Peer Revenue Play credits. Real papers.

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Mathematics

math Watch Mathematics (sign in to use)

Thursday 2 April

Functional Analysis

Lipschitz-free spaces over uniformly discrete metric spaces are 3-Schur

We prove that the Lipschitz-free space over any uniformly discrete metric space has the 3-Schur property

Information Theory · Cross-listed from quant-ph

Transversal non-Clifford gates on almost-good quantum LDPC and quantum locally testable codes

We exhibit nontrivial transversal logical multi-controlled-$Z$ gates on $[\![N,\Theta(N),\tilde\Theta(N)]\!]$ quantum low-density parity-check (qLDPC) codes with soundness $\tilde\Theta(1)$, combining nearly optimal code parameters with fault-tolerant non-Clifford gates on qLDPC and quantum locally testable codes for the first time. Remarkably, our proofs proceed through highly general algebraic arguments. Building on insights from [Li et al.,~arXiv:2603.25831], we develop a general covering space framework for constructing and computing a rich family of cohomological invariant forms on sheaf codes that induce transversal logical multi-controlled-$Z$. To certify their nontriviality, we further demonstrate the existence of two-way product-expanding punctured Reed–Solomon codes, which is striking in light of the many negative examples for the product expansion behavior of ordinary Reed–Solomon codes. This approach directly overcomes the previous obstruction to realizing nontrivial logical operations while simultaneously preserving the code parameters. The claimed almost-good code results follow immediately as examples.
Read the full abstract

Optimization and Control

Scaled Relative Graphs and Dynamic Integral Quadratic Constraints: Connections and Computations for Nonlinear Systems

Scaled relative graphs (SRGs) enable graphical analysis and design of nonlinear systems. In this paper, we present a systematic approach for computing both soft and hard SRGs of nonlinear systems using dynamic integral quadratic constraints (IQCs). These constraints are exploited via application of the S-procedure to compute tractable SRG overbounds. In particular, we show that the multipliers associated with the IQCs define regions in the complex plane. Soft SRG computations are formulated through frequency-domain conditions, while hard SRGs are obtained via hard factorizations of multipliers and linear matrix inequalities. The overbounds are used to derive an SRG-based feedback stability result for Lur'e-type systems, providing a new graphical interpretation of classical IQC stability results with dynamic multipliers.
Read the full abstract

All authors Timo de Groot, Tom Oomen, W.P.M.H. Heemels, Sebastiaan van den Eijnden

Analysis of PDEs

Novel Self-similar Finite-time Blowups with Singular Profiles of the 1D Hou-Luo Model and the 2D Boussinesq Equations: A Numerical Investigation

We present novel self-similar finite-time blowup scenarios for the 1D Hou–Luo model. We numerically demonstrate that solutions that initially satisfy certain derivative degeneracy condition can develop asymptotically self-similar finite-time blowups with singular self-similar profiles that are unbounded at some point. Moreover, this blowup phenomenon exhibits a two-stage feature: the solution first undergoes a local $L^{\infty}$ blowup at some time $\tilde{T}$, then continues in the weak sense beyond $\tilde{T}$ and develops a local $L^p$ blowup at a later time $T>\tilde{T}$ for some $p>0$. A further numerical investigation indicates that both stages are asymptotically self-similar. Finally, we extend our numerical study to the 2D Boussinesq equations and discover similar self-similar finite-time blowups with singular profiles that also exhibit a two-stage feature.
Read the full abstract

Optimization and Control

Bilevel Programming Approach for Image Restoration Problems with Automatically Hyperparameter Selection

In optimization-based image restoration models, the correct selection of hyperparameters is crucial for achieving superior performance. However, current research typically involves manual tuning of these hyperparameters, which is highly time-consuming and often lacks accuracy. In this paper, we concentrate on the automated selection of hyperparameters in the context of image restoration and present a bilevel programming approach that can simultaneously select the optimal hyperparameters and achieve high-quality restoration results. For implementation, we reformulate the bilevel programming problem that incorporates an inequality constraint related to the difference-of-convex functions. Following this, we address a sequence of nonsmooth convex programming problems by employing a feasibility penalty function along with a proximal point term. In this context, the nonsmooth convex programming problem uses the solution of the lower-level problem, which is derived through the alternating direction method of multipliers. Theoretically, we prove that the sequence generated by the algorithm converges to a Karush-Kuhn-Tucker stationary point of the inequality-constrained equivalent bilevel programming problem. We conduct a series of tests on both simulations and real images, which demonstrate that the proposed algorithm achieve superior restoration quality while requiring less computing time compared to other hyperparameter selection methods.
Read the full abstract

All authors Hang Xie, Xuewen Li, Peili Li, Qiuyu Wang

K-Theory and Homology

Chromatic Noshift

The chromatic redshift philosophy, introduced by Ausoni and Rognes, suggests that algebraic $K$-theory raises chromatic height by $1$. We show that the analogue of this philosophy fails in the case of rigid symmetric monoidal stable $\infty$-categories. More precisely, we construct examples of rigid $T(n)$-local categories $C$ where a refinement $\mathrm{Dim}$ of the dimension morphism induces an equivalence $$K(C)\to \mathrm{End}(\mathbf{1}_C)^{BS^1}$$ and for which $K(C)$ therefore vanishes $T(n+1)$-locally. In fact, we prove that this equivalence always holds for $\aleph_1$-Nullstellensatzian rigid $T(n)$-local categories in the sense of Burklund, Schlank and Yuan. We study more in depth the rational version of these results to find a rigid rational additive $1$-category witnessing the failure of redshift at height $0$. Finally, we use our methods to prove and generalize a conjecture of Levy about categorification of ordinary rings.
Read the full abstract

Category Theory

Free rigid commutative algebras

We describe free rigid commutative algebras in $2$-presentably symmetric monoidal $(\infty,2)$-categories as oplax colimits over the $1$-dimensional framed cobordism category. The special case of the $(\infty,2)$-category $\mathrm{Pr}^\mathrm{L}$ itself provides a description of the free symmetric monoidal $(\infty,1)$-category with duals on a given $(\infty,1)$-category, while the case of $\mathrm{Mod}_{\mathcal{V}}(\mathrm{Pr}^\mathrm{L})$ provides a description of a similar object in the $\mathcal{V}$-enriched context, for $\mathcal{V}$ a presentably symmetric monoidal $(\infty,1)$-category. As a byproduct, we obtain new proofs of some results about rigidification of locally rigid categories, as well as a proof that any rigid category over $\mathrm{Sp}$ embeds into a compactly-rigidly generated one.
Read the full abstract

Probability

The on/off Brownian snake

We define what we call an on/off Brownian snake. We use this to construct on/off super Brownian motion recently introduced to the literature by Blath and Jacobi and which is a measure-valued branching process with a dormant state and an active state. Our construction mirrors the construction of super Brownian motion from the Brownian snake by Le Gall. We use the on/off Brownian snake to obtain results concerning the support, range, and expected total mass of on/off super Brownian motion.
Read the full abstract

Number Theory

Towards the $p$-adic Hodge parameters in semistable representations of $\mathrm{GL}_n(\mathrm{Q}_p)$

Let $\rho_p$ be an $n$-dimensional non-critical semistable $p$-adic Galois representation of the absolute Galois group of $\mathbf{Q}_p$ with regular Hodge–Tate weights. Let $\mathbf{D}$ be the associated $(\varphi,\Gamma)$-module over the Robba ring. By combining Ding's and Breuil–Ding's methods for the crystalline case with Qian's computation of higher extension groups of locally analytic generalized Steinberg representations, we capture the full information of the $p$-adic Hodge parameters of $\rho_p$ on the automorphic side by considering several Steinberg subquotients of $\mathbf{D}$ and the “crystalline” Hodge parameters between them. These results also admit geometric and Lie-algebraic reformulations on flag varieties related to the moduli space of Hodge parameters. We then construct an explicit locally analytic representation $\pi_{1}(\rho_p)$ and explicitly describe which Hodge-parameters information of $\rho_p$ it determines. In particular, if the monodromy rank of $\rho_p$ is at most $1$, $\pi_{1}(\rho_p)$ determines $\rho_p$. When $\rho_p$ comes from a $p$-adic automorphic representation, we show that $\pi_{1}(\rho_p)$ is a subrepresentation of the $\mathrm{GL}_n(\mathbf{Q}_p)$-representation globally associated to $\rho_p$, under mild hypotheses. Although it is still difficult to construct an explicit representation $\pi_{1}(\rho_p)$ that determines $\rho_p$, our results provide new evidence for the $p$-adic Langlands program in general semistable cases and demonstrate the broad applicability of Ding's, Breuil–Ding's, and Qian's methods.\;
Read the full abstract

Algebraic Geometry

Genus three Ceresa cycles and limit of archimedean heights

For a one-parameter variation of biextension mixed Hodge structures, Brosnan and Pearlstein showed that the limit of the asymptotic height of the variation is given by a certain limit height of the nilpotent orbit. This limit height depends on the choice of a parameter. In the case of a variation of geometric origin related to Ceresa cycles associated with curves of genus three, after fixing a parameter, we show that this limit height is given by the Deligne splitting of a biextension mixed Hodge structure associated with cycles in the boundary.
Read the full abstract

Differential Geometry

On the equivalence of the integrability obstructions for transitive Lie algebroids

The integrability problem for transitive Lie algebroids can be looked at from different perspectives, revealing an interplay between cohomological methods and homotopical constructions. Mackenzie introduced a cohomological obstruction defined via sheaf-theoretic methods. On the other hand, Crainic and Fernandes used a path space approach and characterized integrability in terms of the monodromy. Recently, Meinrenken formulated the monodromy in terms of a clutching construction. We show that all of these agree. In particular, we identify the monodromy map with the Mackenzie obstruction class through the natural pairing between cohomology and homotopy.
Read the full abstract

Numerical Analysis

Goal oriented error estimation for adaptive sampling of PINNS

Physics-Informed Neural Networks (PINNs) are mesh-free approaches for the numerical approximation of partial differential equations, where a neural network is trained by minimizing a loss function derived from the governing equations and boundary conditions. The Deep Ritz method can be interpreted as a particular variational form of a PINN, where the loss corresponds to the minimization of an energy functional associated with a symmetric positive definite problem. In this work, we study the approximation of the Laplace equation using both the classical PINN formulation and its variational counterpart, the Deep Ritz method, with the objective of accurately estimating prescribed goal functionals. When standard sampling strategies, such as uniform or loss-based sampling, are employed during training, the convergence of the functional error and the attained minimal functional value can be slow. To address this issue, we introduce a functional-oriented importance sampling strategy that can be applied to both PINNs and the Deep Ritz method. The key ingredient is the construction of a reliable and accurate estimator for the error in a given quantity of interest. This estimator is derived using concepts from the Dual Weighted Residual (DWR) framework and is implemented entirely within the neural network setting. It is then used to adaptively guide the sampling of training points in the computational domain, focusing computational effort on regions that have the strongest influence on the functional value. Numerical experiments demonstrate that the proposed adaptive sampling strategy significantly accelerates the convergence of the functional error and improves the minimization of the target functional during training for both PINN and Deep Ritz formulations.
Read the full abstract

Analysis of PDEs

Multiphase cross-diffusion models for tissue structures: modeling, analysis, numerics

Volume-filling cross-diffusion equations for the components of a tissue structure are formally derived from mass conservation laws and force balances for the interphase pressures and viscous drag forces in a multiphase approach. The equations include Maxwell-Stefan, tumor-growth, thin-film solar cell models as well as novel volume-filling population systems. The Boltzmann and Rao entropy structures are explored. If the drag coefficients are all equal to one, the global-in-time existence of bounded weak solutions, their long-time behavior, and the weak-strong uniqueness of solutions to a regularized system are proved using entropy methods. In the general case, the resulting diffusion matrix is positively stable, ensuring local-in-time existence of solutions. Global-in-time existence of weak solutions is proved if the drag coefficients are sufficiently close to each other. This restriction is explained by the fact that the pressure forces are of degenerate type, while the drag forces are nondegenerate in the volume fractions. Numerical simulations are presented in one space dimension to illustrate the solution behavior beyond the entropy regime.
Read the full abstract

Analysis of PDEs

A review of compactness methods for cross-diffusion systems seen as Wasserstein gradient flows

A comprehensive methodology for establishing the existence of gradient flows for cross-diffusion systems with respect to suitable energies is proposed. The approach is based on the construction of piecewise-in-time constant approximations via the Jordan-Kinderlehrer-Otto scheme. Compactness of the approximate sequence is obtained using either the flow interchange technique or the five gradient inequality. These methods are illustrated for both parabolic and hyperbolic-parabolic Busenberg-Travis systems, as well as for several of their variants. This paper reviews the results from the literature and discusses additional properties.
Read the full abstract

Group Theory

Gruenberg-Kegel graphs with four edges of finite solvable cut groups

We prove that there is a unique graph with four edges which is the Gruenberg-Kegel graph of a solvable cut group. This contributes to the classification of the Gruenberg-Kegel graphs of solvable cut groups initiated in [A. Bächle, A. Kiefer, S. Maheshwary, and A. del Río. "Gruenberg-Kegel graphs: Cut groups, rational groups and the prime graph question" Forum Mathematicum, vol. 35, no. 2, 2023, pp. 409-429. https://doi.org/10.1515/forum-2022-0086].
Read the full abstract

Combinatorics

(Even hole, triangle)-free graphs revisited

We revisit a classical paper about (even hole, triangle)-free graphs [Conforti, Cornuéjols, Kapoor and Vu\v sković, Triangle-free graphs that are signable without even holes, Journal of Graph Theory, 34(3), 204–220, 2000]. In fact, the previous study describes a more general class, the so called triangle-free odd signable graphs, and we further generalise the class to the (theta, triangle, wac)-free graphs (not worth defining in an abstract). We exhibit a stronger structure theorem, by precisely describing basic classes and separators. We prove that the separators preserve the treewidth and several properties. Some consequences are a recognition algorithm with running time $O(|V(G)|^4|E(G)|)$, a proof that the treewidth of graphs in the class is at most~4 (improving a previous bound of~5), and a simple criterion to decide if a graph in the class is planar.
Read the full abstract

Optimization and Control

Compactness, Regularity, and Mosco Convergence for a Radially Parametrized Subclass of $C$-GNP Domains

We study a radially parametrized subclass of domains satisfying the Geometric Normal Property with respect to a fixed convex core $C$ ($C$-GNP). These domains are described by a thickness function $d:\pC\to(0,\infty)$. The main contribution is the equivalence \[ \chi_{\Om_n}\to\chi_{\Om}\ \text{in }L^1 \iff d_n\to d\ \text{uniformly} \iff \Om_n\to\Om\ \text{in Hausdorff distance}. \] This equivalence is established via an explicit reconstruction formula $d(c)=\int_0^\infty \chi_{\Om}(c+t\nuC(c))\dd t$, which allows the transfer of convergence of characteristic functions into uniform convergence of geometric parameters. Thus the infinite-dimensional geometric optimization problem is reduced to the optimization of a scalar function on a fixed manifold. We prove compactness of uniformly controlled subclasses. The analysis is extended to non-connected domains with uniform separation. Under the curvature condition $\sup d<\delta_0$ (where $\delta_0$ is the normal tubular radius of $C$), we prove a conditional regularity result for the thickness function when the normal projection admits a $C^{m,\beta}$ inverse parametrization. We also give a sufficient perturbative bilipschitz condition for the radial parametrization. Using the Mosco convergence framework, we prove $\Gamma$-convergence of Dirichlet energies under domain variation for the uniformly controlled radial subclass. We also discuss the relation with the general $C$-GNP class and possible applications in shape optimization.
Read the full abstract

Analysis of PDEs

Local Well-Posedness of a Model for Stress-Driven Growth in the Presence of Nutrients

A model for morphoelastic growth, that is, growth influenced by elastic stress, driven by the absorption of nutrients is considered. The model features a multiplicative decomposition of the deformation gradient into an elastic contribution and a growth tensor. While the evolution of the system is governed by an ordinary differential equation for the growth tensor on a suitable Banach space, which depends on the elastic stresses and the concentration of a nutrient field, the total deformation is given by the solution of a quasi-static equilibrium equation arising from the formal Euler-Lagrange equations of a hyperelastic variational integral. The nutrient concentration is determined by a linear elliptic reaction-diffusion equation which is formulated in Lagrangian coordinates and whose coefficients depend on the growth tensor as well as the deformation gradient accounting for the change of material properties due to elastic deformation and growth. Existence and uniqueness of solutions of this fully coupled system of differential equations is proven via a fixed-point argument.
Read the full abstract

Optimization and Control

Stability of the Timoshenko Beam Equation with One Weakly Degenerate Local Kelvin-Voigt Damping

We consider the Timoshenko beam equation with locally distributed Kelvin-Voigt damping, which affects either the shear stress or the bending moment. The damping coefficient exhibits a singularity, causing its derivative to be discontinuous. By using the frequency domain method and multiplier technique, we prove that the associated semigroup is polynomial stability. Specifically, regardless of whether the local Kelvin-Voigt damping acts on the shear stress or the bending moment, the system decays polynomially with rate $t^{-\frac{1}{2}}$.
Read the full abstract

Logic

The cohesive and stable Ramsey theorems and proof size over a weak base theory

We show that over the weak base theory $\mathrm{RCA}_0^*$, cohesive Ramsey's theorem for pairs $\mathrm{CRT}^2_2$ implies exponential closure of the definable cut $\mathrm{I}^0_1$, which is the intersection of all $\Sigma^0_1$-definable cuts. Consequences include non-elementary proof speedup of $\mathrm{RCA}_0^* + \mathrm{CRT}^2_2$ over $\mathrm{RCA}_0^*$ for $\Pi_1$ sentences and the unprovability of $\mathrm{CRT}^2_2$ in $\mathrm{RCA}_0^* + \mathrm{CAC}$. On the other hand, we show that $\mathrm{RCA}_0^* + \mathrm{SRT}^2_2$, where $\mathrm{SRT}^2_2$ is stable Ramsey's theorem for pairs, is polynomially simulated by $\mathrm{RCA}_0^*$ with respect to proofs of $\forall \Pi^0_3$ sentences. Nevertheless, $\mathrm{SRT}^2_2$ also implies a nontrivial property of $\mathrm{I}^0_1$, specifically closure under functions of quasipolynomial growth rate.
Read the full abstract

Analysis of PDEs

Logarithmically Coupled p-Laplacian Systems: From Discrete Ground States to Rigidity and Critical Continuous Limits

We introduce a new class of logarithmically coupled \(p\)-Laplacian systems on locally finite graphs and in the critical case \(p=N>4\). The coupling poses two challenges: the energy functional is ill-defined on the natural Sobolev space, and loss of the \(L^\infty\)-embedding prevents standard control of logarithmic terms. To resolve this non-separable logarithmic singularity, we develop an exponent calibration technique that converts logarithmic terms into strictly lower-order power estimates. The technique plays an essential role in two critical steps: the existence proof for the regularised problem and the uniform boundedness of regularised ground states that restores compactness. Thus, we prove solvability of the original ill-posed continuous critical equation. In the discrete setting, we identify five new intrinsic structures, including an exact constrained Hessian factorisation and a scalar rigidity index. We complete a discrete–continuous double loop, providing a unified variational framework for logarithmically coupled systems.
Read the full abstract

Analysis of PDEs

Avscon the Schur topology

The aim of the course is to lead to an understanding of homogenisation processes in an operator-theoretic sense. In fact, using solely operator-theoretic means not referring to the particular form of the coefficients, we will identify an operator topology on the level of coefficients that will fully capture the convergence involved in the context of homogenisation. One upshot of this perspective will be that we will obtain homogenisation results for time-dependent partial differential equations (almost) for free.
Read the full abstract

2604.01803 · 2 Apr · This is part of the Avscon Lecture series 2025; Avscon (= A very short course on..) comprises each year of 4 short lecture courses accessible to B.Sc. students or higher; details and registration can be found at https://www.marcus-waurick.de/avscon ; this course on homogenisation and related things has 37 pages

Numerical Analysis

Tree-Adaptive Multiscale Kernel Lasso in Samplet Coordinates

We develop a novel framework for sparse multiscale kernel approximation of large scattered data problems based on a samplet representation. Samplets form a multiresolution analysis of localized discrete signed measures and enable quasi-sparse representations of kernel matrices associated to asymptotically smooth kernels as well as smoothness detection of scattered data. Building on the latter, we introduce an adaptive data site selection strategy based on the native reproducing kernel Hilbert space norm of the samplet expansion coefficients. The selection results in a small set of representative data sites, significantly reducing the effective problem size. On the corresponding reduced kernel subspace, we solve an $\ell^1$-regularized least-squares problem using a trust-region semismooth Newton method in a normal-map formulation, stabilized by an online low-rank SVD on the active set to handle the notorious ill-conditioning of kernel matrices. Numerical experiments in two and three dimensions, including multi-kernel models with varying lengthscales, demonstrate that the proposed approach achieves accurate reconstructions with considerably sparser representations and good computational efficiency.
Read the full abstract

All authors Sara Avesani, Gaia Fumagalli, Michael Multerer, Chiara Segala

Probability

Low-Temperature Asymptotics of the Poincaré and the log-Sobolev Constants for Łojasiewicz Potentials

In this paper, we establish the low-temperature asymptotics of the Poincaré inequality constant for a class of convex potentials satisfying a Łojasiewicz inequality. In addition, we disprove a conjecture previously posed by Chewi and Stromme on the low-temperature asymptotics of the log-Sobolev constant and determine the correct asymptotic behavior in dimension one.
Read the full abstract

Representation Theory

Cuspidal endo-support and strong beta extensions

Let $G$ be an inner form of a general linear group or classical group over a non-archimedean local field of residual characteristic $p$, assumed odd in the classical case. We prove that every smooth representation of $G$ over an algebraically closed field $R$ of characteristic $\ell\neq p$ contains a maximal semisimple character, i.e., one for which the point in the building of the corresponding centralizer is a vertex. Further, for every endo-parameter adapted to $G$, we define its support, which leads also to the notion of cuspidal endo-support of an irreducible representation, and we relate this to its cuspidal support. We also introduce beta extensions for strong facets in the building of a centralizer, and show these are sufficient for the construction of types. These results are used in a subsequent paper to decompose the category of smooth $R$-representations of $G$.
Read the full abstract

All authors David Helm, Robert Kurinczuk, Daniel Skodlerack, Shaun Stevens

Information Theory

Multi-Mode Pinching-Antenna Systems: Polarization-Aware Full-Wave Modeling and Optimization

Millimeter-wave and terahertz communications face a fundamental challenge: overcoming severe path loss without sacrificing spectral efficiency. Pinching antenna systems (PASS) address this by bringing radiators physically close to users, yet existing frameworks treat the waveguide as a mere transmission line, overlooking its inherent multi-mode capabilities and the critical role of polarization. This paper develops the first polarization-aware, full-wave electromagnetic model for multi-mode PASS (MMPASS), capturing spatial radiation patterns, modal polarization states, and polarization matching efficiency from first principles. Leveraging this physically grounded model, we reveal fundamental trade-offs among waveguide attenuation, atmospheric absorption, and geometric spreading, yielding closed-form solutions for optimal PA placement and orientation in single-user scenarios. Extending to multi-user settings, we propose a modular optimization framework that integrates fractional programming with closed-form polarization updates, scaling gracefully to arbitrary numbers of waveguides, PAs, and users. Numerical results show that MMPASS achieves up to a 167% increase in spectral efficiency compared with single-mode PASS. Moreover, when comparing MMPASS with its polarization-ignorant counterpart, polarization awareness alone improves the sum rate by up to 23%. By bridging rigorous electromagnetic theory with scalable optimization, MMPASS establishes a physically complete and practically viable foundation for future high-frequency wireless networks.
Read the full abstract

All authors Dengke Wei, Runxin Zhang, Yulin Shao, Fen Hou, Shaodan Ma

Optimization and Control

Improving Operational Feasibility in Large-Scale Power System Planning

Large-scale power system planning mostly uses linearized, active power only approximations of the power flow equations, ignores many operational constraints, and tests the operational feasibility of the resulting systems only under strongly simplifying assumptions. We propose an approach to obtain solutions to large instances of the alternating current capacity expansion problem via redispatch and reinforcement of an initial solution. The problem formulation considers simultaneous expansion of generators, reactive compensation devices, storage systems, and transmission. Furthermore, it includes operational constraints via startup procedures and capability curves of power sources and simplified stability limits via constraints on voltage angle differences and voltage magnitudes. To obtain initial solutions, we test several established and partly modified power flow approximations and integrate them into an approach for iterative transmission expansion planning, thereby obtaining convex formulations. We demonstrate the approach on large problem instances covering the islands of Great Britain and Ireland at the transmission level, for which we extend the open data source to model reactive power. We find that including transmission losses to determine the initial solution is most decisive, as the amount of redispatch and reinforcements necessary to obtain an alternating current feasible solution is reduced, whereas incorporating reactive power constraints did not lead to further improvements. Our approach ensures an alternating current feasible system under weak assumptions, thus guaranteeing steady-state voltage stability and allowing subsequent dynamic grid simulations, which is instrumental for planning stable future inverter-dominated power systems.
Read the full abstract

All authors Gereon Recht, Oussama Alaya, Benedikt Jahn, Karl-Kiên Cao, Hendrik Lens

Optimization and Control

State-constrained optimal control of the continuity equation and infinite-dimensional viscosity solutions

We study a finite horizon optimal control problem for the continuity equation under a weighted integral state constraint on the mass outside a fixed set. The model is cast in a Hilbert framework for densities. On a suitable invariant compact subset, we prove that the value function is Lipschitz continuous and satisfies, by dynamic programming, the associated infinite dimensional constrained Hamilton Jacobi Bellman equation in viscosity sense (subsolution in the interior, supersolution up to the boundary). We finally prove a comparison principle and uniqueness in the Lipschitz class.
Read the full abstract

Numerical Analysis

Attention Mechanisms Through the Lens of Numerical Methods: Approximation Methods and Alternative Formulations

The attention mechanism is the computational core of modern Transformer architectures, but its quadratic complexity in the input sequence length is the bottleneck for large-scale inference. This has motivated a rapidly growing body of work aimed at accelerating attention through approximation and reformulation. In this survey, we revisit attention mechanisms through the lens of numerical analysis, with a particular emphasis on tools and perspectives from numerical linear algebra. Our goal is twofold: first, we aim to systematically review and classify fast approximation methods according to the numerical principles they exploit. These include sparsity and clustering approaches, low-rank and subspace projection techniques, randomized sketching methods, and tensor-based decompositions. We also discuss kernel-inspired reformulations of attention and recent architectural variants, such as Latent Attention, that modify the standard softmax formulation to improve efficiency. Second, by presenting these developments within a unified mathematical framework, we aim to bridge the gap between disciplines and highlight opportunities for further contributions from computational mathematics, particularly numerical linear algebra, to the design of scalable attention mechanisms.
Read the full abstract

All authors Michel Fabrice Serret, Alice Cortinovis, Yijun Dong, Diana Halikias, Anna Ma, Fabio Matti, Deanna Needell, Katherine J. Pearce, Elizaveta Rebrova, Disha Shur, Rudi Smith, Hai-Xiao Wang, Laura Grigori

Optimization and Control

Day-Ahead Offering for Virtual Power Plants: A Stochastic Linear Programming Reformulation and Projected Subgradient Method

Virtual power plants (VPPs) are an emerging paradigm that aggregates distributed energy resources (DERs) for coordinated participation in power systems, including bidding as a single dispatchable entity in the wholesale market. In this paper, we address a critical operational challenge for VPPs: the day-ahead offering problem under highly intermittent and uncertain DER outputs and market prices. The day-ahead offering problem determines the price-quantity pairs submitted by VPPs while balancing profit opportunities against operational uncertainties. First, we formulate the problem as a scenario-based two-stage stochastic adaptive robust optimization problem, where the uncertainty of the locational marginal prices follows a Markov process and DER uncertainty is characterized by static uncertainty sets. Then, motivated by the outer approximation principle of the column-and-constraint generation (CC&G) algorithm, we propose a novel inner approximation-based projected subgradient method. By exploiting the problem structure, we propose two novel approaches to improve computational tractability. First, we show that under mild modeling assumptions, the robust second-stage problem can be equivalently reformulated as a linear program (LP) with a nested resource allocation structure that is amenable to an efficient greedy algorithm. Furthermore, motivated by the computational efficiency of solving the reformulated primal second-stage problem and the isotonic structure of the first-stage feasible region, we propose an efficient projected subgradient algorithm to solve the overall stochastic LP problem. Extensive computational experiments using real-world data demonstrate that the overall projected subgradient descent method achieves about two orders of magnitude speedup over CC&G while maintaining solution quality.
Read the full abstract