A variational adiabatic hyperspherical finite element R matrix methodology: general formalism and application to H + H2 reaction

The aim of this paper is to present an efficient numerical procedure for the theoretical study of bimolecular reactions. It is based on the R matrix variational formalism and the p-version of the finite element method (p-FEM) for expanding the wave function in a finite basis set, and facilitates the development of an efficient algorithm to invert matrices that significantly reduces the computational time in R matrix calculations. We also utilise the self-consistent finite element method to optimise the elements mesh and provide faster convergence of results. We apply our methodology to the study of the collinear H + H₂ process and evaluate its efficiency by comparing our results with several results previously published in the literature.     

Variational multiscale element-free Galerkin method for 2D Burgers’ equation

A new numerical method, which is based on the coupling between variational multiscale method and meshfree methods, is developed for 2D Burgers’ equation with various values of Re. The proposed method takes full advantage of meshfree methods, therefore, no mesh generation and mesh recreation are involved. Meanwhile, compared with the variational multiscale finite element method, a strong assumption, that is, the fine scale vanishes identically over the element boundaries although non-zero within the elements, is not needed. Subsequently two problems which have an available analytical solution of their own are solved to analyze the convergence behaviour of the proposed method. Finally a 2D Burgers’ equation having large Re is solved and the results have also been compared with the ones computed by two other methods. The numerical results show that the proposed method can indeed obtain accurate numerical results for 2D Burgers’ equation having large Re, which does not refer to the choice of a proper stabilization parameter.     

Adaptive variational curve smoothing based on level set method

This paper presents an adaptive method for variational curve smoothing based on level set implementation. A suitable cost functional is minimized via solving the derived Euler–Lagrangian equation, of which the discretization is conducted on unstructured triangular meshes by employing a simple and effective finite volume scheme. Through adaptive refinement of the mesh, the geometry features of the given curve can be well resolved in a cost-effective way. Various numerical experiments demonstrate the effectiveness and efficiency of the proposed approach.     

The mimetic finite difference method for the 3D magnetostatic field problems on polyhedral meshes

We extend the mimetic finite difference (MFD) method to the numerical treatment of magnetostatic fields problems in mixed div–curl form for the divergence-free magnetic vector potential. To accomplish this task, we introduce three sets of degrees of freedom that are attached to the vertices, the edges, and the faces of the mesh, and two discrete operators mimicking the curl and the gradient operator of the differential setting. Then, we present the construction of two suitable quadrature rules for the numerical discretization of the domain integrals of the div–curl variational formulation of the magnetostatic equations. This construction is based on an algebraic consistency condition that generalizes the usual construction of the inner products of the MFD method. We also discuss the linear algebraic form of the resulting MFD scheme, its practical implementation, and discuss existence and uniqueness of the numerical solution by generalizing the concept of logically rectangular or cubic meshes by Hyman and Shashkov to the case of unstructured polyhedral meshes. The accuracy of the method is illustrated by solving numerically a set of academic problems and a realistic engineering problem.     

非扩散洛伦兹系统的周期轨道

混沌系统的奇怪吸引子是由无数条周期轨道稠密覆盖构成的,周期轨道是非线性动力系统中除不动点之外最简单的不变集,它不仅能够体现出混沌运动的所有特征,而且和系统振荡的产生与变化密切相关,因此分析复杂系统的动力学行为时获取周期轨道具有重要意义.本文系统地研究了非扩散洛伦兹系统一定拓扑长度以内的周期轨道,提出一种基于轨道的拓扑结构来建立一维符号动力学的新方法,通过变分法数值计算轨道显得很稳定.寻找轨道初始化时,两条轨道片段能够被用作基本的组成单元,基于整条轨道的结构进行拓扑分类的方式显得很有效.此外,讨论了周期轨道随着参数变化时的形变情况,为研究轨道的周期演化规律提供了新途径.本研究可为在其他类似的混沌体系中找到并且系统分类周期轨道提供一种可借鉴的方法.     

A robust level set method based on local statistical information for noisy image segmentation

The paper presents an improved local region-based active contour model for image segmentation, which is robust to noise. A data fitting energy functional is defined in terms of curves and the energy terms which are based on the differences between the local average intensity and the global intensity means. Then the energy is incorporated into a level set variational formulation, from which a curve evolution equation is derived for energy minimization. And then the level set function is regularized by Gaussian filter to keep smooth and eliminate the re-initialization. By using the local statistical information, the proposed model can handle with noisy images. The proposed model is first presented as a two-phase level set formulation and then extended to a multi-phase one. Experimental results show desirable performances of the proposed model for both noisy synthetic and real images, especially with high level noise.     

Perturbation theory in terms of electron density

When an inhomogeneous electron gas is subjected to a perturbation, its energy and density both change by small amounts. We calculate the changes in the energy explicitly in terms of the density changes within the density-functional theory of many-electron systems. We also derive the equations for the induced densities, and using these show that a density correct up to order n in terms of the perturbation parameter is sufficient to give energy changes up to order (2n + 1). As a corollary, the even-order energy changes E⁽²ⁿ⁾ are variational with respect to the density changes ?⁽ⁿ⁾. The equations for the induced densities also follow from this corollary. The even-order corollary also gives a variational method of calculating the induced densities. The theory is demonstrated by applying it to calculate the polarizability and hyperpolarizability of the hydrogen, helium and neon atoms.     

A variational approach to connecting orbits in nonlinear dynamical systems

We propose a variational method for determining homoclinic and heteroclinic orbits including spiral-shaped ones in nonlinear dynamical systems. Starting from a suitable initial curve, a homotopy evolution equation is used to approach a true connecting orbit. The procedure is an extension of a variational method that has been used previously for locating cycles, and avoids the need for linearization in search of simple connecting orbits. Examples of homoclinic and heteroclinic orbits for typical dynamical systems are presented. In particular, several heteroclinic orbits of the steady-state Kuramoto–Sivashinsky equation are found, which display interesting topological structures, closely related to those of the corresponding periodic orbits.     

Evolution of the Fermi surface of cuprates on the basis of the spin-polaron approach

A classical field system is considered that consists of two interacting scalar fields, the Higgs real field and a complex scalar field. It is demonstrated that there exists a nontrivial topological solution in this system—a kink carrying a U(1) charge. Certain questions are discussed related to the stability of the solution obtained. An improved variational procedure is proposed for determining topological U(1)-charged configurations.     

Hepatic vessel segmentation using variational level set combined with non-local robust statistics

Hepatic vessel segmentation is a challenging step in therapy guided by magnetic resonance imaging (MRI). This paper presents an improved variational level set method, which uses non-local robust statistics to suppress the influence of noise in MR images. The non-local robust statistics, which represent vascular features, are learned adaptively from seeds provided by users. K-means clustering in neighborhoods of seeds is utilized to exclude inappropriate seeds, which are obviously corrupted by noise. The neighborhoods of appropriate seeds are placed in an array to calculate the non-local robust statistics, and the variational level set formulation can be constructed. Bias correction is utilized in the level set formulation to reduce the influence of intensity inhomogeneity of MRI. Experiments were conducted over real MR images, and showed that the proposed method performed better on small hepatic vessel segmentation compared with other segmentation methods.     

Motion of Curves in Three Spatial Dimensions Using a Level Set Approach

The level set method was originally designed for problems dealing with codimension one objects, where it has been extremely succesful, especially when topological changes in the interface, i.e., merging and breaking, occur. Attempts have been made to modify it to handle objects of higher codimension, such as vortex filaments, while preserving the merging and breaking property. We present numerical simulations of a level set based method for moving curves in R³, the model problem for higher codimension, that allows for topological changes. A vector valued level set function is used with the zero level set representing the curve. Our results show that this method can handle many types of curves moving under all types of geometrically based flows while automatically enforcing merging and breaking.     

Lagrange meshes from nonclassical orthogonal polynomials

The Lagrange-mesh numerical method has the simplicity of a mesh calculation and the accuracy of a variational calculation. A flexible general procedure for deriving an infinity of new Lagrange meshes related to orthogonal or nonorthogonal bases is introduced by using nonclassical orthogonal polynomials. As an application, different Lagrange meshes based on shifted Gaussian functions are constructed. A simple quantum-mechanical example shows that the Lagrange-mesh method may become more accurate than the original variational calculation with a nonorthogonal basis.     

基于任意拉格朗日—欧拉型移动网格的反应流广义黎曼问题方法

提出一种在自由重映移动网格下的广义黎曼问题方法模拟反应流.该方法基于显式的自由重映移动网格广义黎曼问题的解.为保证在时间和空间上的高精度，应用广义黎曼问题方法构造数值通量.为保证反应区的高分辨率，采用变分法生成自适应移动网格.该方法不仅能够保证网格质量，而且能有效地避免任意拉格朗日—欧拉方法中由于显式重映过程而带来的数值误差.包括CJ爆轰及不稳定爆轰的数值实验说明该格式的精确性和鲁棒性，证明这种移动网格下的二阶广义黎曼问题方法可以较好地捕捉反应流的间断与光滑结构.     

An improved Chan–Vese model by regional fitting for infrared image segmentation

In this paper, a regional fitting method is proposed for infrared image segmentation. In our model, the intensity of each pixel in a region is described by using the sum of the class center and the weighted variance of the region, in order to build energy function for encouraging the similarity pixels to be clustered together. The adoption of such way can thereby eliminate the issue associated with the drift of the class center that is existed in Chan–Vese model. Particularly, followed by incorporating energy function into the level set evolution without re-initialization framework, the variational formulation can force the level set function to be closed to object boundaries. Experiments on some representative and real infrared images have demonstrated that our model has higher performance of segmentation in comparison with Chan–Vese model without re-initialization, and some existing methods, including LBF and LCV model.     

Harmonic balance approach to the periodic solutions of the (an)harmonic relativistic oscillator

The first-order harmonic balance method via the first Fourier coefficient is used to construct two approximate frequency-amplitude relations for the relativistic oscillator for which the nonlinearity (anharmonicity) is a relativistic effect due to the time line dilation along the world line. Making a change of variable, a new nonlinear differential equation is obtained and two procedures are used to approximately solve this differential equation. In the first the differential equation is rewritten in a form that does not contain a square-root expression, while in the second the differential equation is solved directly. The approximate frequency obtained using the second procedure is more accurate than the frequency obtained with the first due to the fact that, in the second procedure, application of the harmonic balance method produces an infinite set of harmonics, while in the first procedure only two harmonics are produced. Both approximate frequencies are valid for the complete range of oscillation amplitudes, and excellent agreement of the approximate frequencies with the exact one are demonstrated and discussed. The discrepancy between the first-order approximate frequency obtained by means of the second procedure and the exact frequency never exceeds 1.6%. We also obtained the approximate frequency by applying the second-order harmonic balance method and in this case the relative error is as low 0.31% for all the range of values of amplitude of oscillation A.     

Efficient numerical methods for multiple surfactant-coated bubbles in a two-dimensional stokes flow

We present efficient and highly accurate numerical methods to compute the deformation of surfactant-coated, two-dimensional bubbles in a slow viscous flow. Surfactant acts to locally alter the surface tension and thereby change the nature of the interface motion. In this paper, we restrict our attention to the case of a dilute insoluble surfactant. The convection–diffusion equation for the surfactant concentration on the interface is coupled with the Stokes equations in the fluid domain through a boundary condition based on the Laplace-Young condition. The Stokes equations are first recast as an integral equation and then solved using a fast-multipole accelerated iterative procedure. The computational cost per time-step is only O(N log N) operations, with N being the number of discretization points on the interface. The bubble interfaces are described by a spectral mesh and is advected according to the fluid velocity in such a manner so as to preserve equal arc length spacing of marker points. This equal arc length framework has the dual advantage of dynamically maintaining the spatial mesh and allowing efficient, implicit treatment of the stiffest terms in the dynamics. Several phenomenologically different examples are presented.     

Rotational energy surface and quasiclassical analysis of the rotational energy level cluster formation in the ground vibrational state of PH3

We report and substantiate a method for constructing the rotational energy surface (RES) of a molecule as a pure classical object. For an arbitrary molecule we start from the potential energy surface rather than from a conventional “effective Hamiltonian”. The method is used for constructing the RES of the PH₃ molecule in its ground vibrational state. We have used an ab initio potential energy surface [D. Wang, Q. Shi, Q.-S. Zhu, J. Chem. Phys. 112 (2000) 9624-9631; S.N. Yurchenko, M. Carvajal, P. Jensen, F. Herregodts, T.R. Huet, Chem. Phys. 290 (2003) 59-67.]. The shape of the RES is shown not to change for J from 0 to 120. The procedure of quasiclassical quantization of the RES was also undertaken, yielding a set of quasiclassical critical values of the angular momentum. The results explain the structure of quantum rotational energy levels obtained by variational calculations [S.N. Yurchenko, W. Thiel, S. Patchkovskii, P. Jensen, Phys. Chem. Chem. Phys. 7 (2005) 573-582].     

Direct Minimization of the least-squares spectral element functional – Part I: Direct solver

This paper describes an equivalent but improved least-squares formulation for the numerical approximation of the solution of partial differential equations. Instead of using variational analysis to impose the conditions for minimizing the residual, the residuals are minimized directly, thus leading to a method we will denote by Direct Minimization (DM). DM circumvents setting up the normal equations which consists of matrix–matrix multiplications. Matrix–matrix multiplications are expensive, may lead to loss of accuracy and destroy the sparsity pattern present in the original system. The condition number of the DM formulation is the square root of the condition number which would be obtained if variational analysis was employed. An element-by-element procedure will be presented which allows for parallelization of DM. A computational comparison between DM and the conventional least-squares formulation based on variational analysis will be presented.