生成时间: 2026年09月11日 14:12:15
重点关注论文数量: 3
了解领域论文数量: 5
不相关论文数量: 31
重点关注论文(完整分析)
1. Helmholtz 方程的笛卡尔 PML 截断在高频下具有指数精度
Cartesian PML truncation for the Helmholtz equation is exponentially accurate at high frequency
作者: Martin Averseng, Jeffrey Galkowski, Euan A. Spence
类别: math.AP, math.NA
发布日期: 2026-09-10
arXiv ID: 2609.11343v1
摘要翻译: 对于广泛一类 Helmholtz 散射问题(即那些适合 Sjöstrand–Zworski 黑箱框架的问题),我们证明:用笛卡尔完美匹配层(PML)近似辐射条件所产生的误差,随 PML 的宽度、PML 缩放的强度以及频率呈指数衰减。
Comment: 无
链接: http://arxiv.org/abs/2609.11343v1
1. 研究对象和背景
本文研究频域 Helmholtz 散射问题的数值截断误差, 特别是用笛卡尔完全匹配层 (perfectly matched layer, PML) 近似无穷远辐射条件时产生的误差. 考虑半经典缩放下的 Helmholtz 散射算子 $P$, 它属于 Sjöstrand–Zworski 的黑箱框架 [SZ91]: 在某个球 $Q_{bb}=\prod_{i=1}^d(-R_{bb,i},R_{bb,i})$ 外, $P=-k^{-2}\Delta$; 在球内 $P$ 是某个自伴算子. 该框架涵盖 Dirichlet, Neumann 以及可穿透 Lipschitz 障碍的散射问题, 也涵盖变系数介质的散射问题. 我们考虑方程
$$ (P-1)u=g $$的 outgoing 解, 其中 $k$ 是频率, 方程在半经典缩放下对应于通常的 Helmholtz 方程 $(\Delta+k^2)u=f$. 关键未知量是限制吸收预解 $R_P=\lim_{\epsilon\to0^+}(P-1-i\epsilon)^{-1}$, 它给出满足辐射条件的物理散射解. 在数值上, 为了在无界区域上求解, 人们引入复坐标变换
$$ \gamma_\theta: x\mapsto x+i\tan\theta\,\nabla F(x), \qquad F(x)=\sum_{j=1}^d F_j(x_j), $$其中 $F_j\in C^\infty(\mathbb{R})$ 满足 $F_j\equiv0$ 在 $[-R_{bb,j},R_{bb,j}]$ 上, $F'_j(x)=x$ 当 $|x|\ge R_{lin,j}$, 并且 $\{F''_j>0\}=\{|x|>R_{st,j}\}$, $F_j$ 凸. 该变换在层 $Q_{tr}\setminus Q_{st}$ 中引入复吸收, 然后在截断边界 $\partial Q_{tr}$ 上施加齐次 Dirichlet 条件. 对应的变形算子为
$$ \Delta_\theta=\sum_{j=1}^d\left(\frac{1}{1+i\tan\theta\,F''_j(x_j)}\partial_{x_j}\right)^2, $$而 PML 问题为 $(P_\theta^D-1)u_D=\chi g$, 其中 $P_\theta^D$ 是 $P$ 变形到 $\gamma_\theta$ 后的 Dirichlet 实现. 本文的核心问题是: 当频率 $k$ 很大时, PML 截断解 $u_D$ 与真实散射解 $R_P(\chi g)$ 在未缩放区域内的误差如何随频率衰减. 已有理论背景方面, 径向 PML 早已被证明误差随 PML 宽度和缩放强度指数衰减 [LS98, LS01, HSZ03, BP07], 最近 [GLS23] 又证明了径向 PML 误差随频率指数衰减. 然而, 笛卡尔 PML 在数值分析中使用更广泛, 其分析却更困难: 因为笛卡尔 PML 算子的阻尼每次只出现在一个坐标方向, 一条垂直于缩放方向的射线在 PML 中不会衰减. 此前 [Kim09, KP10a, KP10b] 证明了笛卡尔 PML 误差随 PML 宽度指数小, [BP13] 进一步证明了误差同时随 PML 宽度和缩放强度指数小. 关于频率依赖性, 用传播奇异性方法可以较容易地得到误差超多项式小, 即对任意 $N>0$ 存在 $C_N$ 使得误差 $\le C_Nk^{-N}$; 这一般复吸收情形的结果见 [GGG+24, Theorem 4.7]. 在无散射且缩放函数在 PML 边界为线性的特殊情形, [CX13, Lemma 3.4] 证明了误差随 PML 宽度, 缩放强度和频率指数小; 最近 [WZ26, Theorem 5.6] 将该结果推广到星形 Dirichlet 障碍的散射. 本文填补的空白是: 对一大类黑箱 Helmholtz 散射问题, 在解算子多项式有界的假设下, 证明笛卡尔 PML 截断误差随 PML 宽度, 缩放强度和频率指数衰减. 这一结果把之前径向 PML 的频率指数精度推广到笛卡尔 PML, 并把无散射或星形障碍的特殊结果推广到一般 Lipschitz 障碍和变系数介质, 同时显式处理了笛卡尔 PML 在截断边界附近缺乏强制性的根本困难.
2. 主要定理或主要结果
本文的主要结果是定理 2.1, 它给出笛卡尔 PML 截断误差的指数上界. 首先回顾假设: 取 $\chi\in C_c^\infty(Q_{st})$ 且 $\text{supp}(1-\chi)\cap Q_{bb}=\emptyset$. 假设存在 $J\subset(0,\infty)$ 和 $M>0$, 使得对每个 $k_0>0$ 存在 $C>0$, 满足对一切 $k\in J$, $k\ge k_0$, 有
$$ \|\chi R_P\chi\|_{H\to H}\le Ck^M. $$这是解算子多项式有界的条件; 对非 trapping 散射体包括星形障碍它总是成立, 对一般情形 [LSW21, Theorem 3.3] 说明它发生在“大多数”频率. 在此假设下, 定理断言: 对任意 $\delta,\epsilon>0$, 存在 $C,c,k_1>0$, 使得当 $R_{st,i}+\epsilon 这里 $R_{st,i}$ 是缩放起始位置, $R_{tr,i}$ 是截断位置, 因此 $R_{tr,i}-R_{st,i}$ 是沿第 $i$ 个坐标方向的 PML 有效宽度. 误差上界对 $k$, $\tan\theta$ 和最小层宽均指数小. 定理还强调, 对于 Dirichlet, Neumann 和可穿透 Lipschitz 障碍以及变系数散射, 存在 $C>0$ 使得对 $k>k_0$, 因此 (2.4) 直接控制了 $k$-加权 $H^1$ 误差. 除主定理外, 证明过程中建立了若干关键辅助结果. 定理 3.1 给出自由复缩放预解的估计: 给定 $\delta>0$ 和 $s\in\mathbb{R}$, 存在 $C>0$ 和 $k_0>0$, 使得对 $k\ge k_0$ 和 $\theta\in[\delta,\pi/2-\delta]$, 引理 3.6 给出自由复缩放预解的指数衰减: 给定 $\delta,\epsilon,C_0,k_0>0$, 存在 $C,\tilde c>0$, 使得若 $\chi_1,\chi_2\in C^\infty(\mathbb{R}^d;[0,1])$ 满足 $\text{supp}\,\chi_1\subset Q_{tr}\setminus Q_{lin}$, $\text{supp}\,\chi_2\subset Q_{st}$, 且 $\|\partial^\alpha\chi_i\|_{L^\infty}\le C_0$ ($|\alpha|\le2$), 则对 $R_{st}+\epsilon 引理 3.3 证明射线均匀逃逸到椭圆区域: 对 $\delta>0$, 存在 $T>0$, $\mu>0$, 使得对所有 $\theta\in[\delta,\pi/2-\delta]$ 和所有 $(x_0,\xi_0)\in T^*\mathbb{R}^d$ 满足 $|p_{\theta,\text{fr}}(x_0,\xi_0)-1|\le\mu$, 存在 $0\le t\le T$ 使得 引理 3.7 是一个纯代数不等式: 若 $u,v\in\mathbb{R}^d$ 且 $\langle u,v\rangle\ge0$, $u\neq0$, 则 这些辅助结果与主定理共同构成了对笛卡尔 PML 频率指数精度的完整证明. 需要指出, 从提供的内容中无法确认定理 2.1 证明中所有后续技术细节, 例如 Lemma 5.2 的完整陈述和证明, 但上述结果已充分说明主要结论及其技术核心. 本文证明的核心障碍在于笛卡尔 PML 的截断边界附近缺乏强制性. 对于径向 PML, 在截断边界附近二次型具有强制性, 因此可以较容易地得到先验估计, 见 [GLS23, Lemma 4.4]. 但笛卡尔 PML 只在缩放区域的“角落”处具有强制性, 在沿单个方向缩放的边界附近, 算子并非强制. 这迫使作者采用半经典奇异传播论证, 而不是简单的能量估计或强制性论证. 证明的整体框架是将误差写成可分离形式, 然后分别估计自由复缩放预解的指数衰减和 Dirichlet PML 预解的多项式界. 具体地, 令 $g\in H$, $\chi$ 是支撑在未缩放区域 $Q_{st}$ 的截断函数, 考虑 $u_\theta$ 和 $u_\theta^D$ 分别满足 已知 $u_\theta$ 在未缩放区域等于散射解 $R_P\chi g$, 因此只需证明 $u_\theta$ 与 $u_\theta^D$ 在未缩放区域相差指数小. 定义误差 $E:=\psi u_\theta-u_\theta^D$, 其中 $\text{supp}\,\chi\subset\{\psi\equiv1\}\subset\text{supp}\,\psi\subset Q_{tr}$. 它满足 其中 $\phi\equiv0$ 在黑箱附近, $\phi\equiv1$ 在 $Q_{st}$ 外. 利用自由预解表示散射体外的 Helmholtz 解, 得到 从而 这里 $\tilde g$ 支撑在未缩放区域, 而 $[P_\theta,\psi]$ 支撑在截断边界附近. 于是定理 2.1 归结为两个核心估计: 第一, 自由复缩放预解 $\chi_1(P_{\theta,\text{free}}-1)^{-1}\chi_2$ 随 $\chi_1$ 与 $\chi_2$ 的支撑距离指数衰减; 第二, Dirichlet PML 预解 $(P_\theta^D-1)^{-1}$ 具有多项式界 $\|(P_\theta^D-1)^{-1}\|\le Ck^M$. 第一个核心估计, 即自由复缩放预解的指数衰减, 由引理 3.6 给出. 其证明利用了自由复缩放预解的显式基本解. 具体地, 其中 $\nu:=(d-2)/2$, $H^{(1)}_\nu$ 是第一类 Hankel 函数. 作者用 Schur 检验或 Riesz–Thorin 插值定理将 $L^2\to L^2$ 算子范数归结为对核的积分估计. 然后利用 Hankel 函数在大参数下的渐近展开 得到 $|G_k(z)|\le Ck^{(d+1)/2}|r(z)|^{-(d+1)/2}\exp(-k\Im r(z))$, 只要 $\Im r(z)\ge c$. 关键的一步是估计 $\Im r(\gamma_\theta(x)-\gamma_\theta(y))$. 由 $\gamma_\theta(x)-\gamma_\theta(y)=(x-y)+i\tan\theta(\nabla F(x)-\nabla F(y))$ 以及 $F$ 的凸性, $\langle\nabla F(x)-\nabla F(y),x-y\rangle\ge0$. 此时引理 3.7 给出 结合支撑条件, 对 $x\in\text{supp}\,\chi_1\subset Q_{tr}\setminus Q_{lin}$ 和 $y\in\text{supp}\,\chi_2\subset Q_{st}$, 可以找到某个 $i$ 使得 $|x_i-y_i|\ge c|x-y|$, $|x_i|>R_{lin,i}$, $|y_i|\le R_{st,i}$, 从而由 $F''_i\ge0$ 推出距离下界. 这就建立了指数衰减. 第二个核心估计, 即 $P_{\theta,\text{free}}-1$ 的可逆性与多项式界, 由定理 3.1 给出. 证明使用半经典微局部分析. 算子 $P_{\theta,\text{free}}$ 的半经典主符号为 其虚部 $\Im p_{\theta,\text{fr}}\le0$, 实部为 $\Re p_{\theta,\text{fr}}=\sum_i(1-(\tan\theta F''_i)^2)\xi_i^2/(1+(\tan\theta F''_i)^2)^2$. 引理 3.3 证明从 $|p_{\theta,\text{fr}}-1|\le\mu$ 的相空间点出发, 沿 $\Re p_{\theta,\text{fr}}$ 的 Hamilton 流反向演化会在有限时间 $T$ 内到达 $|p_{\theta,\text{fr}}-1|\ge\mu$ 的椭圆区域. 证明中使用了 $p^i_\theta=(1-(\tan\theta F''_i)^2)/(1+(\tan\theta F''_i)^2)^2\xi_i^2$ 的守恒性以及 Gronwall 不等式. 有了这个逃逸性质, 作者将相空间分解为椭圆区域 $\{|p_{\theta,\text{fr}}-1|\ge\mu\}$ 和传播区域 $\{|p_{\theta,\text{fr}}-1|\le\mu\}$, 分别使用椭圆估计 (引理 3.4) 和传播估计 (引理 3.5). 引理 3.5 的证明通过单位分解和伪微分算子 $\text{Op}_k(a)$ 的局部化, 结合传播奇异性结果 [DZ19, Theorem E.47] 得到 然后对 $P_{\theta,\text{free}}^*$ 应用类似论证, 得到可逆性及定理 3.1 中的估计. 第三个关键技术点是截断边界附近对于 Dirichlet 问题的先验估计. 由于笛卡尔 PML 在边界附近非强制, 作者必须利用传播奇异性论证. 然而, 半经典传播结果在带边界区域尚未有文献写出, 因此作者假设 $F'_i(x)=x$ 当 $|x|\ge R_{lin,i}$, 即缩放函数在 PML 边界外是线性的. 这一假设允许使用反射论证 (类似于 [BP13, Proof of Theorem 5.5] 和 [CX13, Lemma 3.4] 中的二维笛卡尔 PML 分析) 将解延拓过 $\partial Q_{tr}$, 从而绕过带边界传播的问题, 直接使用无边界流形上的半经典传播结果 [DZ19, Appendix E]. 在相空间中, 对截断边界附近的任意点 $(x,\xi)$, 要么算子符号非零而椭圆, 要么反向 Hamilton 轨道在有限时间内逃逸到椭圆区域; 这给出所需的先验估计. 作者在 Remark 2.2 中明确指出, 笛卡尔结构仅在三处被使用: (i) 估计 $(P_{\theta,\text{free}}-1)^{-1}$; (ii) 证明自由 Green 核的指数估计; (iii) 证明截断边界附近的先验估计. 因此, 任何缩放只要能建立这三个估计, 就可以证明类似的 PML 指数误差界. 创新点在于将频率指数精度从径向 PML 和特殊笛卡尔情形推广到一般黑箱散射问题, 并用半经典传播取代了强制性论证, 从而克服了笛卡尔 PML 边界附近非强制的本质困难. 与既有工作相比, 本文的主要推进在于把笛卡尔 PML 截断误差的频率指数精度从特殊情形推广到一大类黑箱 Helmholtz 散射问题. 首先, 与径向 PML 的文献比较. 径向 PML 的误差随 PML 宽度和缩放强度指数衰减早已被证明, 见 [LS98, Theorem 2.1], [LS01, Theorem A], [HSZ03, Theorem 5.8], [BP07, Theorem 3.4]. 最近 [GLS23, Theorem 1.5] 进一步证明了径向 PML 误差随频率指数衰减. 本文处理的是笛卡尔 PML, 它比径向 PML 更难分析, 因为笛卡尔 PML 算子的阻尼每次只出现在一个坐标方向, 一条垂直于缩放方向的射线在 PML 中不被阻尼. 本文证明笛卡尔 PML 同样可以达到频率指数精度, 但针对的是更一般的黑箱散射框架, 而不是径向几何. 在方法上, 径向 PML 可以使用强制性得到先验估计, 而本文必须使用半经典奇异传播, 这体现了技术难度和推进. 其次, 与先前的笛卡尔 PML 工作比较. [Kim09, KP10a, KP10b] 证明了误差随 PML 宽度指数小, [BP13] 进一步证明了误差同时随 PML 宽度和缩放强度指数小. 关于频率依赖性, 一般复吸收通过传播奇异性方法可以得到超多项式小, 即对任意 $N>0$ 误差 $\le C_Nk^{-N}$; 这在光滑变系数散射情形被明确写出, 见 [GGG+24, Theorem 4.7]. 在无散射且缩放函数在 PML 边界为线性的特殊情形, [CX13, Lemma 3.4] 证明了误差随 PML 宽度, 缩放强度和频率指数小. 最近 [WZ26, Theorem 5.6] 将该结果推广到星形 Dirichlet 障碍的散射. 本文与这些工作的区别在于: 不再局限于无散射或星形障碍, 而是考虑 Sjöstrand–Zworski 黑箱框架下的散射问题, 涵盖 Lipschitz Dirichlet 和 Neumann 障碍, 可穿透障碍以及变系数介质; 同时证明误差随频率指数衰减, 而不仅仅是超多项式小. 这一加强依赖于解算子多项式有界的假设, 即 $\|\chi R_P\chi\|_{H\to H}\le Ck^M$. 该假设对非 trapping 散射体包括星形障碍总是成立, 对一般情形 [LSW21, Theorem 3.3] 说明它发生在“大多数”频率. 因此, 本文的适用范围比 [CX13] 和 [WZ26] 更广, 但代价是需要多项式有界假设; 对于不满足该假设的 trapping 频率, 本文结果不直接适用. 第三, 与 [GLS23] 的比较. [GLS23] 证明了径向 PML 的频率指数精度. 本文可以看作笛卡尔 PML 的对应结果. 但证明策略不同: [GLS23] 使用径向 PML 在截断边界附近的强制性, 而本文必须处理笛卡尔 PML 的非强制性, 因此发展了半经典传播论证. 本文的结果在适用范围上更广, 因为黑箱框架涵盖更一般的散射体, 但在 PML 几何上限于笛卡尔型, 并且要求缩放函数在 PML 边界外线性. 这一线性假设是为了使用反射论证, 从而绕过带边界的半经典传播问题. 作者明确指出, 一旦半经典边界传播结果被写出, 该假设可以移除. 因此, 这是当前分析技术的一个暂时性限制, 而不是本质限制. 第四, 与 [GGG+24] 的比较. [GGG+24, Theorem 4.7] 对光滑变系数散射证明了笛卡尔 PML 误差超多项式小. 本文在更强的多项式有界假设下, 将结论加强为指数小, 并且适用于更一般的黑箱散射. 从误差估计的形式看, 本文给出显式的指数因子 $\exp(-c k\tan\theta\min_i\{R_{tr,i}-R_{st,i}\})$, 而超多项式结果只给出 $k^{-N}$ 型衰减. 在频率很高时, 指数小远强于任何多项式小. 但代价是假设了解算子的多项式有界性, 并且要求缩放函数线性. 此外, 本文的常数 $C,c$ 可能依赖于 $M$, $J$, $\delta$, $\epsilon$ 以及几何参数, 但文中没有给出对 $M$ 的显式依赖; 从提供内容中无法确认常数是否对 $M$ 有精确控制. 第五, 与 [WZ26] 的比较. [WZ26, Theorem 5.6] 将 [CX13] 的无散射结果推广到星形 Dirichlet 障碍, 证明了笛卡尔 PML 误差在宽度, 强度和频率上指数小. 本文进一步推广到一般黑箱框架, 包括 Lipschitz 障碍和可穿透障碍以及变系数介质. 在假设方面, [WZ26] 可能依赖于星形几何, 而本文用解算子多项式有界性替代了星形假设. 这使得本文适用于更广泛的散射体, 但多项式有界性假设本身是频率上的限制: 对某些 trapping 频率可能失效. 另外, 本文要求缩放函数在 PML 边界外线性 $F'_i(x)=x$, 这与 [CX13] 和 [WZ26] 中的线性缩放假设类似. 因此, 在缩放剖面选择上, 本文仍有限制. 最后, 关于代价与局限. 本文的证明没有处理带边界的半经典传播, 因此必须假设缩放函数在 PML 边界外线性. 这是技术性代价, 而非物理或数值上的本质限制. 作者在 Remark 2.2 中指出, 笛卡尔结构仅用于自由预解估计, 自由 Green 核指数估计和截断边界先验估计; 任何能建立这三点的缩放都可以得到类似结果. 因此, 本文的方法框架具有可推广性. 另外, 本文假设解算子多项式有界, 这对非 trapping 散射体总是成立, 对一般情形在大多数频率成立. 与 [GGG+24] 的超多项式结果相比, 本文的指数结果更强, 但需要额外假设. 从提供内容中无法确认作者是否讨论了 trapping 频率下的具体行为, 也没有给出数值实验验证. 总体而言, 本文在笛卡尔 PML 的频率指数精度分析上迈出了重要一步, 把适用范围从无散射和星形障碍推进到一般黑箱散射, 同时保留了线性缩放假设作为当前分析技术的限制. Polynomial preserving recoveries of edge element method on Cartesian grids for the time-harmonic Maxwell equations with large wave number 作者: Shuaishuai Lu, Haijun Wu 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11064v1 摘要翻译: 本文考虑了笛卡尔网格上最低阶第一类 Nédélec 棱元法 (EEM),用于求解具有大波数的三维时谐 Maxwell 方程。分别针对棱元解的旋度和解本身,提出了新的多项式保持恢复 (PPR) 算子。在 $κ^3 h^2 C_{\mathrm{sol}}$ 充分小的条件下,对恢复的旋度和恢复的解均证明了二阶超收敛估计,其中 $κ$ 是波数,$h$ 是网格尺寸,$C_{\mathrm{sol}}$ 是与 Maxwell 解算子相关的稳定常数。特别地,分析表明所提出的 PPR 过程不能缓解 EEM 固有的众所周知的污染效应。为减小污染误差,我们进一步提出了一种新的连续内罚棱元法 (CIP-EEM),它加入了一个额外的法向跳跃罚项。结果表明,通过适当选择罚参数,新的 CIP-EEM 能够使相位误差在 $κh$ 方面提高两个阶。数值实验验证了理论超收敛结果,并表明 CIP-EEM 在高频区域能够有效减小污染误差。 Comment: 无 链接: http://arxiv.org/abs/2609.11064v1 本文研究三维时谐 Maxwell 方程在理想导体边界条件下的高频数值逼近。设 $\Omega\subset\mathbb{R}^3$ 为有界、单连通、Lipschitz 多面体,边界 $\Gamma=\partial\Omega$ 连通,波数 $\kappa>0$,源项为 $f$,未知电场 $E$ 满足 其中 $\nu$ 为单位外法向。该模型是计算电磁学中的基本方程,广泛出现于医学成像、无线通信、天线设计、雷达探测与电磁隐身等场景。其弱形式为:求 $E\in V:=H_0(\operatorname{curl};\Omega)$,使得 本文关注最低阶第一类 Nédélec 边元法在笛卡尔网格上的离散。定义 边元解 $E_h\in V_h$ 满足 当波数 $\kappa$ 很大时,除了通常的插值误差外,数值离散还会遭受著名的污染效应。污染源于离散波数与精确波数之间的相位误差,导致数值色散并显著降低离散解精度。已有工作 [11] 表明,在网格条件 $\kappa^3h^2C_{\mathrm{sol}}$ 足够小下,最低阶边元解满足预渐近估计 其中 $h$ 为网格尺寸,$C_{\mathrm{sol}}$ 为 Maxwell 解算子的稳定性常数。右端第一项 $O(\kappa h)$ 为插值误差,第二项 $O(\kappa^3h^2C_{\mathrm{sol}})$ 为污染项。这里的能量范数记为 稳定性分析中引入 $M_\kappa:=\sup_{j\ge1}|(\lambda_j+\kappa^2)/(\lambda_j-\kappa^2)|$,其中 $(\lambda_j,u_j)$ 为正 Maxwell 特征对,并令 $C_{\mathrm{sol}}=\kappa^{-1}M_\kappa$。若 $\kappa^2$ 不是 Maxwell 特征值,则 当 $\Omega$ 为凸多面体时,谱隙论证给出 $\delta_\kappa\lesssim\kappa^{-1}$,从而 $C_{\mathrm{sol}}\gtrsim\kappa^2$。这说明高频下稳定性常数可能很大,网格条件也更苛刻。 对于 Lagrange 有限元,超收敛理论与后处理技术已相当成熟。多项式保持恢复由 Naga 与 Zhang 提出,其基本思想是在局部最小二乘意义下用有限元解重构高阶多项式,再通过微分得到恢复梯度。该方法简单、稳健,已广泛用于超收敛分析与后验误差估计。相比之下,针对 Maxwell 方程的边元超收敛理论远不完整。Monk 首先建立了 Maxwell 问题的超收敛结果,Huang 等对二维和三维 Nédélec 元作了进一步发展。Wang 等将分析推广到任意阶边元并在混合有限元框架下应用 PPR 得到全局恢复超收敛结果;他们还针对二维时谐 Maxwell 方程提出了 curl 恢复方法。然而,已有文献中没有波数显式分析,也未处理大波数区域中污染误差的作用。本文试图填补这一空白:首先,在 $\kappa^3h^2C_{\mathrm{sol}}$ 足够小的条件下,为最低阶第一类 Nédélec 边元构造新的 curl 型多项式保持恢复算子 $G_h$,证明高频预渐近区域中的超收敛估计;其次,构造适用于最低阶第一类边元的函数值恢复算子 $F_h$,因为该边元的 $L^2$ 误差一般不能自然达到二阶收敛;再次,分析恢复过程对污染误差的影响,回答 PPR 是否能消除 Maxwell 边元法中的污染;最后,提出带法向跳量惩罚的新连续内罚边元法,以降低相位误差。需要说明的是,提供的 PDF 文本仅包含摘要、引言与部分预备结果,第 3 至第 5 节的完整定理证明、恢复算子的局部构造与色散分析细节未包含在提供内容中,因此本文只能依据可见部分作可靠概述。 本文的主要结果可分为三部分:curl 恢复的超收敛与污染分析、函数值恢复的超收敛、以及新连续内罚边元法的相位误差改进。 第一,在最低阶第一类 Nédélec 边元、均匀笛卡尔网格、$\Omega$ 为矩形盒的设定下,作者构造了针对 $\operatorname{curl} E_h$ 的多项式保持恢复算子 $G_h$。在条件 $\kappa^3h^2C_{\mathrm{sol}}\le C_0$ 下,其中 $C_0$ 与 $\kappa$ 和 $h$ 无关,且 $\kappa^3h^2C_{\mathrm{sol}}$ 足够小,证明 其中 与已有预渐近估计 $9E-E_h9\lesssim(\kappa h+\kappa^3h^2C_{\mathrm{sol}})(\kappa^{-1}\|\operatorname{curl} E\|_1+|E|_1)$ 相比,恢复算子把插值部分的误差从 $O(\kappa h)$ 改进为 $O(\kappa^2h^2)$,而污染项仍为 $O(\kappa^3h^2C_{\mathrm{sol}})$,阶数没有降低。换言之,PPR 在这种情况下实现了二阶超收敛,但不能消除边元法固有的污染项。 第二,为进一步澄清恢复过程对污染误差的影响,作者证明 该估计表明,恢复量与离散解 curl 之间的差仍继承污染项,因此 (1.3) 中的污染项不能通过恢复过程移除。特别地,$\|G_hE_h-\operatorname{curl} E_h\|$ 只有在条件 $\kappa^3h^2C_{\mathrm{sol}}\lesssim\kappa h$ 下,才能作为 $\|\operatorname{curl}(E-E_h)\|$ 的可靠估计器。在高频区域,PPR 改善插值误差,但不降低内在污染误差。 第三,针对最低阶第一类 Nédélec 边元的 $L^2$ 误差一般不能自然达到二阶的问题,作者构造了函数值恢复算子 $F_h$,并证明相应的超收敛估计。摘要明确指出,对恢复后的解也证明了二阶超收敛估计。然而,提供的 PDF 文本没有包含第 4 节的精确定理、假设和误差表达式,因此从提供内容中无法确认 $F_h$ 的具体定义、证明所需的正则性假设以及最终估计中 $\kappa,h,C_{\mathrm{sol}},C_E$ 的精确组合方式。这里只能确认作者声称获得了函数值层面的恢复超收敛结果。 第四,由于 PPR 改善插值误差但不消除污染,作者进一步提出一种新的连续内罚边元法。经典 CIP-EEM 使用与边元法相同的逼近空间,并在内部单元面的 curl 切向分量跳量上添加惩罚项;该格式在第二类边元空间中已被证明能改善稳定性并减少污染。但作者指出,对于第一类边元空间,仅调整经典 CIP-EEM 的惩罚参数不能显著降低相位误差。为此,新方法额外加入了解的法向分量跳量惩罚项。色散分析表明,通过适当选择惩罚参数,新 CIP-EEM 可以把相位误差在 $\kappa h$ 中改进两个阶。数值实验证实了理论超收敛结果,并显示 CIP-EEM 在高频区域能有效减少污染误差。不过,提供内容未给出新格式的完整双线性形式、惩罚参数最优选取公式与色散关系的详细推导,因此这些具体结果从提供内容中无法确认。 此外,作为支撑性结果,作者在预备部分给出波数显式稳定性、辅助 Maxwell 椭圆问题、离散 Helmholtz 分解、椭圆投影与若干超收敛估计。例如,若 $\kappa h\le\tilde C_0$,则有离散 inf-sup 条件 其中 $\hat a(u,v)=(\operatorname{curl} u,\operatorname{curl} v)-\kappa^2(u,v)+2\kappa^2(P^\perp u,P^\perp v)$。这些估计是后续恢复超收敛证明的核心工具。 本文的技术难点首先来自 Maxwell 方程的向量值特性以及 $\operatorname{curl}$ 算子具有很大的非平凡核。对于高频 Helmholtz 方程,Du 等已证明 PPR 虽然能改善插值误差分量,但污染误差会被恢复量继承,不能仅靠后处理消除。Maxwell 情形更微妙,不能把 Helmholtz 的论证逐字推广。一个直接障碍是,边元空间中的误差可能包含梯度分量,而 $\operatorname{curl}$ 对此分量不敏感;同时,最低阶第一类 Nédélec 元的 $L^2$ 误差一般不能自然达到二阶,这使得函数值恢复需要额外处理。 为处理 curl 的大核,作者系统使用了连续与离散 Helmholtz 分解。对任意 $v\in V$,有 $v=P_0v+P^\perp v$,其中 $P_0v\in\nabla H_0^1(\Omega)\cap V$,$P^\perp v\in H(\operatorname{div}^0;\Omega)\cap V$,且二者 $L^2$ 正交。离散层面,对 $v_h\in V_h$ 有 $v_h=P_0^hv_h+P_h^\perp v_h$,并满足 $(P_h^\perp v_h,\nabla\psi_h)=0$ 对所有 $\psi_h\in U_h^0$,以及 这条离散分解与嵌入结果配合使用:在凸多面体上,对 $v\in X=H_0(\operatorname{curl})\cap H(\operatorname{div}^0)$ 或 $v\in H(\operatorname{curl})\cap H_0(\operatorname{div}^0)$,有 $v\in H^1(\Omega)$ 且 该嵌入把 curl 控制转化为 $H^1$ 控制,是后续对偶论证与恢复误差估计的关键。 第二个核心工具是一类专门设计的辅助 Maxwell 椭圆问题。作者引入 和 对应双线性形式 其中 $\tilde a$ 是强制且对称正定的,因为 $\tilde a(u,u)=9u9^2$。而 $\hat a$ 虽然不定,但通过加入 $2\kappa^2P^\perp$ 项,可以在连续和离散层面建立 inf-sup 条件。连续估计为 离散估计为在 $\kappa h\le\tilde C_0$ 下 离散 inf-sup 的证明是关键技术点:对 $u_h\in V_h$,写 $u_h=u_{1h}+u_{2h}$,其中 $u_{1h}=P_0^hu_h$,$u_{2h}=P_h^\perp u_h$,取测试函数 $v_h=u_{2h}-u_{1h}$。利用离散正交性、$\|P^\perp u_h-P_h^\perp u_h\|\lesssim h\|\operatorname{curl} u_h\|$ 以及 $9v_h9=9u_h9$,得到 当 $\kappa h$ 足够小时,即得 inf-sup 条件。该构造使得原本不定的高频 Maxwell 问题在离散框架中仍具有可控的稳定性。 第三,作者引入两个椭圆投影 $P_h$ 和 $\tilde P_h$: 由上述 inf-sup 和强制形式可得准最优性估计。更重要的是,作者证明了散度自由分量的 $L^2$ 超收敛: 并进一步在 $\kappa h\le\tilde C_0$ 下得到 证明采用对偶论证。以 $P_h$ 为例,取 $w$ 满足 并利用辅助问题的正则性估计 $\kappa\|w\|_1+\|w\|_{H^1(\operatorname{curl})}\lesssim\|P^\perp(u-P_hu)\|$。随后在弱形式中取 $\phi=u-P_hu$,利用 Galerkin 正交性将其替换为 $u-\pi_hu$,再结合插值估计与 $\kappa h\lesssim1$,得到 从而推出所需的 $h$ 阶超收敛。这一引理是恢复函数值达到二阶精度的基础之一。 第四,辅助问题的正则性估计在证明中反复使用。作者证明对任意 $f'\in L^2(\Omega)$,问题 (2.28) 与 (2.29) 的唯一解满足 若 $f'\in H(\operatorname{div}^0;\Omega)$ 且 $\Omega$ 凸,则 若进一步 $\Omega$ 为长方体,则 这些估计为对偶论证提供了必要的正则性提升,并且与 $C_{\mathrm{sol}}$ 的波数显式稳定性一起,使最终超收敛估计可以显式追踪 $\kappa$ 与 $h$。 第五,关于 PPR 算子的局部构造,提供的文本只说明针对 $\operatorname{curl} E_h$ 与解本身分别提出新的多项式保持恢复算子 $G_h$ 和 $F_h$,并说明其思想是在局部最小二乘意义下重构高阶多项式再求导。但第 3 节和第 4 节中 $G_h$ 与 $F_h$ 的精确定义、局部 patch 的选取、权重、多项式空间以及与边元自由度的匹配方式均未包含在提供内容中,因此无法确认其具体实现细节。可以确认的证明框架是:先建立有限元解与精确解边元插值之间的超闭性估计,再结合上述辅助问题、离散 inf-sup 与投影超收敛,推导恢复量的超收敛估计。 第六,污染分析的技术路线是直接估计恢复量与离散解 curl 之间的差 $G_hE_h-\operatorname{curl} E_h$。该估计不试图消除污染,而是揭示污染项在恢复过程中被继承。这一思路与 Du 等在 Helmholtz 方程中的结果相呼应,但 Maxwell 情形需要额外处理向量分解、curl 核与第一类边元函数值恢复。 第七,对于新 CIP-EEM,作者的技术创新是在经典 CIP-EEM 的 curl 切向跳量惩罚之外,再加入法向跳量惩罚。经典格式对第二类边元有效,但对第一类边元调参不能显著降低相位误差。新格式通过色散分析选取惩罚参数,使相位误差在 $\kappa h$ 中改进两个阶。具体色散关系、惩罚参数选取公式以及完整误差分析未在提供内容中给出,因此只能确认作者报告了该理论结论并用数值实验验证了污染减少。 本文与已有工作的比较可以从以下几个维度展开。 首先,与 [11] 中最低阶边元法的波数显式预渐近估计相比,本文在相同的高频框架下工作:均假设 $\kappa^3h^2C_{\mathrm{sol}}$ 足够小。已有估计为 其中 $O(\kappa h)$ 是插值误差,$O(\kappa^3h^2C_{\mathrm{sol}})$ 是污染误差。本文的 curl 恢复估计为 可以看到,恢复算子把插值部分从 $O(\kappa h)$ 改进为 $O(\kappa^2h^2)$,即实现了二阶超收敛;但污染项仍为同一阶 $O(\kappa^3h^2C_{\mathrm{sol}})$。这是本文相对 [11] 的核心推进与同时暴露的局限。进一步,本文通过估计 $\|G_hE_h-\operatorname{curl} E_h\|$ 证明污染项不能被恢复过程移除,并指出 $\|G_hE_h-\operatorname{curl} E_h\|$ 只有在 $\kappa^3h^2C_{\mathrm{sol}}\lesssim\kappa h$ 时才是 $\|\operatorname{curl}(E-E_h)\|$ 的可靠估计器。因此,与单纯的超收敛后处理不同,本文明确刻画了 PPR 在高频预渐近区域的能力边界。 其次,与 Lagrange 有限元中的 PPR 理论以及 Helmholtz 方程的 PPR 分析相比,本文既有继承也有本质区别。Naga 与 Zhang 的 PPR 方法在 Lagrange 有限元中超收敛、后验误差估计和工程应用方面已非常成熟。对于高频 Helmholtz 方程,Du 等证明 PPR 能改善插值误差但污染误差会被恢复量继承,不能仅靠 PPR 消除。本文在 Maxwell 方程中发现了类似现象:PPR 改善插值部分但不降低污染项。但 Maxwell 情形不能通过 Helmholtz 论证的常规推广得到,原因在于解是向量值场,且 $\operatorname{curl}$ 有巨大的非平凡核。本文为此发展了连续与离散 Helmholtz 分解、辅助 Maxwell 椭圆问题、离散 inf-sup 条件以及散度自由分量的 $L^2$ 超收敛估计。这些工具在 Helmholtz 标量情形中没有完全对应的必要性或形式,因此本文的工作不是简单类比,而是针对 $H(\operatorname{curl})$ 结构的技术推进。 第三,与 Maxwell 方程边元超收敛的既有文献相比,本文首次强调并量化了波数显式行为。Monk 的开创性工作建立了 Maxwell 问题的超收敛结果,Huang 等对二维和三维 Nédélec 元作了系统发展,Wang 等将分析扩展到任意阶边元并在混合有限元框架下应用 PPR 得到全局恢复超收敛结果,还针对二维时谐 Maxwell 方程提出了 curl 恢复方法。然而,这些工作没有给出波数显式误差估计,也没有讨论大波数区域中污染误差对恢复过程的影响。本文限制在最低阶第一类 Nédélec 边元、均匀笛卡尔网格和矩形盒区域,但给出了含 $\kappa$、$h$ 与 $C_{\mathrm{sol}}$ 的显式估计,并区分了插值误差与污染误差的贡献。代价是网格与区域假设更强:超收敛分析要求 $\Omega$ 为矩形盒且网格为均匀笛卡尔网格,而许多已有超收敛结果可在更一般的形状正则网格或更一般的区域上讨论。适用范围也更窄:本文只处理最低阶第一类边元,而 Wang 等考虑了任意阶边元。 第四,与经典 CIP-EEM 相比,本文指出了方法适用边元类型上的重要差异。经典 CIP-EEM 使用与边元法相同的逼近空间,并在内部单元面的 curl 切向分量跳量上添加惩罚项;已有工作表明,当边元空间为第二类时,该方法能改善稳定性并减少污染。本文指出,对于第一类边元空间,仅调整经典 CIP-EEM 的惩罚参数不能显著降低相位误差。为此,作者提出在经典格式上额外加入法向分量跳量惩罚。色散分析显示,适当选择惩罚参数后,新 CIP-EEM 可以把相位误差在 $\kappa h$ 中改进两个阶。与经典 CIP-EEM 相比,本文的推进在于将污染减少策略从第二类边元扩展到第一类边元,并通过附加法向跳量惩罚克服了第一类边元中调参无效的问题。代价是格式更复杂,需要引入并优化额外惩罚参数;并且提供内容未给出完整色散关系与参数选取公式,故无法与经典 CIP-EEM 作逐项常数或误差阶的完整比较。 第五,与 DG 方法及其他波数显式边元估计相比,引言提到 DG 方法 [23,33] 和阻抗边界条件下的 EEM [35,37] 也有波数显式误差估计。本文关注的是 PEC 边界条件下最低阶第一类 Nédélec 边元的恢复与污染减少。与这些工作相比,本文的新意不在于提出新的基础离散,而在于建立高频 PPR 超收敛理论并分析其不能消除污染,然后通过修改双线性形式提出新 CIP-EEM。作者并未在提供内容中对这些方法的常数、网格条件或实际精度作全面比较,因此从提供内容中无法确认本文 CIP-EEM 与 DG 或阻抗边界 EEM 在具体误差常数上的优劣。 第六,关于函数值恢复,本文与既有 Maxwell 恢复工作的比较也有特殊之处。最低阶第一类 Nédélec 元的 $L^2$ 误差一般不能自然达到二阶收敛,因此直接对函数值作恢复并非现有标量 Lagrange 理论的直接推广。作者构造了专门的函数值恢复算子 $F_h$ 并报告了二阶超收敛。然而,提供内容没有给出该结果与 Wang 等 [48,49] 的恢复方法在假设、误差阶和适用范围上的详细比较,也没有给出 $F_h$ 的精确构造,因此从提供内容中无法确认其与已有恢复算子在函数值层面的具体差异。 最后,本文的局限也构成与已有工作比较的一部分。超收敛分析局限于均匀笛卡尔网格和矩形盒区域,要求 $\kappa^3h^2C_{\mathrm{sol}}$ 足够小;在凸多面体情形,$C_{\mathrm{sol}}$ 可能按 $\kappa^2$ 增长,因此实际网格条件可能非常苛刻。PPR 虽改善插值误差,但不能消除污染;新 CIP-EEM 虽可改进相位误差两个阶,但其完整误差分析与最优参数选择未在提供内容中展开。作者未显式比较所有前人工作,特别是在复杂几何、非笛卡尔网格、高阶边元以及不同边界条件下的表现,因此本文的适用范围应谨慎限定为最低阶第一类 Nédélec 边元、PEC 边界、均匀笛卡尔网格与大波数预渐近区域。 Quadrature by Matched Asymptotic Expansion (QBMAX): Accelerating Evaluation of Layer Potentials with Boundary Layers 作者: Chaoqi Lin, Xiaoyu Wei, Andreas Klöckner 类别: math.NA 发布日期: 2026-09-09 arXiv ID: 2609.10857v1 摘要翻译: 在存在边界层时,评估具有快速衰减核的层势在计算上变得具有挑战性,正如具有大虚波数的 Helmholtz 层势所发生的那样。我们引入匹配渐近展开求积(Quadrature by Matched Asymptotic Expansion, QBMAX),这是一种高阶方法,通过将核的渐近衰减行为直接纳入势的 Taylor 展开,扩展了展开求积(Quadrature by Expansion, QBX)框架。2D 和 3D 的数值结果均表明,对于所测试的大参数情形,相较于 QBX 最多提升四位数字,而对于较小参数或曲率更强的几何,两种方法的表现可能相近。一个符号加权操作计数模型表明,构造两个展开核的代价相当。两种 Taylor 展开具有相同的形式阶,而 QBMAX 在平直边界模型中减小了截断误差常数,并且在所报告的实验中通常表现出更少的收敛退化。 Comment: 37 pages, 21 figures, 3 tables 链接: http://arxiv.org/abs/2609.10857v1 本文研究带大参数 $k$ 的 modified Helmholtz 或 Yukawa 方程 其中 $\Omega\subset\mathbb{R}^n$ 是有界且充分光滑的区域, $n\in\{2,3\}$, $k>0$ 且通常很大. 这类方程出现在热声散射, 线性化 Poisson-Boltzmann 方程, Navier-Stokes 流, 以及热方程半隐式时间离散等场景; 在最后一种场景中 $k$ 与时间步长成反比. 本文方法还可推广到 rapidly evanescent 的 Helmholtz 方程, 即 $k\in\mathbb{C}$ 且 $\operatorname{Im}(k)\gg 1$; 正文主要围绕实 $k$ 的 Yukawa 方程展开. 边界层现象指当 $k$ 很大时, 基本解在离开边界 $O(1/k)$ 的薄层内快速衰减, 层势的边界附近数值评估因此困难. 积分方程方法利用基本解, 二维时 三维时 其中 $K_0$ 是零阶第二类修正 Bessel 函数. 解可表示为单层和双层势的线性组合 其中 $\alpha,\beta\in\mathbb{R}$, $\sigma$ 为未知密度. 积分方程的优点是把问题降维到边界, 且密度常由第二类 Fredholm 方程确定, 条件数通常良好. 但核 $G_k$ 在 $x=y$ 奇异, 且当目标点 $x$ 靠近或位于边界时, 核既近奇异又快速变化; 对大 $k$, 层势本身在边界外迅速衰减到零, 因而近场和表面评估是主要计算瓶颈. QBX 通过离面局部展开把核奇异性光滑化. 设 $x\in\partial\Omega$, 展开中心 $c_x=x+r_x\nu_x$, 展开线 $l(\tau)=c_x+\tau(x-c_x)$, $\tau\in[0,1]$. 线 QBX 用 近似. 对奇异核, 扩展核可被标准高阶求积准确积分; 求积误差典型估计为 $|\text{Quadrature Error}|\le C_{p,q,\beta}(h/(4r))^{2q}\|\sigma\|_{C^{2q,\beta}}$. 然而对 rapidly decaying Yukawa 核, 固定阶 $p$ 和固定展开半径 $r$ 的 Taylor 多项式不能表示沿展开线的指数衰减 $e^{-k|\cdot|}$. 与 $e^z$ 的 Taylor 余项类似, QBX 截断误差相对于层势在边界上的小量级可以很大; 为保持规定相对精度, 平坦边界模型要求 $r\sqrt{k^2+\omega^2}$ 保持小, 通常需减小 $r$, 这又迫使求积分辨率提高, 在三维中尤其昂贵. 已有方法中, Ockendon 和 Tew 描述了二维 thin-layer 解结构; Fryklund 等把 kernel-split 方法扩展到参数依赖的 modified PDE, 通过递归细化目标附近面板直到核被解析, 二维附加代价为 $O(\log k)$; Stein 和 Barnett 提出 Quadrature by Fundamental Solutions (QFS), 用离面等效源分布表示层势, 后续工作将其用于非齐次 modified Helmholtz 求解, 但 Nyström 离散仍需解析核, 奇异附近面板尺寸应为 $O(1/k)$. 密度插值, 特殊求积, 修改层势表示等主要处理核奇异性, 对快速变化需结合局部自适应细化. 本文要补的空白是: 在 QBX 框架内把核的渐近指数衰减直接编码进线展开, 构造目标特定的匹配渐近展开求积 QBMAX, 在不改变 QBX 总体离散框架的情况下降低截断误差常数, 提升大 $k$ 相对精度, 并给出二维无限平坦边界的显式分析以及二维和三维数值验证. 主要结果分为方法定义, 截断误差恒等式与平坦边界模型, 以及数值表现. 第一, QBMAX 展开. 对层势 $u$, 固定 $x\in\partial\Omega$, 展开中心 $c_x$, 半径 $r=|x-c_x|$, 展开线 $l(\tau)=c_x+\tau(x-c_x)$, $\tau\in[0,1]$. 定义目标特定渐近缩放函数 $u_{ma}(\tau)=\exp(k|l(\tau)-x|)$. 定义 $p$ 阶匹配渐近展开算子 $T_p^{ma}$ 为 当 $\tau=1$ 时 $u_{ma}(1)=1$, 因此直接回到对 $u(x)$ 的近似. 该展开是目标特定的. 对应的扩展核可由核函数 $K$ 替换 $u\circ l$ 得到. 第二, 截断误差. 由带积分余项的 Taylor 定理, 因此 QBMAX 与 QBX 有相同形式阶 $p+1$ 的余项, 但常数因子不同. 引理 3.2 给出 用于构造和计算新展开核. 第三, 二维无限平坦边界模型. 取源为 $\mathbb{R}\times\{0\}$, 密度 $\sigma(y)=e^{-i\omega y_1}$, $\omega\in\mathbb{R}$. 精确层势为 对表面点 $x^*=(x_1,0)$, 展开中心 $c=(x_1,r)$, $l(\tau)=(x_1,r(1-\tau))$, $|l(\tau)-x^*|=r(1-\tau)$. QBX 在 $\tau=1$ 的 $p$ 阶展开为 截断误差为 其中 $\gamma$ 是下不完全 gamma 函数, $\Gamma(p+1,0)=p!$. 除以表面量级 $|u(x^*)|=\pi/\sqrt{k^2+\omega^2}$ 后, QBX 相对误差就是正则化不完全 gamma 函数; 固定 $p$ 时, 当 $r\sqrt{k^2+\omega^2}$ 大, 相对误差趋近 1, 即使绝对误差在 $k\to\infty$ 时为 $O(\pi/k)$. QBMAX 在同模型中的误差为 特别地, 当 $\omega=0$, 即常密度, QBMAX 截断误差为 0, 展开精确. 令 $k'=kr$, $\omega'=\omega r$, 误差比 固定 $p$ 和 $\omega'$, 当 $k'\to\infty$, QBMAX 相对截断误差趋于 0, 而 QBX 相对截断误差保持 $O(1)$. 对 $p\in\{4,6,8\}$ 的绘图给出 $\log_{10}R=1$ 和 $\log_{10}R=4$ 水平集, 对应十倍和一万倍截断误差降低. 第四, 数值结果与成本. 在二维和三维实验中, 对测试的大参数情形, QBMAX 比 QBX 最多提高约四位精度; 对小参数或更强曲率几何, 两者表现相近. 在每个报告的网格加密研究中, 所有 $k$ 使用相同网格; 在此范围内 QBMAX 对 $k$ 的敏感性显著低于 QBX, 观察到的收敛率随 $k$ 增大退化更少, 但收敛率随几何变化. 符号加权操作数模型表明形成 QBX 与 QBMAX 两种扩展核的成本相当. 作者明确指出不声称对任意几何或参数范围所需分辨率在 $k$ 上一致; 一般几何的显式常数未给出. 方法还通过数值例子展示可用于 rapidly evanescent Helmholtz 方程, 并自然推广到 modified biharmonic, modified Stokes 和 Yukawa-Beltrami 等方程. 核心障碍在于 QBX 的线展开对 $(u\circ l)(\tau)$ 本身作 Taylor 展开, 而大 $k$ 的 Yukawa 层势沿展开线含有主导因子 $e^{-kr(1-\tau)}$ 或更一般 $e^{-k|l(\tau)-x|}$. 固定阶 $p$ 的多项式不能捕捉该指数衰减. 在平坦边界模型中, QBX 的相对截断误差由 $\gamma(p+1,r\sqrt{k^2+\omega^2})/p!$ 控制, 当 $r\sqrt{k^2+\omega^2}$ 变大时趋于 1, 所以为达到相对精度必须使 $r\sqrt{k^2+\omega^2}$ 小, 即 $r=O(1/k)$. 但 QBX 求积误差估计含 $(h/(4r))^{2q}$, 减小 $r$ 需要更小的面板尺寸 $h$ 或更多求积点, 在三维中成本很高. 因此直接自适应细化或特殊奇异求积只能处理奇异性, 不能消除由指数衰减带来的相对截断误差常数问题. 作者的真正创新点是匹配渐近展开: 在 QBX 框架中引入目标特定缩放函数 $u_{ma}(\tau)=e^{k|l(\tau)-x|}$, 先把层势沿展开线的主导指数衰减剥离, 对 $(u\circ l)u_{ma}$ 作 Taylor 展开, 再用 $u_{ma}^{-1}$ 恢复原势. 因为 $u_{ma}$ 对每个目标点 $x$ 单独定义, 所以新展开是目标特定的; 它不改变 QBX 的离面展开中心和标准求积积分思想, 但改变扩展核. 等价地, 由引理 3.2, 所以 QBMAX 展开系数是移位微分算子 $(\partial_\tau-kr)^m$ 对层势的作用再乘回 $e^{kr\tau}$. 这把原 Taylor 展开的普通导数替换为带参数 $kr$ 的移位导数, 使展开对残余指数率 $\sqrt{k^2+\omega^2}-k$ 而非 $\sqrt{k^2+\omega^2}$ 敏感. 在 $\omega=0$ 时常密度情形残余率为 0, 因此展开精确; 对固定 $\omega$ 且 $k\to\infty$, 残余率约为 $\omega^2/(2k)$, 趋于 0, 这解释了相对误差的大幅降低. 在分析层面, 作者选择二维无限平坦边界和 Fourier 模态密度作为可精确求解的模型. 利用 $K_0$ 的 Fourier 变换或引理 A.1 得到闭式 对 QBX 和 QBMAX 分别写出有限 Taylor 和的精确表达式, 并用下不完全 gamma 函数表示截断误差, 从而避免仅靠渐近阶而忽略常数. 引入无量纲量 $k'=kr$, $\omega'=\omega r$ 后, 误差比 $R(k',\omega')$ 只依赖这两个量和 $p$, 水平集图直接显示在 $k'$ 大时 QBMAX 相对误差减小. 这个模型虽然理想化, 但抓住边界层的主导指数衰减; 作者用数值实验说明该关系在低到中等曲率几何中仍近似成立. 在实现和成本方面, 作者使用符号加权操作数模型来比较形成 QBX 与 QBMAX 扩展核的运算次数. 由于两个展开都是 $p$ 阶 Taylor 型, 形式阶相同; QBMAX 只是把系数改为移位导数生成, 并附加以 $e^{\pm kr\tau}$ 因子. 操作数模型表明形成两种核的成本相当. 对每个目标点, 扩展核依赖 $k$, 局部法向, 展开半径 $r$ 和 $u_{ma}$, 因此可能需要符号微分或递推来生成高阶系数; 论文没有给出一般几何中常数显式界, 也未声称任意几何下分辨率与 $k$ 无关. 另一个技术点是求积误差与截断误差分离: QBMAX 仍积分一个光滑扩展核, 所以 QBX 的求积误差框架可继续使用; 匹配渐近只改变截断误差常数, 不改变奇异核正则化的总体结构. 作者通过二维和三维数值实验验证相同网格可用于所有测试 $k$, 并报告大参数下最多四位改善, 同时在小参数或强曲率时与 QBX 相当. 此外, 方法可推广到 Helmholtz 的 rapidly evanescent 区以及 modified biharmonic, modified Stokes 和 Yukawa-Beltrami 等, 因为其关键是提取核的主导指数衰减率. 与经典 QBX 比较. 原 QBX 是核独立框架, 通过离面局部展开统一处理奇异和近奇异核, 适用核, 维度和奇异性广. QBMAX 保留这一框架, 但把核的渐近衰减直接编入线展开, 因此不是新的离散方法, 而是对 QBX 扩展核的修改. 在平坦边界模型中, 两者形式阶相同, 但 QBX 相对截断误差由 $\gamma(p+1,r\sqrt{k^2+\omega^2})/p!$ 控制, QBMAX 由 $\gamma(p+1,r(\sqrt{k^2+\omega^2}-k))/p!$ 控制. 当 $k$ 大时, QBX 相对误差趋于 1, 而 QBMAX 趋于 0. 数值上, 对测试大参数, QBMAX 最多提高四位; 小参数或强曲率时无显著优势. 成本由符号操作数模型判断为形成两种扩展核相当. 代价是扩展核目标特定, 依赖 $k$ 和局部几何, 实现更复杂; 且一般几何的常数分析未给出. 与 Fryklund 等 kernel-split 方法比较. 该工作针对参数依赖 modified PDE, 包括 Yukawa, 通过递归细化目标附近面板直到核被充分解析; 二维附加计算代价 $O(\log k)$. 它适用于广泛 $k$, 但通过自适应细化处理快速变化. QBMAX 则把主导指数衰减嵌入展开, 在报告的实验中允许所有 $k$ 使用相同网格, 不依赖目标附近递归面板细化. 作者谨慎指出不声称任意几何或参数范围的所需分辨率在 $k$ 上一致; 因此 QBMAX 的适用范围和稳健性仍需更一般的分析. 在比较中, QBMAX 的优势是保持高阶 QBX 结构和通用性, 并可能同时用于振荡核; 局限是理论显式常数只在平坦边界. 与 QFS 比较. Stein 和 Barnett 的 Quadrature by Fundamental Solutions 用离面等效源分布表示给定层势, 是核独立方法, 与经典 fundamental solutions 方法精神相关, 但源用于近似原势而非直接解边值问题. 后续工作把 QFS 型近场评估并入非齐次 modified Helmholtz 求解器, 用 Nyström 方法和高阶 Alpert 求积离散边界积分算子, 通过边界配点匹配确定源分布. QFS 避免奇异求积规则并降低后续评估成本, 但初始 Nyström 离散必须解析核, 例如奇异附近面板尺寸应为 $O(1/k)$. QBMAX 不构造等效源, 而是在 QBX 扩展核内处理指数衰减; 它仍要求能准确积分正则化后的扩展核. 与 QFS 相比, QBMAX 的更直接竞争点是避免为解析核而把面板细化到 $O(1/k)$; 文中报告相同网格用于所有 $k$, 但没有给出严格的一般复杂度定理. 从提供内容中无法确认两者在实际大规模求解器中的完整墙钟时间和复杂度的定量对比. 与其他近场方法比较. 密度插值, 特殊求积规则和修改层势表示主要针对核奇异性; 要处理大 $k$ 的快速核变化, 通常需要结合局部自适应细化. QBMAX 的匹配渐近思想不同, 它只提取主导指数衰减, 与这些方法可能互补. 与 singularity removal 和函数近似中的 modified Taylor expansions 相比, 作者指出这些思想已出现于其他 context, 但据他们所知, 将其用于 rapidly decaying kernels 和积分方程是新的. 因此本文的推进主要是把这种修改 Taylor 展开系统化到 QBX 近场评估. 适用范围与时间尺度. 该问题的时间尺度由边界层厚度 $O(1/k)$ 决定; QBX 为保持相对截断精度需 $r\sqrt{k^2+\omega^2}$ 小, 相当于 $r=O(1/k)$ 或更小, 从而增加求积成本. QBMAX 展开的是残余率 $\sqrt{k^2+\omega^2}-k$, 对固定 $\omega$ 当 $k\to\infty$ 时它趋于 0, 对常密度精确. 因此 QBMAX 把有效展开参数从 $O(k)$ 降为 $O(\omega^2/k)$ 或更小, 在相同 $r$ 下降低截断误差. 这并不意味着 QBMAX 对所有参数和几何都免分辨率; 强曲率会引入几何导数的高阶项, 可能抵消收益, 数值实验中强曲率时两者表现相近. 作者没有显式展开与某些前人工作的定量比较, 也没有给出一般三维理论误差界; 从提供内容中无法确认这些方面. 局限和代价. QBMAX 需要构造带 $e^{\pm kr\tau}$ 因子的移位导数扩展核, 对高阶 $p$ 可能需要符号计算或递推; 虽然操作数模型表明与 QBX 相当, 但该模型是符号加权计数, 并非实际实现性能. 分析限于二维无限平坦边界和 Fourier 模态, 说明常数减少的机制, 但一般几何, 三维和多个相互作用边界下的严格常数仍开放. 另外, 方法主要针对 modified Helmholtz/Yukawa 大 $k$ 边界层, 也报告了 Helmholtz rapidly evanescent 例子, 并称可推广到 modified biharmonic, modified Stokes 和 Yukawa-Beltrami; 这些推广的详细理论和实现未在提供内容中展开. 因此与已有工作的比较应理解为: 在 QBX 框架内针对快速指数衰减核的相对截断误差常数进行了推进, 而不是声称全面优于所有近场评估或自适应细化方法. Convergence of finite element approximations for the one-dimensional stochastic Burgers equation with additive trace-class noise 作者: Binjie Li, Xiaoping Xie, Qin Zhou 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11584v1 链接: http://arxiv.org/abs/2609.11584v1 中文标题: 具有加性迹类噪声的一维随机 Burgers 方程有限元逼近的收敛性 摘要翻译: 本文研究具有加性迹类噪声的一维粘性随机 Burgers 方程的有限元逼近。对于 (P_2) 有限元空间半离散,我们推导出在 (L^p([0,T] \times Ω;H_{D}^{α,q}))((p,q\in[2,\infty)),(α\in[-1,0]))中关于正则性最优的强误差估计,以及 (L^p(Ω;C([0,T];L^\infty(\mathcal{O})))) 中的一个几乎正则性最优估计。此外,我们推导出终端 $L^q$-范数与时空 $L^p(0,T;L^q(\mathcal{O}))$-范数这两类量的矩的弱误差估计,其弱收敛速率(几乎)是对应强收敛速率的两倍。对于全离散格式,该格式在空间上结合了 (P_2) 有限元方法,在时间上结合了漂移隐式 Euler–Maruyama 格式,我们在 (L^p(Ω;C([0,T];L^\infty(\mathcal{O})))) 的离散类似空间中,在条件 (τ\leqslant h^2) 下,建立了阶为 (τ^{1/2-\varepsilon}) 的强时间收敛速率。数值实验用于展示理论收敛速率。 Approximate Inversion of Discrete Fourier Integral Operators via Hierarchically Semiseparable Matrices 作者: Yingzhou Li, Jingyu Liu 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11520v1 链接: http://arxiv.org/abs/2609.11520v1 中文标题: 基于层次半可分矩阵的离散傅里叶积分算子近似求逆 摘要翻译: 本文提出了一种用于近似离散傅里叶积分算子(FIOs)的逆的新方法。给定 FIO 的一个 $N \times N$ 矩阵表示 $K$,所提出的算法由两个阶段组成。在离线阶段,我们首先构造 $K$ 的蝶形分解(BF)$\tilde{K}$,它能够实现快速正向矩阵-向量乘法。然后,我们利用 $\tilde{K}$ 和 $\tilde{K}^{*}$ 对随机矩阵的快速作用,构造层次半可分(HSS)近似 $\tilde{G} \approx G$,其中 $G = K^{*} K$。最后,我们对 HSS 矩阵 $\tilde{G}$ 应用 ULV 分解,得到近似 $\tilde{F} \approx G^{-1}$。组合这些近似得到近似逆 $K^{-1} \approx \tilde{F} \tilde{K}^{*}$。离线阶段对于一维问题具有复杂度 $O(N \log^{2} N)$,对于二维问题具有复杂度 $O(N^{1.5} \log N)$。在线阶段,所提出的方法对给定输入向量 $u$ 近似 $K^{-1} u$,对于一维和二维问题复杂度均为 $O(N \log N)$。所提出的方法既可以用作直接求解器,也可以用作迭代方法的预条件子。一维和二维 FIO 的数值结果证明了所提出方法的有效性。 A variational physics-informed graph neural network for heterogeneous solid mechanics 作者: Aashay Rajan Yadav, Amiya Prakash Das, Ratna Kumar Annabattula 类别: math.NA, cs.LG 发布日期: 2026-09-10 arXiv ID: 2609.10983v1 链接: http://arxiv.org/abs/2609.10983v1 中文标题: 一种面向非均质固体力学的变分物理信息图神经网络 摘要翻译: 非均质固体中的应力局部化由双材料界面控制,其中位移场保持 $C^0$-连续,而面内应力因刚度失配发生跳跃。基于坐标的物理信息神经网络(PINNs)通过预设正则化宽度或加权界面罚项来表示这一跳跃,使其精度对相间对比度变化的处理方式敏感。本文提出一种变分、无标签的物理信息图神经网络(PI-GNN),其中非均质性由离散化而非试探场承载。该求解器在协调自适应网格图上运行,为每个单元赋予本构行为,并将离散总势能作为单个无权重目标进行最小化,其中仅出现一阶导数。分段线性单元上的离散能量与有限元(FE)Ritz 泛函一致。Dirichlet 条件通过构造强制满足,无需罚项、无需界面权重、无需预设过渡宽度。在二维和三维的小应变弹性与有限应变 Neo-Hookean 超弹性中,使用同一固定架构、优化器和损失函数,von Mises 误差在跨越 $(E_{\mathrm{inc}}/E_{\mathrm{mat}}\in[10^{-2},10^{2}])$ 的刚度对比扫描中保持在 $3.58\%$ 以下;而强形式 PINN 退化至 $5.58\%$,其位移误差达到 $7.66\%$,而 PI-GNN 为 $0.49\%$。训练后的网络将基于能量的 PINN 的 ($σ_{xx}$) 误差减半($5.01\%$ 对 $10.94\%$)。训练成本超过单次 FE 求解一个数量级以上,因此该构造是一种变分一致、无罚项的界面表示,适用于参数化代理模型与逆向识别,而非替代一次性 FE 分析。 A virtual element method for Kelvin–Voigt viscoelasticity 作者: Utkarsh Rajput, Ankit Kumar, Andres E. Rubiano, Ricardo Ruiz-Baier, Sarvesh Kumar 类别: math.NA 发布日期: 2026-09-09 arXiv ID: 2609.10908v1 链接: http://arxiv.org/abs/2609.10908v1 中文标题: 用于 Kelvin–Voigt 粘弹性的虚拟元方法 摘要翻译: 考虑到虚拟元方法(VEM)的计算优势,本文采用协调虚拟元方法对粘弹性的 Kelvin–Voigt 模型进行数值近似。为清晰和简洁起见,我们关注原始形式。空间离散采用虚拟元方法进行,而时间离散则通过 {second-order Crank–Nicolson} 格式处理。我们建立了半离散和全离散问题的适定性,并推导了 {\it a priori} 误差估计。给出了若干具有代表性的数值算例,以验证理论结果并说明所提出形式的有效性。 Structure-preserving operator splitting for 2.5D ideal MHD with an entropy-stable DGSEM and exactly divergence-free compatible finite elements 作者: Siyuan Fan, Guosheng Fu 类别: math.NA 发布日期: 2026-09-09 arXiv ID: 2609.10849v1 链接: http://arxiv.org/abs/2609.10849v1 中文标题: 使用熵稳定 DGSEM 和精确无散度相容有限元的 2.5D 理想 MHD 保结构算子分裂 摘要翻译: 我们针对笛卡尔网格上的 2.5D 理想磁流体力学,发展了一种高阶、全显式算子分裂方法。流体子流由配备逐阶段振荡消除和保正限制器的熵稳定间断 Galerkin 谱元方法推进。磁场–速度子流使用相容有限元以及电场和电流密度的质量集中重构。其离散旋度更新精确保持全局 $H(\mathrm{div})$ 无散度子空间,而理想半离散化则平衡动能、磁能和内能。旋度型人工电阻和方向分辨的速度滤波将被移除的磁能和动能返还给内能;因此,稳定化后的磁场阶段保持正内能,并满足离散熵不等式。两个求解器通过二阶流体–磁–流体 Strang 分裂组合,得到一个无矩阵格式,该格式在接受步上保持全局质量、节点可容许性和磁场散度。光滑 Alfvén 波和平流涡测试揭示了磁场多项式次数中的奇偶收敛模式,验证了二阶时间精度,并表明稳定化保持高阶精度。磁场环、Orszag–Tang、转子、MHD 爆炸波和 Kelvin–Helmholtz 计算展示了对非光滑多维流动的稳健性能,并表明人工电阻抑制网格尺度磁振荡,同时保留已解析结构。 Tolerance-driven close evaluation of the Stokes double layer potential on axisymmetric surfaces 作者: David Krantz, Pritpal Matharu, Anna-Karin Tornberg 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11778v1 链接: http://arxiv.org/abs/2609.11778v1 摘要: We consider boundary integral methods for Stokes mobility and resistance problems involving smooth axisymmetric particles. A primary numerical challenge is the accurate and efficient evaluation of layer potentials at off-surface points close to particle surfaces. We present a tolerance-driven workflow for evaluating the Stokes double layer potential at such evaluation points (targets) to prescribed accuracy while avoiding unnecessary computational cost. For each target–particle interaction, a fast classifier selects the least costly option estimated to meet the tolerance among standard, upsampled, and special quadrature. Geometry-dependent unit-density error indicators are precomputed and tabulated in reduced cylindrical coordinates, then combined on the fly with a local layer-density modifier, making its cost negligible relative to evaluating the potential. Targets requiring special quadrature are treated using a stabilized version of singularity swap surface quadrature: the periodic azimuthal integral is evaluated first using translated singularity swap quadrature to prevent severe cancellation near the surface, followed by adaptive Gauss–Legendre quadrature in the polar direction guided by error indicators. We integrate this workflow into a boundary integral solver with precomputed quadrature by expansion for on-surface self-interactions and demonstrate the workflow’s performance for challenging configurations of spheroidal particles. Numerical results show that target classification is highly accurate. The prescribed tolerance is met for nearly all target–particle interactions and the error remains within a modest factor of the tolerance in the few remaining cases. Although the experiments focus on the Stokes double layer potential for spheroids, the off-surface framework applies to general smooth axisymmetric surfaces and can be easily adapted to other Stokes layer potentials. Some theoretical and practical results on noisy signals recovery 作者: M. Makurin, Yu. Malykhin, K. Ryutin, V. Temlyakov 类别: math.NA, math.FA 发布日期: 2026-09-10 arXiv ID: 2609.11765v1 链接: http://arxiv.org/abs/2609.11765v1 摘要: We discuss some theoretical results and their applications to specific practical problems from wireless communications. We assume that we know the noisy version of the signal, which is sparse with respect to a given system of elements (dictionary), at a finite number of points and we want to approximately recover it. This problem of recovery of a noisy signal is closely related to the problem of establishing the Lebesgue-type inequalities for the corresponding algorithms and it motivates us to prove such inequalities. Under certain conditions on a dictionary (RIP-type condition, coherence condition) we obtain different kinds of the Lebesgue-type inequalities for the OMP and its version WOMP algorithms. The most important feature of our new theoretical result is the assumption that the dictionary has the RIP-type property instead of the assumption that it is the Riesz basis, which was used in the previous results. Our approach allows us to treat redundant (overcomplete) systems, which is important in applications.
We consider the setting of sparse recovery for highly-coherent dictionaries that appear in OFDM setting for wireless communication. Our main example is the oversampled Fourier dictionary and the recovery of the frequency response of a sparse channel. We develop a general approach to such problems and we prove that under certain conditions the OMP algorithm recovers all dictionary elements with large coefficients, and estimate its accuracy (in NMSE metric) in terms of signal-to-noise ratio. Verified Linear Programming through Tolerance-Aware Precision Boosting 作者: Ernesto Casablanca, Martin Sidaway, Paolo Zuliani, Sadegh Soudjani 类别: math.OC, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11721v1 链接: http://arxiv.org/abs/2609.11721v1 摘要: Linear programming plays a fundamental role in computer science, with applications in optimization, formal verification, SMT solving, and numerous other domains. When exactness guarantees are required, numerical inaccuracies arising from floating-point arithmetic can compromise soundness, whereas exact rational arithmetic often incurs significant computational overhead. In this paper, we investigate the numerical stability of the simplex algorithm and establish conditions under which a precision-boosting floating-point implementation provably produces the same pivot decisions and final basis as an exact rational implementation, from which the final result is reconstructed and certified exactly. Our analysis shows that correctness depends on both arithmetic precision and a careful handling of numerical tolerances. Based on these results, we develop a tolerance-aware, precision-boosting simplex algorithm with formal correctness guarantees. Finally, we introduce a delta-complete termination criterion that allows the algorithm to terminate once certified upper and lower bounds on the optimal objective differ by at most a user-specified threshold delta, providing a certified, user-controlled optimality gap. $hp$-adaptive trees for graph signal approximation 作者: Giacomo Elefante, Wolfgang Erb, Michael Multerer 类别: eess.SP, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11701v1 链接: http://arxiv.org/abs/2609.11701v1 摘要: Tree-encoded partitionings of graphs are fundamental tools for the decomposition and approximation of graph signals. For the efficient approximation of such graph signals, we develop strategies based on $hp$-refinement by combining domain decomposition with an improved local approximation using polynomials of higher degree. In this way, from a given graph partitioning tree, a more efficient subtree is extracted in which the cost of the signal approximation is considerably reduced by still maintaining the same total error. To this end, we interpret the refinement process as a binary knapsack problem to determine an enhanced partitioning tree. We further study an a-posteriori strategy which prunes the partitioning tree by optimizing the polynomial degrees over the subdomains. To make polynomial basis systems accessible for general graphs or high-dimensional data, we propose local embeddings of graphs into low dimensional Euclidean spaces. We underpin the efficiency of our algorithms with extensive numerical tests which carefully assess the impact of the applied refinements and optimization strategies. Numerical Study of Reported Multiple Solutions to a Cauchy Problem for the 2d Incompressible Euler Equations 作者: Allen Tesdall, Richard Sanders 类别: math.AP, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11694v1 链接: http://arxiv.org/abs/2609.11694v1 摘要: Bressan and Shen (2021) jointly with Bressan and Murray (2020) proposed a simple example of a Cauchy problem on $\mathbb{R}^2$, characterized by two key parameters, intended to yield nonunique weak solutions of the incompressible Euler equations. They suggest that a limiting process applied to two specific approximations of their example initial data yields two distinct solutions. Here we provide evidence coming from a carefully designed numerical procedure to support if, or when, this intriguing claim is valid. Some of the numerical challenges encountered in the present study are: (i) the proposed initial data is unbounded in a neighborhood of the origin, (ii) the proposed problem is posed on all of $\mathbb{R}^2$, (iii) several essential symmetries present in the true solution must be captured, (iv) the proposed example depends on two parameters which may be critically linked. We demonstrate that an extremely large computational domain is mathematically necessary to capture the core symmetries of the exact free-space problem. Accordingly we implement a nested-grid strategy specifically designed for its particular suitability and efficacy for the problem at hand. Our strategy also allows us to achieve extreme grid refinement in the region where solution accuracy is most critical. Within our nested-grid framework, we design an underlying discretization scheme that is high-order, robust, and fast.
We consider a broad collection of allowable parameter pairs, but among all of them we find only one that may satisfy the suggested claim. Quasilinear multiplication in the real Cayley–Dickson tower 作者: Harrison Lemley 类别: math.RA, cs.SC, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11588v1 链接: http://arxiv.org/abs/2609.11588v1 摘要: Direct evaluation of the defining product in the real Cayley–Dickson algebra $A_n$, of dimension $N=2^n$, has quadratic arithmetic complexity. This paper gives a uniform algorithm for multiplication using $O(N\log N)$ real arithmetic operations and $O(N)$ auxiliary storage. The algorithm reduces multiplication to the alternating product on the imaginary subspace, then evaluates that product by a two-call recursion over one fixed quadratic coefficient extension. For $n\ge1$, the resulting bilinear algorithm uses at most $(9n-15)2^{n-1}+10$ input-dependent real multiplications, and for $n\ge3$, the specified arithmetic schedule uses $(34n-83)2^{n-1}+50$ real operations in total. Under this counting convention, the quasilinear schedule uses fewer operations than direct multiplication for $N\ge16$ and than the uniform Cariow–Cariowa method for $N\ge32$. The algorithm is implemented in the MIT-licensed C11 library fastCD, with a NumPy-backed Python interface, and its results are checked against an independent implementation of the defining recursion. In single-core benchmarks against direct multiplication and the uniform Cariow–Cariowa method, the quasilinear implementation had the lowest mean time of the three at every tested dimension $N\ge32$, for both single and batched products, and was roughly $16$ times faster than direct multiplication at $N=1024$. Mean Square Error Analysis of Stochastic Runge–Kutta Integrators 作者: Xuda Ye 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11528v1 链接: http://arxiv.org/abs/2609.11528v1 摘要: We analyze the mean square error of stochastic Runge–Kutta integrators for overdamped Langevin dynamics whose potential is convex outside a bounded region. A decomposition splits the local error into a mean-zero term and a smaller remainder, and the discrete Poisson equation turns their moments into a bound on the error of a time average. We carry this out for two stochastic Runge–Kutta integrators proposed by Yang and Wang, both of strong order $\frac32$ and weak order 2, evaluating the gradient of the potential and no higher derivative. We show that their laws approach the law of the exact solution at second order in Wasserstein-1 distance, up to a logarithm, uniformly in the number of steps. For a test function with bounded derivatives up to third order, we prove that the mean square error over $N$ steps with step size $h$ is $\mathcal O ( \frac{1}{Nh} + h^4 )$, which is the optimal order in the discretization. Experiments measure the strong and weak orders and the sampling bias on a nonconvex potential in $\mathbb R^2$, and compare the integrators on a diffusion model of CIFAR-10. Space-time finite element interpolation of nonsmooth functions satisfying boundary conditions 作者: Thomas Führer, Gregor Gantner, Roberto González, Michael Karkulik 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11496v1 链接: http://arxiv.org/abs/2609.11496v1 摘要: We construct Scott-Zhang-type space-time quasi-interpolation operators on tensor- product meshes for parabolic settings. Specifically, these are linear and uniformly bounded projections from the classical parabolic energy space onto a space-time tensor-product finite element space that preserve discrete Dirichlet boundary conditions on the lateral space-time boundary. We apply our operator in the context of space-time first-order system least-squares finite elements for parabolic equations to establish a quasi-optimal method for inhomogeneous Dirichlet boundary conditions. PH2T-splines, Part I: A Reasonable Mesh Assumption 作者: Bingru Huang, Yue Xi 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11443v1 链接: http://arxiv.org/abs/2609.11443v1 摘要: This paper is the first in a three-part series on the construction of polynomial splines with the highest order of smoothness over hierarchical T-meshes, referred to as $\PHtwoT$-splines. For splines of bi-degree $(d,d)$, we study suitable refinement conditions for the subsequent basis construction, which requires dimensional stability of the underlying spline space. We present two groups of examples, considering unrestricted hierarchical refinement and refinement without vanishable T $l$-edges, respectively. The first setting permits new edges without additional degrees of freedom. In the second group, every refinement level excludes vanishable T $l$-edges and increases the dimension. Nevertheless, the dimension is unstable in both groups. We then introduce template translations to describe each refined region as a union of translates of a fixed template in the cell-index grid. Together with the known stability result under $(d-1)\times(d-1)$ template refinement, these examples justify this condition as a reasonable mesh assumption for the subsequent $\PHtwoT$-spline construction. Control Strategies for Multi-Species Wasserstein Gradient Flows 作者: Dante Kalise, Lucas M. Moschen, Grigorios A. Pavliotis 类别: math.OC, math.AP, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11422v1 链接: http://arxiv.org/abs/2609.11422v1 摘要: We stabilize stationary states of coupled multi-species McKean–Vlasov equations by feedback control, extending a spectral approach based on the Wasserstein Hessian and a Riccati equation from a single density to a product of Wasserstein spaces. We give necessary and sufficient conditions for local exponential stabilization at a prescribed rate when only one species is actuated, together with the minimum number of control inputs required. For two species, the condition reduces to a rank condition on the Fourier coefficients of the cross-interaction kernel over the eigenspaces of the unactuated species’ linearized dynamics. A two-community noisy Kuramoto model and a cell-sorting model illustrate the results. Deep operator learning for efficient sampling from invariant measures of stochastic differential equations 作者: Lin Guo, Li Lei, Jingtong Zhang 类别: math.NA, cs.LG 发布日期: 2026-09-10 arXiv ID: 2609.11376v1 链接: http://arxiv.org/abs/2609.11376v1 摘要: We introduce an amortized neural sampler that combines operator learning with flow methods for sampling. It maps SDE coefficient functions to pushforwards from a reference measure to the invariant measures, enabling efficient sampling across families of stochastic differential equations. Our framework shifts traditional sampling cost to an initial training phase, after which new SDE instances require only one encoder pass and a few ODE solver steps, independent of mixing time. To handle problems in high dimensions, we use Lagrangian trajectory sensors for the coefficient functions and cross attention in the architecture. We also theoretically establish the expressivity and resolution invariance of our framework. Experiments on 1D and 2D SDE families show competitive accuracy with substantial speedups over MCMC in regimes with slow mixing, transfer across sensor counts, and demonstration results on a 64D interacting particle SDE where traditional grid approaches are infeasible. The tubal Arnoldi method for tensor function approximation 作者: Fatima Bouyghf, Mohamed El Ghomari, Alaa El Ichi 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11369v1 链接: http://arxiv.org/abs/2609.11369v1 摘要: This paper describes Krylov subspace methods based on the tensor t-product for approximating quantities associated with third-order tensor functions. Specifically, it introduces the tensor-tubal Arnoldi method, which projects highdimensional tensor problems onto lower-dimensional tensor Krylov subspaces. The derived method exploits novel algebraic properties of the t-product to address large-scale tensor computations efficiently. Applications include the evaluation of parameter-dependent tensor functions and the solution of multidimensional ordinary differential equations. Numerical experiments illustrate the effectiveness and accuracy of the proposed methods compared with other approaches. Scalable Approximate Selected Inversion Based on Single-Level Incomplete $LDL^T$ Factorization and Spectral Corrections for Large Sparse Systems 作者: Tahamina Akter, Matthias Bollhöfer, Olaf Schenk 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11363v1 链接: http://arxiv.org/abs/2609.11363v1 摘要: This article introduces four parallel numerical techniques for computing entries of the inverse of large sparse symmetric systems, all grounded in incomplete $LDL^T$ (ILDL) factorizations: (1) the selected inversion method (SelInv), which applies the $LDL^T$ factorization to recover entries of the matrix inverse within the sparsity pattern of the computed factors; (2) an approximate inversion method based on a truncated Neumann series expansion applied to the inverse of the L factor (NInv), providing an alternative at the cost of reduced accuracy; (3) a Mix approximation that merges the best of both SelInv and NInv; and (4) Mix-SPAI, which applies sparse approximate inverse (SPAI) refinement on the output of the Mix method to improve entry-level accuracy. To further improve accuracy while maintaining a stable sparsity pattern, we additionally employ a low-rank correction based on eigenvector updates, providing an alternative to tightening the drop tolerance. We report the performance of the proposed numerical techniques on a comprehensive collection of sparse matrices from scientific and industrial applications. Time-optimal elevator control with higher-order state constraints: analysis and computation of boundary contacts 作者: Dmitry Karamzin, Aleksandra Zhukova, Roman Chertovskih, Pedro A. Aguiar 类别: math.OC, math.DS, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11348v1 链接: http://arxiv.org/abs/2609.11348v1 摘要: In this paper, we consider the classical time-optimal elevator problem in the presence of higher-order state constraints. The main contribution is a constructive indirect solution framework based on a Pontryagin Maximum Principle specifically formulated for higher-order constrained systems. First, we derive specialized optimality conditions and analyze the resulting structure of extremal trajectories. Second, we show that the infinite-dimensional optimal control problem can be transformed into a finite-dimensional system of nonlinear algebraic equations involving switching times, boundary-contact times, and multiplier parameters. This provides a computationally efficient procedure for calculating candidate optimal trajectories using standard nonlinear equation solvers. Third, we characterize the geometry of boundary contacts and explain the emergence of contact-chattering phenomena in higher-order constrained systems. In particular, we show that normal extremals cannot evolve along the constraint boundary for a positive amount of time, although sequences of boundary contacts may accumulate indefinitely. The proposed approach is demonstrated on a fourth-order elevator model, and the computed solutions are independently verified using a direct IPOPT-based optimization method. Minimax-Optimal Early Stopping for Continuous-Time SGD via the Discrepancy Principle 作者: Tim Jahn, Loucas Pillaud-Vivien, Adrien Schertzer 类别: math.NA, math.OC, math.ST 发布日期: 2026-09-10 arXiv ID: 2609.11273v1 链接: http://arxiv.org/abs/2609.11273v1 摘要: We study early stopping for a continuous-time model of stochastic gradient descent (SGD) in ill-posed linear inverse problems. We consider an a posteriori stopping rule based on the discrepancy principle, which stops the dynamics once the residual reaches the noise level. Unlike deterministic gradient flow, the stochastic dynamics exhibits persistent multiplicative fluctuations whose amplitude scales with the residual itself.
Under a source condition on the initial error and a polynomial bound on the spectrum of the empirical covariance operator, we prove that, with high probability, the error at this random stopping time achieves the minimax rate over the corresponding source class, up to a logarithmic factor. Our results therefore establish the discrepancy principle as an adaptive regularization strategy for continuous-time SGD that is minimax-optimal up to a logarithmic factor, despite the persistent fluctuations induced by stochastic sampling. To our knowledge, this is the first convergence-rate result for a continuous-time model of stochastic gradient descent in inverse problems under an a posteriori stopping rule: existing analyses, including those of variance-reduced variants, bound the stopping time but do not quantify the error attained there. A note on the convergence analysis of Laguerre approximations for analytic functions 作者: Thomas Caussade, David P. Hewett 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11259v1 链接: http://arxiv.org/abs/2609.11259v1 摘要: In a recent paper (H. Wang, Math. Comp. 93, 2861-2884, 2024), Wang has presented a number of results concerning the convergence of approximations based on generalised Laguerre polynomials, claiming to have provided “the first rigorous proof of root-exponential convergence of Laguerre approximations for analytic functions”. In this note we argue that the proofs of the main results of Wang’s paper are incomplete, because they rely on taking a limit under an integral sign, and this is not properly justified. We explain how the proofs could be completed using the dominated convergence theorem, provided that certain nontrivial inequalities involving generalised Laguerre polynomials and their weighted Cauchy transforms hold. We make a related conjecture, which, if true, would complete the proofs of Wang’s results. This conjecture remains unproven, but appears plausible, based on numerical experiments. A Geometrically Parameterized Quasi-Stationary 3D Model for High-Frequency Induction Tube Welding 作者: Dmitry Ivanov, John Inge Asperheim, Bjørnar Grande, Prerana Das 类别: physics.comp-ph, cs.CE, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11250v1 链接: http://arxiv.org/abs/2609.11250v1 摘要: A three-dimensional multiphysics finite element framework for the simulation of high-frequency induction welding of tubes is presented. The model couples a time-harmonic magnetic scalar potential formulation with a stabilized quasi-stationary advection-diffusion heat transport equation, enabling accurate prediction of electromagnetic and thermal fields under industrial operating conditions. The framework incorporates parameterized geometry generation and semi-automated, physics-tailored mesh construction and is implemented using the open-source tools GetDP and Gmsh. Validation against measurements from a commercial induction welding line for AISI 304 stainless steel tubes demonstrates good agreement with operating data. The validated model is subsequently applied to investigate the influence of impeder material by comparing a conventional FeNiZnV ferrite with the soft magnetic composite Ferrotron 559H for the induction welding of AISI 304 stainless steel tubes. Sparse Approximation via Polynomial Equations 作者: Matija Tomić, Raphaël Widdershoven, Lieven De Lathauwer 类别: eess.SP, math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11215v1 链接: http://arxiv.org/abs/2609.11215v1 摘要: We consider the problem of finding sparse solutions of an underdetermined linear system $Ax=b$. In contrast to conventional approaches based on greedy algorithms or convex relaxation, we reformulate sparse approximation as a structured system of polynomial equations and connect with the literature on tensor methods. We develop an eigenvalue decomposition based method that formally guarantees recovery of all sparse solutions if there is more than one. We also develop two optimization-based methods achieving favorable computational complexity. The new methods allow explicit control of the target sparsity. Numerical experiments illustrate the performance and compare to basis pursuit (denoising) and orthogonal matching pursuit. Mixed Hybrid High-Order Methods for Sixth-Order Problems: Fully Non-conforming and $C^0$-Conforming Discretizations 作者: Arbaz Khan, Ajay Kumar, Harpal Singh 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11167v1 链接: http://arxiv.org/abs/2609.11167v1 摘要: We consider a class of sixth-order elliptic partial differential equations in two and three dimensions subject to simply supported and Cahn–Hilliard-type boundary conditions. Based on the Ciarlet–Raviart reformulation, we recast the sixth-order problem as an equivalent mixed system involving second- and fourth-order equations and establish its well-posedness under suitable assumptions. For the resulting mixed formulation, we propose and analyse two discretization frameworks: a fully non-conforming hybrid high-order (HHO) method on general polytopal meshes and a $C^0$-conforming HHO–finite element method on simplicial meshes. We prove stability and derive optimal-order a priori error estimates for the primary variables under appropriate regularity assumptions. For the fully non-conforming HHO method, we further derive a reliable residual-based a posteriori error estimator. Numerical experiments confirm the theoretical convergence rates and illustrate the robustness of the proposed methods for a range of mobility parameters. Dirichlet-Neumann Waveform Relaxation Method for Hyperbolic PDE with Time Delay in Multiple Subdomains 作者: Bankim Chandra Mandal, Deeksha Tomer 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11120v1 链接: http://arxiv.org/abs/2609.11120v1 摘要: Hyperbolic partial differential equations (PDEs) with time delay are essential mathematical tools used to model numerous physical systems where wave propagation or oscillations depend heavily on their historical states. As these applications scale in physical complexity, efficient parallel computing techniques become essential; however, developing highly scalable parallel solvers for delayed PDEs remains a significant computational challenge. To address this gap, this study extends the Dirichlet-Neumann Waveform Relaxation (DNWR) method to a multi-subdomain framework explicitly designed for solving time-delayed hyperbolic PDEs in parallel. We advance the underlying mathematical framework and provide a rigorous convergence analysis for both one-dimensional and two-dimensional spatial configurations. A central theoretical contribution of this work is the establishment of finite-step convergence for the multi-domain DNWR algorithm. These theoretical guarantees are firmly corroborated through comprehensive numerical experiments. Furthermore, a detailed comparative analysis against alternative domain decomposition techniques namely NNWR, OSWR, and Classical SWR, highlight the advantages of the proposed method. Finally, this approach provides a robust, highly parallelizable solver capable of efficiently handling complex hyperbolic PDEs with time delay. Neumann-Neumann Waveform Relaxation Method for Hyperbolic PDE with Time Delay 作者: Deeksha Tomer 类别: math.NA, math.AP 发布日期: 2026-09-10 arXiv ID: 2609.11098v1 链接: http://arxiv.org/abs/2609.11098v1 摘要: Hyperbolic PDEs with time delay are essential for modeling a wide range of real-world applications. Since closed-form solutions are often not attainable, developing accurate and efficient numerical methods becomes crucial. In this work a novel substructuring waveform relaxation method namely Neumann-Neumann Waveform Relaxation (NNWR) has been implemented for numerically solving hyperbolic delay PDEs in the case of non-uniform subdomain partitioning. The convergence behavior is first analyzed using Fourier analysis to obtain an estimate, followed by a demonstration of finite-step convergence through Laplace transform techniques. Various test cases are presented as numerical illustrations. Additive reduced basis preconditioners for large-scale parametrized PDEs 作者: Xinnan Fan, Qixiao Hu, Shiquan Zhang 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11097v1 链接: http://arxiv.org/abs/2609.11097v1 摘要: We introduce a class of additive reduced basis preconditioners designed to accelerate the iterative solution of large-scale linear systems arising from discretized parametrized PDEs. The main idea is to regularize the inherently singular reduced-order approximation by adding simple correction terms: either a scaled identity correction or a projected correction based on a basic preconditioner. This yields nonsingular preconditioners under explicit and easily checked conditions. The construction and application of the preconditioners are integrated into an FGMRES framework through an offline strategy that dynamically builds the reduced-basis component by proper orthogonal decomposition at each FGMRES step. We establish sufficient conditions for the nonsingularity of the preconditioners and derive error bounds for the preconditioned Richardson iteration. Numerical results for convection-diffusion, anisotropic vortex, Stokes, and Helmholtz problems are provided to verify the efficiency and convergence of the proposed ARB preconditioners. The method consistently converges in a few iterations and substantially reduces online solve time, supporting its efficiency for multi-query engineering scenarios. Adaptive Diagonally Implicit Runge-Kutta Methods Devoid of Order Reduction for Semilinear ODEs 作者: Steven B. Roberts, Abhijit Biswas, David Shirokoff, Benjamin Seibold 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11069v1 链接: http://arxiv.org/abs/2609.11069v1 摘要: Diagonally implicit Runge-Kutta (DIRK) methods are a prominent class of numerical methods for solving stiff systems of ordinary differential equations (ODEs). Stiffness does not only impose stability challenges on Runge-Kutta methods; it can also degrade the order of convergence. This so-called order reduction phenomenon occurs when assumptions used for classical convergence analysis, e.g., an asymptotically small step size, fail to hold. In a prior paper by the authors, sharp order conditions and global error bounds for Runge-Kutta methods were developed, which hold uniformly with respect to stiffness when applied to a wide class of semilinear ODEs. In this work, those conditions are leveraged to construct the first DIRK methods of order four and five which satisfy these conditions and thus do not exhibit order reduction. Numerical results demonstrate that for a broad class of relevant nonlinear test problems, these new methods successfully mitigate order reduction, accurately estimate local error via an embedding for adaptive step size control, and can outperform classical DIRK methods. Computational Trade-Offs Between Newton and Picard Solvers for Mean Field Game PDE Systems 作者: Mathieu Lauriere, Andrew Shi 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11050v1 链接: http://arxiv.org/abs/2609.11050v1 摘要: We study computational trade-offs between two solvers for the same semi-implicit finite-difference discretization of forward-backward partial differential equation (PDE) systems arising in mean field games (MFGs). The Picard method uses an outer fixed-point iteration that alternates a forward Fokker-Planck solve and a backward Hamilton-Jacobi-Bellman solve. The Newton method instead applies Newton’s method directly to the coupled nonlinear space-time system. Across one- and two-dimensional MFG benchmarks, we find that the Picard method has much lower wall-clock cost when it converges, but may fail at sufficiently low viscosity and may require strong damping under temporal shocks. With parameter continuation, the Newton method is more robust in these regimes, at the cost of larger coupled linear systems. We relate these trade-offs to the residuals, Jacobian blocks, and sparsity structures produced by separable, local nonseparable, and nonlocal Hamiltonians. We also test a hybrid continuation strategy on a two-dimensional double-well MFG benchmark, using Picard iterations for the inexpensive part of the viscosity descent before switching to Newton continuation. A Second-Order Method for Locating Critical Points of Prescribed Index on Energy Landscapes 作者: Qiang Du, Baoming Shi 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.11038v1 链接: http://arxiv.org/abs/2609.11038v1 摘要: Exploring high dimensional energy landscape is a challenging problem in many applications where both ground states and transition states offer important information about the underlying physical systems. To facilitate the computational exploration, we propose a cubic-regularized second-order method for locating critical points of prescribed index, including both energy minima (with an index $0$) and transition states (saddle points with positive index). At each step of the iteration, the proposed method involves several ingredients, such as updating the eigen-directions of the Hessian according to the prescribed index, reflecting the gradient and the Hessian along selected unstable eigen-directions, if any, and constructing the next step by solving a cubic-regularized subproblem. Under suitable assumptions, we show that, when the eigenspace and the cubic subproblem are solved exactly, the iteration converges locally and quadratically to a critical point of the prescribed index. We further establish convergence results in the inexact setting, where both the eigenspace computation and the solution of the cubic subproblem are performed approximately. We show that the resulting iteration retains local linear convergence, with the convergence rate depending on the accuracy of the inexact computations. Moreover, we prove that the fixed points of the algorithm are precisely critical points of the prescribed index. In addition, we introduce an adaptive strategy for updating the cubic regularization parameter to improve robustness and computational efficiency of the proposed algorithm. Numerical experiments confirm the predicted convergence behavior, demonstrate the effectiveness of the adaptive strategy, and illustrate the capability of the proposed method for locating critical points and constructing solution landscapes. Full-window branch discovery and loss-selected EnKF continuation for data assimilation 作者: Angxiu Ni 类别: math.NA, math.DS, nlin.CD 发布日期: 2026-09-10 arXiv ID: 2609.10949v1 链接: http://arxiv.org/abs/2609.10949v1 摘要: We develop a framework for offline full-window branch discovery, optionally followed by online continuation with an ensemble Kalman filter (EnKF). Three mechanisms drive the branch search: adjoint path-kernel (APK) differentiation balances kernel differentiation and correction-stabilized path perturbation, shifting the optimization from exploration to exploitation; an optimized Gaussian initial law broadens the search over initial-state basins; and loss-weighted mixing across independent runs recombines successful path components. We may then select an interior state using a local loss and continue online with an EnKF. In 40-dimensional Lorenz-96 experiments, the mean offline path RMSE of APK is 4.3 times smaller than that of population weak-$\mathrm{4D\text{-}Var}_x$. The resulting APK-EnKF method has a mean online RMSE 64 times smaller than that of ordinary EnKF. Stress-divergence, Laplacian, and rotational forms of the incompressible Navier–Stokes equations with variable viscosity 作者: Benjamin Cisternas, Douglas R. Q. Pacheco, Felipe Galarce, Ernesto Castillo 类别: math.NA 发布日期: 2026-09-10 arXiv ID: 2609.10919v1 链接: http://arxiv.org/abs/2609.10919v1 摘要: In the Navier–Stokes equations, incompressibility allows rewriting the viscous term in various forms leading to distinct numerical properties and flow descriptions. Furthermore, models accounting for non-Newtonian, thermal or turbulent effects often break the constant-viscosity assumption, thereby producing additional consistency terms. In this context, the present work compares the classical symmetric-gradient diffusion term with more recent variable-viscosity generalizations of the Laplacian and rotational forms. We discuss, analyze and test their differences with respect to implementation, efficiency, numerical stability and outflow boundary conditions. With a focus on time-dependent flows, we consider second-order implicit-explicit (IMEX) temporal discretizations aimed at improving efficiency and numerical stability. Through a rigorous stability analysis, we show how selected explicit treatments can bypass algorithmic nonlinearities without inducing CFL conditions. Our numerical results highlight important differences between the three viscous formulations—especially in the presence of outflow boundaries, for which the generalized Laplacian form proves more suitable in diffusion-dominated regimes. %(as widely known for constant viscosity). Preserving Structure in Variational Data Assimilation of Hyperbolic Conservation Laws via Regularization 作者: Yao Xiao, Anne Gelb 类别: math.NA 发布日期: 2026-09-09 arXiv ID: 2609.10906v1 链接: http://arxiv.org/abs/2609.10906v1 摘要: Hyperbolic conservation laws pose a challenging setting for variational data assimilation, since the Gaussian assumption imposes a smoothness that smears jump discontinuities. A sparsity promoting regularization can help to mitigate this problem by incorporating a structural prior into the variational objective. Care must be taken when choosing this prior because it serves to counterbalance both a structurally misspecified background-error covariance, which introduces spurious oscillations over smooth regions, as well as spatially sparse observations, which leave too little data to constrain the state. Moreover, state variables of hyperbolic conservation laws typically do not have piecewise constant structure, which is an assumption made when using standard sparsity-promoting operators, such as total variation. Indeed the underlying {\em variability} of hyperbolic conservation law solutions are neither known in advance nor fixed. Hence using higher order total variation is also not suitable. Here we introduce a new regularization term for data assimilation, built on the sparsity promoting residual transform that neither commits to a fixed smoothness order nor requires the underlying variability to be known a priori. We solve the resulting nonconvex objective within a three-dimensional variational framework using generalized sparse Bayesian learning. This approach replaces a global (\ell _{1}) penalty with a hierarchical conditional Gaussian prior on the residual transform, where a learned hyper-prior yields location-specific weights attached to each spatial point. Numerical experiments across scalar, shallow-water, and Euler test problems show that this regularization term resolves discontinuities more accurately than total variation, and that its advantage grows as the background-error covariance becomes misspecified and the observations become sparser. A Simple Complexity Lower Bound for Solving $Ax=b$ 作者: Ohad Shamir 类别: math.OC, math.NA 发布日期: 2026-09-09 arXiv ID: 2609.10874v1 链接: http://arxiv.org/abs/2609.10874v1 摘要: In this note, we provide a short and direct proof that approximately solving $Ax=b$ to relative error $\varepsilon$, where $A$ has condition number $κ$ and unrestricted dimension, requires $Ω(κ\log(1/\varepsilon))$ matrix-vector multiplications in the worst case, even for randomized algorithms. This essentially recovers the lower bound of Dereziński, Epperly and Meyer [2026] for this setting, whose elegant and more general approach inspired us to seek a short direct proof. A straightforward reduction implies the classical $Ω(\sqrtκ\log(1/\varepsilon))$ lower bound for optimizing strongly convex quadratic functions, applicable to randomized algorithms. Stability Framework for the Singularity of the Euler Equations on $\mathbb{R}^3$ 作者: Valentin Duruisseaux, Adarsh Ganeshram, Robert J. George, Anima Anandkumar 类别: math.AP, math.NA 发布日期: 2026-09-09 arXiv ID: 2609.10860v1 链接: http://arxiv.org/abs/2609.10860v1 摘要: In a companion numerical study, we found a high-precision singular profile for the Euler equations on the unbounded domain $\mathbb{R}^3$. The present manuscript complements that study by establishing in detail a framework for proving (nonlinear) stability of the approximate self-similar profile, reducing the analysis to a large but finite collection of explicit estimates and computable constants. Conditional on rigorous certification of the estimates and constants appearing in the argument, the framework closes the stability proof and, crucially, allows the stable rescaled profile to be reconstructed as an admissible solution in the original variables that becomes singular in finite physical time. With the stability and reconstruction mechanisms established, the remaining work is largely quantitative: certifying the explicit constants and, where necessary, sharpening selected analytic estimates to close the stability proof. A Robust All-Mach Six-Equation Diffuse-Interface Method for Multiphase Flows with Surface Tension 作者: Ghanshyam Bharate, J. C. Mandal 类别: physics.flu-dyn, math.NA, physics.comp-ph 发布日期: 2026-09-09 arXiv ID: 2609.10727v1 链接: http://arxiv.org/abs/2609.10727v1 摘要: A robust finite-volume framework is presented for the simulation of compressible multiphase flows with surface tension across a wide range of Mach numbers. The method is based on a two-pressure, six-equation diffuse interface model incorporating viscous, gravitational, and capillary effects through the continuum surface force formulation. To consistently account for capillary-induced pressure jumps, an HLLC Riemann solver is developed using generalized Riemann invariant analysis, and the instantaneous pressure relaxation procedure is modified to preserve the Laplace pressure jump during phase equilibration. To overcome the excessive numerical diffusion of conventional approximate Riemann solvers in the low-Mach regime, a robust low-Mach correction is proposed by extending our previous formulation with a modified scaling strategy that remains stable in regions of strong pressure variation. The resulting method retains accuracy from nearly incompressible flows to compressible regimes while preserving the robustness of the six-equation formulation. The numerical framework is validated using a series of benchmark problems involving surface tension, viscosity, gravity, and compressibility. The results demonstrate accurate prediction of interface dynamics, capillary pressure, and low-Mach flow features, while significantly reducing numerical dissipation without compromising stability. The proposed methodology provides an efficient and reliable approach for the simulation of complex multiphase flows spanning a broad range of flow regimes.3. 研究方法、关键技术和核心工具
4. 与之前工作的比较
2. 笛卡尔网格上大波数时谐 Maxwell 方程的棱元法多项式保持恢复
1. 研究对象和背景
2. 主要定理或主要结果
3. 研究方法、关键技术和核心工具
4. 与之前工作的比较
3. 匹配渐近展开求积(QBMAX):加速带边界层的层势计算
1. 研究对象和背景
2. 主要定理或主要结果
3. 研究方法、关键技术和核心工具
4. 与之前工作的比较
了解领域论文(摘要翻译)
1. 具有加性迹类噪声的一维随机 Burgers 方程有限元逼近的收敛性
摘要翻译
2. 基于层次半可分矩阵的离散傅里叶积分算子近似求逆
摘要翻译
3. 一种面向非均质固体力学的变分物理信息图神经网络
摘要翻译
4. 用于 Kelvin–Voigt 粘弹性的虚拟元方法
摘要翻译
5. 使用熵稳定 DGSEM 和精确无散度相容有限元的 2.5D 理想 MHD 保结构算子分裂
摘要翻译
不相关论文(基本信息)
1. 容差驱动的轴对称曲面上 Stokes 双层位势的近场计算
2. 关于含噪信号恢复的一些理论结果与实用结果
3. 通过容差感知的精度增强实现的可验证线性规划
4. 用于图信号逼近的 $hp$ 自适应树
5. 二维不可压缩Euler方程Cauchy问题已报道多重解的数值研究
6. 实 Cayley–Dickson 塔中的拟线性乘法
7. 随机 Runge–Kutta 积分器的均方误差分析
8. 满足边界条件的非光滑函数的时空有限元插值
9. PH2T样条,第一部分:一个合理的网格假设
10. 多物种 Wasserstein 梯度流的控制策略
11. 深度算子学习用于从随机微分方程的不变测度中高效采样
12. 用于张量函数逼近的管状 Arnoldi 方法
13. 面向大型稀疏系统的基于单层不完全 $LDL^T$ 分解与谱校正的可扩展近似选择性求逆
14. 带高阶状态约束的时间最优升降舵控制:边界接触的分析与计算
15. 基于偏差原则的连续时间 SGD 极小极大最优早停
16. 关于解析函数的 Laguerre 逼近收敛性分析的一点注记
17. 用于高频感应管材焊接的几何参数化准静态三维模型
18. 基于多项式方程的稀疏逼近
19. 六阶问题的混合杂交高阶方法:完全非协调与 $C^0$-协调离散化
20. 多子域中含时滞双曲偏微分方程的 Dirichlet-Neumann 波形松弛方法
21. 求解带时滞双曲型偏微分方程的 Neumann-Neumann 波形松弛方法
22. 大规模参数化偏微分方程的加性降基预条件子
23. 面向半线性常微分方程的无阶数降低自适应对角隐式 Runge-Kutta 方法
24. 平均场博弈偏微分方程系统中 Newton 与 Picard 求解器的计算权衡
25. 一种定位能量景观上给定指标临界点的二阶方法
26. 面向数据同化的全窗口分支发现与损失选择的 EnKF 延拓
27. 变黏度不可压缩 Navier–Stokes 方程的应力-散度形式、Laplacian 形式与旋转形式
28. 通过正则化在双曲守恒律的变分数据同化中保持结构
29. 求解 $Ax=b$ 的一个简单复杂度下界
30. $\mathbb{R}^3$ 上 Euler 方程奇点的稳定性框架
31. 一种用于带表面张力多相流的鲁棒全马赫数六方程扩散界面方法