Searcharxiv⌕ Search

arXiv subjects

Till Preuster

Publications and source records attributed to Till Preuster.

11 recordsLinked to original sources

Cost controllability and predictive control of input-constrained monotone systems on Hilbert spaces

Cost controllability, relating controllability and a performance measure like the stage cost in model predictive control (MPC), is a key tool in the stability and suboptimality analysis of closed-loop systems. We establish cost controllability for maximal monotone control systems subject to input constraints, a class encompassing finite- and infinite-dimensional port-Hamiltonian systems with colocated input and output. To this end, we employ a recently proposed saturated feedback to derive state-dependent bounds on the closed-loop cost. Our main result is that cost controllability holds globally with semiglobal exponential rate. We first establish this result in continuous time and then extend it to sampled-data systems. Whereas a qualitative detectability-like property is sufficient for asymptotic convergence of the controller, cost controllability requires quantitative rates. We provide two verifiable conditions ensuring cost controllability with explicit exponential rates: coercivity of the sum of system operator and output injection, and exponential detectability of the linear part. Moreover, under the coercivity condition, the resulting MPC scheme without terminal conditions renders the equilibrium globally asymptotically stable with semiglobal exponential rate already for prediction horizon two. We illustrate our results by a nonlinear finite-dimensional system, a semilinear heat equation and a nonlinearly damped wave equation.

math.OC↗

Abstract second-order boundary control systems

Many physical systems, such as vibrating elastic bodies, acoustic media and electromagnetic fields, are governed by second-order evolution equations $\ddot x+D\dot x+Sx=0$ and interact with their environment through their boundary. The classical analysis via the operator matrix $\mathcal{A}=\left[\begin{smallmatrix}0&I\\-S&-D\end{smallmatrix}\right]$ requires $S$ to be self-adjoint, which typically fixes a boundary condition a priori and thus excludes boundary control. We therefore use $S$ only to define the energy space and govern the dynamics by a maximal operator $S_{\max}\supset S$ whose domain carries no boundary condition. Based on a factorization $S=A^*A$, we construct a boundary triplet for the associated operator matrix and define boundary inputs and outputs of Robin type, parametrized by an operator $P$ on the boundary space. We show that the resulting system is impedance passive and internally well-posed if $P$ is a contraction, and that for $D=0$ this covers all boundary conditions under which the free dynamics generate a contraction semigroup. Moreover, we show that the position-momentum formulation (e.g., $\left[\begin{smallmatrix}0&I\\Δ&0\end{smallmatrix}\right]$ for the wave equation) and the strain-momentum formulation (e.g., $\left[\begin{smallmatrix}0&\nabla\\\operatorname{div}&0\end{smallmatrix}\right]$) describe the same open system, including its boundary inputs and outputs. We illustrate our results by an $n$-dimensional wave equation and Maxwell's equations.

math.AP↗

Dual companions and improved regularity of Lagrange multipliers in optimal control

We investigate optimal control problems whose state equation takes the form $\mathcal{A}x-\mathcal{B}u=b$, where $\mathcal{A}:\mathcal{X}\to\mathcal{P}'$ is an isomorphism from a Banach state space $\mathcal{X}$ onto the dual of a reflexive Banach space $\mathcal{P}$. The adjoint equation involves the Banach space adjoint $\mathcal{A}':\mathcal{P}\to\mathcal{X}'$, so that, in general, state and multiplier are of different regularity. For particular classes of partial differential equations, the typically weaker regularity of the multiplier is improved a posteriori by bootstrapping. In this work, we provide an abstract foundation for this methodology on the operator level. More precisely, we show that whenever the state cost extends to a differentiable function on a Hilbert space $\mathcal{H}$ in which $\mathcal{X}$ is densely embedded, as for $L^2$-tracking costs on Sobolev-type state spaces, the multiplier automatically lies in a subspace of $\mathcal{P}$, the dual companion, on which $\mathcal{A}'$ restricts to an isomorphism onto $\mathcal{H}'$. This yields an optimality system with balanced regularity, where state and multiplier solve equations of the same type. In applications, the dual companion can be identified by integration by parts in time or space. We illustrate the framework by linear elliptic and parabolic equations, Cauchy problems governed by general $C_0$-semigroups, and a wave equation with accumulated-state observation.

math.OC↗

Nonlinear controlled port-Hamiltonian systems: Existence of (optimal) solutions

We investigate the existence of solutions of reversible and irreversible port-Hamilto\-nian systems. To this end, we utilize the associated exergy, a function that is composed of the system's Hamiltonian and entropy, to prove global existence in time for bounded control functions. Then, we rigorously verify our existence conditions for two examples, the gas-piston system and a network of heat exchangers. Last, we explore model predictive control tailored to irreversible port-Hamiltonian systems by means of a numerical case study with a heat exchanger network.

math.OC↗

Optimization-based control by interconnection of nonlinear port-Hamiltonian systems

In this paper, we develop a control-by-interconnection approach for the stabilization of nonlinear port-Hamiltonian systems. Motivated by model predictive control, the controller is realized as a continuous-time primal-dual gradient flow associated with a finite-horizon, control-constrained optimal control problem. By exploiting its port-Hamiltonian structure, the optimization dynamics are interconnected with the nonlinear plant. Introducing a time-scale parameter for the optimization dynamics reveals a singularly perturbed closed-loop system, whose reduced dynamics correspond to the optimal finite-horizon feedback, while the boundary-layer dynamics capture convergence of the primal-dual variables to the optimality system. Using a composite Lyapunov function and singular-perturbation arguments, we prove asymptotic stability of the coupled plant-controller dynamics, which is local for sufficiently fast optimization dynamics and becomes global under additional assumptions. Numerical experiments for a nonlinear port-Hamiltonian oscillator illustrate the results.

math.OC↗

Stabilize-then-optimize: Feedback transformations as preconditioners in optimal control

Many numerical algorithms for optimal control leverage an elimination of the state via the control-to-state map such as condensed approaches or preconditioned conjugate gradient methods for the optimality system. As such, the norm of the control-to-state map directly enters the convergence estimates for these methods, e.g., via the condition number of the associated linear system. In this work we show that using feedback transformations one may reformulate the optimal control problem to decrease the norm of the (feedbacked) control-to-state map, leading to a drastic improvement of the involved condition numbers. We illustrate the abstract approach for ordinary and partial differential equations such as parabolic, hyperbolic or elliptic equations. For each of these problem classes we provide a constructive method to improve solution operator norms via feedbacks. Further, we showcase the efficacy of the method by means of various numerical examples with elliptic, parabolic and hyperbolic partial differential equations.

math.OC↗

Coupling optimization algorithms and monotone control systems: Suboptimal model predictive control as an operator splitting scheme

We propose a framework for suboptimal model predictive control (MPC) based on the interconnection of monotone dynamical systems, such as port-Hamiltonian systems. In contrast to classical MPC formulations, where the optimizer is treated as an instantaneous mapping, we model both the plant and the optimizer as dynamical systems and couple them through a structured interconnection. This leads to a continuous-time closed-loop formulation governed by (quasi-)monotone operators. Within this setting, we establish well-posedness of the coupled optimizer-plant dynamics and provide a unified interpretation of suboptimal MPC schemes. In particular, we reveal a direct connection between iterative optimization algorithms and dynamical control systems theory by showing that standard suboptimal MPC algorithms can be understood as time discretizations of the underlying continuous-time dynamics via operator splitting methods.

math.OC↗

Stabilization of monotone control systems with input constraints

We present a stabilizing output-feedback controller for nonlinear finite and infinite-dimensional control systems governed by monotone operators that respects given input constraints. In particular, we show under a detectability-like assumption that a saturated version of the classical output feedback controller in passivity-based control achieves control-constrained stabilization as long as the control corresponding to the desired equilibrium is in the interior of the control constraint set. We illustrate our findings using a heat equation, a wave equation, and a finite-dimensional nonlinear port-Hamiltonian system.

math.OC↗

Extension theory via boundary triplets for infinite-dimensional implicit port-Hamiltonian systems

The solution of constrained linear partial-differential equations can be described via parametric representations of linear relations. To study these representations, we provide a novel definition of boundary triplets for linear relations in range representations where the associated boundary map is defined on the domain of the parameterizing operators rather than the relation itself. This allows us to characterize all boundary conditions such that the underlying dynamics is represented by a self-adjoint, skew-adjoint or maximally dissipative relation. The theoretical results are applied to a class of implicit port-Hamiltonian systems on one-dimensional spatial domains. More precisely, we explicitly construct a boundary triplet which solely depends on the coefficient matrices of the involved matrix differential operators and we derive the associated Lagrangian subspace. We exemplify our approach by means of the Dzektser equation, the biharmonic wave equation, and an elastic rod with non-local elasticity condition.

math.AP↗

Jet space extensions of infinite-dimensional Hamiltonian systems

We analyze infinite-dimensional Hamiltonian systems corresponding to partial differential equations on one-dimensional spatial domains formulated with formally skew-adjoint Hamiltonian operators and nonlinear Hamiltonian density. In various applications, the Hamiltonian density can depend on spatial derivatives of the state such that these systems can not straightforwardly be formulated as boundary port-Hamiltonian system using a Stokes-Dirac structure. In this work, we show that any Hamiltonian system of the above class can be reformulated as a Hamiltonian system on the jet space, in which the Hamiltonian density only depends on the extended state variable itself and not on its derivatives. Consequently, well-known geometric formulations with Stokes- Dirac structures are applicable. Additionally, we provide a similar result for dissipative systems. We illustrate the developed theory by means of the the Boussinesq equation, the dynamics of an elastic rod and the Allen-Cahn equation.

math.AP↗