Numerical Analysis
New submissions
[ showing up to 2000 entries per page: fewer  more ]
New submissions for Thu, 20 Feb 20
 [1] arXiv:2002.07859 [pdf, ps, other]

Title: A strong law of large numbers for scrambled net integrationAuthors: Art B. OwenSubjects: Numerical Analysis (math.NA); Statistics Theory (math.ST); Computation (stat.CO)
This article provides a strong law of large numbers for integration on digital nets randomized by a nested uniform scramble. This strong law requires a square integrable integrand and it holds along a sequence of sample sizes of the form $n=mb^k$ for $m=1,\dots,M$ and $k\ge0$ and a base $b\ge2$.
 [2] arXiv:2002.07879 [pdf, ps, other]

Title: A note on devising HDG+ projections on polyhedral elementsSubjects: Numerical Analysis (math.NA)
In this note, we propose a simple way of constructing HDG+ projections on polyhedral elements. The projections enable us to analyze the LehrenfeldSch\"{o}berl HDG (HDG+) methods in a very concise manner, and make many existing analysis techniques of standard HDG methods reusable for HDG+. The novelty here is an alternative way of constructing the projections without using $M$decomposition as a middle step. This extends our previous results [S. Du and F.J. Sayas, SpringerBriefs in Mathematics (2019)] (elliptic problems) and [S. Du and F.J. Sayas, arXiv:1903.11766] (elasticity) to polyhedral meshes.
 [3] arXiv:2002.07989 [pdf, ps, other]

Title: Implicit bias with RitzGalerkin method in understanding deep learning for solving PDEsSubjects: Numerical Analysis (math.NA)
This paper aims at studying the difference between RitzGalerkin (RG) method and deep neural network (DNN) method in solving partial differential equations (PDEs) to better understand deep learning. To this end, we consider solving a particular Poisson problem, where the information of the righthand side of the equation f is only available at n sample points while the bases (neuron) number is much larger than n, which is common in DNNbased methods. Through both theoretical study and numerical study, we show the RG method solves this particular problem by a piecewise linear function because RG method considers the discrete sampling points as linear combinations of Dirac delta functions. However, we show that DNNs solve the problem with a much smoother function based on previous study of FPrinciple (Xu et al., (2019) [15] and Zhang et al., (2019) [17]), that is, DNN methods implicitly impose regularity on the function that interpolates the discrete sampling points. Our work shows that with implicit bias of DNNs, the traditional methods, e.g., FEM, could provide insights into understanding DNNs.
 [4] arXiv:2002.08038 [pdf, other]

Title: Statistical Inversion Using Sparsity and Total Variation Prior And Monte Carlo Sampling Method For Diffuse Optical TomographySubjects: Numerical Analysis (math.NA); Image and Video Processing (eess.IV)
In this paper, we formulate the reconstruction problem in diffuse optical tomography (DOT) in a statistical setting for determining the optical parameters, scattering and absorption, from boundary photon density measurements. A special kind of adaptive Metropolis algorithm for the reconstruction procedure using sparsity and total variation prior is presented. Finally, a simulation study of this technique with different regularization functions and its comparison to the deterministic Iteratively Regularized Gauss Newton method shows the effectiveness and stability of the method.
 [5] arXiv:2002.08145 [pdf, other]

Title: First order leastsquares formulations for eigenvalue problemsSubjects: Numerical Analysis (math.NA)
In this paper we discuss spectral properties of operators associated with the leastsquares finite element approximation of elliptic partial differential equations. The convergence of the discrete eigenvalues and eigenfunctions towards the corresponding continuous eigenmodes is studied and analyzed with the help of appropriate $L^2$ error estimates. A priori and a posteriori estimates are proved.
 [6] arXiv:2002.08187 [pdf, other]

Title: High order numerical simulations for the polymer selfconsistent field theory using the adaptive virtual element and spectral deferred correction methodsComments: 18 pages, 10 figuresSubjects: Numerical Analysis (math.NA)
In this paper, we propose an efficient and accurate numerical method, which combine the high order adaptive virtual element method (VEM) and spectral deferred correction (SDC) scheme together, to simulate the selfconsistent field theory (SCFT) model on arbitrary domains. The VEM is very flexible in handling rather general polygonal elements, which is more suitable to discrete the model defined on complex geometric domain, and the SDC scheme can efficiently improve the approximation order of contour derivative. Moreover, an adaptive method equipped with a new marking strategy is developed to efficiently solve the SCFT for strong segregation systems. Results indicate that our approaches can efficiently produce highaccuracy numerical results.
Crosslists for Thu, 20 Feb 20
 [7] arXiv:2002.07868 (crosslist from quantph) [pdf, ps, other]

Title: Highprecision quantum algorithms for partial differential equationsComments: 38 pagesSubjects: Quantum Physics (quantph); Numerical Analysis (math.NA)
Quantum computers can produce a quantum encoding of the solution of a system of differential equations exponentially faster than a classical algorithm can produce an explicit description. However, while highprecision quantum algorithms for linear ordinary differential equations are well established, the best previous quantum algorithms for linear partial differential equations (PDEs) have complexity $\mathrm{poly}(1/\epsilon)$, where $\epsilon$ is the error tolerance. By developing quantum algorithms based on adaptiveorder finite difference methods and spectral methods, we improve the complexity of quantum algorithms for linear PDEs to be $\mathrm{poly}(d, \log(1/\epsilon))$, where $d$ is the spatial dimension. Our algorithms apply highprecision quantum linear system algorithms to systems whose condition numbers and approximation errors we bound. We develop a finite difference algorithm for the Poisson equation and a spectral algorithm for more general secondorder elliptic equations.
 [8] arXiv:2002.07886 (crosslist from math.AP) [pdf, other]

Title: Numerical study of ZakharovKuznetsov equations in two dimensionsSubjects: Analysis of PDEs (math.AP); Numerical Analysis (math.NA)
We present a detailed numerical study of solutions to the (generalized) ZakharovKuznetsov equation in two spatial dimensions with various power nonlinearities. In the $L^{2}$subcritical case, numerical evidence is presented for the stability of solitons and the soliton resolution for generic initial data. In the $L^2$critical and supercritical cases, solitons appear to be unstable against both dispersion and blowup. It is conjectured that blowup happens in finite time and that blowup solutions have some resemblance of being selfsimilar, i.e., the blowup core forms a rightward moving selfsimilar type rescaled profile with the blowup happening at infinity in the critical case and at a finite location in the supercritical case. In the $L^{2}$critical case, the blowup appears to be similar to the one in the $L^{2}$critical generalized Kortewegde Vries equation with the profile being a dynamically rescaled soliton.
 [9] arXiv:2002.08044 (crosslist from math.OC) [pdf, other]

Title: Relaxed GaussNewton methods with applications to electrical impedance tomographyComments: 43 pages, 29 figuresSubjects: Optimization and Control (math.OC); Numerical Analysis (math.NA)
As secondorder methods, GaussNewtontype methods can be more effective than firstorder methods for the solution of nonsmooth optimization problems with expensivetoevaluate smooth components. Such methods, however, often do not converge. Motivated by nonlinear inverse problems with nonsmooth regularization, we propose a new GaussNewtontype method with inexact relaxed steps. We prove that the method converges to a set of connected critical points given that the linearisation of the forward operator for the inverse problem is sufficiently precise. We extensively evaluate the performance of the method on electrical impedance tomography (EIT).
 [10] arXiv:2002.08110 (crosslist from cs.MS) [pdf, other]

Title: hyper.deal: An efficient, matrixfree finiteelement library for highdimensional partial differential equationsComments: 33 pages, 18 figuresSubjects: Mathematical Software (cs.MS); Numerical Analysis (math.NA)
This work presents the efficient, matrixfree finiteelement library hyper.deal for solving partial differential equations in two to six dimensions with highorder discontinuous Galerkin methods. It builds upon the lowdimensional finiteelement library deal.II to create complex lowdimensional meshes and to operate on them individually. These meshes are combined via a tensor product on the fly and the library provides new specialpurpose highly optimized matrixfree functions exploiting domain decomposition as well as shared memory via MPI3.0 features. Both nodelevel performance analyses and strong/weakscaling studies on up to 147,456 CPU cores confirm the efficiency of the implementation. Results of the library hyper.deal are reported for highdimensional advection problems and for the solution of the VlasovPoisson equation in up to 6D phase space.
 [11] arXiv:2002.08180 (crosslist from physics.geoph) [pdf, other]

Title: Fluid flow through anisotropic and deformable double porosity media with ultralow matrix permeability: A continuum frameworkSubjects: Geophysics (physics.geoph); Numerical Analysis (math.NA)
Fractured porous media or double porosity media are common in nature. At the same time, accurate modeling remains a significant challenge due to bimodal pore size distribution, anisotropy, multifield coupling and various flow patterns. The purpose of this study is to formulate a comprehensive coupled flow and geomechanics model of anisotropic and deformable double porosity media with ultralow matrix permeability. Fluid in fissures is modeled with the generalized Darcy's law with an equivalent permeability upscaled from the detailed geological characterizations while the liquid in much less permeable matrix follows a low velocity nonDarcy flow characterized by threshold values and nonlinearity, and fluid mass transfer is dependent on shape factor, phase pressure difference, and interface permeability. The geomechanics relies on a thermodynamically consistent effective stress derived from the energy balance equation, and it is modeled following poroelastic theory. Scaling analysis is performed to drop negligible force density terms under reasonable parameters' ranges which also guarantee no violation of entropy inequality. The discussion revolves around generic double porosity media. Numerical simulation of the initial boundary value problem reveals the capability of this framework to capture the crucial role of coupling, anisotropy and ultralow matrix permeability in dictating the pressure and displacement fields.
 [12] arXiv:2002.08235 (crosslist from cs.CE) [pdf, other]

Title: Physicsinformed Neural Networks for Solving Nonlinear Diffusivity and Biot's equationsSubjects: Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Numerical Analysis (math.NA)
This paper presents the potential of applying physicsinformed neural networks for solving nonlinear multiphysics problems, which are essential to many fields such as biomedical engineering, earthquake prediction, and underground energy harvesting. Specifically, we investigate how to extend the methodology of physicsinformed neural networks to solve both the forward and inverse problems in relation to the nonlinear diffusivity and Biot's equations. We explore the accuracy of the physicsinformed neural networks with different training example sizes and choices of hyperparameters. The impacts of the stochastic variations between various training realizations are also investigated. In the inverse case, we also study the effects of noisy measurements. Furthermore, we address the challenge of selecting the hyperparameters of the inverse model and illustrate how this challenge is linked to the hyperparameters selection performed for the forward one.
Replacements for Thu, 20 Feb 20
 [13] arXiv:1611.09646 (replaced) [pdf, ps, other]

Title: How to overcome the CourantFriedrichsLewy condition of explicit discretizations?Authors: Denys Dutykh (LAMA)Comments: 21 pages, 7 figures, 8 references. Other author's papers can be found at this http URLJournalref: In: Numerical Methods for Diffusion Phenomena in Building Physics (2019), Springer, ChamSubjects: Numerical Analysis (math.NA); Computational Physics (physics.compph)
 [14] arXiv:1804.02095 (replaced) [pdf, other]
 [15] arXiv:1804.04215 (replaced) [pdf, other]

Title: Optimal Unbiased Estimation for Expected Cumulative CostComments: 35 pages, 5 figuresSubjects: Numerical Analysis (math.NA); Optimization and Control (math.OC); Probability (math.PR)
 [16] arXiv:1810.09344 (replaced) [pdf, other]

Title: Reduced Basis Greedy Selection Using Random Training SetsComments: 20 pagesSubjects: Numerical Analysis (math.NA)
 [17] arXiv:1901.03958 (replaced) [pdf, other]

Title: On the Convergence of the Laplace Approximation and NoiseLevelRobustness of Laplacebased Monte Carlo Methods for Bayesian Inverse ProblemsComments: 49 pages, 10 figuresSubjects: Numerical Analysis (math.NA); Computation (stat.CO)
 [18] arXiv:1906.09462 (replaced) [pdf, other]

Title: A hybrid Hermite WENO scheme for hyperbolic conservation lawsComments: 38 pagesJournalref: Journal of Computational Physics, 405 (2020), 109175Subjects: Numerical Analysis (math.NA)
 [19] arXiv:1907.06223 (replaced) [pdf, ps, other]

Title: A fast direct solver for two dimensional quasiperiodic multilayered media scattering problemsSubjects: Numerical Analysis (math.NA)
 [20] arXiv:1907.10550 (replaced) [pdf, other]

Title: Exploiting variable precision in GMRESSubjects: Numerical Analysis (math.NA)
 [21] arXiv:1907.12431 (replaced) [pdf, ps, other]

Title: Bayesian Approach to Inverse Timeharmonic Acoustic Scattering with Phaseless Farfield DataSubjects: Numerical Analysis (math.NA)
 [22] arXiv:2001.04452 (replaced) [pdf, ps, other]

Title: Error analysis for timefractional semilinear parabolic equations using upper and lower solutionsAuthors: Natalia KoptevaSubjects: Numerical Analysis (math.NA)
 [23] arXiv:2001.10957 (replaced) [pdf, other]

Title: Approximation of Stieltjes ordinary differential equationsComments: PreprintSubjects: Numerical Analysis (math.NA)
 [24] arXiv:1904.03507 (replaced) [pdf, ps, other]

Title: LowRank Approximability and Entropy Area Laws for Ground States of Unbounded HamiltoniansAuthors: Mazen AliComments: 31 pages, 2 figuresSubjects: Analysis of PDEs (math.AP); Mathematical Physics (mathph); Functional Analysis (math.FA); Numerical Analysis (math.NA)
 [25] arXiv:1906.09289 (replaced) [pdf, other]

Title: ControlTheoretic Models of Environmental CrimeComments: Expanded Remarks 3.2 and 4.1, fixed minor typosSubjects: Optimization and Control (math.OC); Numerical Analysis (math.NA)
 [26] arXiv:1911.01234 (replaced) [pdf, other]

Title: An Approximate Message Passing Algorithm for Rapid ParameterFree Compressed Sensing MRIComments: 5 pages, 5 figures, submission to IEEE International Conference on Image Processing (ICIP) 2020Subjects: Signal Processing (eess.SP); Information Theory (cs.IT); Numerical Analysis (math.NA)
[ showing up to 2000 entries per page: fewer  more ]
Disable MathJax (What is MathJax?)
Links to: arXiv, form interface, find, math, recent, 2002, contact, help (Access key information)