt, V. Ivan, I. Havoutis, and S. Vijayakumar, “Sparsity-inducing optimal control via differential dynamic programming,” in IEEE International Conference on Robotics and Automation, 2021, pp. 8216–8222.DOI
- [3]J. Yu, J. Yan, Z. Jin, C. Hu, and Z. Wang, “Multi-priority reactive motion control for safe and coordinated dual-arm manipulation in dynamic environments,” IEEE Robotics and Automation Letters, vol. 11, no. 6, pp. 7596–7603, 2026.DOI
- [4]K. Pfeiffer, Q. Zhang, Y. Chen, G. O. Boateng, Y. Wang, V. Bonnet, and A. Kheddar, “Integrated Hierarchical Decision-Making in Inverse Kinematic Planning and Control,” in Proceedings of Robotics: Science and Systems, Sydney, Australia, July 2026.arxiv.org/abs/2412.01324
- [5]S. Boyd, N. Parikh, E. Chu, B. Peleato, and J. Eckstein, “Distributed optimization and statistical learning via the alternating direction method of multipliers,” Foundations and Trends in Machine Learning, vol. 3, pp. 1–122, 01 2011.
- [6]I. Mahajan, K. Nguyen, S. Schoedel, E. Nedumaran, M. Mata, B. Plancher, and Z. Manchester, “Code generation and conic constraints for model-predictive control on microcontrollers with conic-timypc,” in IEEE International Conference on Robotics and Automation (ICRA), 2026.arxiv.org/abs/2403.18149
- [7]Y. Shirai, A. Raghunathan, and D. K. Jha, “Hierarchical contact-rich trajectory optimization for multi-modal manipulation using tight convex relaxations,” in 2025 IEEE International Conference on Robotics and Automation and Automation (ICRA), 2025, pp. 9199–9205.DOI
- [8]V. Gonçalves, P. Fraisse, A. Crosnier, and B. Adorno, “Parsimonious kinematic control of highly redundant robots,” IEEE Robotics and Automation Letters, vol. 1, no. 1, pp. 65–72, 2016.
- [9]E. Candès, M. Wakin, and S. Boyd, “Enhancing sparsity by reweighted ℓ1 minimization,” Journal of Fourier Analysis and Applications, vol. 14, no. 5, pp. 877–905, 2007.arxiv.org/abs/0711.1612
- [10]P. E. Gill, W. Murray, and M. A. Saunders, “Snopt: An sqp algorithm for large-scale constrained optimization,” SIAM Rev., vol. 47, no. 1, pp. 99–131, jan. 2005.
- [11]H. Ferreau, C. Kirches, A. Potschka, H. Bock, and M. Diehl, “qpOASES: A parametric active-set algorithm for quadratic programming,” Mathematical Programming Computation, vol. 6, no. 4, pp. 327–363, 2014.DOI
- [12]R. Schwan, Y. Jiang, D. Kuhn, and C. N. Jones, “PIQP: A proximal interior-point quadratic programming solver,” in IEEE Conference on Decision and Control, 2023, pp. 1088–1093.arxiv.org/abs/2304.00290
- [13]K. Koh, S.-J. Kim, and S. Boyd, “An interior-point method for large-scale ℓ1-regularized logistic regression,” Journal of Machine Learning Research, vol. 8, no. 54, pp. 1519–1555, 2007.
- [14]J. Gondzio and A. Grothey, “A new unblocking technique to warmstart interior point methods based on sensitivity analysis,” SIAM Journal on Optimization, vol. 19, no. 3, pp. 1184–1210, 2008.DOI
- [15]B. Stellato, G. Banjac, P. Goulart, A. Bemporad, and S. Boyd, “OSQP: an operator splitting solver for quadratic programs,” Mathematical Programming Computation, vol. 12, no. 4, pp. 637–672, 2020.
- [16]A. Jordana, S. Kleff, A. Meduri, J. Carpentier, N. Mansard, and L. Righetti, “Structure-exploiting sequential quadratic programming for model-predictive control,” IEEE Transactions on Robotics, vol. 41, pp. 4960–4974, 2025.DOI
- [17]H. Li, Q. Zhang, A. Cui, and J. Peng, “Minimization of fraction function penalty in compressed sensing,” IEEE Transactions on Neural Networks and Learning Systems, vol. 31, no. 5, pp. 1626–1637, 2020.DOI
- [18]R. Fletcher, S. Leyffer, and P. L. Toint, “On the global convergence of a filter–sqp algorithm,” SIAM Journal on Optimization, vol. 13, no. 1, pp. 44–59, 2002.DOI
- [19]K. Pfeiffer and A. Kheddar, “Sequential hierarchical least-squares programming for prioritized non-linear optimal control,” Optimization Methods and Software, vol. 39, no. 4, pp. 1104–1142, 2024.DOI
- [20]K. Pfeiffer, A. Escande, P. Gerondet, and A. Kheddar, “The hierarchical newton’s method for numerically stable prioritized dynamic control,” IEEE Transactions on Control Systems Technology, vol. 31, no. 4, pp. 1622–1635, 2023.arxiv.org/abs/2303.04428
- [21]T. V. Dang, K. V. Ling, and J. Maciejowski, “Banded null basis and admm for embedded mpc,” IFAC-PapersOnLine, vol. 50, no. 1, pp. 13 170–13 175, 2017, 20th IFAC World Congress.DOI
- [22]K. Pfeiffer, A. Escande, and L. Righetti, “νIPM-HLSP: an efficient interior-point method for hierarchical least-squares programs,” Optimization and Engineering, vol. 25, no. 2, pp. 759–794, 2024.DOI
- [23]P. A. Knight, D. Ruiz, and B. Uçar, “A symmetric preserving algorithm for matrix scaling,” SIAM Journal on Matrix Analysis and Applications, vol. 35, no. 3, pp. 931–955, 2014.DOI
- [24]E. Ghadimi, A. Teixeira, A. Shames, and M. Johansson, “Optimal parameter selection for the alternating direction method of multipliers (admm): Quadratic problems,” IEEE Transactions on Automatic Control, 09 2014.
- [25]J. Carpentier, G. Buondonno, J. Mirabel, F. Lamiraux, O. Stasse, and N. Mansard, “The Pinocchio C++ library—a fast and flexible implementation of rigid body dynamics algorithms and their analytical derivatives,” in IEEE/ISICE International Symposium on System Integration, 2019, pp. 614–619.DOI
- [26]J. Englsberger, C. Ott, and A. Albu-Schäffer, “Three-dimensional bipedal walking control based on divergent component of motion,” IEEE Transactions on Robotics, vol. 31, no. 2, pp. 355–368, Apr. 2015.DOI