Analysis of PDEs
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Thursday 12 March
This article is concerned with the inverse problem on determining the temporal component of the source term in a coupled time-fractional diffusion system by a single point observation. Under a non-degeneracy condition on the known spatial component, we establish the Lipschitz stability by observing all solution components by a series representation of the mild solution. To reduce the observation data, we prove the strict positivity of some fractional integral of the solution to the homogeneous problem by a modified Picard iteration. This, together with a coupled Duhamel's principle, lead us to the uniqueness of the inverse problem by observing any single solution component under a specific structural constraint on the unknown. Numerically, we propose an iterative regularizing ensemble Kalman method (IREKM) for the simultaneous recovery of the temporal sources. Through extensive numerical tests, we demonstrate its accuracy, robustness against noise, and scalability with respect to the number of components. Our findings highlight the essential roles of the non-degeneracy condition, measurement configuration, and fractional structural constraints in ensuring reliable reconstructions. The proposed framework provides both rigorous theoretical guarantees and a practical algorithmic approach for multi-component source identification in fractional diffusion systems.
Numerical Analysis
Cayley Commutator-free Methods for Krotov-Type Algorithms in Quantum Optimal Control
This paper presents a class of structure-preserving numerical methods for quantum optimal control problems, based on commutator-free Cayley integrators. Starting from the Krotov framework, we reformulate the forward and backward propagation steps using Cayley-type schemes that preserve unitarity and symmetry at the discrete level. This approach eliminates the need for matrix exponentials and commutators, leading to significant computational savings while maintaining higher-order accuracy. We first recall the standard linear setting and then extend the formulation to nonlinear Schrödinger and Gross-Pitaevskii equations using a Cayley-polynomial interpolation strategy. Numerical experiments on state-transfer problems illustrate that the CF-Cayley method achieves the same accuracy as high-order exponential or Cayley-Magnus schemes at substantially lower cost, especially for longtime or highly oscillatory dynamics. In the nonlinear regime, the structure-preserving properties of the method ensure stability and norm conservation, making it a robust tool for large-scale quantum control simulations. The proposed framework thus bridges geometric integration and optimal control, offering an efficient and reliable alternative to existing exponential-based propagators.
Read the full abstract
Numerical Analysis
Non-uniform $\alpha$-Robust Alikhanov Mixed FEM with Optimal Convergence for the Time-Fractional Allen–Cahn Equation
We investigate a mixed finite element method for the spatial discretization of a time-fractional Allen–Cahn equation defined on a convex polyhedral domain, combined with a nonuniform Alikhanov scheme for the temporal approximation. Under suitable regularity assumptions on the initial data that are weaker than those typically imposed in the literature, we establish regularity results for the solution and its flux. We then derive optimal $L^2$-error estimates, up to a logarithmic factor, for both the solution and the flux. The estimates are robust with respect to the fractional order $\alpha$, in the sense that the associated constants remain bounded as $\alpha \to 1^{-}$. Numerical experiments are presented to confirm the theoretical findings.
Read the full abstract
Statistics Theory · Cross-listed from stat.AP
On the Unit Teissier Distribution: Properties, Estimation Procedures and Applications
The Teissier distribution, originally proposed by Teissier [31], was designed to model mortality due to aging in domestic animals. More recently, Krishna et al. [19] introduced the Unit Teissier (UT) distribution on the interval (0, 1) through the transformation $X=e^{-Y}$, where $Y$ follows the Teissier distribution. In their work, the authors derived several fundamental properties of the UT distribution and investigated parameter estimation using maximum likelihood, least squares, weighted least squares and Bayesian methods. Building upon this work, the present paper develops additional theoretical and inferential results for the UT distribution. In particular, closed-form expressions for single moments of order statistics and L-moments are obtained, and characterization results based on truncated moments are established. Furthermore, several alternative parameter estimation methods are considered, including maximum product of spacings, Cramér-von Mises, Anderson-Darling, right-tail Anderson-Darling, percentile and L-moment estimation, while the estimation methods previously studied by Krishna et al. [19] are also included for comparison. Extensive simulation studies under various parameter settings and sample sizes are conducted to assess and compare the performance of the estimators. Finally, the flexibility and practical utility of the UT distribution are demonstrated using a real dataset.
Read the full abstract
Mathematical Physics
Systems of partial differential equations describing pseudo-spherical or spherical surfaces
In this paper, we study systems of nonlinear partial differential equations which describe surfaces of constant curvature. From the flatness condition of connection 1-forms, we present a classification of systems of Camassa-Holm-type equations of the form \begin{equation*} \left\{ \begin{aligned} u_{t} - u_{xxt} &= F(x, t, u, u_{x}, \dots, \partial ^{m} u/\partial _x^{m}, v, v_{x}, \dots, \partial ^{n} v/\partial _x^{n}), \\ v_{t} - v_{xxt} &= G(x, t, u, u_{x}, \dots, \partial ^{m} u/\partial _x^{m}, v, v_{x}, \dots, \partial ^{n} v/\partial _x^{n}), \end{aligned} \right. \end{equation*} with $m,n\geq2$, for $F$ and $G$ smooth functions, describing pseudospherical or spherical surfaces. We also establish classification results for a special type of third-order system. Applications of the results provide new examples of such systems, such as the Song-Qu-Qiao system, the two-component Camassa-Holm system with cubic nonlinearity, and the modified Camass-Holm-type system. Moreover, we construct the nonlocal symmetry and a non-trivial solutions for the two-component Camassa-Holm system with cubic nonlinearity from the gradients of spectral parameters.
Read the full abstract
Combinatorics · Cross-listed from nlin.SI
Hankel Determinants from Quadratic Orthogonal Pairs for Hyperelliptic Functions and Their Applications
As argued by Hone in the paper [Commun. Pure Appl. Math., 74(11):2310–2347, 2021], a “mismatch" problem remained unsolved while he was investigating continued fraction expansions and Hankel determinants from hyperelliptic curves. In this paper, by introducing a new notion called quadratic orthogonal pairs for hyperelliptic functions, we resolve the corresponding problem. As further applications, we give a thorough treatment of the initial value problems for two discrete integrable systems, i.e. the bilateral Somos-4 and Somos-5 recurrences.
Read the full abstract
Numerical Analysis
Compact LABFM: a framework for meshless methods with spectral-like resolving power
Meshless methods are often used in numerical simulations of systems of partial differential equations (PDEs), particularly those which involve complex geometries or free surfaces. Here we present a novel compact scheme based on the local anisotropic basis function method (LABFM), a meshless method which provides approximations to spatial operators to arbitrary polynomial consistency. Our approach mimics compact finite-differences by using implicit stencils to optimise the resolving power of each operator, whilst retaining diagonal dominance of the resulting global sparse linear system. The new method is demonstrated to provide improved approximations by a series of convergence tests and resolving power analysis, before solutions to canonical PDEs are computed. Significant gains in accuracy are observed, in particular for solutions containing high wavenumber components. Our compact meshfree method provides a pathway to high-order simulations of PDEs in complex geometries with spectral-like resolving power, and has the potential to lead to a step-change in the accuracy of numerical solutions to such problems.
Read the full abstract
Combinatorics
An Improved Interpolation Theorem and Disproofs of Two Conjectures on 2-Connected Subgraphs
We prove that any \(2\)-connected graph \(G\) on \(n\) vertices with minimum degree \(\delta(G) \ge \frac{n}{4}+2\) contains a \(2\)-connected subgraph of order \(k\) for every integer \(k\) with \(4 \le k \le n\). This improves a previous result of Yin and Wu. In \cite{YinWu-DAM-2026}, Yin and Wu proposed two conjectures. The first states that for any \(2\)-connected graph \(G\) of order \(n\) and size \(m\), there exists a \(2\)-connected subgraph of order \(k\) for each \(k \in \{4, \dots, n\}\) whenever \(m \ge \frac{1}{2} n^{3/2}\). The second conjecture asserts that the same conclusion holds under the alternative condition \(\delta(G) \ge \sqrt{n}\). In this paper, we construct counterexamples that completely disprove the first conjecture. Furthermore, using the existence of \((v, k, 2)\)-Symmetric Balanced Incomplete Block designs (i.e., SBIBDs), we disprove the second conjecture for all \(n \in \{8, 14, 22, 32, 74, 112, 158\}\). Finally, we propose a conjecture of our own: for any \(2\)-connected graph \(G\) on \(n\) vertices with \(\delta(G) \ge \frac{n}{k}\), where \(k \ge 3\) and \(n\) is sufficiently large, \(G\) contains a \(2\)-connected subgraph of every order from \(4\) to \(n\).
Read the full abstract
Combinatorics
Lattice point enumeration of some arbor polytopes
The $n$-dimensional lattice polytopes $\mathcal{Q}_{n,k}$ obtained by intersecting the $n$th dilate of the standard $n$-dimensional simplex in $\mathbb{R}^n$ with the half-spaces $x_i \le 1$ for $1 \le i \le k$ form an interesting special case of Chapoton's arbor polytopes. They interpolate between the $n$th dilate of the standard $n$-dimensional simplex and the standard $n$-dimensional cube in $\mathbb{R}^n$. This paper provides an explicit combinatorial interpretation of the $h^\ast$-polynomial of $\mathcal{Q}_{n,k}$, as the ascent enumerator of certain words, and partly confirms some of Chapoton's conjectures on the lattice point enumeration of arbor polytopes in this special case. More specifically, the Ehrhart polynomial of $\mathcal{Q}_{n,k}$ is shown to be magic positive, by means of a new combinatorial parking model for cars, and the real-rootedness of its $h^\ast$-polynomial is deduced. The polynomial whose coefficients count the lattice points of $\mathcal{Q}_{n,k}$ by the number of their nonzero coordinates is shown to be gamma-positive and a combinatorial interpretation of the $h^\ast$-polynomial of any arbor polytope is conjectured.
Read the full abstract
Representation Theory
Automorphism groups and derivation algebras of Hamiltonian Lie algebras
In this paper, we compute the automorphism group and derivation algebra of the Hamiltonian Lie algebra $\mathcal{H}_{N}$ and its derived subalgebra $\mathcal{H}_{N}'$, where $N$ is an even positive integer. The automorphism groups are shown to be $\mathbf{GSp}_{N}(\mathbb{Z})\ltimes (\mathbb{\mathbb{K}}^{\times})^{N}$ for both Lie algebras and we prove that all derivations are inner for the Hamiltonian Lie algebra, also we compute the full derivation space for the derived subalgebra of Hamiltonian Lie algebra. Finally we compute the second cohomology group of Hamiltonian Lie algebra.
Read the full abstract
Algebraic Geometry
The genus of configuration curves of planar linkages is generically odd
A one-degree-of-freedom graph is a graph obtained from a minimally rigid graph in the plane and removing an edge. For such graph, the set of realisations with fixed edge length, modulo rotations and reflections, is an algebraic curve. The genus of a connected component for generic edge lengths is a number that depends only on the graph. We prove that this genus is always odd, unless it is zero. The proof is based on tropical geometry.
Read the full abstract
Logic
The magmatic universe revisited: we define ordered pairs, relations, numbers and a special form of Separation
This is a companion article to \cite{Tz24}. We address the following two questions: 1) Can we define in the magmatic universe $M$ of \cite{Tz24} counterparts, or just analogues, of some very basic set-theoretic objects which are missing from $M$, specifically ordered pairs, binary relations, especially functions, as well as natural and ordinal numbers? 2) Are there restricted forms of the Separation and, perhaps, Replacement schemes that hold in $M$? We show the following: 1) Magmatic analogues of ordered pairs can indeed be defined by means of certain magmas called “magmatic pairs”. However when we use them to generate relations and especially functions, some unsurmountable problems come up. These problems are due to the peculiarity of the elements of magmas to be distinguished into “intended” and “collateral” ones, a distinction due to their inherent relation of dependence. So magmatic functions are defined under very special conditions. 2) A certain class of formulas, called “magmatic formulas” is isolated, and the scheme of Separation restricted to these formulas, called “Magmatic Separation Scheme” (MSS), is proven to hold in $M$. On the other hand Replacement fails badly, and this is due to its functional form.
Read the full abstract
Numerical Analysis
T-systems: a theory of orthonormal functions with a tridiagonal differentiation matrix
The starting point of this paper is that a spectral method is essentially a combination of an orthonormal basis of the underlying Hilbert space with Galerkin conditions. The choice of an orthonormal basis depends on a number of desirable features which we explore in the context of spectral methods for time-dependent partial differential equations in a single space dimension. A central role in ensuring many of the above features is played by the differentiation matrix of the underlying orthonormal system. In particular, it is beneficial if this matrix is skew-symmetric and tridiagonal. While orthonormal systems with this feature have been characterised in A. Iserles & M. Webb, “Orthogonal systems with a skew-symmetric differentiation matrix”, Found. Comput. Maths, 19 (2019), 1191–1221, employing Fourier transforms, in this paper we provide an alternative characterisation using the differential Lanczos algorithm, which can be implemented constructively. It is valid for inner products that obey an `integration-by-parts condition', inclusive of $L_2$ and Sobolev norms on the real line. Motivated by quest for integration methods that conserve Hamiltonian energy, we conclude the paper replacing inner products by more general sesquilinear forms and presenting preliminary results. Here the Fourier transform characterisation generalises to spectral theory of Schrödinger operators and the differential Lanczos algorithm generalises to the differential Arnoldi algorithm.
Read the full abstract
Logic
An observer-based approach to the sorites paradox and the logic derived from that
We approach the sorites paradox (SP) through an observer-based and time-dependent approach to truth of vague assertions. Formally the approach gives rise to a semantics, called fluxing-object semantics (FOS), because it involves models that contain “fluxing objects”, that is, entities changing with time and observer. The models are equipped with agents (observers) and a linear and discrete time axis for time. The changing entities are represented by partial functions of time and agent, and this partiality causes truth-value gaps. If we interpret a truth-value gap as a third truth value, then FOS becomes a three-valued logic, which, quite interestingly, is proved identical to strong Kleene three-valued logic. The sorites phenomena can be represented in a structure of FOS as special objects that change imperceptibly with respect to an observer and with respect to a particular attribute. In this account the key point that eliminates the paradoxical character of sorites is the partiality of functions. When the observer is fixed the partiality corresponds to interrupted watching on his part. The interruption creates watching gaps during which the attributed property of the object as understood by the observer may change, without violating the imperceptibility condition. Interrupted watching has, according to experts on visual attention and focusing capabilities of humans, a firm physiological justification. The relationship of watching gaps with horizon crossing is also discussed.
Read the full abstract
Number Theory
On the discrete mean square of certain hybrid sum involving $a_{\mathbb{K}}(n)$
Let $\mathbb{K}$ be a non-normal algebraic number field of cubic degree given by the polynomial $x^{3}+ax^{2}+bx+c$ of discriminant $D_{\mathbb{K}}<0$. For sufficiently large $x$, we establish an asymptotic formula for the hybrid sum $$\sum\limits_{\substack{n= \sum_{i=1}^{8}n_{i}^{2}\leq x\\ (n_{1},n_{2},n_{3},n_{4},n_{5},n_{6},n_{7},n_{8})\in \mathbb{Z}^{8} }} a_{\mathbb{K}}^{2}(n)$$ with a tight error term.
Read the full abstract
Number Theory
Iwasawa Main Conjecture for ordinary semistable elliptic curves over global function fields
Let $A$ be an ordinary elliptic curve over a global function field $K$ of characteristic $p$, assumed semistable at every place, and let $L/K$ be a $\mathbb{Z}_p^d$-extension ramified only at finitely many places where $A$ has ordinary reduction. Building on the framework of [Tan26] (arXiv:2603.10576), we prove the Iwasawa Main Conjecture for $A$ over $L$, subject to a technical $\mu$-invariant hypothesis that is already detected after specialization to the unramified $\mathbb{Z}_p$-extension. The principal new input is a `$\chi$-formula' that compares appropriate $\chi$-isotypic characteristic ideals of Selmer modules with the corresponding specializations of the $p$-adic $L$-function. Finally, to show that our $\mu$-hypothesis is non-vacuous, we prove, for $p>3$, that the hypothesis holds on a Zariski open dense locus in the moduli of semistable elliptic curves.
Read the full abstract
Combinatorics
Upper bound of some character ratios and large genus asymptotic behavior of Hurwitz numbers
In [14] we found the large genus asymptotics of Hurwitz numbers for the Riemann sphere with a fixed number of general profiles and some (2,1^{d-2}) profiles. In this paper, motivated from [3], we generalize these results to Hurwitz numbers of an arbitrary compact Riemann surface with a fixed number of general profiles and some (r,1^{d-r}) profiles.
Read the full abstract
Combinatorics
On large genus asymptotics of certain Hurwitz numbers
In this paper, based on the value of central character on the transposition, we find structure and large genus asymptotics of certain Hurwitz numbers.
Representation Theory
Construction of Local Arthur Packets for Metaplectic Groups and the Adams Conjecture
In this article, we explicitly construct local Arthur packets for metaplectic groups over non-Archimedean local fields of characteristic zero. Our construction is a generalization of Atobe's construction of local Arthur packets for classical groups. As a result, we prove that the local Arthur packets are multiplicity free. Moreover, we generalize Moeglin's earlier work about the Adams conjecture to metaplectic groups.
Read the full abstract
Optimization and Control · Cross-listed from cs.LG
Hybrid Energy-Aware Reward Shaping: A Unified Lightweight Physics-Guided Methodology for Policy Optimization
Deep reinforcement learning for continuous control often suffers from high variance, low energy efficiency, and poor generalization under distribution shift, as purely data-driven exploration ignores available physical structure. This paper proposes Hybrid Energy-Aware Reward Shaping (H-EARS), which encodes dominant energy terms – assumed known a priori – directly as reward potentials at O(n) per-step computation. H-EARS decomposes the shaping potential into task-oriented and energy-based components, supplemented by an action regularization term that deliberately modifies the optimization objective to enforce energy-efficient control. A complete theoretical foundation is established: functional independence of shaping and regularization, energy-based gradient enrichment under positive-definite Hessian conditions, convergence guarantees under function approximation, and approximate potential error bounds. Across four continuous control benchmarks and four baseline algorithms, H-EARS achieves consistent gains in convergence speed, policy stability, and final performance. High-fidelity vehicle simulations validate applicability in safety-critical settings under extreme road conditions.
Read the full abstract
Representation Theory
Asymptotic behavior of modular representations over abelian $p$-groups
In this paper, we prove some results on the asymptotic behavior arising in modular representation theory over abelian $p$-groups. First, we embed the representation ring of a cyclic $p$-group into a real algebra of functions. Second, we calculate the asymptotic order of the dimension of the core of $n$-th tensor power of a direct sum of syzygies and cosyzygies of the trivial module, which is of the form $C\gamma^nn^\alpha$. This result leads to a negative answer to a question by Benson and Symonds, that is, the dimension of the core of $M^{\otimes n}$ for certain $\Omega$-algebraic module $M$ is not eventually recursive. Third, we give a systematic way of computing the core series of $\Omega$-algebraic modules. Finally, we show the existence of a transcendental core series, which comes from iterated syzygy modules of the trivial representation.
Read the full abstract
Dynamical Systems
Relaxed Newton's Method as a Family of Root-finding Methods: Dynamics and Convergence
Relaxed Newton's method is a one-parameter family of root-finding methods that generalizes the classical Newton's method. When viewed as a rational map on the Riemann sphere, this family exhibits rich and subtle global dynamics that depend both on the underlying polynomial and on the relaxation parameter. In this paper, we investigate the complex dynamical behavior of relaxed Newton maps associated with complex polynomials. We first characterize rational maps that arise as relaxed Newton maps in terms of the multipliers of their fixed points. Our main results identify several explicit classes of polynomials for which relaxed Newton's method is convergent for all parameters $h$, in the sense that the Fatou set consists precisely of the basins of attraction of the roots. We further show that this favorable behavior does not hold in general: for any fixed relaxation parameter $h\in \{z:|z-1|<1\}$, there exists a generic cubic polynomial for which the relaxed Newton map fails to be convergent. Additional results include a complete characterization of when the Julia set is a straight line, an analysis of symmetry groups arising from rotational invariance, and sufficient conditions ensuring that all immediate basins of attraction are unbounded.
Read the full abstract
Mathematical Physics
Autoparallels and the Inverse Problem of the Calculus of Variations
We prove that autoparallel curves associated with a torsion-free but not necessarily metric-compatible affine connection can be derived from an action principle. We explicitly construct the action functional and show by standard variational techniques that it produces the desired equations. Our analysis is based on systematically solving the inverse problem of the calculus of variation and the associated Helmholtz conditions. This demonstrates that the dynamics of autoparallels admit a consistent variational formulation even in the presence of non-metricity. Our results provide a variational framework for particle motion in metric-affine geometries and thereby contribute to the mathematical foundations of the geodesic principle in relativistic gravity.
Read the full abstract
Dynamical Systems
Regular and chaotic Welander oscillations in a four-dimensional conceptual model for the Atlantic Meridional Overturning Circulation
The Atlantic Meridional Overturning Circulation (AMOC) is a key component of the Earth's climate. Evidence indicates a twentieth-century weakening, and enhanced freshwater input to the subpolar North Atlantic may further reduce overturning strength. We present and study a conceptual four-dimensional, single-hemisphere box model with three compartments: a tropical surface box, a subpolar surface box, and a large deep-water box. Advective exchange couples the surface boxes and vertical exchange with the deep ocean is represented by a smooth convective-adjustment scheme. A comprehensive bifurcation analysis reveals an equilibrium structure with up to four coexisting overturning states, together with regimes of bistability and tristability. We identify families of periodic solutions and chaotic attractors with a clear timescale separation: a millennial oscillation is modulated by faster decadal-to-centennial variability arising from episodic shutdowns of subpolar convection. As prescribed freshwater fluxes increase, shutdown events become more frequent and the background overturning weakens. Additionally, for certain values of freshwater influx, the dynamics become chaotic, producing an irregular on-off switching of convection.
Read the full abstract
Dynamical Systems
Statistical regularity and linear response of Mather measures for Tonelli Lagrangian systems
We study the statistical regularity of Mather measures associated with $C^1$ perturbations of a Tonelli Lagrangian. When the unperturbed Mather measure is supported on a quasi-periodic torus with a Diophantine frequency, we establish Hölder continuity of the perturbed Mather measure with respect to the perturbation parameter. The Hölder exponent is shown to depend explicitly on the Diophantine index of the frequency. We also discuss the possibility of achieving Lipschitz regularity using KAM theory.
Read the full abstract
Probability
A note on geometric {\alpha}-stable processes and the existence of ground states for associated Schrödinger operators
In this paper, we establish the existence of transition density for geometric $\alpha$-stable processes by using the property of self-decomposability–a fundamental concept in the theory of Lévy processes. In contrast to traditional and analytic methods that often rely on the $L^{1}$-integrability of the characteristic function, our approach is purely probabilistic and focuses on the structural regularity of the Lévy measure. As an application, we prove the existence of ground states for Schrödinger operators associated with recurrent geometric stable processes.
Read the full abstract
Dynamical Systems · Cross-listed from cs.MA
Feedback-Coupled Memory Systems: A Dynamical Model for Adaptive Coordination
This paper develops a dynamical framework for adaptive coordination in systems of interacting agents referred to here as Feedback-Coupled Memory Systems (FCMS). Instead of framing coordination as equilibrium optimization or agent-centric learning, the model describes a closed-loop interaction between agents, incentives, and a persistent environment. The environment stores accumulated coordination signals, a distributed incentive field transmits them locally, and agents update in response, generating a feedback-driven dynamical system. Three main results are established. First, under dissipativity, the closed-loop system admits a bounded forward-invariant region, ensuring dynamical viability independently of global optimality. Second, when incentives depend on persistent environmental memory, coordination cannot be reduced to a static optimization problem. Third, within the FCMS class, coordination requires a bidirectional coupling in which memory-dependent incentives influence agent updates, while agent behavior reshapes the environmental state. Numerical analysis of a minimal specification identifies a Neimark-Sacker bifurcation at a critical coupling threshold ($\beta_c$), providing a stability boundary for the system. Near the bifurcation threshold, recovery time diverges and variance increases, yielding a computable early warning signature of coordination breakdown in observable time series. Additional simulations confirm robustness under nonlinear saturation and scalability to populations of up to $N = 10^{6}$ agents making it more relevant for real-world applications. The proposed framework offers a dynamical perspective on coordination in complex systems, with potential extensions to multi-agent systems, networked interactions, and macro-level collective dynamics.
Read the full abstract
Numerical Analysis
Deep Domain Decomposition Method for Solving the Variational Inequality Problems
By integrating physics-informed neural network (PINN) techniques with domain decomposition method, a deep domain decomposition method is presented for solving elliptic variational inequality problems. Based on the Ritz variation method, the elliptic variational inequality problem is firstly reformulated as an optimization problem, and then the subproblem in each subdomain is solved by using the Ritz-PINN method, which the parameters in the network are updated by the Adam optimizer, and the residual-adaptive training by introducing a residual-adaptive dataset update strategy to gradually guide the model to learn more complex regions. Additionally, the impact of overlapping regions on the performance of the new algorithm is explored. Numerical results demonstrate the effectiveness of the proposed algorithm, the mean square error can be reached 1.0e-07, and the number of iterations is independent of grid length h under uniform overlap conditions.
Read the full abstract
Optimization and Control
Forward and Backward Reachability Analysis of Closed-loop Recurrent Neural Networks via Hybrid Zonotopes
Recurrent neural networks (RNNs) are widely employed to model complex dynamical systems due to their hidden-state structure, which inherently captures temporal dependencies. This work presents a hybrid zonotope-based approach for computing exact forward and backward reachable sets of closed-loop RNN systems with ReLU activation functions. The method formulates state-pair sets to compute reachable sets as hybrid zonotopes without requiring unrolling. To improve scalability, a tunable relaxation scheme is proposed that ranks unstable ReLU units across all layers using a triangle-area score and selectively applies convex relaxations within a fixed binary limit in the hybrid zonotopes. This scheme enables an explicit tradeoff between computational complexity and approximation accuracy, with exact reachability as a special case. In addition, a sufficient condition is derived to certify the safety of closed-loop RNN systems. Numerical examples demonstrate the effectiveness of the proposed approach.
Read the full abstract
Numerical Analysis
Deep Ritz Physics-Informed Neural Network Method for Solving the Variational Inequality
Variational inequalities are widely applied in mechanical engineering, fluid penetration, transportation, and other fields. In this paper, a Deep Ritz method based on Physics-Informed Neural Networks (PINNs) is proposed to enhance the accuracy and efficiency of solving elliptic variational inequalities. The Ritz variational method is firstly utilized to transform the variational inequality problem into an optimization problem. Then Bayesian optimization is employed to tune the weights of the loss function, and a residual-based adaptive dataset update strategy is introduced to improve the convergence and accuracy of the model. Numerical experiments show that the proposed method can effectively approximate the analytical solution.