Efficient computation of the Gauss-Newton direction when fitting NURBS using ODR

We consider a subproblem in parameter estimation using the Gauss-Newton algorithm with regularization for NURBS curve fitting. The NURBS curve is fitted to a set of data points in least-squares sense, where the sum of squared orthogonal distances is minimized. Control-points and weights are estimated. The knot-vector and the degree of the NURBS curve are kept constant. In the Gauss-Newton algorithm, a search direction is obtained from a linear overdetermined system with a Jacobian and a residual vector. Because of the properties of our problem, the Jacobian has a particular sparse structure which is suitable for performing a splitting of variables. We are handling the computational problems and report the obtained accuracy using different methods, and the elapsed real computational time. The splitting of variables is a two times faster method than using plain normal equations.     

A dynamical approach to accelerating numerical integration with equidistributed points

We show how ideas originating in the theory of dynamical systems inspire a new approach to numerical integration of functions. Any Lebesgue integral can be approximated by a sequence of integrals with respect to equidistributions, i.e. evenly weighted discrete probability measures concentrated on an equidistributed set. We prove that, in the case where the integrand is real analytic, suitable linear combinations of these equidistributions lead to a significant acceleration in the rate of convergence of the approximate integral. In particular, the rate of convergence is faster than that of any Newton-Cotes rule.     

Efficient simulation of unsaturated flow using exponential time integration

We assess the performance of an exponential integrator for advancing stiff, semidiscrete formulations of the unsaturated Richards equation in time. The scheme is of second order and explicit in nature but requires the action of the matrix function φ(A) = A⁻¹(e^A − I) on a suitability defined vector v at each time step. When the matrix A is large and sparse, φ(A)v can be approximated by Krylov subspace methods that require only matrix-vector products with A. We prove that despite the use of this approximation the scheme remains second order. Furthermore, we provide a practical variable-stepsize implementation of the integrator by deriving an estimate of the local error that requires only a single additional function evaluation. Numerical experiments performed on two-dimensional test problems demonstrate that this implementation outperforms second-order, variable-stepsize implementations of the backward differentiation formulae.     

Efficient numerical modelling of hydro-mechanical effects in fractures using a hybrid dimensional approach and non-conformal mesh coupling

The investigation of subsurface fluid flow in porous and fractured media is of interest in particular to determine the properties of matter and heat transport below the earth surface and the characteristics of the underground storage of fluids. Classical approaches such as extended diffusion equations are lacking the possibilities to capture key phenomena such as inverse water-level fluctuations (Noordbergum effect) obtained during pumping tests performed on aquifers [1, 2, 4]. In order to efficiently model hydro-mechanical effects, this contribution presents an implementation of the newly introduced hybrid-dimensional approach [3] in the DUNE-PDELab environment. (© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Efficient decoupling technique applied to the numerical time integration of advanced interaction models for railway dynamics

Railway interaction is characterised by the coupling between the train and the track introduced through the wheel/rail contact. The introduction of the flexibility in the wheelset and the track through the finite element (FE) method in the last four decades has permitted to study high-frequency phenomena such as rolling noise and squeal, whose origin lies in the strongly non-steady state and non-linear behaviour of the contact forces that arise from the small contact area. In order to address models with a large number of degrees of freedom, innovative Eulerian-modal models for wheelsets with rotation and cyclic tracks have been developed in recent years. The aim of this paper is to extend the resulting formulation to an uncoupled linear matrix equation of motion that allows solving each equation independently for each time step, considerably reducing the associated computational cost. The decoupling integration method proposed is compared in terms of computational performance with Newmark and Runge-Kutta schemes, commonly used in vehicle dynamics, for simulations with the leading wheelset negotiating a tangent track and accounting the rail roughness.     

A stochastic approach to global error estimation in ODE multistep numerical integration

In order to assess the quality of approximate solutions obtained in the numerical integration of ordinary differential equations related to initial-value problems, there are available procedures which lead to deterministic estimates of global errors. The aim of this paper is to propose a stochastic approach to estimate the global errors, especially in the situations of integration which are often met in flight mechanics and control problems. Treating the global errors in terms of their orders of magnitude, the proposed procedure models the errors through the distribution of zero-mean random variables belonging to stochastic sequences, which take into account the influence of both local truncation and round-off errors. The dispersions of these random variables, in terms of their variances, are assumed to give an estimation of the errors. The error estimation procedure is developed for Adams-Bashforth-Moulton type of multistep methods. The computational effort in integrating the variational equations to propagate the error covariance matrix associated with error magnitudes and correlations is minimized by employing a low-order (first or second) Euler method. The diagonal variances of the covariance matrix, derived using the stochastic approach developed in this paper, are found to furnish reasonably precise measures of the orders of magnitude of accumulated global errors in short-term as well as long-term orbit propagations.     

Construction of fully symmetric numerical integration formulas of fully symmetric numerical integration formulas

The paper develops a construction for finding fully symmetric integration formulas of arbitrary degree 2k+1 inn-space such that the number of evaluation points isO((2n)^k)/k!),n . Formulas of degrees 3, 5, 7, 9, are relatively simple and are presented in detail. The method has been tested by obtaining some special formulas of degrees 7, 9 and 11 but these are not presented here.     

Some algorithms for numerical quadrature using the derivatives of the integrand in the integration interval

Some quadrature formulae using the derivatives of the integrand are discussed. As special cases are obtained generalizations of both the ordinary and the modified Romberg algorithms. In all cases the error terms are expressed in terms of Bernoulli polynomials and functions.     

Principles of verified numerical integration

We describe methods for the numerical calculation of integrals with verified error bounds. The problems range from integration over an interval to integration of parameter-dependent integrands over the whole d-variate space. It is argued, why we use bounds for the integrands in the complex plane as a tool for bounding the error in our own integration software.     

Comparison of numerical integration formulas

In this paper we study the mean square error of numerical integration, when the integrand is a random stationary process. We obtain exact asymptotic errors of classical quadrature formulas and give lower and upper bounds for the least mean square error.     

Efficient error control in numerical integration of ordinary differential equations and optimal interpolating variable-stepsize peer methods

Automatic global error control of numerical schemes is examined. A new approach to this problem is presented. Namely, the problem is reformulated so that the global error is controlled by the numerical method itself rather than by the user. This makes it possible to find numerical solutions satisfying various accuracy requirements in a single run, which so far was considered unrealistic. On the other hand, the asymptotic equality of local and global errors, which is the basic condition of the new method for efficiently controlling the global error, leads to the concept of double quasi-consistency. This requirement cannot be satisfied within the classical families of numerical methods. However, the recently proposed peer methods include schemes with this property. There exist computational procedures based on these methods and polynomial interpolation of fairly high degree that find the numerical solution in a single run. If the integration stepsize is sufficiently small, the error of this solution does not exceed the prescribed tolerance. The theoretical conclusions of this paper are supported by the numerical results obtained for test problems with known solutions.     

The Riemann approach to stochastic integration using non-uniform meshes

In this note we shall prove that the stochastic integral with respect to a semimartingale can be defined by Riemann's approach. However in this approach we use non-uniform meshes instead of the usual uniform meshes.     

A numerical scheme for a kinetic model for mixtures in the diffusive limit using the moment method

In this article, we consider a multi‐species kinetic model which leads to the Maxwell–Stefan equations under a standard diffusive scaling (small Knudsen and Mach numbers). We propose a suitable numerical scheme which approximates both the solution of the kinetic model in rarefied regime and the one in the diffusion limit. We prove some a priori estimates (mass conservation and nonnegativity) and well‐posedness of the discrete problem. We also present numerical examples where we observe the asymptotic‐preserving behavior of the scheme.     

A comparison of numerical integration programs

A comparative study of 15 general-purpose numerical integrators is reported. Each of the programs in the study is tested on a battery of 110 test integrals displaying a wide variety of integrand behaviour and general comments on the performance of the programs are presented.     

Automatic differentiation of numerical integration algorithms

Automatic differentiation (AD) is a technique for automatically augmenting computer programs with statements for the computation of derivatives. This article discusses the application of automatic differentiation to numerical integration algorithms for ordinary differential equations (ODEs), in particular, the ramifications of the fact that AD is applied not only to the solution of such an algorithm, but to the solution procedure itself. This subtle issue can lead to surprising results when AD tools are applied to variable-stepsize, variable-order ODE integrators. The computation of the final time step plays a special role in determining the computed derivatives. We investigate these issues using various integrators and suggest constructive approaches for obtaining the desired derivatives.

     

The centroid method of numerical integration

Summary The midpoint method of integration of a function of one variable is perhaps the simplest method of numerical integration, although it is often not mentioned in textbooks. It is here generalized to any number of dimensions and the generalization is called thecentroid method. This again is a very simple method and it can be conveniently used, for example, for the integration of a function of several variables over any non-pathological region. The numerical examples include the integration of multinormal integrands.     

Efficient integration of flexible multibody dynamics

The modelling of flexible multibody dynamics as finite dimensional Hamiltonian system subject to holonomic constraints constitutes a general framework for a unified treatment of rigid and elastic components. Internal constraints, which are associated with the kinematic assumptions of the underlying continuous theory, as well as external constraints, representing the interconnection of different bodies by joints, can be accounted for in a likewise systematic way. The discrete null space method developed in [0] provides an energy-momentum conserving integration scheme for the DAEs of motion of constrained mechanical systems. It relies on the elimination of the constraint forces from the discrete system along with a reparametrisation of the nodal unknowns. The resulting reduced scheme performs advantageously concerning different aspects: the constraints are fulfilled exactly, the condition number of the iteration matrix is independent of the time step and the dimension of the system is reduced to the minimal possible number saving computational costs. A six-body-linkage possessing a single degree of freedom is analysed as an example of a closed loop structure. (© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Efficient curve fitting: An application of multiobjective programming

Curve fitting is an interesting and important subject in mathematics and engineering. It has been studied extensively and a number of approaches, mostly based on polynomials and piecewise polynomials, have been employed. In the usual setting, some data points are given and one wants to find a polynomial function with the minimum violations measured by a norm in the given data points. In these approaches, norms are applied to aggregate all violations as a scalar.