生成时间: 2026年09月01日 09:33:16
重点关注论文数量: 5
了解领域论文数量: 7
不相关论文数量: 31
重点关注论文(完整分析)
1. 使用FE-HMM方法的局部受限非线性涡流问题的磁场保形多尺度公式
Magnetic Field Conforming Multiscale Formulations for Locally-Confined Nonlinear Eddy Current Problems Using the FE-HMM Method
作者: Innocent Niyonzima, Gérard Meunier, Antoine Marteau, Ruth V. Sabariego, Olivier Chadebec, Nicolas Galopin, Christophe Geuzaine
类别: math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30542v1
摘要翻译: 磁性复合材料常用于电能转换,其中往往包含相互绝缘的铁磁夹杂物以减少涡流损耗。针对这些复合材料的数值模型必须足够稳健,以处理由非线性磁性夹杂物和胞元中显著受限涡流的存在而可能引起的收敛性问题。本文提出了一种周期框架下磁性复合材料的$\mathbf{h}$协调多尺度公式。所提方法使用异质多尺度方法(HMM),并包含两个细观尺度问题:一个用于升尺度化均匀化磁通密度$\mathbf{B}_M$的磁准静态问题,以及一个用于升尺度化宏观增量磁阻率$(\partial \mathbf{B}_M/\partial \mathbf{H}_M)$的静磁问题。此外,该方法在宏观和细观尺度上都采用松弛的Newton–Raphson迭代,以缓解与$\mathbf{h}$协调公式中非线性BH本构关系相关的众所周知的NR收敛问题。利用具有线性和非线性BH曲线的二维和三维理想化周期软磁复合材料评估了该公式的准确性和性能。此外,论文证明了磁准静态细观尺度问题可以被一个磁体 a 所替代。
Comment: 31 pages, including the bibliography, 14 figures, 1 algorithm, 1 table, paper accepted for publication (but not published yet) in SIAM Multiscale Modeling and Simulations
链接: http://arxiv.org/abs/2608.30542v1
1. 研究对象和背景
本文研究磁性复合材料中局部受限涡流问题的多尺度数值建模. 研究对象是周期性复合材料, 其微观结构含有铁磁夹杂物, 夹杂物彼此绝缘以降低涡流损耗. 在低频电磁设备中, 控制方程为磁准静态 (MQS) 近似下的 Maxwell 方程组: $\partial_t \mathbf{b}^{\varepsilon} + \operatorname{curl} \mathbf{e}^{\varepsilon} = 0$, $\operatorname{curl} \mathbf{h}^{\varepsilon} = \mathbf{j}^{\varepsilon}$, $\operatorname{div} \mathbf{b}^{\varepsilon} = 0$, 其中 $\varepsilon$ 表示多尺度特征, 本构关系为 $\mathbf{j}^{\varepsilon} = \sigma^{\varepsilon} \mathbf{e}^{\varepsilon} + \mathbf{j}_{s}$, $\mathbf{b}^{\varepsilon} = B(\mathbf{h}^{\varepsilon}) = \mu(\mathbf{h}^{\varepsilon}) \mathbf{h}^{\varepsilon}$. 求解域 $\Omega$ 分为导电区 $\Omega_c$ 和非导电区 $\Omega_c^C$, 边界满足 $\mathbf{n} \times \mathbf{h}^{\varepsilon}|_{\Gamma} = 0$, 初始条件 $\mathbf{h}^{\varepsilon}(\mathbf{x},0) = \mathbf{h}^{\varepsilon}_0(\mathbf{x})$. 弱形式为: 求 $\mathbf{h}^{\varepsilon} \in C^1([0,T]; V^I(\Omega))$ 使得 $(\partial_t B(\mathbf{h}^{\varepsilon}), \mathbf{v})_{\Omega} + (\rho \operatorname{curl} \mathbf{h}^{\varepsilon}, \operatorname{curl} \mathbf{v})_{\Omega_c} = 0$ 对所有 $\mathbf{v} \in V^0(\Omega)$ 成立, 其中 $\rho = 1/\sigma$, $V^I(\Omega)$ 和 $V^0(\Omega)$ 是 $H(\operatorname{curl}; \Omega)$ 中带有电流约束和零切向边界的空间. 本文针对的问题类别是所谓’局部受限涡流问题’, 即导电域在周期单元内是不连通的 (disconnected conducting domain), 此时宏观问题为静磁问题, 而介观单元内仍保持磁准静态. 这类问题在经典渐近均匀化中因均匀化电导率趋于零而难以处理. 此外, $\mathbf{h}$-conforming 格式中非线性 $B$-$H$ 律直接导致 Newton-Raphson (NR) 迭代的收敛困难, 且当集肤深度 $\delta$ 与微结构尺寸 $l_c$ 同量级时, 尺度分离假设 $l_c \ll \delta$ 被破坏. 本文旨在填补如下空白: 提出一个能够在 3D 周期框架下处理非线性、局部受限涡流、且允许瞬态模拟的 $\mathbf{h}$-conforming 多尺度格式, 并避免 NR 不收敛问题.
2. 主要定理或主要结果
本文的主要结果如下: (1) 提出基于 HMM 的 $\mathbf{h}$-conforming 多尺度格式, 在宏观尺度求解磁静力型均匀化方程, 在介观尺度构造两个辅助问题: 一个磁准静态 (MQS) 问题用于升尺度均匀化磁通密度 $\mathbf{B}_M$, 一个磁静力问题用于升尺度宏观增量磁阻率 $\partial \mathbf{B}_M / \partial \mathbf{H}_M$. 具体地, 宏观有效关系通过体积平均定义: $\mathbf{B}_M = \frac{1}{|\Omega_m|} \int_{\Omega_m} \mathbf{b}_m \, d\Omega_m$, 而增量磁阻率由静磁单元问题解得 (提供文本中未给出完整闭式公式). (2) 通过量纲分析归一化 Maxwell 方程 (3.1)–(3.4), 证明了在涡流问题中尺度分离条件 (3.10) 中的 $l_{c,i} = \varepsilon L_{c,i} \ll \delta_c$ 可能不满足, 从而说明经典渐近展开需要沿 [6,7] 的思路重新推导. (3) 针对局部受限涡流问题, 提出了将介观 MQS 问题在特定条件下简化为静磁问题的判据: 当单元内磁功率 $P_m^{mag}$ 远大于涡流损耗功率 $P_m^{eddy}$ (即 $P_m^{eddy}/P_m^{mag} \ll 1$) 时, 两者给出相同的均匀化磁通密度; 否则动态 MQS 解与静磁解不同 (见第 3.2 节及图 2–3). 文中通过数值算例’演示’了这一判据, 但未给出严格的定理证明. (4) 在宏观和介观两个尺度上均采用松弛 Newton-Raphson 迭代, 以缓解 $\mathbf{h}$-conforming 格式中因非线性 $B$-$H$ 曲线导致的 NR 不收敛. (5) 数值验证: 对 2D 和 3D 理想化周期软磁复合材料, 在线性和非线性 $B$-$H$ 曲线下, 将 HMM 结果与细尺度参考解对比, 验证了准确性; 并做了强扩展和弱扩展性能测试. 需要说明的是, 由于提供文本不完整, 具体误差阶、收敛率定理编号等细节无法从提供内容中确认.
3. 研究方法、关键技术和核心工具
本文方法的核心障碍有两点: (i) 局部受限涡流导致均匀化电导率消失, 经典渐近展开不适用; (ii) $\mathbf{h}$-conforming 格式中非线性 $B$-$H$ 律使 Newton-Raphson 迭代鲁棒性差. 作者的真正创新点在于: 针对局部受限涡流问题, 不再试图升尺度材料参数, 而是采用 HMM 框架直接升尺度电磁场量, 即通过介观单元问题的解来构造宏观磁通密度和增量磁阻率. 具体工具包括: (1) 无量纲化与渐近分析: 利用特征长度 $L_c$, 集肤深度 $\delta_c = \sqrt{2/(\sigma_c \omega_c \mu_c)}$, 波长 $\lambda_c$ 等, 将 Maxwell 方程归一化为 (3.4), 然后忽略位移电流得到 (3.6). 通过分析各项系数, 指出当 $\delta_c$ 与 $l_c$ 同量级时, 标准渐近展开的阶数假设失效. 这一分析为后续提出新的升尺度策略提供了理论依据. (2) 双介观问题构造: 在 HMM 框架下, 每个宏观点 $\mathbf{x}$ 处定义周期单元 $\Omega_m$, 求解带周期边界条件或绝缘边界的单元问题. 其中 MQS 单元问题包含时间导数项 $\partial_t \mathbf{b}_m$, 用于捕获单元内涡流对 $\mathbf{B}_M$ 的影响; 静磁单元问题则用于计算恒定磁化下的响应, 从而获得增量磁阻率 $\partial \mathbf{B}_M / \partial \mathbf{H}_M$. 文中提到通过边缘平均或面平均等技术准确升尺度 $\mathbf{H}_M$, 并引入涡流产生的磁化 $\mathbf{M} = \frac{1}{|\Omega_m|} \int_{\Omega_m} \frac{1}{2}(\mathbf{x}-\mathbf{x}_c) \times \mathbf{j}_m \, d\Omega_m$ 来修正宏观磁场, 避免直接取 $\mathbf{h}_m$ 平均值带来的误差. (3) 松弛 Newton-Raphson (NR) 方案: 在宏观与介观非线性系统求解中, 采用松弛因子或阻尼 NR 迭代. 文中并未给出具体松弛参数公式, 但强调这一做法可显著提升收敛性. (4) 数值离散: 在导体内使用一阶 Nédélec 元 $\mathrm{ND}_1(T) = \{\mathbf{a} + \mathbf{b} \times \mathbf{x}\}$, 在非导体区域使用 $P_1$ 节点元表示梯度场; 通过全局基函数和树-余树规范处理电流约束与拓扑. 强/弱扩展性测试表明算法可并行高效实现. 需要注意的是, 提供文本未给出具体时间离散格式和松弛参数取值, 这些细节无法从当前内容中确认.
4. 与之前工作的比较
与已有工作的比较: (1) 与 Bottauscio 等人的工作相比: 他们基于渐近展开、MsFEM 和 VMS 求解 2D/3D 非线性问题, 能处理周期复合材料的静态和动态问题, 但对局部受限涡流动态问题不准确. 本文的 HMM 格式专门针对这类问题, 在介观保留 MQS 从而捕获涡流. (2) 与 [58,60,59,48,73] 的 HMM 工作相比: 这些工作已用于 2D 叠片和理想软磁复合材料, 最近扩展到 3D 非线性和多谐波稳态, 但不能模拟瞬态过程. 本文的 $\mathbf{h}$-conforming 格式直接基于时间域, 支持瞬态模拟. (3) 与 [51,50] 的多尺度 $\mathbf{b}$-和 $\mathbf{h}$-conforming 格式相比: 它们处理了 2D 问题包括绕组, 并提出了适合受限涡流的升尺度技术, 但仅限线性问题. 本文将其扩展到非线性 $B$-$H$ 曲线, 并解决 NR 收敛问题. (4) 与 Schöbinger 和 Hollaus 的非线性叠片 MsFEM 相比: 他们引入了模型降阶, 但方法固有地限于叠层复合材料. 本文适用于任意周期单元 (至少理想化软磁复合材料). (5) 与 Ren 等人的 LOD/VMS 方法相比: 他们处理 2D 非线性涡流并扩展到拓扑优化; 本文面向 3D 更一般的微结构. (6) 与平均场方法相比: 平均场方法不能正确应对具有强受限涡流的 3D 非线性问题, 本文 HMM 能够处理. (7) 与经典均匀化方法 [36,38,37] 相比: 那些方法用于叠片和绕组, 不适用于任意周期/非周期单元. 本文在周期框架下对任意单元构造介观问题. (8) 关于尺度分离失效: 本文明确指出现有方法通常隐含假设 $l_c \ll \delta_c$, 而本文通过量纲分析指出在涡流问题中该条件可能被破坏, 并提出了基于功率比 $P_m^{eddy}/P_m^{mag}$ 的判定准则. 这一对比是本文的重要贡献. 代价: 本文方法需要在每个宏观点求解两个介观非线性问题, 计算成本较高; 但通过强/弱扩展性展示可并行化. 另外, 作者未显式讨论非周期或随机分布夹杂物; 文中说明只考虑周期性设置.
2. 带非线性跳跃条件的抛物型界面问题的浸入界面方法
An Immersed Interface Method for Parabolic Interface Problems with Nonlinear Jump Conditions
作者: So-Hsiang Chou, Patrick Nyadjo Fonga
类别: math.NA
发布日期: 2026-08-30
arXiv ID: 2608.30062v1
摘要翻译: 我们针对一维非线性抛物型界面问题发展了一种浸入界面有限差分方法,其跳跃条件为 [ [u]_α=λu^+u^-。] 该方法将 Crank–Nicolson 浸入界面离散与非线性界面条件的 (s) 参数约化相结合,从而将非线性耦合约化为一个标量二次方程。我们还讨论了将牛顿迭代与浸入界面离散相结合的两种不同观点,即 IIM–Newton 和 Newton–IIM 格式。文中给出了数值实验,以说明该方法的性质和精度。
Comment: 17 pages,6figures
链接: http://arxiv.org/abs/2608.30062v1
1. 研究对象和背景
本文研究一维非线性抛物型界面问题(nonlinear parabolic interface problem, NPIP)的浸入界面有限差分方法. 研究对象为如下方程(对应原文(1)-(5)):
$$ Lu := u_t - (\beta u_x)_x + q u = f, \quad x \in \Omega^- \cup \Omega^+, \ t \in (0,T], $$$$ u(a,t) = \xi(t), \quad u(b,t) = \eta(t), $$$$ [u]_\alpha = \lambda u^+ u^-, \quad [\beta u_x]_\alpha = 0, $$$$ u(x,0) = g(x), \quad x \in \Omega^- \cup \Omega^+. $$其中 $\alpha$ 为固定界面点, $\Omega^- = (a,\alpha)$, $\Omega^+ = (\alpha,b)$, 扩散系数 $\beta = \beta(x) > 0$ 为分段常数: $\beta = \beta^-$ 在 $\Omega^-$ 内, $\beta = \beta^+$ 在 $\Omega^+$ 内. 记号 $[u]_\alpha = u^+ - u^-$, $[\beta u_x]_\alpha = \beta^+ u_x^+ - \beta^- u_x^-$, 其中 $u^\pm = \lim_{x \to \alpha^\pm} u(x,t)$. 非线性界面条件 $[u]_\alpha = \lambda u^+ u^-$ 耦合了界面两侧的解迹, 是问题的主要非线性来源.
这类问题广泛出现于涉及界面、膜与非均匀介质的扩散与输运模型中. 在数值方法层面, 处理间断系数界面问题的常用手段包括拟合有限差分、浸入界面方法(IIM)、浸入有限元方法以及相关非拟合网格离散. 其中由 LeVeque 和 Li 引入的浸入界面方法(参考文献[4])通过在界面附近非规则网格点处添加修正项, 在笛卡尔网格上保持了高阶精度. 后续对椭圆与抛物型界面问题的 IIM 发展可见文献[1,5,6], 浸入有限元方法的相关讨论见文献[7].
经典 IIM 的修正项构造依赖于跳跃条件被显式给定. 当界面条件为非线性时, 跳跃量 $[u]_\alpha$ 不再是已知数据, 而是依赖于未知解自身在界面两侧的迹 $u^+$ 和 $u^-$. 这使得修正项无法直接构造, 也使得非线性与离散格式的耦合变得复杂. 已有的基于拟合网格的有限差分方法(文献[2])和非线性跳跃的 $s$ 参数化思想(文献[3])提供了些线索, 但尚未在浸入界面框架下系统发展.
本文试图填补的空白是: 在经典 IIM 框架中纳入非线性跳跃条件 $[u]_\alpha = \lambda u^+ u^-$, 使得界面修正仍然保持局部性和稀疏性, 同时不破坏二阶全局精度. 作者将 $s$ 参数化引入 IIM, 将解分解为 $u = u_0 + s u_1$, 从而把非线性界面耦合约化为一个标量二次方程. 该分解把非线性界面条件与线性体方程分离, 使得复杂的非线性问题在时间方向可以通过 Crank-Nicolson 结合两个线性 IIM 问题来处理. 此外, 作者还对比了两种牛顿迭代与 IIM 离散结合的顺序(IIM-Newton 与 Newton-IIM), 并给出理论上的局部截断误差分析. 这些工作为非线性界面问题的非拟合网格高精度求解提供了新的视角和可行性验证.
2. 主要定理或主要结果
本文的主要结果可归纳为以下几个方面.
线性问题的浸入界面离散与截断误差. 对于带一般线性跳跃条件 $A(t) u^+ - B(t) u^- = w(t)$, $[\beta u_x]_\alpha = v(t)$ 的线性抛物型界面问题(7)-(11), 作者给出了 θ-方法(12)在非规则网格点 $x_j$ 和 $x_{j+1}$ 处的修正项 $C_j$ 和 $C_{j+1}$, 具体公式为(16)和(20). 这些修正项通过泰勒展开并匹配至 $O(h^2)$ 项而得到. 在 $A=B=1$ 时退化为标准已知跳跃情形, 修正项相应简化. 在引理1中, 作者证明了空间截断误差的局部阶: 在规则点 $i \notin \{j,j+1\}$ 有 $\tau_i = O(h^2)$; 在两个非规则点有 $\tau_j = O(h)$ 和 $\tau_{j+1} = O(h)$. 进一步, 作者利用一维问题中非规则缺陷仅支撑在两个网格点以及离散逆算子(离散Green函数)的 $O(h)$ 量级, 说明 $O(h)$ 的局部截断误差并不破坏全局二阶精度. 这是本文对经典 IIM 中“局部低阶但全局高阶”现象的严格论证.
非线性界面条件的 $s$ 参数化. 对于非线性界面条件 $[u]_\alpha = \lambda u^+ u^-$, 作者采用 $s$ 参数化: 设 $u = u_0 + s u_1$, 其中 $u_0$ 满足齐次跳跃 $[u_0]_\alpha = 0$, $[\beta u_{0,x}]_\alpha = 0$(或给定 $v$), 而 $u_1$ 对应单位跳跃 $[u_1]_\alpha = 1$. 将分解代入非线性跳跃条件后, 得到关于 $s$ 的标量二次方程. 该方程的解 $s$ 唯一确定了界面跳跃量, 从而将非线性问题转化为两个线性 IIM 子问题. 这种处理避免了迭代耦合, 也使得 Crank-Nicolson 时间离散能直接适用.
稳态椭圆模型的 IIM-Newton 格式. 为清晰展示非线性代数结构, 作者先考虑稳态椭圆问题(21)-(24). 非线性跳跃可改写为 $w = \lambda (u^-)^2/(1-\lambda u^-)$, 它与未知迹 $u^-$ 相关. 在离散层面, 系统形如 $A_h U = F_h + \Phi(U)$, 其中修正向量 $\Phi(U)$ 仅在两个非规则点 $I$ 和 $I+1$ 处非零. 采用二次拉格朗日外推 $\tilde U^- = U_{I-2}\ell_{I-2}(\alpha) + U_{I-1}\ell_{I-1}(\alpha) + U_I\ell_I(\alpha)$ 来逼近 $u^-(\alpha)$, 并定义离散非线性跳跃 $\mu(U) = \lambda(\tilde U^-)^2/(1-\lambda\tilde U^-)$. 修正项可写为 $\Phi_I = \kappa_I \mu(U) + r_I$, $\Phi_{I+1} = \kappa_{I+1} \mu(U) + r_{I+1}$, 其中 $\kappa_i$ 和 $r_i$ 由线性 IIM 格式确定. 于是 Jacobian $G'(U) = A_h - \Phi'(U)$ 与线性矩阵仅差少数几项, 并且 $\Phi'(U)$ 只涉及 $\mu$ 对 $U_{I-2}, U_{I-1}, U_I$ 的导数, 公式(32)-(34)给出了这些偏导数. 因此 Newton 迭代保持稀疏性, 且避免了全域稠密 Jacobian.
Newton-IIM 的困难. 作者讨论了另一种结合顺序: 先在连续层面对非线性问题线性化, 再对校正方程应用 IIM. 文中指出这需要离散带有分布源和跳跃的校正方程, 在标准 IIM 框架内会产生实际困难, 而 IIM-Newton 则是更自然的选择. 该讨论虽未给出严格分析, 但为非线性 IIM 的实现提供了重要指导.
数值验证. 在椭圆实验中, 作者构造了带 $\lambda = 1.5, \beta^- = 2, \beta^+ = 1$ 的问题, 精确解为(39), 其中常数 $A$ 满足二次方程 $\lambda r A^2 = (r+1)A$, $r = \beta^-/\beta^+$. 方程存在两个解: 连续解 $u_c$ 与间断解 $u_d$. 数值试验显示, 初始迭代取 $t u_d(x_i)$ 时, $0 \le t \le 0.67$ 收敛到 $u_c$, 更大的 $t$ 则可能收敛到 $u_d$, 说明解的多重性与迭代敏感性. 在抛物实验中, 采用 Crank-Nicolson 结合 $s$ 参数化, 并利用精确辅助解提供界面迹, 在离散最大范数下获得了二阶精度. 文中虽未列出误差表, 但明确说明数值结果与前述二阶全局精度相容.
总的来说, 本文的主要结果是一套将非线性跳跃条件 $[u]_\alpha = \lambda u^+ u^-$ 纳入经典浸入界面有限差分框架的完整构造, 包括线性修正项、$s$ 参数化、IIM-Newton 迭代、截断误差分析和数值实验. 它首次在 IIM 文献中系统展示了非线性跳跃条件如何在不牺牲精度和稀疏性的情况下被处理.
3. 研究方法、关键技术和核心工具
本文的方法具有以下技术特点, 这些特点共同解决了非线性跳跃条件带来的核心障碍.
核心障碍: 跳跃未知导致的构造困难
在经典 IIM 中, 跳跃 $[u]_\alpha = w(t)$ 和 $[\beta u_x]_\alpha = v(t)$ 是已知函数, 因此非规则点处的修正项可直接从界面数据算出. 而在非线性问题中, $[u]_\alpha = \lambda u^+ u^-$ 依赖于未知解迹, 修正项无法预先给定. 如果直接迭代求解非线性代数系统, 则每次迭代都要重新计算修正项, 且 Jacobian 可能充满, 破坏稀疏性.
$s$ 参数化: 将非线性耦合化为标量方程
作者采用文献[3]提出的 $s$ 参数化, 但在 IIM 框架下重新诠释. 将解 $u$ 分解为 $u = u_0 + s u_1$, 并定义 $u_0$ 和 $u_1$ 满足特定的线性跳跃条件. 则 $u^+$ 和 $u^-$ 都是 $s$ 的线性函数, 代入非线性条件 $[u]_\alpha = \lambda u^+ u^-$ 后, 得到关于 $s$ 的二次方程. 解出 $s$ 后, 界面跳跃量 $w$ 就确定了. 于是可以像线性情形一样构造修正项. 这相当于把非线性问题在界面处等价地转化为一个代数方程, 其系数来自辅助线性问题的解. 这种处理的创新在于, 它不需要在每次迭代中更新整个离散系统, 而是几乎无额外代价地确定跳跃.
线性修正项的精细推导
对于线性问题, 作者在非规则点 $x_j$ 和 $x_{j+1}$ 处以界面 $\alpha$ 为展开点做泰勒展开: 在 $x_j$ 处展开 $u_{j-1}, u_j$ 利用 $u^-$ 及其导数, 展开 $u_{j+1}$ 利用 $u^+$ 及其导数, 并通过界面条件 $u^+ = (B u^- + w)/A$ 和 $u_x^+ = (v + \beta^- u_x^-)/\beta^+$ 消去 $u^+$ 和 $u_x^+$. 为确定 $u_{xx}^+$, 还需要对 $A u^+ - B u^- = w$ 关于时间求导, 结合原方程在两侧的限制, 得到 $u_{xx}^+$ 的表达式, 其中包含 $A_t, B_t, w_t$ 以及 $q, f$ 的差. 将这些展开代入差分算子并匹配 $u^-$, $u_x^-$, $u_{xx}^-$ 的系数, 便得到修正项公式(16). 类似地, 在 $x_{j+1}$ 处得到(20). 这一推导是标准 IIM 技巧在带时变系数跳跃条件上的推广, 其难点在于正确处理时间导数与界面一致性的关系.
局部截断误差与全局精度的论证
引理1证明了 $\tau_i = O(h^2)$ 在规则点成立, $\tau_j = O(h)$, $\tau_{j+1} = O(h)$ 在非规则点成立. 由于差分算子系数的量级为 $O(h^{-2})$, 泰勒展开中 $O(h^3)$ 的剩余项产生 $O(h)$ 的局部截断. 为了说明这不破坏全局二阶精度, 作者采用了一维离散 Green 函数的表示: 非规则点只有两个, 而 $G_h$ 对单位缺陷的响应为 $O(h)$ 量级, 因此 $O(h)$ 的缺陷贡献为 $O(h) \cdot O(h) = O(h^2)$. 这个论证是 IIM 文献中的常见思想, 但作者在这里给出了清晰的文字说明.
迹近似与拉格朗日外推
为了在离散系统中表述 $\mu(U)$, 需要逼近 $u^-(\alpha)$. 作者采用三个相邻节点 $U_{I-2}, U_{I-1}, U_I$ 的二次拉格朗日外推, 公式给出了基函数 $\ell_{I-2}, \ell_{I-1}, \ell_I(\alpha)$. 这是简单的一侧外推, 对于光滑解具有 $O(h^3)$ 的逼近精度, 与格式的整体精度匹配. 但作者也注明, 在抛物验证实验中, 为了隔离离散误差, 界面迹直接从精确辅助解求得, 而没有采用外推.
IIM-Newton: 先离散后线性化的策略
对稳态问题, 作者构造非线性残量 $G(U) = A_h U - F_h - \Phi(U)$. 关键的观察是 $\Phi(U)$ 只依赖于标量 $\mu(U)$, 且 $\Phi'(U)$ 的稀疏模式非常有限: 只涉及 $U_{I-2}, U_{I-1}, U_I$. 这样 Newton 迭代的每个线性系统都可以用高效的稀疏求解器处理. 对这个策略, 作者还给出了 $d\mu/d\tilde U^-$ 的表达式和偏导数公式(32)-(34), 便于实际实现. 与直接对非线性系统应用牛顿法相比, IIM-Newton 避免了对整体 Jacobian 的稠密填充, 是本文方法实用性的核心.
4. 与之前工作的比较
Newton-IIM 的讨论 作者还分析了另一种顺序: 先在连续问题中对解 $u$ 做牛顿线性化, 得到校正方程, 再对校正方程应用 IIM. 但由于校正方程含有 $\delta u$ 的跳跃项和非线性导数项, 其界面条件系数依赖于当前迭代值, 在标准 IIM 框架中难以构造一致的修正项. 作者指出, IIM-Newton 是更可行的选择. 这个讨论虽然没有进行严格的数值比较, 但为后续研究提供了方向.
数值实验中的技术处理
在椭圆问题中, 精确解已知, 作者利用它验证了牛顿迭代对不同初始值的敏感性. 在抛物问题中, 采用 Crank-Nicolson 与 $s$ 参数化相结合, 在每个时间层上先根据已知辅助解计算 $s$, 再组装线性方程组. 这种验证方式避免了额外迹重构误差, 使实验能够聚焦于 IIM 离散和 $s$ 参数化本身. 这体现了作者在实验设计上的细致考虑.
总体而言, 本文的方法论创新在于通过 $s$ 参数化和 IIM-Newton 策略, 将非线性界面问题转化为少量线性 IIM 问题和一个标量方程, 同时保持了离散矩阵的稀疏性与二阶收敛精度. 这些工具在文献中多为零散出现, 而本文进行了系统整合和数值验证.
4. 与之前工作的比较
本文的比较对象包括经典 IIM 及其在非线性跳跃上的应用, 以及 $s$ 参数化的历史. 下面从几个维度展开.
与经典 IIM(LeVeque-Li 等)的比较
经典 IIM(参考文献[4])专门处理预设跳跃条件, 修正项直接由已知 jump 数据构造. 本文将其推广到非线性跳跃 $[u]_\alpha = \lambda u^+ u^-$, 这是 IIM 文献中的首次系统尝试. 经典 IIM 中的修正项在本文中保留了线性框架, 但把跳跃量 $w$ 替换为未知迹的非线性函数 $\mu(U)$. 这种替代在数学上并无本质困难, 但在理论分析和数值实现上引入了新的非线性耦合. 本文通过 $s$ 参数化避免了迭代耦合, 从而使非线性处理几乎不增加额外成本. 需要指出的是, 经典 IIM 通常还会分析椭圆问题的最优收敛性, 而本文只给出了空间截断误差的局部阶和启发式解释, 并未证明非线性问题的全局 $L^\infty$ 误差估计. 这是与最新 IIM 收敛性理论(如针对界面问题的 $H^2$ 或 $L^2$ 误差估计)相比的一大简化.
与 $s$ 参数化来源(文献[3])的关系
摘要和引言中明确说明 $s$ 参数化由文献[3]提出. 从上下文看, 文献[3]很可能是针对界面条件为 $[u] = \lambda u^+ u^-$ 的某种方法, 但未必是浸入界面有限差分. 本文的贡献在于将 $s$ 参数化嵌入 IIM 的离散框架, 并利用它构造出局部化的非线性修正项. 因此, 本文不是提出新的参数化, 而是展示了该参数化与 IIM 的相容性. 作者没有详细比较文献[3]中的具体格式, 但从本文的数值实验可以看出, $s$ 参数化在 IIFDM 中能顺利计算.
与拟合有限差分方法(文献[2])的对比
文献[2]研究非线性跳跃的拟合有限差分方法. 拟合方法要求网格线与界面对齐, 通常需要移动网格或使用曲线坐标. 本文采用非拟合笛卡尔网格, 界面可以位于网格单元内部, 这更符合 IIM 的初衷. 虽然本文是一维问题, 但 IIM 的优势在于容易推广到多维和复杂界面, 而拟合方法在多维中网格生成较复杂. 此外, 拟合方法往往需要显式处理界面两侧的连续性, 而 IIM 通过修正项自动满足. 但本文也付出了代价: 非规则点处的截断误差只有 $O(h)$, 需要依赖离散 Green 函数的特殊性质来保证整体二阶精度, 这在多维情形中不一定成立.
两种牛顿顺序的系统讨论
本文明确区分了 IIM-Newton 和 Newton-IIM, 并指出 Newton-IIM 的标准实现会遇到困难. 据我所知, 这种对“线性化与离散化顺序”的细致讨论在此前 IIM 文献中较少出现. IIM-Newton 保持了 Jacobian 的稀疏性, 而 Newton-IIM 在连续层面线性化后, 校正方程的界面跳跃仍然是非线性的, 导致修正项难以构造. 作者没有进行数值对比, 但这为实践者提供了重要警示. 然而, 也可以考虑将 Newton-IIM 与不同的离散策略(例如虚拟节点法)结合, 或许能克服困难, 这超出了本文范围.
与处理其他非线性界面条件的方法比较
本文只考虑了乘积型非线性 $\lambda u^+ u^-$. 在材料科学和电化学中, 还可能出现 $[u] = g(u^+,u^-)$ 的复杂函数. 本文的 $s$ 参数化依赖于跳跃关于 $u^+ u^-$ 的特殊代数形式, 对于更一般的非线性, 未必能约化为二次方程. 文献中也有用 Schur complement 或迭代法处理一般非线性跳跃, 但本文的方法更直接. 因此, 本文的适用范围有一定限制, 但为类似乘积型非线性提供了模板.
局限性与代价
本文的所有分析均在一维和分段常数 $\beta$ 下进行. 对于变系数 $\beta(x)$ 和多维情形, 修正项推导会更复杂, 且离散 Green 函数的 $O(h)$ 性质未必成立. 此外, 作者在数值验证抛物问题时使用了精确解的迹, 而没有在整体计算中实现一致的迹重构. 作者承认, 迹恢复系统分析留待后续工作. 这意味着在实际应用中, 需要额外的同侧重构或耦合求解, 可能带来额外误差. 从误差理论角度看, 本文只给出了空间截断误差的渐近阶, 没有给出非线性问题解的存在唯一性、牛顿迭代收敛性等理论分析. 这些都是在未来工作中需要补全的.
从整体上看, 本文的推进在于将经典 IIM 扩展到非线性跳跃, 并且提供了有效的参数化手段和两种离散-线性化顺序的洞察. 与已发表的 IIM 扩展工作相比, 本文的贡献是概念清晰、实现简单, 但理论分析尚不完善. 作者尚未显式展开与更近期 IIM 变体(如高阶 IIM, 隐式 IIM)的比较, 上述比较仅基于本文引用的文献信息.
3. 界面问题中来自大局部误差的二阶精度:离散 Green 函数分析
Second-Order Accuracy from Large Local Errors in Interface Problems: A Discrete Green’s Function Analysis
作者: So-Hsiang Chou, Patrick Nyadjo Fonga
类别: math.NA
发布日期: 2026-08-30
arXiv ID: 2608.30020v1
摘要翻译: 界面问题的有限差分方法通常在界面附近表现出大的局部截断误差,特别是当系数或解导数存在不连续时。然而,许多此类格式仍能达到二阶精度,这一现象无法由标准的逐点相容性论证完全解释。
Comment: 23 pages
链接: http://arxiv.org/abs/2608.30020v1
1. 研究对象和背景
本文研究一维和二维界面问题中有限差分格式的收敛性, 重点关注局部截断误差很大但全局误差仍达到二阶精度的现象. 设一维界面问题为
$$ -\frac{d}{dx}\left(\beta(x)\frac{du}{dx}\right)=f,\quad x\in(a,\alpha)\cup(\alpha,b),\quad u(a)=u(b)=0, $$其中 $\beta(x)>0$ 在界面 $x=\alpha$ 处可能间断. 界面上施加跳跃条件 $[u]_\alpha=w$, $[\beta u_x]_\alpha=v$. 在均匀网格 $x_i=a+ih$ 上, 数值解 $U_i$ 满足守恒型通量格式
$$ -\beta_{i-\frac12}U_{i-1}+\left(\beta_{i-\frac12}+\beta_{i+\frac12}\right)U_i-\beta_{i+\frac12}U_{i+1}=h^2 r_i^h, $$其中 $\beta_{i+\frac12}$ 为数值面系数. 当界面 $\alpha$ 位于两个相邻网格点 $x_I 本文主要结果如下. 对一维守恒型通量格式 (1.2), 设 $A_h$ 为三对角矩阵, $L_h=h^{-2}A_h$, 离散 Green 函数 $G_{ij}=(L_h^{-1})_{ij}$. 定理 2.1 和命题 2.2 给出显式逆公式: 若 $s_k=1/\beta_{k-\frac12}$ 且 $S(c,d)=\sum_{k=c}^d s_k$, 则 从而 此外, $G_{ij}>0$, $G_{ij}=G_{ji}$. 加权 Green 增量具有两平台结构: 对每个固定源点 $j$, 数值通量 $\beta_{i+\frac12}(G_{i+1,j}-G_{ij})$ 在 $i Green 函数满足关键估计 (推论 2.3): 存在仅依赖于 $b-a$ 和 $\beta_{\min}=\min_i\beta_{i+\frac12}$ 的常数 $C$, 使得2. 主要定理或主要结果
设界面截断误差满足 $\tau_I=O(1)$, $\tau_{I+1}=O(1)$, 且 $\tau_I+\tau_{I+1}=O(h)$. 那么误差可分解为
$$ e_i=G_{iI}(\tau_I+\tau_{I+1})+(G_{i,I+1}-G_{iI})\tau_{I+1}, $$进而得到 $G_{iI}(\tau_I+\tau_{I+1})=O(h^2)$ 和 $(G_{i,I+1}-G_{iI})\tau_{I+1}=O(h^2)$, 故 $e_i=O(h^2)$. 这解释了个别 $O(1)$ 局部误差如何通过守恒抵消和 Green 函数增量变成 $O(h^2)$ 全局误差.
本文定义一类可容许的守恒界面格式: 这些格式具有共同的切割面通量、正的面系数、界面相邻节点各自的截断误差为 $O(1)$, 并满足组合一致性条件 $\tau_I+\tau_{I+1}=O(h)$. 对于该类中所有格式, Green 函数论证保证最大范数下的二阶收敛.
加权谐波离散是该类的一个重要成员: 当 $\alpha=x_I+\theta h$, $0<\theta<1$ 时, 切割区间上的有效系数为
$$ \widehat{\beta}_H=\left(\frac{\theta}{\beta_-}+\frac{1-\theta}{\beta_+}\right)^{-1}. $$并结合跳跃修正后, 相邻截断误差满足 $\tau_I+\tau_{I+1}=O(h)$.
对于二维笛卡尔网格上平行于坐标轴的平坦界面, 本文证明相同的守恒抵消机制沿与界面相交的网格行/列逐行保持, 并指出完整的多维 Green 函数分析所需的额外条件.
3. 研究方法、关键技术和核心工具
本文的核心技术障碍是: 界面附近局部截断误差为 $O(1)$, 经典逐点相容性分析无法解释为何全局误差仍为二阶. 作者的主要创新在于将误差分析完全建立在离散 Green 函数的精确结构上, 而不是依赖逐点一致性. 证明框架包括以下几个关键技术步骤.
首先, 对一般守恒型通量三对角算子 $A_h$, 作者利用离散通量表示 $q_i=\beta_{i-\frac12}(U_i-U_{i-1})$, 将方程 $A_hU=e_j$ 化为 $q_i-q_{i+1}=\delta_{ij}$. 通过求解分段常数的离散通量并利用边界条件 $U_0=U_{n+1}=0$, 得到 $A_h^{-1}$ 的显式公式 (定理 2.1). 这一推导避免了逐点局部截断误差分析, 直接刻画了逆算子的全局性质.
其次, 通过 $L_h=h^{-2}A_h$ 得到 Green 函数 $G_{ij}$ 的显式表示, 并由此导出加权增量 $\beta_{i+\frac12}\Delta_i G(j)$ 的两平台结构. 这一结构是后续误差分解的基础: 它表明 Green 函数在相邻源点之间的变化是 $O(h^2)$, 而 Green 函数本身是 $O(h)$. 推论 2.3 中的最大范数估计 $\|L_h^{-1}\|_\infty\le C$ 保证了离散解的稳定性, 使得局部误差向全局误差的传播被严格控制.
第三, 误差表示 $e_i=-\sum_j G_{ij}\tau_j$ 将局部截断误差与 Green 函数关联. 对于只有两个相邻大误差 $\tau_I,\tau_{I+1}$, 作者采用恒等式
$$ G_{iI}\tau_I+G_{i,I+1}\tau_{I+1}=G_{iI}(\tau_I+\tau_{I+1})+(G_{i,I+1}-G_{iI})\tau_{I+1}, $$将问题转化为估计 $G_{iI}$ 和相邻列差 $G_{i,I+1}-G_{iI}$. 前者为 $O(h)$, 后者为 $O(h^2)$, 因此只需证明 $\tau_I+\tau_{I+1}=O(h)$ 即可得到 $O(h^2)$ 误差. 这一分解巧妙地隔离了“大局部误差”和“抵消”两个要素.
第四, 作者将上述分析提炼为可容许格式族的充分条件, 并针对加权谐波离散在附录 A 中进行了详细的局部一致性和通量平衡计算, 直接验证 $\tau_I=O(1)$, $\tau_{I+1}=O(1)$ 且 $\tau_I+\tau_{I+1}=O(h)$.
最后, 对二维平坦界面, 作者将一维机制沿网格行/列方向逐行应用. 由于界面临近节点上的截断误差在垂直于界面的方向上满足相同的守恒抵消, 结合 Green 函数在行方向上的张量性质 (或逐行椭圆估计), 可以保持二阶收敛. 不过, 作者也明确指出, 完整的多维 Green 函数分析仍需要额外的技术, 如角点处理和更复杂的 Green 函数结构.
4. 与之前工作的比较
与已有工作的比较可以从几个维度展开.
首先, Beale 和 Layton [1] 早已通过离散椭圆估计和 Green 函数估计指出, 界面附近降低的局部相容性仍可导致二阶解精度, 但他们没有给出离散 Green 函数的显式结构, 也没有将机制归结为守恒通量下的相邻误差抵消. 本文在一维守恒型通量格式中构造了精确的 Green 函数, 首次显式展示了两平台结构, 并将抵消机制表述为 $\tau_I+\tau_{I+1}=O(h)$ 这一简洁条件. 这比 Beale-Layton 的分析更直接, 也更适合指导格式设计.
其次, 与浸入界面方法 (LeVeque-Li [7], Li-Ito [8]) 相比, 这些方法通常需要显式修正跳跃条件, 局部截断误差在大多数点达到高阶, 但界面附近仍可能较大. 本文不侧重于修正技巧, 而是从离散 Green 函数的角度说明, 即使在界面附近保留 $O(1)$ 的局部误差, 只要满足守恒抵消, 全局仍可二阶. 因此, 本文为“无需精确局部修正”的格式提供了理论依据.
第三, 与 Ewing-Iliev-Lazarov [2] 等修正有限体积方法相比, 这些方法强调通过修改界面通量来保持精度. 本文的分析表明, 更本质的不是单个通量的精确性, 而是相邻两节点截断误差之和的一阶小量. 加权谐波平均在 [3] 中已被用作一维界面格式, 本文将其纳入一个更一般的可容许格式族, 并直接从 Green 函数机制证明了其二阶收敛性.
第四, 在稳定性与条件数方面, 本文明确指出 $\|L_h^{-1}\|_\infty$ 与 $h$ 无关, 仅依赖下椭圆界, 而谱条件数 $\kappa_2(L_h)=O(\beta_{\max}/\beta_{\min} h^{-2})$. 这澄清了“大局部误差”不是由不稳定造成的. 以往文献常将界面格式的精度问题与矩阵条件数混为一谈, 本文通过显式逆算子区分了这两者.
第五, 关于多维推广, 作者承认本文仅对笛卡尔网格上平行于坐标轴的平坦界面给出了行/列方向的机制, 完整的多维 Green 函数分析仍未完成. 这是本文的局限: 对于复杂界面或非笛卡尔网格, 需要额外的几何处理, 而本文的显式两平台结构在一般情况下可能不再成立. 与浸入界面方法相比, 本文方法在几何灵活性上有所不足, 但其理论清晰性为理解和设计守恒型界面格式提供了有力工具.
总体而言, 本文相对于以往工作的推进在于: (1) 提供了离散 Green 函数的显式公式和两平台结构, 将误差机制从黑箱估计转化为可验证的结构性质; (2) 给出了可容许格式族的充分条件, 使格式设计与理论分析直接挂钩; (3) 将加权谐波离散纳入统一框架. 代价是限制在一维和特殊二维情形, 且对界面几何有较高要求. 从提供内容中无法确认作者是否将方法推广到曲界面或非均匀网格, 也未与浸入界面方法进行数值比较.
4. 一种用于局部周期椭圆问题高阶均匀化的张量神经网络方法
A Tensor Neural Network Method for High-Order Homogenization of Locally Periodic Elliptic Problems
作者: Huaijia Zhang, Haochen Liu, Xia Ji, Hehu Xie
类别: math.NA, math.AP
发布日期: 2026-08-29
arXiv ID: 2608.29176v1
摘要翻译: 我们针对形式为 $-\nabla\cdot(A(x,x/\varepsilon)\nabla u_\varepsilon)=f$ 的局部周期椭圆多尺度问题,发展了一种高阶张量神经网络(TNN)方法。由于系数同时依赖于慢变量 $x$ 和快周期变量 $y=x/\varepsilon$,高阶胞元问题和宏观校正子方程比经典情形 $A=A(y)$ 更为复杂,且校正子依赖于参数 $x$。我们推导了一个可计算的高阶双尺度展开,并在无边界层情形(包括周期区域和理想边界匹配构型)下证明了部分展开的 $H^1$ 收敛估计。该证明利用了校正子递推结构的递归相容性以及 $H^{-1}$ 中的零均值振荡估计。随后,我们为高维校正子问题构建了 TNN 框架。其张量积结构允许对胞元问题、均匀化系数、宏观源项和损失函数进行确定性的一维求积,从而避免了蒙特卡洛积分误差。数值实现假设系数分量和组装数据具有有限或受控的张量积表示;这一计算假设与分析中使用的通用矩阵值系数类别是分开的。针对标量局部周期系数的实验显示了高阶校正子的准确性。$H^1$ 半范数误差与所证明的估计一致,而逐点归一化的 $L^2$ 误差则显示出形式展开所预测的名义高阶行为。
Comment: 36 pages, 9 figures, 8 tables
链接: http://arxiv.org/abs/2608.29176v1
1. 研究对象和背景
本文研究如下局部周期椭圆多尺度问题:
$$ -\nabla\cdot\left(A\left(x,\frac{x}{\varepsilon}\right)\nabla u_\varepsilon(x)\right)=f(x),\quad x\in\Omega, $$其中 $0<\varepsilon\ll1$ 为微观尺度,$A(x,y)$ 是 $n\times n$ 对称矩阵值函数,对快变量 $y$ 以 $\mathbb{Y}=(0,1)^n$ 为周期,且满足一致椭圆性条件 (2.2)。关键未知量是 $u_\varepsilon$。当系数仅依赖快变量时即经典周期均匀化,其理论已经成熟;本文关注更一般的局部周期情形,其中系数还显式依赖慢变量 $x$,这使得基元胞问题、均匀化系数以及高阶校正子都成为 $x$ 的参数化函数,结构上更加复杂。在周期性设置中,已知均匀化方程由有效张量 $A^*(x)=\int_\mathbb{Y} A(x,y)(I+\nabla_y\chi(x,y))dy$ 给出,且低阶校正子满足简单的胞元问题;但局部周期情形下的高阶展开需要递归求解一系列带参数 $x$ 的胞元问题和宏观校正方程,且高阶校正子与低阶校正子耦合。已有理论工作如 [13,14,15] 主要关注定性均匀化和定量算子估计,但缺乏可计算的高阶渐近层级构造及其数值实现。从数值角度看,直接离散多尺度问题时网格必须解析 $\varepsilon$ 尺度振荡,代价巨大;传统多尺度有限元或异质多尺度方法虽能处理低阶问题,但在高阶局部周期情形下,辅助校正子同时依赖慢快变量且需递归求解,网格法需要反复在大量宏观采样点求解胞元问题或直接在 $\Omega\times\mathbb{Y}$ 上离散,随维数和展开阶数升高而迅速变得不现实。近年来神经网络方法被引入多尺度问题和均匀化,但多数聚焦于首阶均匀化或依赖随机采样计算高维积分,难以保证高阶展开所需的积分精度。本文旨在填补这一空白:一方面构造可直接计算的高阶双尺度展开并给出无边界层设定下的 $H^1$ 收敛证明;另一方面提出张量神经网络 (TNN) 框架,利用张量积结构将高维积分化归为确定性一维求积,从而精确求解高阶校正子层级。文中将分析部分的假设(一般光滑一致椭圆矩阵系数)与数值实现部分的假设(系数及数据具有有限或可截断的张量积表示)明确分离,为后续工作提供了清晰的框架。对于有界区域上的 Dirichlet 问题,高阶双尺度展开通常不满足边界条件,需要边界层校正,本文刻意隔离了边界层问题,在周期性域或理想边界匹配配置下分析。这种设定便于集中研究高阶代数结构与递归计算,而边界层尾巴对内部误差的影响以及有效边界条件的推导留待未来。
2. 主要定理或主要结果
设 $\Omega$ 为 $\mathbb{T}^n$ 或满足理想边界匹配的配置,$A(x,y)$ 对 $y$ 为 $Y$-周期、对称且一致椭圆 (2.2)。令能量空间 $V$ 满足 Poincaré 不等式 (2.25),所有胞元函数取零 $Y$-均值。作者通过 Algorithm 1 构造递归正确子层级,定义部分展开
$$ U_k^*=\sum_{j=0}^{k-1}\varepsilon^j u_j(x,x/\varepsilon)+\varepsilon^k u_k^*(x,x/\varepsilon), $$以及完整展开
$$ U_k=\sum_{j=0}^k \varepsilon^j u_j(x,x/\varepsilon). $$在无边界层设定下,对任意 $k\ge 1$,存在与 $\varepsilon$ 无关的常数 $C$(可能依赖 $k$、$f$ 和 $A$ 的正则性)使得
$$ \|\nabla(u_\varepsilon-U_k^*)\|_{L^2(\Omega)} \le C \varepsilon^{k}. $$这里我们按标准双尺度展开的 $H^1$ 误差阶写出;由于 PDF 摘录未包含定理的完整陈述,精确的指数和常数需参考原文。证明的关键步骤包括:引理 2.1 验证所有递归胞元问题右端零 $Y$-均值,保证可解性;引理 2.2 建立级联恒等式
$$ L_0u_0=0,\quad L_0u_1^*+L_1u_0=0,\quad L_0u_{m+1}^*+L_1u_m+L_2u_{m-1}=\delta_{m1}f,\quad m\ge1; $$引理 2.3 给出零均值振荡估计:若 $g(x,y)$ 对 $y$ 周期且 $\int_Y g(x,y)dy=0$,则 $\int_\Omega g(x,x/\varepsilon)\varphi(x)dx$ 可被 $\varepsilon$ 乘以某范数控制,从而在 $H^{-1}$ 中获得 $\varepsilon$ 阶增益。结合这些估计与 V-椭圆性,得到上述 $H^1$ 误差界。
数值方面,TNN 将所有胞元问题、均匀化系数和宏观源项的高维积分化为确定性一维求积,实验使用标量局部周期系数,计算的各阶正确子与理论一致;$H^1$ 半范数误差符合证明的估计,点归一化 $L^2$ 误差呈现名义高阶行为(后者为实验观察,非严格证明)。此外,Remark 2.1 给出了一维精确边界匹配机制,说明在一维情形下通过规范选取可使截断展开精确满足 Dirichlet 边界条件。
3. 研究方法、关键技术和核心工具
本文的核心技术障碍在于局部周期系数导致正确子层级全部参数化地依赖慢变量,且高阶展开的误差分析需要处理快速振荡残差。作者的主要创新点有三:
(1) 递归构造的可计算格式:通过分解 $D_\varepsilon=\nabla_x+\varepsilon^{-1}\nabla_y$,定义算子 $L_0,L_1,L_2$ (2.3)-(2.5),将级联方程匹配得到正确子递归关系。每个正确子 $u_k$ 被拆分为胞元部分 $\hat u_k$、振荡校正 $\delta u_k$ 和宏观部分 $\tilde u_k$,其中 $\delta u_k=\chi(x,y)\cdot\nabla_x\tilde u_{k-1}$,而 $\tilde u_k$ 由可解性条件决定。这种拆分使得每一步只需处理标准椭圆问题,且宏观方程具有相似的均匀化结构。
(2) 收敛性证明的三个关键引理:引理 2.1 利用周期积分和宏观方程 (2.15)、(2.23) 验证递归胞元右端的零均值性;引理 2.2 将 $L_0,L_1,L_2$ 作用结果合并为极简残差;引理 2.3 是零均值振荡估计,对满足 $\int_Y g=0$ 的周期函数 $g$,它给出了 $\int_\Omega g(x,x/\varepsilon)\varphi(x)dx$ 的 $\varepsilon$ 阶上界,从而在 $H^{-1}$ 范数中赢得一个 $\varepsilon$ 因子。这个技巧避免了直接取 $L^2$ 模带来的 $\varepsilon^{-1}$ 损失,是获得最优 $H^1$ 误差的关键。
(3) 数值方法:TNN 采用张量积结构函数 $\Phi(x)=\sum_{\alpha=1}^r \prod_{i=1}^n \phi_{\alpha,i}(x_i)$,其积分可分解为 $\int_\Omega \Phi dx = \sum_{\alpha} \prod_i \int_0^1 \phi_{\alpha,i}(x_i)dx_i$,因此所有高维积分(胞元问题弱形式、均匀化系数、宏观源项、损失函数)都可用一维 Gauss 求积精确计算,完全避免了 Monte Carlo 误差。由于高阶正确子的计算需要反复使用低阶结果,高精度求积阻止了误差在层级间的累积,这是保证高阶收敛的实现基础。作者还指出,数值实现假设系数条目具有有限或可控的张量积表示,这一假设独立于分析中的一般矩阵值系数类,因此理论结果不依赖此假设。
此外,作者在 Remark 2.1 中给出了一维精确边界匹配的机制:当 $\Omega=(0,1)$ 且 $\varepsilon=1/N$ 时,若取振荡正确子在端点 $y=0$ 为零,并将附加常数吸收进宏观正确子,则所有截断展开精确满足 Dirichlet 边界条件,故无边界层。这个观察依赖于递归胞元方程的相容性,且仅在一维成立,多维边界层仍需单独处理。
4. 与之前工作的比较
与已有工作相比,本文的推进体现在几个方面。
首先,对局部周期椭圆问题的高阶展开:经典周期均匀化理论 [1-7] 中 $A=A(y)$,胞元问题不依赖 $x$,高阶展开只需逐个求解常数系数胞元问题;而局部周期情形 [13-15] 已有定性均匀化和定量算子估计,但多数停留在首阶或低阶,且未给出可计算的高阶递归格式。本文首次系统构造了 $A(x,y)$ 情形下的完整高阶展开,并证明了在无边界层设定下的 $H^1$ 收敛性,为高阶均匀化提供了理论保证。
其次,与数值方法的比较:传统多尺度有限元 (MsFEM) [18] 和异质多尺度方法 (HMM) [19,20] 在低维和低阶问题中表现良好,但处理局部周期高阶问题时需要在每个宏观样本点重新解胞元问题,或直接在 $\Omega\times Y$ 上离散,计算量随维数和阶数增长迅速。本文的 TNN 方法利用张量积结构,将高维积分降阶为多个一维积分,不需要 Monte Carlo 采样,也不需要宏观和微观网格的耦合,因此在中等维度下可扩展性更强。代价是要求系数和数据具有张量积表示,否则需额外的张量化预处理,其误差本文未分析。
第三,与神经网络求解器比较:现有方法如 [21] 用于多尺度椭圆问题,[22] 提出神经均匀化 PINN,[23] 学习均匀化本构律,[24] 将 PINN 用于一阶两尺度渐近均匀化。这些方法大多依赖随机采样(collocation 点或 Monte Carlo 积分)来估计损失,积分误差较大;而本文使用 TNN 的确定性一维求积,可以达到机器精度,特别适合高阶展开中大量积分的递归计算。这是本文在机器学习方法中的关键创新。
第四,边界层处理:本文刻意回避了有界域 Dirichlet 问题的边界层,只在周期域或特殊的一维精确匹配配置下证明收敛。因此,对于一般的多维 Dirichlet 问题,本文结果不能直接使用,需要补充边界层校正项,这正是作者在结论中提到的未来工作。相比 [1,16,17] 中边界层分析,本文未涉及。
第五,误差估计方面:本文证明了 $H^1$ 半范数估计,但未给出 $L^2$ 估计;实验显示点归一化 $L^2$ 误差呈现名义高阶行为,但作者指出 $L^2$ 模不是规范不变的,且其收敛阶未从理论上保证。因此,$L^2$ 理论是未来研究方向。与 [8-12] 的定量均匀化相比,本文的高阶展开需要更强的正则性假设(例如胞元解的 $x$-正则性),这在文中未详细列出,但从证明推断需要适当的高阶可积性。
综上,本文在理论、算法和数值实现上都有明显推进,但适用范围限于无边界层设定,且数值实现依赖张量积表示。
5. 三维麦克斯韦组合场积分方程的高阶精度连续性强制的Nyström离散化
High-Order-Accurate Continuity Enforcing Nyström Discretization of 3D Maxwell Combined Field Integral Equations
作者: Bernd Hofmann, Reza Molavi, Constantine Sideris
类别: math.NA, physics.comp-ph
发布日期: 2026-08-28
arXiv ID: 2608.28876v1
摘要翻译: 在电场积分方程(EFIE)的Nyström配置离散化中,表面散度作用于可能跨贴片边界不连续的表面密度,这降低了精度和收敛性。我们表明,这不仅影响EFIE,而且影响该算子出现的每一种公式,无论是出现在方程本身还是出现在散射场计算中,并提出一种针对光滑曲面的高阶精度连续性强制方案,作为直接和间接EFIE、磁场积分方程(MFIE)以及正则化组合场积分方程(CFIE)的补救措施。该方案包含两个要素:i) 我们展示了如何通过基于Chebyshev的Nyström方案对方程进行离散化,该方案允许闭式求积规则。ii) 由于未知量和测试向量采用贴片局部曲线基表示,连续性通过基变换来强制执行:我们构建仅由曲线几何描述组装而成的稀疏映射矩阵。这样做,我们恢复了EFIE的精度,使其可以与MFIE以等权重组合形成CFIE。针对典型和现实几何体的散射数值研究表明,所有考虑的单种公式和组合公式都受益于连续性强制,具体表现为更好的条件数、迭代求解器迭代次数减少以及散射场精度提高几位有效数字,尽管减少了未知量总数。
Comment: 无
链接: http://arxiv.org/abs/2608.28876v1
1. 研究对象和背景
本文研究三维时谐电磁散射问题中,理想导体(PEC)目标表面的积分方程数值离散。设时谐因子为 $\mathrm{e}^{-\mathrm{i}\omega t}$,背景介质介电常数 $\varepsilon$、磁导率 $\mu$,波数 $k = \omega\sqrt{\mu\varepsilon}$,波阻抗 $\eta = \sqrt{\mu/\varepsilon}$。散射体表面 $\Gamma\subset\mathbb{R}^3$ 为光滑、有界、可简单或多重连通的二维流形,单位法向为 $\hat{\boldsymbol{n}}$。入射场 $(\boldsymbol{e}^{\mathrm{ex}}, \boldsymbol{h}^{\mathrm{ex}})$ 照射 PEC 目标后,需求解表面感应电流密度 $\boldsymbol{j}$(直接法)或非物理表面密度 $\boldsymbol{a}, \boldsymbol{b}, \boldsymbol{c}$(间接法),再通过表示公式计算散射场 $(\boldsymbol{e}^{\mathrm{sc}}, \boldsymbol{h}^{\mathrm{sc}})$。
核心积分算子包括第一类向量算子 $\mathcal{T}$ 和超奇异算子 $\mathcal{K}$:
$$ \mathcal{T}\boldsymbol{j} = \mathrm{i}k\iint_\Gamma g_k(\boldsymbol{r},\boldsymbol{r}')\boldsymbol{j}(\boldsymbol{r}')\,\mathrm{d}\boldsymbol{r}' + \mathrm{i}k^{-1}\nabla\iint_\Gamma g_k(\boldsymbol{r},\boldsymbol{r}')\,\nabla'_\Gamma\cdot\boldsymbol{j}(\boldsymbol{r}')\,\mathrm{d}\boldsymbol{r}', $$$$ \mathcal{K}\boldsymbol{j} = \iint_\Gamma \nabla g_k(\boldsymbol{r},\boldsymbol{r}') \times \boldsymbol{j}(\boldsymbol{r}')\,\mathrm{d}\boldsymbol{r}', $$其中自由空间格林函数 $g_k(\boldsymbol{r},\boldsymbol{r}') = \mathrm{e}^{\mathrm{i}k|\boldsymbol{r}-\boldsymbol{r}'|}/(4\pi|\boldsymbol{r}-\boldsymbol{r}'|)$。直接 EFIE 和 MFIE 分别取 $\mathcal{T}$ 和 $\hat{\boldsymbol{n}}\times\mathcal{K}$ 的切向迹,并存在腔体谐振频率处的零空间导致解不唯一;经典补救是形成组合场积分方程(CFIE),即 EFIE 与 MFIE 的线性组合。本文还考虑文献[2]提出的正则化 CFIE(CFIE-R),通过引入正则化算子
$$ \mathcal{R}_w\boldsymbol{j} = \iint_\Gamma g_w(\boldsymbol{r},\boldsymbol{r}')\boldsymbol{j}(\boldsymbol{r}')\,\mathrm{d}\boldsymbol{r}',\quad w = \mathrm{i}k/2, $$使得 CFIE 变为第二类 Fredholm 积分方程,改善迭代求解性能。
数值离散方面,作者采用基于 Chebyshev 插值的 Nyström-配置格式(见文献[11,12,13])。每个曲面片上的未知密度按协变基展开,例如
$$ \boldsymbol{j}(u,v) = \frac{1}{D}\left[j^u(u,v)\boldsymbol{e}_u(u,v) + j^v(u,v)\boldsymbol{e}_v(u,v)\right], $$其中 $D$ 为参数域到物理域的广义 Jacobi 行列式,$j^u, j^v$ 为未知标量函数。Nyström 方法直接在配置点(求积节点)上离散,矩阵元素只需核函数求值,避免了 MoM 中昂贵的四维积分。然而,由于表面散度 $\nabla'_\Gamma\cdot\boldsymbol{j}$ 作用在分片光滑但可能跨片不连续的密度上,在片边界会产生虚假线电荷,导致 EFIE 及其相关公式精度退化。已有方法如局部校正 Nyström 或 GP-DIM 需在每个片上求解最小二乘问题,而间断 Galerkin 方法增加迭代次数且缺乏预处理策略。本文旨在填补以下空白:为三维 PEC 散射的所有直接/间接 EFIE、MFIE、CFIE 及正则化 CFIE 提出一种统一的高阶精度连续性强制 Nyström 离散格式,在不改变算子离散核心的前提下,通过构造稀疏映射矩阵显式强制密度在片间连续,从而恢复精度并改善条件数和迭代次数。
2. 主要定理或主要结果
本文的主要结果可归纳如下:
连续性强制阵的构造:针对 Clenshaw-Curtis 闭合求积(包含边界点),设两个相邻曲面片 $a,b$ 在公共边界点 $\boldsymbol{p}$ 处的密度分别为
$$ \boldsymbol{j}(\boldsymbol{p}) = \frac{1}{D^a_{ij}}\left[j^a_{u,ij}\boldsymbol{e}^a_{u,ij} + j^a_{v,ij}\boldsymbol{e}^a_{v,ij}\right] = \frac{1}{D^b_{kl}}\left[j^b_{u,kl}\boldsymbol{e}^b_{u,kl} + j^b_{v,kl}\boldsymbol{e}^b_{v,kl}\right], $$则通过内积可建立系数变换关系(公式(42)-(47)),例如
$$ j^b_{u,kl} = \alpha_{u,ij}j^a_{u,ij} + \alpha_{v,ij}j^a_{v,ij},\quad \alpha_{u,ij} = \frac{D^b_{kl}}{D^a_{ij}}\boldsymbol{e}^{b,u}_{kl}\cdot\boldsymbol{e}^a_{u,ij}. $$由此构造宽稀疏矩阵 $\boldsymbol{C}\in\mathbb{R}^{N'\times N}$,使得所有未知量 $\boldsymbol{j}$ 与独立未知量 $\boldsymbol{j}'$ 满足 $\boldsymbol{j} = \boldsymbol{C}^{\mathrm{T}}\boldsymbol{j}'$,其中 $N' = 2P(N_u-2)(N_v-2)+2U < N$,$P$ 为曲面片数,$N_u=N_v$ 为每片每方向求积点数,$U$ 为边界唯一点数。类似地构造测试向量侧的矩阵 $\boldsymbol{C}$(用于逆变测试)和 $\tilde{\boldsymbol{C}}$(用于协变测试),并满足关键恒等式 $\boldsymbol{C}\boldsymbol{C}^{\mathrm{T}} = \boldsymbol{I}$。
离散方程组的具体形式:表 IV 中列出了所有配方经连续性强制后的离散方程,例如直接 CFIE:
$$ \boldsymbol{C}\left(\pm\boldsymbol{I}/2 + \boldsymbol{nK} + \bar{\boldsymbol{T}}\right)\boldsymbol{C}^{\mathrm{T}}\boldsymbol{j}' = \boldsymbol{C}\left(\boldsymbol{nh}^{\mathrm{ex}} + \eta^{-1}\bar{\boldsymbol{e}}^{\mathrm{ex}}\right), $$其中 $\boldsymbol{nK}$ 为离散的 $\hat{\boldsymbol{n}}\times\mathcal{K}$,$\bar{\boldsymbol{T}}$ 为用协变测试的离散 $\mathcal{T}$。关键性质为 $\boldsymbol{C}(\pm\boldsymbol{I}/2)\boldsymbol{C}^{\mathrm{T}} = \pm\boldsymbol{I}/2$,即第二类算子的恒等部分保持不变,连续性强制只作用于紧致部分。
数值结果:对球、超椭球、空天飞机和汽车模型进行平面波散射测试,使用 GMRES 无重启求解。以 $N=3072$ 未知量($16\times16$ 每片)离散半径 $r=1\,\mathrm{m}$ 的球,连续性强制后未知量降至 $N'=2704$。结果表明:
- 对 EFIE,连续性强制使条件数改善约 4 个数量级,电流密度误差改善超过 2 个数字,电场和磁场误差改善约 6 个数字,达到与连续性强制 MFIE 相当的误差水平。
- 对 MFIE,条件数和迭代数未改善,但磁场误差改善约 1 个数字,电场误差改善最多 3 个数字(因为电场计算涉及 $\mathcal{T}$ 算子)。
- 对 CFIE,连续性强制显著改善条件数、迭代数和所有误差,达到 EFIE 和 MFIE 的最低误差水平且无谐振崩溃。
- 对 CFIE-R,正则化本身已改善条件数和迭代数,但误差受限于 MFIE 精度;只有同时采用连续性强制后,才达到与连续性强制 CFIE 相同的误差水平,并获得最低条件数和迭代数。
- 随未知量加密(每片 $10\times10$ 至 $24\times24$),连续性强制版本在相同或更少未知量下始终优于未强制版本,且对于 EFIE 的 $\boldsymbol{T}$ 和 $\bar{\boldsymbol{T}}$ 两种测试方式均有效。
适用范围与限制:方案要求共形四边形网格(片间节点匹配),且每片两个参数方向的求积点数应相同(特殊几何如环面可不同)。所有连续性强制矩阵仅由曲率几何描述构造,含 $O(N)$ 非零元,应用代价可忽略。
3. 研究方法、关键技术和核心工具
本文的核心技术障碍在于 Nyström 离散中表面散度作用于分片不连续密度时产生虚假线电荷,导致 EFIE 及其衍生公式精度退化。作者并非简单地在方程中引入额外线积分(间断 Galerkin 思路),而是通过闭合求积规则和基变换显式强制密度连续。
关键工具一:Chebyshev 插值与闭合求积。采用 Fejér 第一规则(开放)或 Clenshaw-Curtis 规则(闭合)。对未知函数 $j^{u/v}(u,v)$,使用 Chebyshev 多项式插值(表 III):
$$ j^{u/v}(u,v) = \sum_{m=0}^{N-1}\sum_{n=0}^{N-1}\gamma_{mn}T_m(u)T_n(v), $$其中系数由求积点上的值通过离散正交性确定。闭合的 Clenshaw-Curtis 规则包含边界点 $u=\pm1$,使得未知量在片边界上重复出现,这是后续构造连续性矩阵的前提。同时,利用插值多项式表示,导数矩阵 $\boldsymbol{D}^u, \boldsymbol{D}^v$ 可精确作用于密度,得到表面散度对应的电荷系数 $\boldsymbol{\sigma} = \boldsymbol{D}\boldsymbol{j}$。奇异积分通过变量替换和 Chebyshev 正交性得到高精度校正权重。
关键工具二:算子离散的闭合性。作者证明其离散策略具有“闭合”性质:算子将切向量场映射为切向量场,且用逆变测试向量 $D\boldsymbol{e}^{u/v}$ 可恢复其协变分量。因此复合算子(如 $\hat{\boldsymbol{n}}\times\mathcal{R}_w(\hat{\boldsymbol{n}}\times\mathcal{T}\boldsymbol{j})$)可直接用矩阵乘积离散,例如 $\boldsymbol{nR}\,\boldsymbol{nT}\boldsymbol{j}$,而无须中间转换。具体地,$\hat{\boldsymbol{n}}\times\mathcal{T}$ 离散为 $\boldsymbol{nT} = \mathrm{i}k\boldsymbol{nTA} + \mathrm{i}k^{-1}\tilde{\boldsymbol{D}}\boldsymbol{S}\boldsymbol{D}$,其中 $\boldsymbol{nTA}$ 通过恒等式 $D\boldsymbol{e}_u\cdot(\hat{\boldsymbol{n}}\times\boldsymbol{a}) = -\boldsymbol{e}_v\cdot\boldsymbol{a}$ 等得到;$\bar{\boldsymbol{T}}$(协变测试)与 $\boldsymbol{nT}$ 仅差符号置换。$\hat{\boldsymbol{n}}\times\mathcal{K}$ 离散为 $\boldsymbol{nK} = \boldsymbol{N} - \boldsymbol{L}$,其中 $\boldsymbol{L} = \boldsymbol{G}\boldsymbol{D}^{xyz}\boldsymbol{S}^{xyz}$ 涉及度量张量 $\boldsymbol{G}$。这些矩阵的确切定义(公式(14)-(30))保证了离散算子与连续算子的对应关系。
关键工具三:连续性强制矩阵的构造。这是本文真正的创新点。对每个公共点,选择参考片上的两个未知量作为唯一未知量,通过坐标变换系数(由协变基内积和 Jacobi 行列式比给出)将其他片上的未知量表示为参考片未知量的线性组合,组装为 $\boldsymbol{C}^{\mathrm{T}}$。测试向量侧采用类似的变换但取平均以对称化,得到 $\boldsymbol{C}$ 和 $\tilde{\boldsymbol{C}}$。关键性质 $\boldsymbol{C}\boldsymbol{C}^{\mathrm{T}} = \boldsymbol{I}$ 保证了第二类算子离散后仍含真正单位阵,从而保持谱性质。作者还仔细区分了密度连续性与表面散度平方可积性:强制的是密度本身连续,而表面散度仍允许片间跳跃,这恰好符合 div-conforming 条件(只需 $\nabla_\Gamma\cdot\boldsymbol{j}\in L^2$)。
关键工具四:混合配置方案。受 MoM 中 primal-dual 基(RWG 与 BC 函数)启发,作者构造了混合离散:EFIE 部分用协变测试($\bar{\boldsymbol{T}}$),MFIE 部分用逆变测试($\boldsymbol{nK}$),从而在 Nyström 框架中模拟了 BC 函数的效果,但无需构造额外基函数。数值实验表明该混合方案在未强制连续性时略优于纯逆变测试,但强制连续性后优势消失,说明连续性恢复是更关键的因素。
证明与数值验证框架。作者并未给出严格的收敛性定理,而是通过大量数值实验(球、超椭球、空间飞行器、汽车)系统地比较了各配方在连续性强制前后的条件数、GMRES 迭代数、散射场及电流密度误差。误差以相对平均误差 $\aleph$ 衡量(公式(57)),参考解对球采用 Mie 级数。实验覆盖频率变化(0.15-0.25 GHz)和未知量加密(每片 $10\times10$ 至 $24\times24$)两种场景。
4. 与之前工作的比较
本文与既有工作的比较可从多个维度展开:
与 MoM 高阶离散的比较。传统 MoM 通过 div-conforming 基函数(如 RWG)自然满足法向连续,但矩阵元素需计算四维积分,高阶 MoM 进一步增加成本。而 Nyström 方法只需核求值,效率高但缺乏内建连续性。本文通过稀疏矩阵显式强制连续,在保留 Nyström 效率的同时恢复了与 MoM 相当的精度。与文献[24,25]中通过奇异值分解构造连续性基函数的方法相比,本文的矩阵是显式解析构造的,不含 SVD,且数量级为 $O(N)$,计算和存储代价低得多。
与间断 Galerkin Nyström 方法比较。文献[15,22,23]考虑在 Nyström 离散中引入线积分处理不连续,但会显著增加迭代次数且缺乏预处理策略。本文直接强制连续性,避免了额外线积分,并且由于 $\boldsymbol{C}\boldsymbol{C}^{\mathrm{T}}=\boldsymbol{I}$,保持了第二类 CFIE 的优良谱性质,条件数和迭代数均优于 DG 方案。
与局部校正 Nyström 和 GP-DIM 比较。这些方法需在每个片上求解小规模 LSE 或最小二乘问题来校正近场奇异积分,可能限制精度和效率。本文采用 Chebyshev 插值与闭合求积的解析校正(基于离散正交性),无需逐片求解,因此精度高且成本低。
与既有连续性强制方案比较。文献[23]仅考虑直接 EFIE,本文将其推广到所有直接/间接 EFIE、MFIE、CFIE 及 CFIE-R。具体来说,本文构造了三种独立的稀疏映射矩阵(未知量侧 $\boldsymbol{C}$,测试侧 $\boldsymbol{C}$ 和 $\tilde{\boldsymbol{C}}$),分别对应逆变和协变测试,适用于所有配方。数值结果表明,不仅 EFIE 受益,MFIE 和 CFIE 也获得了精度提升,尽管 MFIE 本身不含表面散度,但散射场计算中的 $\mathcal{T}$ 算子仍会造成误差损失。
正则化 CFIE 的左右正则化差异。作者注意到直接 CFIE-R 中 $\mathcal{R}_w$ 作为左正则化子作用于已微分离散后的算子,而间接 CFIE-R 中作为右正则化子作用于未知密度 $\boldsymbol{c}$ 之后才施加表面散度。理论上右正则化可能平滑密度从而改善连续性,但数值实验发现间接 CFIE-R 并未因此获益,反而直接 CFIE-R 在连续性强制后达到最佳精度。该观察表明连续性强制与正则化的相互作用并非直观,作者对此未给出理论解释,但指出了开放问题。
与 primal-dual 混合 MoM 的比较。MoM 中常用 RWG 展开未知量并测试 EFIE,而用 BC 函数测试 MFIE 以保证 CFIE 的谱等价性(见文献[7,8])。本文在 Nyström 框架中提出混合配置,即 EFIE 用协变测试、MFIE 用逆变测试。数值结果表明该混合方案在未强制连续时效果有限,强制连续后无明显优势,从而说明在 Nyström 框架中,连续性强制比混合测试更重要。
局限性与代价。主要限制是要求共形网格且两参数方向求积点数相同,这在实际 CAD 模型中通常可满足,但牺牲了某些非共形网格的灵活性。然而,由于未知量数从 $N$ 降至 $N'$(例如 $3072\to2704$),实际计算代价并未增加,反而因迭代减少而更加高效。此外,作者明确指出连续性强制并不保证表面散度连续,这符合 div-conforming 的最小要求,但若某些应用需要更强的正则性,该方案需进一步扩展。
与理论分析的关系。本文未提供严格的误差估计或稳定性证明,而是以数值实验为主导。因此,关于连续性强制如何影响离散算子谱的机制,仍停留在经验层面。从提供内容中无法确认作者是否对收敛阶进行理论推导,也未与 Calderón 预处理等方法进行直接比较。
了解领域论文(摘要翻译)
1. Fokker–Planck 方程的保持结构的细致平衡主方程离散
Structure-Preserving Detailed-Balance Master-Equation Discretizations for Fokker–Planck Equations
作者: Satish Chandran, Yiwei Wang
类别: math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30121v1
链接: http://arxiv.org/abs/2608.30121v1
摘要翻译
中文标题: Fokker–Planck 方程的保持结构的细致平衡主方程离散
摘要翻译: 我们直接从 Fokker–Planck 方程的能量耗散律出发,发展了一种细致平衡主方程离散的变分–马尔可夫构造。我们不对微分算子进行离散,而是通过两个方向性跳跃率来表示相邻网格点之间的质量传递。在每条边上,离散能量耗散律确定净通量,而细致平衡固定速率比;然后通过从相邻网格点值局部重构边量来唯一确定这两个速率。对数平均重构恢复了经典的 Scharfetter–Gummel/Wang–Peskin–Elston 速率,而其他重构则产生其他可逆格式。所得的半离散系统保持质量、保持非负性、满足细致平衡,并耗散离散自由能。相同的边构造可推广到状态依赖和退化的迁移率、非局部相互作用能以及高维问题。相应的隐式和半隐式全离散格式是线性的、质量守恒的、保持正性的且能量稳定的,仅对核非负半定的非局部相互作用能需要时间步长限制。数值实验证实了所有细致平衡格式的预期空间收敛阶,并验证了它们在非连续势、饱和、非局部相互作用和二维问题中的平衡精度、正性保持和自由能衰减。
2. 稀疏傅里叶神经算子中的选择、表示与执行
Selection, Representation, and Execution in Sparse Fourier Neural Operators
作者: Abdul Qadir Ibrahim, Martin Burger
类别: cs.LG, math.NA
发布日期: 2026-08-30
arXiv ID: 2608.30070v1
链接: http://arxiv.org/abs/2608.30070v1
摘要翻译
中文标题: 稀疏傅里叶神经算子中的选择、表示与执行
摘要翻译: 稀疏表示通常被认为能使模型更小并降低推理成本。对于傅里叶神经算子(FNOs),这些目标并不等价,也不总是对齐:移除学习到的算子的一部分可能使底层变换和稠密计算保持不变,而改变模型评估的网格则可能引入自身的开销。因此,我们区分表示中的稀疏性、存储参数中的稀疏性、理论操作计数中的稀疏性以及实测运行时间中的稀疏性,并对通向稀疏FNOs的几种途径进行了实证研究,分别测试了它们之间每个转换。细化执行网格降低了理论成本但不降低实测延迟,而添加修正项恢复精度,但代价是使模型更慢。即使在普通执行下减少83%的参数,仍然比稠密基线慢。这些结果促使我们对有用稀疏性作出更严格的定义:部署的算子必须保持解的准确性,并将其缩小的支持映射到真正更便宜的执行路径上。
3. DPG对流离散的Fortin算子
Fortin operators for DPG advection discretizations
作者: Pablo Cortés Castillo, Jay Gopalakrishnan
类别: math.NA
发布日期: 2026-08-30
arXiv ID: 2608.30043v1
链接: http://arxiv.org/abs/2608.30043v1
摘要翻译
中文标题: DPG对流离散的Fortin算子
摘要翻译: 我们为对流方程的间断Petrov-Galerkin(DPG)离散构造了Fortin算子,其中对流向量$β$是分片常数且无散度的,空间维数$N$的单纯形网格上,试探空间的任意多项式次数$k \ge 1$。在每个单元上,由面泡和内部泡构造一个最小的检验空间,并随后进行扩充。证明了Fortin算子关于形状正则网格族一致有界,其范数为基于$L_q$(其中$1 < q < \infty$)的自然$β$加权分片检验图范数,这里$q$是试探指数$p$的共轭指数。当$k \ge 2$时,假设了一个非特征面条件,而最低阶情形不需要该条件,并且允许特征面。作为应用,我们证明了实用的全离散残差最小化方法在DPG能量范数下对每个$1 < p < \infty$都是拟最优的,其拟最优常数仅由Fortin算子决定;其可计算残差是全局可靠且有效的后验误差估计子;并且扩充检验空间和改变检验范数可以改善最低阶情形的结果。
4. 几何流的高阶能量稳定 BGN 参数有限元方法
High-order energy-stable BGN parametric finite element methods for geometric flows
作者: Shu Ma, Qiqi Rao
类别: math.NA
发布日期: 2026-08-30
arXiv ID: 2608.29877v1
链接: http://arxiv.org/abs/2608.29877v1
摘要翻译
中文标题: 几何流的高阶能量稳定 BGN 参数有限元方法
摘要翻译: 我们为平面曲线的曲线缩短流和曲线扩散,以及闭 genus-$0$ 曲面的平均曲率流和表面扩散,构造了 Barrett–Garcke–Nürnberg (BGN) 参数有限元方法的高阶 Runge–Kutta 推广。在每个时间片 $I_m=(t_m,t_{m+1}]$ 上,连续方程通过映射 $\X^{m,t}:Γ^m\toΓ^t$ 置于左端点曲面 $Γ^m=Γ^{t_m}$ 上,其目标是 $t$ 时刻的演化曲面。对于曲线,该公式通过从 $Γ^0$ 到 $Γ^m$ 的调和拉回得到。对于曲面,在每个时间片上分别构造保向调和微分同胚,共形性产生权重 $\frac12|\nabla_{Γ^m}\X^{m,t}|^2$。在 Runge–Kutta 内部时间处评估时间片方程,并应用质量集中参数有限元,得到公共域 $Γ^m$ 上的方程组,其中 $Γ^{t_m+c_iτ_m}$ 作为中间目标几何。这种内部时间离散化构造了四种流的高阶 BGN 结构推广。对于具有非负权重的代数稳定表,每个具有非退化中间构型的精确阶段解,在所有使得该解存在的正时间步长上,都满足离散曲线长度或曲面面积的单调递减。Radau IIA 实验表明,四种流均具有能量衰减,曲线测试中具有 BGN 型网格重分布。Hausdorff 自收敛结果表现出与相应设计阶一致的高阶行为。
5. ENPINN: 能量范数引导的梯度增强PINN用于具有尖锐梯度的广义输运问题
ENPINN: Energy-Norm-Guided Gradient-Enhanced PINNs for Generalized Transport Problems with Sharp Gradients
作者: Subhendu Maity, Pratibhamoy Das, Arihant Patawari, Ameya D. Jagtap
类别: math.NA, math-ph
发布日期: 2026-08-30
arXiv ID: 2608.29499v1
链接: http://arxiv.org/abs/2608.29499v1
摘要翻译
中文标题: ENPINN: 能量范数引导的梯度增强PINN用于具有尖锐梯度的广义输运问题
摘要翻译: 物理信息神经网络(PINNs)已成为求解偏微分方程(PDEs)的传统数值方法的一种无网格替代方案。然而,它们在捕捉尖锐梯度方面的能力有限,在解析边界层和内部层时可能导致显著误差。在这里,我们引入一种能量范数增强的PINN(ENPINN),它将梯度信息和变分结构纳入损失函数,以提高层主导解的解析分辨率。我们首先考察两种相关公式:弱损失PINNs(WLPINNs),它将测试函数纳入传统PINN残差;以及梯度增强PINNs(gPINNs),它用PDE残差的空间导数增强损失。通过分析这些公式,我们识别出它们在解析陡峭解梯度方面的局限性,并激发了对ENPINN的系统构建。我们从理论上建立了能量范数误差如何依赖于ENPINN损失,并表明适当修改的残差导数项对于准确捕捉边界层至关重要。我们进一步证明了具有任意小能量误差的神经网络逼近的存在性,并推导了相应的导数界,为所提出的框架提供了理论基础。通过系统比较ENPINN与现有PINN变体在对流-扩散-反应问题(表现出陡峭梯度)上的性能,评估了ENPINN的性能。数值实验包括一个燃烧模型、一个耦合多尺度系统、一个具有内部层的二维Burgers方程,以及一个三维时间相关问题。
6. 广义牛顿流非残差VMS逼近的动态子尺度和无条件半离散稳定性
Dynamic subscales and unconditional semi-discrete stability for non-residual VMS approximations of generalised Newtonian flows
作者: G. R. Barrenechea, E. Castillo, R. Codina, J. Pastene, P. Vega
类别: math.NA
发布日期: 2026-08-29
arXiv ID: 2608.29457v1
链接: http://arxiv.org/abs/2608.29457v1
摘要翻译
中文标题: 广义牛顿流非残差VMS逼近的动态子尺度和无条件半离散稳定性
摘要翻译: 我们分析了一种具有动态子尺度的不可压缩广义牛顿Navier–Stokes流的非残差变分多尺度有限元格式。假设表观粘度关于剪切率有界且Lipschitz连续,涵盖了若干正则化的流变定律。该方法通过正交投影分离未 resolved 的压力-散度、对流和压力梯度贡献。对于线性化半离散问题,我们证明了在各向异性VMS范数下的适定性和无条件稳定性估计。该估计控制了粘性耗散、子尺度能量、离散散度以及一个耦合的加速度-对流-压力平衡。对于非线性格式,通过不动点论证,在适当的正则性和小性假设下得到了存在性和最优阶先验界。该分析还提供了离散压力-对流平衡的弱时间控制。一个关键点是,与压力相关的动态子尺度提供了控制投影压力梯度误差所需的时间导数贡献,而这一机制在相应的拟静态设置中是不可用的。
7. 基于图卷积的可学习多重网格框架
A Learnable Multigrid Framework via Graph Convolutions
作者: Xiuping Wang, Li Luo
类别: math.NA
发布日期: 2026-08-29
arXiv ID: 2608.29082v1
链接: http://arxiv.org/abs/2608.29082v1
摘要翻译
中文标题: 基于图卷积的可学习多重网格框架
摘要翻译: 本文提出了一种新颖框架,将可学习图卷积与几何多重网格方法相结合,用于求解偏微分方程(PDEs)。PDEs的离散化首先表示为图结构,从而能够应用图卷积来增强多重网格性能。通过将图卷积纳入多重网格组件(如光滑算子和网格间转移算子),我们开发了一种可学习多重网格,能够根据底层问题特征自适应优化其性能。在该框架中,图卷积直接嵌入多重网格循环内部,从而将整个多重网格求解器有效转化为一种专门的神经网络架构,而非将经典求解器与代理模型相结合。该可学习多重网格框架在参数数量方面是轻量级的,并且只需极少的训练开销即可获得良好性能。数值实验证明了所提方法在求解若干具有挑战性问题时的有效性,显示出与传统多重网格方法相比更优的收敛速度。此外,还研究了学习参数在不同问题设置(包括变化的源项、系数、几何形状和网格尺寸)上的泛化能力,并采用了适当的权重共享和迁移学习策略。
不相关论文(基本信息)
1. 量子信号处理中的约束极小极大逼近
Constrained minimax approximation for quantum signal processing
作者: Yulong Dong, James B. Larsen, Lin Lin, Rahul Sarkar
类别: quant-ph, math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30937v1
链接: http://arxiv.org/abs/2608.30937v1
摘要: Quantum signal processing (QSP) provides a simple and efficient framework for implementing polynomial transformations using quantum circuits. Its classical design stage leads to a constrained minimax approximation problem: find a polynomial of prescribed parity that approximates a target function uniformly on a fitting set while remaining bounded in magnitude by one on the domain $[0,1]$, which can be viewed as a semi-infinite constraint. Discretization converts the problem into a linear program, but feasibility at a set of finitely many sampled points does not ensure feasibility on the whole domain, especially when an optimal approximant reaches the boundary of the feasible set. We investigate two approaches to address this difficulty. A Remez exchange method combined with active-set constraint enforcement is efficient on many tested instances, but its stability depends on the target and problem geometry. We then introduce nonlinear Fourier retraction, which uses QSP completion and phase synthesis to turn a nearly feasible polynomial into phase factors for a feasible QSP polynomial without increasing the degree. Across representative problems, retraction largely preserves approximation accuracy and remains effective on instances where the Remez heuristic is unstable. The resulting workflow connects classical minimax approximation and semi-infinite optimization with nonlinear Fourier analysis, and is implemented in the qsppack software package.
2. 大型开放量子系统的良态迭代方法
Well-conditioned iterative methods for large open quantum systems
作者: Gaspard Beugnot, Paul Gregory, Rémi Robin, Antoine Tilloy
类别: quant-ph, math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30860v1
链接: http://arxiv.org/abs/2608.30860v1
摘要: Markovian open quantum systems are well modeled by the Lindblad Master Equation (ME) $\frac{\mathrm{d}}{\mathrm{d} t} ρ_t = \mathcal{L} ρ_t$, where $\mathcal{L}$ is a linear (super-)operator and $ρ_t$ is the system state, a positive matrix. When designing or characterizing a quantum system, one is usually interested in the steady state $ρ_\infty$ (such that $\mathcal{L} ρ_\infty = 0$), the first few excited states, and trajectories $t\mapsto ρ_t$. In finite dimension, $ρ_t$ is an $n\times n$ matrix, $\mathcal{L}$ thus typically costs $n^4$ to store explicitly as a dense matrix, and $O(n^6)$ to diagonalize or invert exactly, making standard linear algebraic techniques expensive for large systems. However, $\mathcal{L}$ usually costs only $O(n^3)$ to apply. This makes iterative methods appealing, but they do not work without a good preconditioner. In this article, our main observation is that a part of the Lindblad equation, corresponding to the so-called no-jump evolution $\mathcal{S}$, can be inverted efficiently. Using this inverse map, we introduce an auxiliary completely positive trace-preserving (CPTP) map $Φ$ whose fixed point is directly related to $ρ_\infty$, all the other eigenvalues having smaller magnitude. The map $Φ$ is thus well suited to iterative methods, and $ρ_\infty$ can be found in a few Arnoldi iterations. Using the same inverse map $\mathcal{S}^{-1}$ as preconditioner, we compute the low-lying spectrum efficiently via shift-invert Arnoldi, and, as a proof of concept, build an implicit time integrator that is competitive on stiff systems in the low-precision regime. For the steady-state and low excited states problems, our methods scale like $O(n^3)$ per iteration and offer state-of-the-art performance on CPU and GPU.
3. 基于神经网络的端口哈密顿系统模型降阶
Model reduction of port-Hamiltonian systems via neural networks
作者: Silke Glas, Alexander Heinlein, Harald Monsuur, Hongliang Mu
类别: math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30788v1
链接: http://arxiv.org/abs/2608.30788v1
摘要: In this paper, we consider structure-preserving model reduction of port-Hamiltonian (pH) systems which extend classical Hamiltonian systems with dissipation and an input-output port. These pH systems are often used in multi-physics systems, as the interconnection of one or more \pH systems results again in a pH system. If particularly the system matrices associated with the interconnection and/or dissipation of a pH system are state-dependent, then the evaluation of standard reduced-order models (ROMs) may depend on the dimension of the original full-order model, resulting in high computational costs. To circumvent these high costs, we propose to use structure-preserving neural networks. In particular, we perform two steps: (1) we use the generalized manifold Galerkin projection to project the pH system onto the reduced space; then (2) we train a neural network to learn the map from the reduced-order state to the reduced-order interconnection and dissipation system matrices. To ensure that the resulting ROM is again a pH system, the architecture of the neural network is chosen such that the skew-symmetry and positive semi-definiteness of the reduced-order systems matrices are maintained. In a numerical example, we consider a nonlinear mass-spring-damper system with state-dependent system matrices. The numerical results show that the proposed method achieves a significant computational speed-up compared to the original \ROM with comparable accuracy.
4. Vasudeva Murthy的松弛系统的积分不变量:分析与数值验证
Integral Invariants of Vasudeva Murthy’s Relaxation Systems: Analysis and Numerical Validation
作者: Sudipta Sahu, Rathan Samala
类别: math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30770v1
链接: http://arxiv.org/abs/2608.30770v1
摘要: Vasudeva Murthy’s relaxation approach [A.S. Vasudeva Murthy, J. Comput. Appl. Math., 203(2), pp. 437-443, 2007], originally proposed for the Jin-Xin relaxation model, provides an alternative formulation with invariant properties that is consistent and retains the semilinear structure incomparison to the standard one. In this work, Vasudeva Murthy’s relaxation approach for various relaxation systems are proposed such as the shallow water equations, the Broadwell model, the Euler equations with heat transfer and two-dimensional Jin-Xin model. For proposed relaxation models, the associated integral invariants are rigorously established at the theoretical level. The main advantage of the integral invariant is that it provides a conserved quantity for the relaxation system by incorporating the coupled contributions of the solution variables in vector form. To validate the analytical results, numerical simulations are carried out for each model using three second-order numerical schemes: CS-EBT2, a semi-implicit second-order central finite-volume scheme for hyperbolic systems with relaxation source terms [S. Sahu, E. Macca, and R. Samala, J. Comput. Phys., 563, 115100, 2026]; UCS2, a finite-volume central relaxation-type scheme [S. F. Liotta, V. Romano, and G. Russo, SIAM J. Numer. Anal., 38(4), 1337-1356, 2000]; and IMEX-RK2, a second-order Implicit-Explicit Runge-Kutta scheme [Pareschi and Russo, J. Sci. Comput., 25, 129-155, 2005]. Numerical results, compared to exact or finely resolved reference solutions, confirm that the models preserve integral invariants, remain stable under CFL restrictions, and exhibit robust and accurate behavior across all benchmark systems tested.
5. Poisson约束的JKO格式的高效原始-对偶分裂方法及其在Poisson-Nernst-Planck模型中的应用
Efficient primal–dual splitting methods for a Poisson-constrained JKO scheme for Poisson-Nernst-Planck models
作者: Wei Wu, Jin Zeng, Zhen Zhang, Chaozhen Wei
类别: math.NA, math.OC, physics.comp-ph
发布日期: 2026-08-31
arXiv ID: 2608.30693v1
链接: http://arxiv.org/abs/2608.30693v1
摘要: The Poisson–Nernst–Planck (PNP) equations strongly couple ionic transport and electrostatic interactions through the Poisson equation, posing substantial numerical challenges under small permittivity and complex potential boundary conditions. Underlying these equations is a natural Wasserstein gradient-flow structure, in which the Poisson equation serves as a local realization of the nonlocal electrostatic interaction energy. Exploiting this structure, we formulate each time step as a constrained convex minimization problem where the ionic continuity equations and the Poisson equation are incorporated as linear constraints, allowing the concentrations, fluxes, and electrostatic potential to be updated simultaneously. The variational structure of the scheme intrinsically guarantees the dissipation of the original free energy, mass conservation, and nonnegativity of ionic concentrations under general electrostatic boundary conditions. Moreover, the framework is structurally modular: extending from classical to modified PNP models with steric interactions and concentration-gradient corrections requires only modifying the energy functional, while all structure-preserving properties are automatically retained. To efficiently solve the resulting large-scale constrained problems, we develop preconditioned and transformed primal–dual algorithms equipped with tailored fast dual solvers, namely DCT-based direct and Schur-complement iterative methods, that exploit the coupled block structure of the PDE constraints. Numerical experiments on classical and modified PNP systems demonstrate the accuracy and structure-preserving properties of the scheme, and show that the proposed algorithms converge reliably in strongly coupled small-permittivity regimes without significant growth in computational cost.
6. Signorini问题的障碍正则化对称Nitsche方法
A Barrier-Regularized Symmetric Nitsche Method for the Signorini Problem
作者: Peter Hansbo, Mats G. Larson
类别: math.NA
发布日期: 2026-08-31
arXiv ID: 2608.30470v1
链接: http://arxiv.org/abs/2608.30470v1
摘要: We introduce and analyze a barrier-regularized symmetric Nitsche method for the scalar Signorini problem. Applying a logarithmic barrier to the nonnegative slack variable in an augmented Lagrangian and then eliminating that variable yields a smooth positive-part operator and the perturbed complementarity relation on a primal-dual central path, with barrier parameter $μ=γs$. For every $s>0$, the discrete problem is continuously differentiable, uniquely solvable, and has a symmetric positive definite Newton matrix when the Nitsche parameter is sufficiently large. Under a mild barrier-feasibility condition, the continuous logarithmic energy has a unique minimizer $u_μ$ for every $μ>0$, its gap is positive almost everywhere, and $\|u-u_μ\|_{H^1(Ω)}\lesssimμ^{1/2}$. Under the Sobolev regularity used for finite element approximation, this minimizer satisfies the (L^2) central-path boundary law. If the obstacle is locally the trace of an $H^2$-function, we also prove a uniform local $H^2$ bound in the interior of each planar contact face, on neighborhoods that may contain a free-boundary point of the limiting Signorini solution. The method is exactly consistent and quasi-optimal relative to $u_μ$, with constants independent of $s$. If $u_μ$ is uniformly bounded in $H^r(Ω)$, $3/2 From multi-layered problems to multiple two-layered problems: a novel frequency-time hybrid multiple-scattering integral equation solver 作者: Shuai Pan, Tao Yin, Lu Zhang 类别: math.NA 发布日期: 2026-08-31 arXiv ID: 2608.30287v1 链接: http://arxiv.org/abs/2608.30287v1 摘要: This paper proposes a novel frequency-time hybrid multiple scattering (FTH-MS) integral equation solver for time-dependent wave equation problems in general multi-layered media, with remarkable scalability with respect to the number of layers $N$. In light of the finite speed of wave propagation, the new methodology provides an innovative multiple-scattering idea of re-modeling the original $N$-layered problem into a sequence of $N-1$ two-layered sub-problems, for which the main advantages lie in that (i) each sub-problem enjoys much simpler wave scattering properties compared with the complicated problem in a multi-layered medium, (ii) it enables to develop high-accuracy solver utilizing Fourier transform and frequency-domain boundary integral equation (BIE) method; and (iii) numerical evaluation of the sub-problems in each multiple scattering step can be parallelized. Both multiplicative- and additive-type strategies are developed and equivalence results, which indicate that the $M$-th order multiple scattering sums can provide equivalent representations of the solutions up to a certain time $T(M)$, are rigorously derived. Owing to the existed result of exponential convergence of the perfectly-matched-layer (PML) truncation for two-layered problem, all the sub-problems is numerically resolved by means of the FTH method based on the Fourier transform and the PML-BIE method whose numerical evaluation is addressed utilizing the Chebyshev-based rectangular-polar solver with high accuracy. Numerical examples are presented to validate the efficiency and accuracy of the proposed method. The specular ellipse method for scalar ordinary differential equations: exactness and accuracy up to fourth order 作者: Kiyuob Jung 类别: math.NA 发布日期: 2026-08-31 arXiv ID: 2608.30280v1 链接: http://arxiv.org/abs/2608.30280v1 摘要: This paper introduces a family of one-step implicit methods for solving scalar ordinary differential equations. At each step, the update uses a scaled angular mean of two vector-field evaluations, and the positive scale may vary from step to step. The leading terms of the local truncation error can be expressed in terms of the derivative of the signed curvature of the scaled solution graph. We prove that the proposed method reproduces the exact solution at the mesh points when the solution graph has constant signed curvature under a fixed positive scaling and each implicit update is unique. When this special geometric condition is not satisfied, we establish second-order consistency and convergence for positive scale sequences satisfying suitable uniform conditions. Furthermore, third- and fourth-order consistency and convergence can be achieved by choosing the scale to cancel the relevant curvature terms in the local truncation error. Using only the given problem data, we classify when these improvements are possible and determine the corresponding scale choices. An example shows that the proposed fourth-order method can yield smaller errors than the classical fourth-order Runge–Kutta method at the same step size. Parameter-Robust Subspace Correction with Multiple Semidefinite Penalties 作者: Subhransu S. Bhattacharjee 类别: math.NA 发布日期: 2026-08-31 arXiv ID: 2608.30265v1 链接: http://arxiv.org/abs/2608.30265v1 摘要: Independently weighted semidefinite penalties arise in augmented-Lagrangian and constrained formulations. This paper characterizes when an exact additive subspace-correction preconditioner remains uniformly effective over all nonnegative penalty weights on a fixed finite-dimensional space. Robustness holds precisely when the correction spaces decompose every joint kernel generated by a nonempty subset of penalties. If one condition fails, a computable constant determines the exact first-order decay of the smallest preconditioned eigenvalue along the associated parameter ray, and the condition number grows linearly; none of the subset conditions can be discarded in general. Filtered decompositions provide computable sufficient bounds on parameter-ordering cones, while distributive kernel lattices permit a single common splitting. Exact-additive computations confirm the characterization and predicted rates. Separate Scott-Vogelius experiments produce stable multilevel iteration counts over the tested weights and mesh levels. The analysis does not establish mesh-uniformity. Mesh-Uniform Power Stability of Two-Relaxation-Time Vector Lattice Boltzmann Schemes with Reversible Boundaries 作者: Jin Zhao 类别: math.NA 发布日期: 2026-08-31 arXiv ID: 2608.30253v1 链接: http://arxiv.org/abs/2608.30253v1 摘要: We consider the collision-transport operator for vector-valued two-relaxation-time lattice Boltzmann schemes linearized about a uniform rest state. Assume that the equilibrium blocks are positive definite and that the link-even and link-odd relaxation parameters satisfy s_+ + s_- = 2 and 0 < s_- < 2. If the homogeneous transport is unitary in the equilibrium metric and reversible under velocity exchange, then the powers of the amplification operator are bounded uniformly with respect to the number and arrangement of lattice nodes. The admissible transports include periodic transport, vector halfway bounce-back, coordinate-aligned specular reflection, tangential orthogonal involutions, and compatible multi-channel scattering. The proof reduces the population equation to a two-step macroscopic recurrence generated by a contraction. An inclusion of the numerical range in an ellipse, combined with the Crouzeix-Palencia theorem, yields a dimension-independent estimate for the companion operator and hence the population bound. For a three-coefficient off-midpoint boundary interpolation, we give an exact rational D2N5 example whose finite-domain amplification matrix has a real eigenvalue larger than one, although the interpolation coefficients and bulk parameters are admissible. Thus coefficient convexity alone does not ensure stability for this boundary family. Checkerboard Shells: A Position-Only Thin-Shell Discretization with Scale-Compatible Completion 作者: Junjie Song, Xuanyu Wu, Zhifeng Zhang, Bingtao Hu, Zhaoxi Hong, Xiuju Song, Yixiong Feng, Jianrong Tan 类别: math.NA 发布日期: 2026-08-31 arXiv ID: 2608.30195v1 链接: http://arxiv.org/abs/2608.30195v1 摘要: Checkerboard edge-midpoint geometry provides an exact planar Varignon parallelogram for every spatial quadrilateral, allowing a local tangent frame and normal to be recovered directly from nodal positions even when the raw quadrilateral is warped. Building on this property, we develop a position-only thin-shell discretization with no independent director, rotation, or strain variables. The connected edge-midpoint surface is taken as the physical midsurface: the first fundamental form is evaluated on planar B faces, while a W-centered second fundamental form is constructed from variations of neighboring B-face normals, so membrane and bending share the same geometric carrier. A variational-kernel analysis shows that smooth second-order consistency does not eliminate lattice-scale blind modes. After quotienting out the raw checkerboard gauge, the B metric has one physical membrane blind direction and the symmetric W curvature has two curvature blind directions. We introduce a quotient-minimal membrane compatibility coordinate $X_M$ and an objective reference-relative curvature coordinate $X_W^{rel}$, placed consistently in the $O(t)$ membrane and $O(t^3)$ bending sectors. The formulation admits an explicit midpoint quotient, complete flat blind-mode classification, rigid-motion objectivity, reference-state consistency, and an $O(h^2)$ near-isometry approximation result for aligned generalized cylinders. Numerical tests show second-order curvature convergence, targeted removal of the membrane defect, and a sub-percent, refinement-decaying influence of $X_W^{rel}$. Linear and nonlinear shell benchmarks further demonstrate flat bending, curved-shell membrane-bending coupling, thickness sensitivity, large rotation, nonlinear pinching, and localized ovalization within a single position-only framework. Spectra of Non-Self-Adjoint Almost Mathieu Matrices and the Scottish Flag Operator 作者: Simon Becker, Izak Oltman 类别: math.SP, math.NA 发布日期: 2026-08-31 arXiv ID: 2608.30191v1 链接: http://arxiv.org/abs/2608.30191v1 摘要: For $N\geq 3$ and a potential phase $\vartheta\in\mathbb{R}$, we study the non-self-adjoint almost Mathieu matrix obtained by multiplying the discrete Laplacian by a complex phase with angle $\varphi\in\mathbb{R}$, $A_N(\varphi,\vartheta)=e^{i\varphi}(S+S^{-1})/2+\operatorname{diag}(\cos(2πj/N+\vartheta))_{j\in\mathbb{Z}/N\mathbb{Z}}$, where $S e_j=e_{j+1}$ is the periodic shift on $\mathbb{C}^N$. We derive a Chambers formula and isolate the part $Q_{N,\varphi}$ of the characteristic polynomial that depends only on $N$ and $\varphi$, but not on $\vartheta$ or on a change of boundary conditions for the shift operator. We then show, for every $N$, that the zeros of $Q_{N,\varphi}$ lie on the two perpendicular lines $e^{i\varphi/2}\mathbb{R}\cup e^{i(\varphi/2+π/2)}\mathbb{R}$. For even $N$, the same property holds for the matrices $A_N(\varphi,\vartheta)$ with $\vartheta\in 2π\mathbb{Z}/N$, and we compute their limiting eigenvalue measure explicitly. For $\varphi\in[-π,π]$, the eigenvalue distribution approximates elliptic-integral densities with masses $1-|\varphi|/π$ and $|\varphi|/π$, and maximal radii $2|\cos(\varphi/2)|$ and $2|\sin(\varphi/2)|$, respectively. At $\varphi=π/2$, the central polynomial $Q_{N,\varphi}$ factors into positive quartic factors. This proves that the Scottish flag matrix, after Trefethen and Chapman, has its spectrum on the two diagonal lines of the saltire. A variant of the block preconditioner for indefinite complex symmetric linear systems 作者: Mehdi Makhdomi, Davod Khojasteh Salkuyeh 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29985v1 链接: http://arxiv.org/abs/2608.29985v1 摘要: In this paper, we propose an efficient preconditioner for solving indefinite complex symmetric linear systems within a block preconditioning framework. We analyze the convergence of the corresponding iterative method and investigate several spectral properties of the preconditioned matrix, including eigenvalue distributions and eigenvector structures. The new preconditioner is used to accelerate the convergence of the flexible version of GMRES. Numerical experiments are presented to illustrate the effectiveness of the proposed preconditioner, and comparisons with existing block preconditioners demonstrate its superior performance. Advancements in Spectral Collocation Methods for High-Order Eigenvalue Problems 作者: Sameh Gana 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29932v1 链接: http://arxiv.org/abs/2608.29932v1 摘要: This paper focuses on computing spectral solutions for high-order eigenvalue problems using an efficient discretization method based on Chebfun spectral discretization algorithms and domain truncation. We solve several numerical eigenvalue problems, demonstrating both the accuracy and computational efficiency of the proposed approach. Wigner-Eckart Factorization of the Polyatomic Boltzmann Collision Operator 作者: Torsten Keßler, Bas W. T. Gieling, René R. Hiemstra, Michael R. A. Abdelmalik 类别: math.NA, physics.comp-ph 发布日期: 2026-08-30 arXiv ID: 2608.29825v1 链接: http://arxiv.org/abs/2608.29825v1 摘要: We extend the Wigner-Eckart factorization of the spectral Boltzmann collision operator to polyatomic gases with continuous internal energy. Because internal energies are invariant under spatial rotations, the SO(3) reduction survives the Borgnakke-Larsen energy exchange, and the twelve-dimensional collision integral collapses onto a nine-dimensional kinematic core. The core splits into a sparse geometric tensor, evaluated exactly, and a dense physical tensor, integrated by singularity-resolving Gauss rules with an auxiliary Laplace representation of the fractional energy couplings. The quadrature attains near machine precision at the fractional exponents of real gases. The collision invariants are embedded exactly, preserving the translational-internal energy exchange. The factorization compresses the operator by three to nearly four orders of magnitude and accelerates its evaluation 40-fold over dense formulations. The method is validated against the exact monatomic limit, Landau-Teller relaxation, and an analytic frozen-channel Prandtl number, and it matches a published calibration of the same kernel for N2, CO, and H2. Optimal error estimates of sequential finite element method for nonlinear thermo-poroelasticity problems 作者: Fang Wang, Fan Chen, Changqing Lv, Chenguang Zhou 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29799v1 链接: http://arxiv.org/abs/2608.29799v1 摘要: This study introduces and analyzes a three-step sequential decoupling algorithm designed to address nonlinear, fully coupled quasi-static thermo-poroelasticity systems incorporating convective transport. The finite element method is employed for spatial discretization and the backward Euler method for temporal discretization. The proposed sequential method has a higher computing efficiency than the fully implicit nonlinear numerical scheme, since it does not require any internal iterations. The well-posedness of the numerical solution is discussed by introducing a cut-off operator and the stability analysis of the algorithm is performed. Rigorous analysis yields optimal convergence order estimates for both spatial and temporal discretizations. In order to confirm the theoretical results and the effectiveness of the suggested approach, numerical experiments are finally carried out. Efficient Hybrid WENO Schemes for Special Relativistic Hydrodynamics with Adaptive Characteristic Reconstruction 作者: Rakesh Kumar, Biswarup Biswas, Asha Kumari Meena, Harish Kumar 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29776v1 链接: http://arxiv.org/abs/2608.29776v1 摘要: Special relativistic hydrodynamics (SRHD) equations arise in the modeling of high-speed fluid flows encountered in astrophysical phenomena such as jets, supernova explosions, and gamma-ray bursts. Owing to their highly nonlinear hyperbolic nature, solutions often develop strong discontinuities, making the design of stable and accurate numerical schemes challenging. Although Weighted Essentially Non-Oscillatory (WENO) schemes are widely used for such problems, component-wise WENO reconstruction may produce spurious oscillations near discontinuities. On the other hand, characteristic-wise WENO reconstruction provides accurate non-oscillatory solutions for systems of conservation laws, but it involves the computation of eigenvectors in each cell, which leads to high computational cost. In this work, we intend to develop hybrid schemes which maintain the non-oscillatory feature of characteristic-wise WENO while being less costly. We propose three hybrid schemes, namely the H1-WENO, H2-WENO, and H3-WENO schemes, based on a new troubled-cell indicator constructed from the smoothness indicators of the WENO scheme. The proposed troubled-cell indicator effectively distinguishes smooth and discontinuous regions, allowing the hybrid schemes to employ inexpensive reconstructions in smooth regions and the characteristic-wise WENO reconstruction only near discontinuities. Numerical experiments demonstrate that the proposed schemes retain the accuracy and robustness of characteristic-wise WENO methods while significantly reducing the computational cost. In particular, the H1-WENO scheme achieves an approximately 30–40% improvement in computational efficiency compared to the standard WENO scheme in 2D test cases. Constraint Preserving AFD-WENO Schemes for Relativistic Hydrodynamics with General Equations of State 作者: Pramodit Mishra, Shubham Upadhyay, Rakesh Kumar, Biswarup Biswas 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29654v1 链接: http://arxiv.org/abs/2608.29654v1 摘要: We develop a high-order physical-constraint-preserving (PCP) alternative finite difference weighted essentially non-oscillatory (AFD-WENO) scheme for the special relativistic hydrodynamics equations with general equations of state. The proposed scheme comprises two key limiters: a state limiter, which acts after the WENO state interpolation step, and a flux limiter, which acts on the final high-order fluxes. The state limiter ensures that the interpolated states are physically admissible, while the flux limiter ensures that the numerical fluxes are physically admissible. The resulting scheme is rigorously proved to satisfy the physical constraints. Incorporating multiple WENO interpolation techniques, including an improved adaptive-order formulation (WENO-AOI), the method is validated through extensive one- and two-dimensional numerical benchmarks with various equations of state. The numerical results demonstrate high-order accuracy, sharp resolution of discontinuities, and robust stability in extreme relativistic regimes. Tensor Orthogonal Subspace Split: Theory and Applications 作者: Jifei Miao, Juan Han, Michael K. Ng, Kit Ian Kou 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29638v1 链接: http://arxiv.org/abs/2608.29638v1 摘要: Tensor representations have emerged as a fundamental paradigm for modeling multidimensional data by preserving intrinsic correlations across multiple modes. This paper proposes a novel theoretical framework, termed Tensor Orthogonal Subspace Split (TOSS), which explicitly splits a tensor, along a prescribed mode, into two orthogonal components: a dominant component lying in a prescribed subspace and a residual component lying in the corresponding orthogonal complement. We first present the general formulation of TOSS and systematically investigate its fundamental properties. As an important and practically meaningful special case, we further introduce the rank-one TOSS, which imposes a separable rank-one structure along the splitting mode and admits a clear geometric interpretation. This formulation naturally captures dominant consistent patterns while effectively isolating orthogonal residual component. The proposed framework establishes a unified theoretical foundation for tensor-domain orthogonal split and opens new avenues for structured tensor modeling across diverse applications. Building upon the developed TOSS theory, hyperspectral image restoration and color video background modeling are considered as two representative tasks, for which corresponding optimization models are formulated. Efficient algorithms are developed to solve the resulting problems. Extensive experimental results validate the effectiveness and superiority of the proposed approaches. Analysis and Approximation of Stochastic Multiscale Subdiffusion Driven by Fractional Gaussian Noise 作者: Jincheng Dong, Ning Du, Xu Guo, Mengmeng Liu, Xiangcheng Zheng 类别: math.NA 发布日期: 2026-08-30 arXiv ID: 2608.29533v1 链接: http://arxiv.org/abs/2608.29533v1 摘要: This paper investigates a stochastic multiscale subdiffusion model driven by fractional Gaussian noise, where the multiscale Abel kernel with variable exponent $α(t)\in(0,1)$ is used to capture multiscale and crossover behavior in anomalous diffusion. The main difficulties of this model lie in the complexity of the multiscale Abel kernel (e.g. non-monotonicity and non-coercivity) and the low regularity caused by the noise. Concerning these issues, we prove the well-posedness and regularity of the mild solutions by means of solution operator approach and a perturbation technique for multiscale Abel kernel. Then both the semidiscrete-in-time and fully-discrete numerical schemes are proposed and analyzed under the low-regularity numerical analysis framework, with proved temporal and spatial convergence rates. Numerical experiments are presented to substantiate the theoretical results. Nechvile-Transformed Spacecraft Dynamics and Propellant Computation in the 3-Body Problem 作者: Michael J. Dixon, Isaac M. Ross 类别: math.OC, eess.SY, math.NA 发布日期: 2026-08-29 arXiv ID: 2608.29271v1 链接: http://arxiv.org/abs/2608.29271v1 摘要: The uncontrolled equations of motion in the Nechvile frame for the restricted three-body problem have been well-known since at least the 1960s. It would seem that adding an external force to these equations is quite trivial: simply add an external force per mass term to the acceleration equations. Here we show that the last statement is not true. In fact, we show that the additive generic external force must be multiplied by the inverse of $(1 + e \cosθ)^3$ where $e$ is the relative eccentricity of the primaries and $θ$ is the true anomaly of the rotating frame located at the barycenter. Furthermore, when this result is combined with the mass flow rate equation, it generates several surprising results due to the mismatch between the resulting quadratic term and the cubic term in the equations of motion. This leads to a corresponding modification of the rocket equation itself. A Birkhoff-theoretic solution to an illustrative cislunar space mission problem shows propellent savings of 80% with the use of the correct cost functional. The popular quadratic cost utilizes more than $2X$ the minimum propellant consumption. Entropy-Stable and Physical-Constraint-Preserving DGSEM for Symmetry-Reduced General-Relativistic Hydrodynamics on Stationary Spacetimes 作者: Guosheng Fu, Jian-Guo Liu 类别: math.NA, gr-qc, physics.comp-ph 发布日期: 2026-08-29 arXiv ID: 2608.29229v1 链接: http://arxiv.org/abs/2608.29229v1 摘要: We develop an entropy-stable and physical-constraint-preserving discontinuous Galerkin spectral element method for symmetry-reduced general-relativistic hydrodynamics on prescribed stationary spacetimes. Using a local orthonormal transformation, the fluid variables are expressed in a form for which the relativistic hydrodynamic algebra and the admissible set are independent of the spatial metric, while the spacetime geometry enters through stationary coefficients. This separation allows entropy-conservative special-relativistic fluxes to be combined with a compatible discretization of the geometric source terms. On affine tensor-product meshes, the resulting DGSEM is conservative and satisfies a semidiscrete entropy inequality, while the transformed variables provide a convex framework for physical-constraint preservation.
For practical stabilization, we use a geometry-only causal speed that is sufficient for both classical local Lax–Friedrichs entropy dissipation and the physical-constraint-preserving Lax–Friedrichs splitting. The fully discrete method combines this stabilization with SSP Runge–Kutta time stepping, oscillation elimination, and conservative local-orthonormal-state scaling. Numerical experiments cover smooth and strongly shocked special-relativistic flows, an axisymmetric jet, stationary Michel accretion, Schwarzschild Bondi–Hoyle flow, and four Kerr accretion cases. The results demonstrate the designed high-order accuracy in smooth regimes and robust performance for demanding relativistic flows on curved stationary backgrounds. Random attractors and almost-sure stability under discretization of a stochastic autoparametric system 作者: Chuchu Chen, Jialin Hong, Yibo Wang 类别: math.DS, math.NA 发布日期: 2026-08-29 arXiv ID: 2608.29149v1 链接: http://arxiv.org/abs/2608.29149v1 摘要: For a stochastic autoparametric block-and-pendulum system, the long-time dynamics exhibit two fundamental features: the almost-sure stability of the single mode solution, characterized by its Lyapunov exponent, and the global asymptotic dynamics when this single mode solution loses stability. This naturally raises the question of whether these dynamical features are preserved under discretization, since such preservation is essential for the resulting discrete system to faithfully capture the qualitative behavior of the continuous system. To address this question, we first establish the existence of a random attractor for the continuous system subject to multiplicative stochastic excitation, providing a rigorous characterization of the global asymptotic dynamics. We then propose a numerical discretization that induces a discrete random dynamical system and prove the convergence of its random attractor to the continuous one as the step size tends to zero. In addition, we show that the numerical Lyapunov exponent of the single mode solution has the same sign as its continuous counterpart for sufficiently small step sizes, thus preserving the corresponding almost-sure stability or instability classification. These results demonstrate that the proposed discretization captures both the global asymptotic dynamics and the stability characteristics of the underlying stochastic autoparametric system. D-RADI: A Low-rank ADI Algorithm for Solving Large-scale Discrete-time Algebraic Riccati Equations 作者: Umair Zulfiqar 类别: math.NA, eess.SY 发布日期: 2026-08-29 arXiv ID: 2608.29131v1 链接: http://arxiv.org/abs/2608.29131v1 摘要: The low-rank alternating direction implicit (ADI) method is an efficient numerical technique for solving several types of large-scale matrix equations that admit low-rank solutions. The discrete-time algebraic Riccati equation (DARE) is an important matrix equation with applications in state estimation, controller design, and filter design. In the literature, the low-rank Cholesky factor ADI method for Stein equations has been used within Newton iterations to solve large-scale DAREs. However, no dedicated low-rank ADI solver is available for such DAREs. To address this gap, this paper presents a low-rank ADI solver for large-scale DAREs. We also propose an efficient approach to generate ADI shifts automatically, which makes the proposed solver fully autonomous for solving DAREs. The effectiveness of the proposed solver is compared with MATLAB’s \texttt{idare} on a moderate-order problem. Efficiency and accuracy are further demonstrated on large-scale DAREs of order $10^6$. Numerical results confirm that the solver is efficient, accurate, and fully autonomous. Convergence analysis of Parametric Probabilistic Manifold Decomposition 作者: Jiaming Guo, Dunhui Xiao 类别: math.NA 发布日期: 2026-08-29 arXiv ID: 2608.29105v1 链接: http://arxiv.org/abs/2608.29105v1 摘要: This paper presents a convergence analysis for a newly developed nonlinear model reduction method: parametric probabilistic manifold decomposition (PPMD)~\cite{guo2026parametric}. In addition, existing analyzes of nonlinear reduced order models typically treat subspace reduction, manifold representation, regression, and nonlinear reconstruction as separate components and often remain at the level of discrete state vectors. To the best of our knowledge, no theory tracks the complete error propagation in a data-dependent model whose basis, residual geometry, spectral coordinates, parameter maps, and lifting operator are all learned from the same numerical solution data.
We develop a coupled perturbation analysis for the entire PPMD procedure. A trajectory geometry induced by the spatial discretization and temporal quadrature connects discrete trajectory vectors isometrically with the corresponding PDE norm. Population spectral objects are introduced to align the empirical residual coordinates and derive a uniform coordinate error estimate, whose propagation through the Hilbert-valued kernel lifting estimator is then quantified. Combining these results with the full order discretization error, weighted low-rank approximation, parameter regression, and residual representation defect yields deterministic and high-probability trajectory error bounds and consistency in probability in the continuous PDE trajectory space. The theory identifies how the principal errors interact and which components limit the accuracy of the nonlinear reduced order model. Gappy probabilistic manifold decomposition for nonlinear field reconstruction 作者: Qihan Feng, Jiaming Guo, Jiarun Meng, Dunhui Xiao 类别: math.NA 发布日期: 2026-08-29 arXiv ID: 2608.29014v1 链接: http://arxiv.org/abs/2608.29014v1 摘要: This paper proposes gappy probabilistic manifold decomposition (Gappy PMD), a nonlinear method for reconstructing high-dimensional fields from extremely sparse measurements. Gappy PMD reconstructs the field on the nonlinear manifold learned by probabilistic manifold decomposition (PMD). We further propose a differentiable point selection method for reduced-order model (ROM)-based field reconstruction (DPS). Using differentiable meshless interpolation within the ROM-based reconstruction framework, DPS makes the full-field reconstruction error differentiable with respect to the sampling locations and directly optimizes these locations. In addition, a theoretical error analysis for Gappy PMD is also given. It splits the squared reconstruction error into two orthogonal parts: one normal to the reconstruction manifold and the other induced by sparse sampling and observation noise. Under a stability condition on the sampling operator, this error vanishes with the PMD approximation error and the noise. The Gappy PMD is evaluated on three numerical test cases: flow past a cylinder, lid-driven cavity flow, and backward-facing step flow. For the same reduced dimension and sampling points, Gappy PMD attains mean relative $L^2$ errors one to two orders of magnitude below Gappy POD. Optimizing the sampling points with DPS further improves reconstruction accuracy and robustness. A Sharp Unitarily Invariant Norm Bound for the Off-Diagonal Block Perturbation of a Hermitian Matrix 作者: Lei-Hong Zhang, Ren-Cang Li 类别: math.NA 发布日期: 2026-08-29 arXiv ID: 2608.29009v1 链接: http://arxiv.org/abs/2608.29009v1 摘要: Let be two partitioned Hermitian matrices, where $\widetilde A$ is obtained from $A$ by simply dropping the off-diagonal blocks, and let $η$ be the gap between the spectra ${\rm eig}(H_1)$ of $H_1$ and ${\rm eig}(H_2)$ of $H_2$. Define, for $δ\ge 0$ and $ε\ge 0$, and let $V=A-\widetilde A$ and $ε_2=\|E\|_2=\|V\|_2$, the matrix spectral norm. Li and Li [{\em Linear Algebra Appl.}, 395:183–190, 2005] established a sharp spectral-norm bound on the changes in the eigenvalues of $A$: where $\pmbλ(A)$ is the vector whose components are the eigenvalues of $A$ in descending order and similarly for $\pmbλ(\widetilde A)$. The goal of this paper is to resolve the question: how far an extension of this result in the form remains valid for some or all unitarily invariant norms $\|\cdot\|_{\rm UI}$? Two results are obtained: (a) the extension holds for any $Q$-norm, a subclass of unitarily invariant norms that encompasses the Schatten $p$-norm for $2\le p\le\infty$ (particularly, the Frobenius norm and the spectral norm included), and (b) the extension holds for any unitarily invariant norm if ${\rm rank}(E)\le 1$. It is demonstrated that the equality is attained on the $2\times 2$ matrix $A$. Minimizing the Arithmetic and Communication Complexity of Jacobi’s Method for Eigenvalues and Singular Values: Part Two – Parallel Algorithms 作者: James Demmel, Hengrui Luo, Ryan Schneider, Yifu Wang 类别: math.NA 发布日期: 2026-08-28 arXiv ID: 2608.28952v1 链接: http://arxiv.org/abs/2608.28952v1 摘要: This paper presents several parallel versions of Jacobi’s method for the symmetric eigenvalue problem and the SVD. A continuation of [Demmel, Luo, Schneider, & Wang 2025], we develop parallel Jacobi algorithms whose arithmetic cost is optimal and whose bandwidth or latency can match the corresponding lower bounds of parallel matrix multiplication. Our focus is a standard distributed-memory setting with variable processor layouts, including both 2D and 2.5D processor grids. In the 2D case, we demonstrate that a standard implementation of parallel Jacobi achieves a perfect speedup in arithmetic cost – i.e., complexity $O(n^3/P)$ when done with $P$ processors – while hitting the 2D matrix-multiplication lower bound for bandwidth and (nearly) the lower bound for latency. By employing a 2.5D processor grid and leveraging 2.5D matrix multiplication, equivalently by increasing the memory per processor, we demonstrate that parallel Jacobi can achieve even lower bandwidth/latency, though we also prove that these costs cannot simultaneously match the best-known bounds for parallel matrix multiplication in any Jacobi algorithm. Finally, we extend our results to one-sided Jacobi SVD. A GPU-Accelerated Blocked Adaptive Randomized Range Finder Based on an Implicit Householder QR Decomposition 作者: Carolin Penke, Andreas Herten 类别: math.NA, cs.MS 发布日期: 2026-08-28 arXiv ID: 2608.28941v1 链接: http://arxiv.org/abs/2608.28941v1 摘要: Low-rank methods can reduce the memory and computational requirements of deep neural network training in approaches such as GaLore. Randomized range finders offer an attractive alternative to singular value decompositions, particularly when the required rank is determined adaptively from a prescribed approximation tolerance. We introduce a blocked adaptive randomized range finder based on an implicit Householder QR decomposition and an optimized hybrid CPU–GPU implementation. The proposed method avoids explicit reorthogonalization. Numerical experiments show that it preserves orthogonality and approximation accuracy in regimes where block Gram–Schmidt without reorthogonalization becomes unstable. The blocked formulation exposes matrix–matrix operations and enables overlap of CPU panel factorization with GPU updates. On an NVIDIA GH200, the overlapped implementation reduces the runtime for the largest tested matrix from 9.91 seconds on the CPU to 0.407 seconds. The method provides a stable and efficient building block for low-rank approximation on heterogeneous systems with applications in computational science and engineering. A General Superconvergence Result for Cubature on Triangulated Polygonal Domains 作者: Jonah A. Reeger 类别: math.NA 发布日期: 2026-08-28 arXiv ID: 2608.28883v1 链接: http://arxiv.org/abs/2608.28883v1 摘要: Cubature rules, which approximate definite integrals as a linear combination of a set of function values, are ubiquitous and necessary for computational methods in the physical sciences. A superconvergence result for cubature rules on polygonal domains is developed, whereby a rule that is exact for all bivariate polynomials of a fixed even degree realize an extra order of convergence under a decrease in the spacing between nodes. Separable Nonnegative Matrix Factorization Using Powered Ratio-of-Norms Regularization 作者: Matthew McCarver, Jing Qin 类别: math.NA, cs.LG, math.OC 发布日期: 2026-08-28 arXiv ID: 2608.28799v1 链接: http://arxiv.org/abs/2608.28799v1 摘要: Separable nonnegative matrix factorization (SNMF) has been widely used for low-rank representation and clustering of nonnegative data, owing to its ability to produce part-based and interpretable decompositions. In particular, SNMF is closely related to graph clustering and community detection. To enhance sparsity and identifiability of the learned factors, we propose an $\ell_1^p/\ell_2$-regularized SNMF model based on a powered ratio-of-norms regularizer. The resulting formulation is nonconvex and nonsmooth, which poses significant challenges for optimization. To address this, we develop efficient algorithms based on the difference-of-convex function algorithm (DCA) and the alternating direction method of multipliers (ADMM). The proposed methods decompose the original problem into tractable subproblems, leveraging closed-form proximal operators associated with the powered norm terms. We establish descent and limiting criticality properties for the DCA scheme and convergence under standard assumptions for the ADMM scheme. Extensive numerical experiments on synthetic datasets and hand gesture classification tasks demonstrate that the proposed approach achieves competitive or improved performance in anchor identification and classification accuracy compared with existing SNMF methods, while maintaining competitive computational efficiency.7. 从多层问题到多个两层问题:一种新的频时混合多重散射积分方程求解器
8. 标量常微分方程的镜面椭圆方法:精确性与四阶精度
9. 带多重半正定罚项的参数鲁棒子空间校正
10. 带可逆边界的双松弛时间向量格子玻尔兹曼格式的网格一致幂稳定性
11. 棋盘格壳:一种具有尺度兼容补全的仅位置薄壳离散化
12. 非自伴几乎Mathieu矩阵与苏格兰旗算子的谱
13. 不定复对称线性方程组块预条件子的一种变体
14. 高阶特征值问题谱配点法的进展
15. 多原子玻尔兹曼碰撞算子的Wigner-Eckart分解
16. 非线性热孔隙弹性问题序贯有限元方法的最优误差估计
17. 带自适应特征重构的狭义相对论流体动力学高效混合WENO格式
18. 具有一般状态方程的相对论流体动力学中的约束保持AFD-WENO格式
19. 张量正交子空间分裂:理论与应用
20. 分数高斯噪声驱动的随机多尺度次扩散的分析与逼近
21. Nechvile变换下的航天器动力学与三体问题中的推进剂计算
22. 定常时空上对称约化广义相对论流体动力学的熵稳定且保持物理约束的DGSEM
23. 随机自参数系统离散化下的随机吸引子与几乎必然稳定性
24. D-RADI:一种求解大规模离散时间代数Riccati方程的低秩ADI算法
25. 参数化概率流形分解的收敛性分析
26. 用于非线性场重构的间隙概率流形分解
27. Hermitian矩阵非对角块扰动的酉不变范数的一个精确界
28. 最小化雅可比方法用于特征值和奇异值的算术与通信复杂度:第二部分——并行算法
29. 一种基于隐式Householder QR分解的GPU加速分块自适应随机范围查找器
30. 三角剖分多边形区域上求积的一般超收敛结果
31. 使用幂化范数比正则化的可分离非负矩阵分解