Showing 1–12 of 12 results for author: Lakoba, T I

Efficient simulations of Hartree--Fock equations by an accelerated gradient descent method

Authors: Y. Ohno, A. Del Maestro, T. I. Lakoba

Abstract: We develop convergence acceleration procedures that enable a gradient descent-type iteration method to efficiently simulate Hartree--Fock equations for atoms interacting both with each other and with an external potential. Our development focuses on three aspects: (i) optimization of a parameter in the preconditioning operator; (ii) adoption of a technique that eliminates the slowest-decaying mode… ▽ More We develop convergence acceleration procedures that enable a gradient descent-type iteration method to efficiently simulate Hartree--Fock equations for atoms interacting both with each other and with an external potential. Our development focuses on three aspects: (i) optimization of a parameter in the preconditioning operator; (ii) adoption of a technique that eliminates the slowest-decaying mode to the case of many equations (describing many atoms); and (iii) a novel extension of the above technique that allows one to eliminate multiple modes simultaneously. We illustrate performance of the numerical method for the 2D model of the first layer of helium atoms above a graphene sheet. We demonstrate that incorporation of aspects (i) and (ii) above into the ``plain" gradient descent method accelerates it by at least two orders of magnitude, and often by much more. Aspect (iii) -- a multiple-mode elimination -- may bring further improvement to the convergence rate compared to aspect (ii), the single-mode elimination. Both single- and multiple-mode elimination techniques are shown to significantly outperform the well-known Anderson Acceleration. We believe that our acceleration techniques can also be gainfully employed by other numerical methods, especially those handling hard-core-type interaction potentials. △ Less

Submitted 27 February, 2024; originally announced February 2024.

Comments: main text (39 pages); supplement appended (7 pages)

Journal ref: Phys. Rev. E 110, 055304 (2024)

Study of instability of the Fourier split-step method for the massive Gross--Neveu model

Authors: Taras I. Lakoba

Abstract: Stability properties of the well-known Fourier split-step method used to simulate a soliton and similar solutions of the nonlinear Dirac equations, known as the Gross--Neveu model, are studied numerically and analytically. Three distinct types of numerical instability that can occur in this case, are revealed and explained. While one of these types can be viewed as being related to the numerical i… ▽ More Stability properties of the well-known Fourier split-step method used to simulate a soliton and similar solutions of the nonlinear Dirac equations, known as the Gross--Neveu model, are studied numerically and analytically. Three distinct types of numerical instability that can occur in this case, are revealed and explained. While one of these types can be viewed as being related to the numerical instability occurring in simulations of the nonlinear Schrödinger equation, the other two types have not been studied or identified before, to the best of our knowledge. These two types of instability are {\em unconditional}, i.e. occur for arbitrarily small values of the time step. They also persist in the continuum limit, i.e. for arbitrarily fine spatial discretization. Moreover, one of them persists in the limit of an infinitely large computational domain. It is further demonstrated that similar instabilities also occur for other numerical methods applied to the Gross--Neveu soliton, as well as to certain solitons of another relativistic field theory model, the massive Thirring. △ Less

Submitted 2 November, 2019; originally announced November 2019.

Comments: 37 pages; 12 multi-panel figures

Instability of the finite-difference split-step method on the background of localized solutions of the generalized nonlinear Schrödinger equation

Authors: Taras I. Lakoba

Abstract: We consider numerical instability that can be observed in simulations of localized solutions of the generalized nonlinear Schrödinger equation (NLS) by a split-step method where the linear part of the evolution is solved by a finite-difference discretization. Properties of such an instability cannot be inferred from the von Neumann analysis of the numerical scheme. Rather, their explanation requir… ▽ More We consider numerical instability that can be observed in simulations of localized solutions of the generalized nonlinear Schrödinger equation (NLS) by a split-step method where the linear part of the evolution is solved by a finite-difference discretization. Properties of such an instability cannot be inferred from the von Neumann analysis of the numerical scheme. Rather, their explanation requires tools of stability analysis of nonlinear waves, with numerically unstable modes exhibiting novel features not reported for "real" unstable modes of nonlinear waves. For example, modes that cause numerical instability of a standing soliton of the NLS are supported by the sides of the soliton rather than by its core. Furthermore, we demonstrate that both properties and analyses of the numerical instability may be substantially affected by specific details of the simulated solution; e.g., they are substantially different for standing and moving solitons of the NLS. △ Less

Submitted 13 October, 2014; originally announced October 2014.

Investigating early warning signs of oscillatory instability in simulated phasor measurements

Authors: Goodarz Ghanavati, Paul D. H. Hines, Taras I. Lakoba

Abstract: This paper shows that the variance of load bus voltage magnitude in a small power system test case increases monotonically as the system approaches a Hopf bifurcation. This property can potentially be used as a method for monitoring oscillatory stability in power grid using high-resolution phasor measurements. Increasing variance in data from a dynamical system is a common sign of a phenomenon kno… ▽ More This paper shows that the variance of load bus voltage magnitude in a small power system test case increases monotonically as the system approaches a Hopf bifurcation. This property can potentially be used as a method for monitoring oscillatory stability in power grid using high-resolution phasor measurements. Increasing variance in data from a dynamical system is a common sign of a phenomenon known as critical slowing down (CSD). CSD is slower recovery of dynamical systems from perturbations as they approach critical transitions. Earlier work has focused on studying CSD in systems approaching voltage collapse; In this paper, we investigate its occurrence as a power system approaches a Hopf bifurcation. △ Less

Submitted 18 December, 2013; originally announced December 2013.

Comments: 5 pages, 7 figures, IEEE Power & Energy Society General Meeting 2014

Instability of the finite-difference split-step method on the background of a soliton of the nonlinear Schrodinger equation

Authors: T. I. Lakoba

Abstract: We consider the implementation of the split-step method where the linear part of the nonlinear Schrödinger equation is solved using a finite-difference discretization of the spatial derivative. The von Neumann analysis predicts that this method is unconditionally stable on the background of a constant-amplitude plane wave. However, simulations show that the method can become unstable on the backgr… ▽ More We consider the implementation of the split-step method where the linear part of the nonlinear Schrödinger equation is solved using a finite-difference discretization of the spatial derivative. The von Neumann analysis predicts that this method is unconditionally stable on the background of a constant-amplitude plane wave. However, simulations show that the method can become unstable on the background of a soliton. We present an analysis explaining this instability. Both this analysis and the instability itself are substantially different from those of the Fourier split-step method, which computes the spatial derivative by spectral discretization. We also found that the modes responsible for the numerical instability are supported by the sides of the soliton, in contrast to unstable modes of linearized nonlinear wave equations, which (the modes) are supported by the soliton's core. △ Less

Submitted 2 August, 2012; originally announced August 2012.

Comments: 25 pages, 8 figures

Spurious localized highest-frequency modes in Schrödinger-type equations solved by finite-difference methods

Authors: Taras I. Lakoba

Abstract: High-frequency solutions of one or several Schrödinger-type equations are well known to differ very little from the plane wave solutions $\exp[\pm ik x]$. That is, the potential terms impact the envelope of a high-frequency plane wave by only a small amount. However, when such equations are solved by a finite-difference method, the highest-frequency solutions may, under certain conditions, turn ou… ▽ More High-frequency solutions of one or several Schrödinger-type equations are well known to differ very little from the plane wave solutions $\exp[\pm ik x]$. That is, the potential terms impact the envelope of a high-frequency plane wave by only a small amount. However, when such equations are solved by a finite-difference method, the highest-frequency solutions may, under certain conditions, turn out to be localized. In this letter we explain this numerical artifact. △ Less

Submitted 11 April, 2012; originally announced April 2012.

Comments: 4 pages, 6 figure panels (2 figures)

Stability analysis of the split-step Fourier method on the background of a soliton of the nonlinear Schrödinger equation

Authors: Taras I. Lakoba

Abstract: We analyze a numerical instability that occurs in the well-known split-step Fourier method on the background of a soliton. This instability is found to be very sensitive to small changes of the parameters of both the numerical grid and the soliton, unlike the instability of most finite-difference schemes. % on the background of a monochromatic wave, considered earlier in the literature. Moreover,… ▽ More We analyze a numerical instability that occurs in the well-known split-step Fourier method on the background of a soliton. This instability is found to be very sensitive to small changes of the parameters of both the numerical grid and the soliton, unlike the instability of most finite-difference schemes. % on the background of a monochromatic wave, considered earlier in the literature. Moreover, the principle of ``frozen coefficients", in which variable coefficients are treated as ``locally constant" for the purpose of stability analysis, is strongly violated for the instability of the split-step method on the soliton background. Our analysis explains all these features. It is enabled by the fact that the period of oscillations of the unstable Fourier modes is much smaller than the width of the soliton. △ Less

Submitted 29 August, 2010; originally announced August 2010.

Comments: 28 pages, 5 figures This is the original manuscript submitted to the journal Numerical methods for PDEs. A revised version, in whose title the word "Stability" is changed to "Instability", is scheduled to appear in that journal at the end of 2010

MSC Class: 65M12; 65M70

Convergence conditions for iterative methods seeking multi-component solitary waves with prescribed quadratic conserved quantities

Authors: T. I. Lakoba

Abstract: We obtain local (i.e., linearized) convergence conditions for iterative methods that seek solitary waves with prescribed values of quadratic conserved quantities of multi-component Hamiltonian nonlinear wave equations. These conditions extend the ones found for single-component solitary waves in [J. Yang and T.I. Lakoba, Stud. Appl. Math. {\bf 120}, 265--292 (2008)]. We also show that, and why,… ▽ More We obtain local (i.e., linearized) convergence conditions for iterative methods that seek solitary waves with prescribed values of quadratic conserved quantities of multi-component Hamiltonian nonlinear wave equations. These conditions extend the ones found for single-component solitary waves in [J. Yang and T.I. Lakoba, Stud. Appl. Math. {\bf 120}, 265--292 (2008)]. We also show that, and why, these convergence conditions coincide with dynamical stability conditions for ground-state solitary waves. △ Less

Submitted 25 September, 2009; originally announced September 2009.

Comments: 20 pages, 1 multi-part figure

Conjugate gradient method for finding fundamental solitary waves

Authors: Taras I. Lakoba

Abstract: The Conjugate Gradient method (CGM) is known to be the fastest generic iterative method for solving linear systems with symmetric sign definite matrices. In this paper, we modify this method so that it could find fundamental solitary waves of nonlinear Hamiltonian equations. The main obstacle that such a modified CGM overcomes is that the operator of the equation linearized about a solitary wave… ▽ More The Conjugate Gradient method (CGM) is known to be the fastest generic iterative method for solving linear systems with symmetric sign definite matrices. In this paper, we modify this method so that it could find fundamental solitary waves of nonlinear Hamiltonian equations. The main obstacle that such a modified CGM overcomes is that the operator of the equation linearized about a solitary wave is not sign definite. Instead, it has a finite number of eigenvalues on the opposite side of zero than the rest of its spectrum. We present versions of the modified CGM that can find solitary waves with prescribed values of either the propagation constant or power. We also extend these methods to handle multi-component nonlinear wave equations. Convergence conditions of the proposed methods are given, and their practical implications are discussed. We demonstrate that our modified CGMs converge much faster than, say, Petviashvili's or similar methods, especially when the latter converge slowly. △ Less

Submitted 18 March, 2009; originally announced March 2009.

Comments: 44 pages, submitted to Physica D

Accelerated Imaginary-time Evolution Methods for the Computation of Solitary Waves

Authors: Jianke Yang, Taras I. Lakoba

Abstract: Two accelerated imaginary-time evolution methods are proposed for the computation of solitary waves in arbitrary spatial dimensions. For the first method (with traditional power normalization), the convergence conditions as well as conditions for optimal accelerations are derived. In addition, it is shown that for nodeless solitary waves, this method converges if and only if the solitary wave is… ▽ More Two accelerated imaginary-time evolution methods are proposed for the computation of solitary waves in arbitrary spatial dimensions. For the first method (with traditional power normalization), the convergence conditions as well as conditions for optimal accelerations are derived. In addition, it is shown that for nodeless solitary waves, this method converges if and only if the solitary wave is linearly stable. The second method is similar to the first method except that it uses a novel amplitude normalization. The performance of these methods is illustrated on various examples. It is found that while the first method is competitive with the Petviashvili method, the second method delivers much better performance than the first method and the Petviashvili method. △ Less

Submitted 21 November, 2007; originally announced November 2007.

Comments: To appear in Stud. Appl. Math. (28 pages, 6 figures)

A generalized Petviashvili iteration method for scalar and vector Hamiltonian equations with arbitrary form of nonlinearity

Authors: T. I. Lakoba, J. Yang

Abstract: The Petviashvili's iteration method has been known as a rapidly converging numerical algorithm for obtaining fundamental solitary wave solutions of stationary scalar nonlinear wave equations with power-law nonlinearity: \ $-Mu+u^p=0$, where $M$ is a positive definite self-adjoint operator and $p={\rm const}$. In this paper, we propose a systematic generalization of this method to both scalar and… ▽ More The Petviashvili's iteration method has been known as a rapidly converging numerical algorithm for obtaining fundamental solitary wave solutions of stationary scalar nonlinear wave equations with power-law nonlinearity: \ $-Mu+u^p=0$, where $M$ is a positive definite self-adjoint operator and $p={\rm const}$. In this paper, we propose a systematic generalization of this method to both scalar and vector Hamiltonian equations with arbitrary form of nonlinearity and potential functions. For scalar equations, our generalized method requires only slightly more computational effort than the original Petviashvili method. △ Less

Submitted 12 June, 2007; originally announced June 2007.

Comments: to appear in J. Comp. Phys.; 35 pages

A mode elimination technique to improve convergence of iteration methods for finding solitary waves

Authors: T. I. Lakoba, J. Yang

Abstract: We extend the key idea behind the generalized Petviashvili method of Ref. \cite{gP} by proposing a novel technique based on a similar idea. This technique systematically eliminates from the iteratively obtained solution a mode that is "responsible" either for the divergence or the slow convergence of the iterations. We demonstrate, theoretically and with examples, that this mode elimination tech… ▽ More We extend the key idea behind the generalized Petviashvili method of Ref. \cite{gP} by proposing a novel technique based on a similar idea. This technique systematically eliminates from the iteratively obtained solution a mode that is "responsible" either for the divergence or the slow convergence of the iterations. We demonstrate, theoretically and with examples, that this mode elimination technique can be used both to obtain some nonfundamental solitary waves and to considerably accelerate convergence of various iteration methods. As a collateral result, we compare the linearized iteration operators for the generalized Petviashvili method and the well-known imaginary-time evolution method and explain how their different structures account for the differences in the convergence rates of these two methods. △ Less

Submitted 12 June, 2007; originally announced June 2007.

Comments: to appear in J. Comp. Phys.; 24 pages