https://opus4.kobv.de/opus4-trr154

Domain Decomposition for Gas Network Control

Falk Hante; Martin Hernandez; Günter Leugering; Martin Schmidt; Enrique Zuazua — Mon, 01 Jun 2026 23:52:41 +0200

We present domain decomposition techniques for efficient and scalable simulation and optimization of partial differential equations on network domains. Motivated by transient gas network applications, the state-of-the-art of space-and-time decompositions for systems derived from Euler's equations are presented. Moreover, combinatorial aspects of switching valves are addressed in a decomposition framework for mixed-integer and ODE-constrained problems and a convergence analysis is presented. Motivated by stochastic approximation, we also propose and explore randomized decomposition approaches. Some of these advanced solution strategies can be combined and regarded as variants of certain penalty alternating direction methods. We present the potential applications of these techniques by summarizing numerical results from the literature for transient GasLib instances modeling a realistic gas transport network.

An Adjoint Calculus for Optimal Control of Entropy Solutions of the Generalized Riemann Problem for Hyperbolic Systems of Conservation Laws with Shocks

Jannik Breitkopf; Stefan Ulbrich — Fri, 29 May 2026 13:51:33 +0200

We derive an adjoint gradient representation for the objective functional of an optimal control problem of the unique entropy solution to the Generalized Riemann Problem (GRP) for strictly hyperbolic systems of conservation or balance laws. The GRP is an initial value problem where the initial state is piecewise C^1 with exactly one discontinuity. In a recent work, differentiability properties of the solution operator of the GRP were derived. In particular, the differentiability of a class of tracking-type functionals was shown. We build on these results to derive an adjoint gradient representation for such functionals. The adjoint problem is a system of linear transport equations with a discontinuous coefficient and discontinuous terminal state. We prove the existence and uniqueness of a piecewise Lipschitz continuous solution to the adjoint problem which satisfies suitable interior boundary conditions along all shock curves. The adjoint state generally contains infinitely many discontinuities but its total variation remains uniformly bounded in time.

A discrete gradient scheme for preserving QSR-dissipativity

Attila Karsai; Philipp Schulze — Thu, 21 May 2026 12:27:22 +0200

The notion of dissipative dynamical systems provides a formal description of processes that cannot generate energy internally. For these systems, changes in energy can only occur due to an external energy supply or dissipation effects. Unfortunately, dissipative properties tend to deteriorate in numerical computations, especially in nonlinear systems. Discrete gradient methods can help mitigate this problem. In this paper, we present a class of structure-preserving time discretization schemes based on discrete gradients for a special class of systems that are dissipative with respect to a quadratic supply rate.

Structure-Preserving Discretization and Model Reduction for Energy-Based Models

Robert Altmann; Attila Karsai; Philipp Schulze — Thu, 21 May 2026 12:27:21 +0200

We investigate discretization strategies for a recently introduced class of energy-based models. The model class encompasses classical port-Hamiltonian systems, generalized gradient flows, and certain systems with algebraic constraints. Our framework combines existing ideas from the literature and systematically addresses temporal discretization, spatial discretization, and model order reduction, ensuring that all resulting schemes are dissipation-preserving in the sense of a discrete dissipation inequality. For this, we use a Petrov-Galerkin ansatz together with appropriate projections. Numerical results for a nonlinear circuit model and the Cahn-Hilliard equation illustrate the effectiveness of the approach.

Nonlinear energy-based systems: modeling, control and numerical realization

Attila Karsai — Thu, 21 May 2026 12:26:44 +0200

Dynamical systems consist of ordinary and partial differential equations and are among the most prominent approaches to model physical processes. They describe the evolution of the system in terms of the current system state and external control inputs. To improve model accuracy, it can be beneficial to explicitly include properties such as energy conservation or energy dissipation, which are usually present in the real-world phenomena, in the mathematical problem description. Ideally, these properties should be kept in mind whenever one interacts with the model. The focus of this thesis is on two kinds of interactions: the discretization of such models, and their control. Discretization techniques are necessary whenever the system evolution is to be approximated using computational methods. Similarly fundamental is the numerical realization of control inputs that lead to desired outcomes. Both aspects require special attention to retain an energy-based perspective. For this, the first step is usually to encode the energy properties in the algebraic description of the model. This description must strike a balance between general applicability and its corresponding benefits. The next step is to leverage the algebraic description in further analysis. While the field is well-developed for linear systems, the nonlinear case often poses additional difficulties. First, there seems to be no clear consensus about what model class to use to describe nonlinear physical phenomena. Second, many discretization methods that preserve the energy-based viewpoint in the linear case are not trivial to generalize to nonlinear systems. Third, studying the behavior of energy-optimal controls and finding control laws that can be realized via energy-based models is usually more difficult. In this thesis, these points are addressed. We give an overview of energy-based model classes used for nonlinear phenomena in both finite and infinite dimensions and investigate their relationship. Moreover, using a modified Petrov--Galerkin method and discrete gradients, we present multiple structure-preserving discretization schemes. Furthermore, we show that similar to the linear case, energy-optimal controls steer the associated trajectories to the submanifold of the state space where no dissipation is present. Finally, we combine an optimal feedback law characterized by the Hamilton--Jacobi--Bellman equation with output feedback to state an energy-based feedback controller. Our theoretical results are illustrated using numerical experiments.

121 Patchworked Curves of Degree Seven

Thu, 21 May 2026 12:19:59 +0200

The 121 real schemes, i.e., ambient isotopy classes, of smooth real plane algebraic curves of degree seven were classified by Viro (1984). By constructing one patchwork of the dilated triangle 7⋅Δ2 for each real scheme, we provide an explicit method for constructing polynomials realizing each real scheme. In particular, every real scheme of degree seven can be realized as a T-curve; this settles a question raised by Itenberg and Viro (1996).

Fast Isotopy Computation for T-Curves

Thu, 21 May 2026 12:18:54 +0200

A T-curve of degree d is given by a regular unimodular triangulation of d⋅Δ2 together with a sign distribution on its lattice points. By Viro's Patchworking Theorem, this determines the ambient isotopy type (a.k.a. real scheme) of a smooth real plane projective algebraic curve of the same degree. We present a near-quadratic time algorithm for extracting that isotopy type from the triangulation and the signs. Through a GPU-accelerated implementation, this allows one to compute billions of real schemes per second, enabling exhaustive enumeration at scale. This algorithm was essential for our recent construction of all 121 real schemes of degree seven by T-curves.

Clash of MINLP Relaxations: Piecewise Linear vs. Global Parabolic

Adrian Göß — Wed, 13 May 2026 08:50:15 +0200

Solving mixed-integer nonlinear programs (MINLPs) typically relies on constructing relaxations that are easier to tackle than the original problem. Recently, global parabolic (PARA) relaxations were introduced, featuring separable quadratic functions – paraboloids – as global under- or overestimators of general nonlinear constraint functions. So far, the paraboloids are all computed at once by solving a mixed-integer linear program (MIP). For small tolerances or wide function domains, the corresponding MIP grows in size and is eventually intractable, preventing a meaningful comparison with established relaxation techniques. We therefore propose a novel iterative method to compute PARA approximations that succeeds on all tolerance-domain combinations where the original one has failed. The computational study is preceded by a thorough theoretical explanation and analysis. Finally, the improved method enables a computational comparison with piecewise linear (PWL) relaxations in terms of runtime on general MINLP instances. The results show that the modern solver SCIP can solve PWL relaxations faster when the olerance is high, shifting strongly in favor of PARA for tighter tolerances. We attribute the effect to the difference in the corresponding problem size: PWL relaxations introduce binary variables to identify the active linear piece and their number grows with decreasing tolerance. PARA, on the other hand, does not require additional variables such that the dimension is maintained. For problems with at least one (co)sine constraint, the effect significantly amplifies. Thereby, for medium tolerances, PARA relaxations outperform SCIP stand-alone. Applied problems like alternating current optimal power flow (AC-OPF) feature such constraint types, leaving PARA a viable relaxation strategy.

Approximating the Network Design Problem for Potential-Based Flows

Max Klimm; Marc E. Pfetsch; Martin Skutella; Lea Strubberg — Tue, 05 May 2026 16:31:59 +0200

We develop efficient algorithms for a fundamental network design problem arising in potential-based flow models, which are central to many energy transport networks (e.g., hydrogen and electricity). In contrast to classical network flow problems, the nonlinearities inherent in potential-based networks introduce significant new challenges. We address these challenges through intricate reductions to classical combinatorial optimization problems, such as (constrained) shortest path problems, enabling the application of well-established algorithmic techniques to compute exact and approximate solutions efficiently. Finally, we complement these algorithmic results with matching complexity results concerning the hardness and non-approximability of the considered problem variants.

Extending the Continuum of Six-Colorings

Sebastian Pokutta; Christoph Spiegel; Max Zimmer; Konrad Mundinger — Mon, 23 Mar 2026 10:10:02 +0100

We present two novel six-colorings of the Euclidean plane that avoid monochromatic pairs of points at unit distance in five colors and monochromatic pairs at another specified distance $d$ in the sixth color. Such colorings have previously been known to exist for $0.41 < \sqrt{2} - 1 \le d \le 1 / \sqrt{5} < 0.45$. Our results significantly expand that range to $0.354 \le d \le 0.657$, the first improvement in 30 years. Notably, the constructions underlying this were derived by formalizing colorings suggested by a custom machine learning approach.

Neural Parameter Regression for Explicit Representations of PDE Solution Operators

Sebastian Pokutta; Max Zimmer; Konrad Mundinger — Mon, 23 Mar 2026 10:09:50 +0100

We introduce Neural Parameter Regression (NPR), a novel framework specifically developed for learning solution operators in Partial Differential Equations (PDEs). Tailored for operator learning, this approach surpasses traditional DeepONets (Lu et. al, 2021) by employing Physics-Informed Neural Network (Raissi et. al, 2019) techniques to regress Neural Network (NN) parameters. By parametrizing each solution based on specific initial conditions, it effectively approximates a mapping between function spaces. Our method enhances parameter efficiency by incorporating low-rank matrices, thereby boosting computational efficiency and scalability. The framework shows remarkable adaptability to new initial and boundary conditions, allowing for rapid fine-tuning and inference, even in cases of out-of-distribution examples.

Neural Discovery in Mathematics: Do Machines Dream of Colored Planes?

Sebastian Pokutta; Christoph Spiegel; Max Zimmer; Aldo Kiem; Konrad Mundinger — Mon, 23 Mar 2026 10:09:36 +0100

We demonstrate how neural networks can drive mathematical discovery through a case study of the Hadwiger-Nelson problem, a long-standing open problem at the intersection of discrete geometry and extremal combinatorics that is concerned with coloring the plane while avoiding monochromatic unit-distance pairs. Using neural networks as approximators, we reformulate this mixed discrete-continuous geometric coloring problem with hard constraints as an optimization task with a probabilistic, differentiable loss function. This enables gradient based exploration of admissible configurations that most significantly led to the discovery of two novel six-colorings, providing the first improvement in thirty years to the off-diagonal variant of the original problem (Mundinger et al., 2024a). Here, we establish the underlying machine learning approach used to obtain these results and demonstrate its broader applicability through additional numerical insights.

Convergence Analysis of a Fully Discrete Observer for Data Assimilation of the Barotropic Euler Equations

Aidan Chaumet; Jan Giesselmann — Wed, 18 Mar 2026 09:40:35 +0100

We study the convergence of a discrete Luenberger observer for the barotropic Euler equations in one dimension, for measurements of the velocity only. We use a mixed finite element method in space and implicit Euler integration in time. We use a modified relative energy technique to show an error bound comparing the discrete observer to the original system's solution. The bound is the sum of three parts: an exponentially decaying part, proportional to the difference in initial value, a part proportional to the grid sizes in space and time and a part that is proportional to the size of the measurement errors as well as the nudging parameter. The proportionality constants of the second and third parts are independent of time and grid sizes. To the best of our knowledge, this provides the first error estimate for a discrete observer for a quasilinear hyperbolic system, and implies uniform-in-time accuracy of the discrete observer for long-time simulations.

Mixing of Gases in Stationary Networks: Properties and Optimization

Pascal Börner; Marc E. Pfetsch; Stefan Ulbrich — Fri, 13 Mar 2026 02:34:10 +0100

This paper deals with the mixing of gases in stationary networks. We first derive a model and pressure law for mixing, for which there is empirical evidence for its accuracy. The model is based on the change of the speed of sound in gas mixtures. We then consider stationary gas networks. The existence result for solutions of single gas flow on networks is extended to mixtures. Further, we establish an easy to check criterion for uniqueness of solutions on the network. This results in a uniqueness proof of solutions on all networks with a mixture of natural gas and low hydrogen percentages. We then develop a solution algorithm that alternates between the solution of a single gas problem and update of the mixture ratios. In a computational study, different model variants and their impact on performance are compared. Moreover, the increased complexity of solving stationary gas transport problems with mixing is evaluated.

Chance-Constrained Linear Complementarity Problems

René Henrion; Martin Schmidt — Wed, 04 Mar 2026 19:49:22 +0100

We study linear complementarity problems (LCPs) under uncertainty, which we model using chance constraints. Since the complementarity condition of the LCP is an equality constraint, it is required to consider relaxations, which naturally leads to optimization problems in which the relaxation parameters are minimized for given probability levels. We focus on these optimization problems and first study the continuity of the related probability functions and the compactness of the feasible sets. This leads to existence results for both types of models: one with a joint chance constraint and one with separate chance constraints for both uncertainty-affected conditions of the LCP. For both, we prove the differentiability of all probability functions and derive respective gradient formulae. For the separate case, we prove convexity of the respective optimization problem and use the gradient formulae to derive necessary and sufficient optimality conditions. In a small example of a Cournot oligopoly among energy producers, we finally illustrate our theoretical findings.

A Catalog of Gas Network Models: PDEs, Coupling Conditions, and Numerical Schemes

Tue, 24 Feb 2026 13:23:15 +0100

This document aims to provide a concise and clear introduction to the topic of gas flow modeling. We present several models for gas flow, organized into hierarchies based on complexity. We discuss in detail the modeling of individual components such as valves and compressors. Network model classes based on purely algebraic relations and energy-based port-Hamiltonian models are included, along with a brief overview of basic numerical methods for hyperbolic balance laws and port-Hamiltonian systems. We do not claim completeness and refer in many places to the existing literature.

Gas Mixtures on Networks: Modeling, Simulation and Optimization

Thu, 19 Feb 2026 13:37:59 +0100

This chapter addresses mathematical models for isothermal mixtures of hydrogen and natural gas, motivated by the need for reliable simulation tools in future low-carbon energy systems. We analyze several classes of mixture models and investigate their convergence properties in the regime of strong interaction between constituents, covering stationary and instationary single-pipe settings as well as network flows. Since mixture models critically depend on the choice of pressure law, we compare the industry-standard GERG equation of state with simplified alternatives that preserve convex energies and reduce computational costs. For network applications, we discuss consistent coupling conditions across model classes, explore optimization of steady flows using the algebraic Weymouth formulation, and provide numerical evidence for its applicability in relevant operating regimes. The study reveals when simplified models are justified and outlines key open challenges for the modeling of gas mixtures.

A posteriori existence of strong solutions to the Navier-Stokes equations in 3D

Aaron Brunk; Jan Giesselmann; Tabea Tscherpel — Wed, 18 Feb 2026 16:32:54 +0100

Global existence of strong solutions to the three-dimensional incompressible Navier--Stokes equations remains an open problem. A posteriori existence results offer a way to rigorously verify the existence of strong solutions by ruling out blow-up on a certain time interval, using only numerical solutions. In this work we present such a result for the Navier--Stokes equations subject to periodic boundary conditions, which makes use of a version of the celebrated blow-up criterion in the critical space $L^\infty(L^3)$ by Iskauriaza, Serëgin and Shverak (2003). Our approach is based on a conditional stability estimate in $L^2$ and $L^3$. The a posteriori criterion that, if satisfied, verifies existence of strong solutions, involves only negative Sobolev norms of the residual. We apply the criterion to numerical approximations computed with mixed finite elements and an implicit Euler time discretisation. A posteriori error estimates allow us to derive a fully computable criterion without imposing any extra assumptions on the solution. While limited to short time intervals, with sufficient computational resources in principle the criterion might allow for a verification over longer time intervals than what can be achieved by theoretical means.

On the Existence of Steady States for Blended Gas Flow with Non-Constant Compressibility Factor on Networks

Simone Göttlich; Michael Schuster; Alena Ulke — Thu, 12 Feb 2026 13:54:35 +0100

In this paper, we study hydrogen-natural gas mixtures transported through pipeline networks. The flow is modeled by the isothermal Euler equations with a pressure law involving a non-constant, composition-dependent compressibility factor. For a broad class of such compressibility models, we prove the existence of steady-state solutions on networks containing compressor stations. The analysis is based on an implicit representation of the pressure profiles and a continuity argument that overcomes the discontinuous dependence of the gas composition on the flow direction. Numerical examples illustrate the influence of different compressibility models on the resulting states.

On the Lipschitz continuity of the spherical cap discrepancy around generic point sets

Holger Heitsch; René Henrion — Wed, 04 Feb 2026 12:20:29 +0100

The spherical cap discrepancy is a prominent measure of uniformity for sets on the d-dimensional sphere. It is particularly important for estimating the integration error for certain classes of functions on the sphere. Building on a recently proven explicit formula for the spherical discrepancy, we show as a main result of this paper that this discrepancy is Lipschitz continuous in a neighbourhood of so-called generic point sets (as they are typical outcomes of Monte-Carlo sampling). This property may have some impact (both algorithmically and theoretically for deriving necessary optimality conditions) on optimal quantization, i.e., on finding point sets of fixed size on the sphere having minimum spherical discrepancy.

Continuous stochastic gradient and spherical radial decomposition

Wed, 04 Feb 2026 12:01:01 +0100

In this paper, a new method is presented for solving chance-constrained optimization problems. The method combines the well-established Spherical-Radial Decomposition approach with the Continuous Stochastic Gradient method. While the Continuous Stochastic Gradient method has been successfully applied to chance-constrained problems in the past, only the combination with the Spherical-Radial Decomposition allows to avoid smoothing of the integrand. In this chapter, we prove this fact for a relevant class of chance-constrained problems and apply the resulting method to the capacity maximization problem for gas networks.

Control and Optimization under Uncertainty in the Context of Gas Network Operation

Michael Schuster — Mon, 02 Feb 2026 16:29:57 +0100

The transition to renewable energy and the increasing role of hydrogen as a future energy carrier pose major challenges for the operation and optimization of gas transport networks. A rigorous mathematical understanding of the topic is necessary as efficient and reliable operation requires advanced methods to deal with uncertainty, nonlinear dynamics, and complex network topologies. At the same time, optimal control theory – in particular the turnpike phenomenon – provides powerful tools to simplify long-term optimization problems and to connect dynamic models with their stationary counterparts. This habilitation thesis focuses on establishing new results in three related areas. For optimization problems with probabilistic constraints, an approach for approximating probabilities based on kernel density estimation is presented and analyzed, yielding necessary and sufficient conditions for the convergence of solutions of approximated stochastic optimization problems. In the modeling of gas transport networks, the existence and uniqueness results of a mixing model are presented. Furthermore, a finite-time turnpike result for an optimal control problem governed by the wave equation as well as an integral turnpike result for an optimal control problem governed by the transport equation under uncertainty is established. On the application side, the problem of optimal placement of compressor stations in stationary gas networks is analyzed for both deterministic and random gas demand. Results on the optimal number of compressor stations and their locations on gas networks are presented. Further, the turnpike phenomenon is applied to an optimal control and design problem for gas networks, allowing steady-state models to replace the gas dynamics in the long-term planning. For the corresponding stationary problem, a fast and efficient algorithm for identifying the optimal network topology with low control cost is provided.

Synchronization of velocities in pipeline flow of blended gas

Martin Gugat — Mon, 02 Feb 2026 16:28:37 +0100

We consider the pipeline flflow of blended gas. The flow is governed by a coupled system where for each component we have the isothermal Euler equations with an additional velocity coupling term that couples the velocities of the different components. Our motivation is hydrogen blending in natural gas pipelines, which will play a role in the transition to renewable energies. We show that with suitable boundary conditions the velocities of the gas components synchronize exponentially fast, as long as the L2-norm of the synchronization error is outside of a certain interval where the size of the interval is determined by the order of the interaction terms. This indicates that in some cases for a mixture of ncomponents it is justifified to use a flux model where it is assumed that all components flow with the same velocity. For the proofs we use an appropriately chosen Lyapunov function which is based upon the idea of relative energy.

Potential-Based Flows - An Overview

Marc Pfetsch; Martin Schmidt; Martin Skutella; Johannes Thürauf — Thu, 29 Jan 2026 16:41:41 +0100

Potential-based flows provide an algebraic way to model static physical flows in networks, for example, in gas, water, and lossless DC power networks. The flow on an arc in the network depends on the difference of the potentials at its end-nodes, possibly in a nonlinear way. Potential-based flows have several nice properties like uniqueness and acyclicity. The goal of this paper is to provide an overview of the current knowledge on these models with a focus on optimization problems on such networks. We cover basic properties, computational complexity, monotonicity, uncertain parameters, and the corresponding behavior of the network as well as topology optimization.

Boundary Stabilization of Quasi-Linear Hyperbolic Systems with Varying Time-Delay

Martin Gugat — Tue, 27 Jan 2026 16:17:11 +0100

In this paper we consider the boundary feedback stabilization of a quasi-linear hyperbolic system of balance laws. At one end of the space interval, there is a reflecting boundary condition. At the other end a stabilizing feedback law with a varying time-delay is prescribed. We present sufficient conditions for the exponential stability of the system. We show that exponential stabilization is possible if the product of the length of the interval and an upper bound for the source term is sufficiently small. We also show that if the product of the length of the interval and a lower bound for the source term is sufficiently large, the system is unstable. Our analysis is based on Lyapunov functions with weights that are given by hyperbolic functions that generalize the well-known exponential weights. Compared with previous contributions, we obtain conditions that can be verified more easily in terms of the system parameters. Our results show that for sufficiently short space intervals, and also with varying time-delay, exponential stabilization is possible with appropriately chosen feedback gains that depend on the maximal value of the time-delay and the maximal absolute value of its derivative.