Improving the Performance of Numerical Algorithms for the Bethe‐Salpeter Eigenvalue Problem

PAMM ◽  
2018 ◽  
Vol 18 (1) ◽  
Author(s):  
Peter Benner ◽  
Andreas Marek ◽  
Carolin Penke
Keyword(s):  
Eigenvalue Problem ◽  
Numerical Algorithms

K-L Expansion of a Random Process in Wavelet Bases

1998 ◽  
Author(s):  
Om P. Agrawal ◽  
Shantaram S. Pai
Keyword(s):  
Eigenvalue Problem ◽  
Wavelet Transforms ◽  
Series Representation ◽  
Numerical Algorithms ◽  
Analytical Techniques ◽  
Random Processes ◽  
Random Coefficients ◽  
Engineering Systems ◽  
Wavelet Bases ◽  
Orthogonal Wavelets

Abstract Random processes play a significant role in stochastic analysis of mechanical systems, structures, fluid mechanics, and other engineering systems. In this paper, a numerical method for series representation of random processes, with specified mean and correlation functions, in wavelet bases is presented. In this method, the Karhunen-Loeve expansion approach is used to represent a process as a linear sum of orthonormal eigenfunctions with uncorrelated random coefficients. The correlation and the eigenfunctions are approximated as truncated linear sums of compactly supported orthogonal wavelets. The eigenfunctions satisfy an integral eigenvalue problem. Using the above approximations, the integral eigenvalue problem is converted to a matrix (finite dimensional) eigenvalue problem. Numerical algorithms are discussed to compute one- and two-dimensional wavelet transforms of certain functions, and the resulting equations are solved to obtain the eigenvalues and the eigenfunctions. The scheme provides an improvement over other existing schemes. Two examples are considered to show the feasibility and effectiveness of this method. Numerical studies show that the results obtained using this method compare well with analytical techniques.

scholarly journals Numerical Algorithms for Computing an Arbitrary Singular Value of a Tensor Sum

Axioms ◽  
2021 ◽  
Vol 10 (3) ◽  
pp. 211
Author(s):  
Asuka Ohashi ◽  
Tomohiro Sogabe
Keyword(s):  
Eigenvalue Problem ◽  
Large Scale ◽  
Numerical Algorithms ◽  
Direct Methods ◽  
Coefficient Matrix ◽  
Lanczos Method ◽  
Singular Value ◽  
Preconditioned Conjugate Gradient ◽  
Schur Decomposition ◽  
Pcg Method

We consider computing an arbitrary singular value of a tensor sum: T:=In⊗Im⊗A+In⊗B⊗Iℓ+C⊗Im⊗Iℓ∈Rℓmn×ℓmn, where A∈Rℓ×ℓ, B∈Rm×m, C∈Rn×n. We focus on the shift-and-invert Lanczos method, which solves a shift-and-invert eigenvalue problem of (TTT−σ˜2Iℓmn)−1, where σ˜ is set to a scalar value close to the desired singular value. The desired singular value is computed by the maximum eigenvalue of the eigenvalue problem. This shift-and-invert Lanczos method needs to solve large-scale linear systems with the coefficient matrix TTT−σ˜2Iℓmn. The preconditioned conjugate gradient (PCG) method is applied since the direct methods cannot be applied due to the nonzero structure of the coefficient matrix. However, it is difficult in terms of memory requirements to simply implement the shift-and-invert Lanczos and the PCG methods since the size of T grows rapidly by the sizes of A, B, and C. In this paper, we present the following two techniques: (1) efficient implementations of the shift-and-invert Lanczos method for the eigenvalue problem of TTT and the PCG method for TTT−σ˜2Iℓmn using three-dimensional arrays (third-order tensors) and the n-mode products, and (2) preconditioning matrices of the PCG method based on the eigenvalue and the Schur decomposition of T. Finally, we show the effectiveness of the proposed methods through numerical experiments.

scholarly journals Numerical algorithms without saturation for the Schroedinger equation of hydrogen atom

2018 ◽  
pp. 215-218
Author(s):  
С.Д. Алгазин
Keyword(s):  
Eigenvalue Problem ◽  
Unit Sphere ◽  
Separation Of Variables ◽  
Numerical Algorithms ◽  
Numerical Results ◽  
Mathematical Apparatus ◽  
Entire Space ◽  
Lyman Alpha ◽  
The Laplace Operator ◽  
Reduced Problem

Математически проблема сводится к задаче на собственные значения для оператора Лапласа во всем пространстве с кулоновским потенциалом. Для численного решения этой задачи применяется новый математический аппарат, разработанный автором. Инверсией относительно единичной сферы задача сводится к проблеме собственных значений в проколотом в центре единичном шаре. Граничное условие в бесконечности (нулевое) переходит в центр шара. В шаре можно исключить периодическую переменную $\varphi$ и построить дискретизацию, наследующую свойство разделения переменных дифференциального оператора ($h$-матрица). По $\varphi$ выбиралось 11 точек. Клетки $\Lambda_0$, $\Lambda_1$, $\Lambda_2$, $\Lambda_3$, $\Lambda_4$ и $\Lambda_5$ в $h$-матрице соответствуют линиям Lyman, Balmer, Paschen, Brackett, Pfund и Humphreys. Из рассмотрения, представленных расчетов видим, что $\alpha$-линия Lyman определена с точностью $5.43\%$. Таким образом, совпадение результатов расчетов с теоретическими значениями удовлетворительное. Mathematically, the problem under consideration is reduced to the eigenvalue problem for the Laplace operator in the entire space with the Coulomb potential. The new mathematical apparatus developed by the author is applied to the numerical solution of the reduced problem. This problem is reduced to the eigenvalue problem in the unit ball punctured at the center after inversion with respect to the unit sphere. The null boundary condition at infinity is transformed to the condition at the center of the unit sphere. In the sphere it is possible to split off the periodic variable $\varphi$ and to construct the discretization inheriting the property of the separation of variables of the differential operator (the $h$-matrix). Eleven points is chosen based on the values of $\varphi$. The blocks $\Lambda_0$, $\Lambda_1$, $\Lambda_2$, $\Lambda_3$, $\Lambda_4$, and $\Lambda_5$ of the $h$-matrix correspond to the Lyman, Balmer, Paschen, Brackett, Pfund, and Humphreys lines. From the obtained numerical results, it follows that the Lyman-alpha line is determined with the accuracy equal to 5.43\%. Thus, the coincidence of the numerical results with the theoretical values is satisfactory.

Eigenvalue Problem for the Second Order Differential Equation with Nonlocal

2006 ◽  
Vol 11 (1) ◽  
pp. 13-32 ◽  
Author(s):  
B. Bandyrskii ◽  
I. Lazurchak ◽  
V. Makarov ◽  
M. Sapagovas
Keyword(s):  
Differential Equation ◽  
Eigenvalue Problem ◽  
Variable Coefficient ◽  
Order Differential Equation ◽  
Integral Condition ◽  
Second Order ◽  
Computational Results ◽  
Discrete Method ◽  
Complex Eigenvalues ◽  
Nonlocal Integral Condition

The paper deals with numerical methods for eigenvalue problem for the second order ordinary differential operator with variable coefficient subject to nonlocal integral condition. FD-method (functional-discrete method) is derived and analyzed for calculating of eigenvalues, particulary complex eigenvalues. The convergence of FD-method is proved. Finally numerical procedures are suggested and computational results are schown.

Hypothesis on the Solvability of Parabolic Equations with nonlocal Condition

2002 ◽  
Vol 7 (1) ◽  
pp. 93-104 ◽  
Author(s):  
Mifodijus Sapagovas
Keyword(s):  
Parabolic Equation ◽  
Parabolic Equations ◽  
Existence And Uniqueness ◽  
Nonlocal Condition ◽  
Nonlocal Conditions ◽  
Numerical Algorithms ◽  
Uniqueness Of The Solution

Numerous and different nonlocal conditions for the solvability of parabolic equations were researched in many articles and reports. The article presented analyzes such conditions imposed, and observes that the existence and uniqueness of the solution of parabolic equation is related mainly to ”smallness” of functions, involved in nonlocal conditions. As a consequence the hypothesis has been made, stating the assumptions on functions in nonlocal conditions are related to numerical algorithms of solving parabolic equations, and not to the parabolic equation itself.

Sign in / Sign up

Export Citation Format

Share Document