Calculation Of Higher Order Derivatives Research Articles

Computing the Mittag-Leffler function of a matrix argument

It is well-known that the two-parameter Mittag-Leffler (ML) function plays a key role in Fractional Calculus. In this paper, we address the problem of computing this function, when its argument is a square matrix. Effective methods for solving this problem involve the computation of higher order derivatives or require the use of mixed precision arithmetic. In this paper, we provide an alternative method that is derivative-free and works entirely using IEEE standard double precision arithmetic. If certain conditions are satisfied, our method uses a Taylor series representation for the ML function; if not, it switches to a Schur-Parlett technique that will be combined with the Cauchy integral formula. A detailed discussion on the choice of a convenient contour is included. Theoretical and numerical issues regarding the performance of the proposed algorithm are discussed. A set of numerical experiments shows that our novel approach is competitive with the state-of-the-art method for IEEE double precision arithmetic, in terms of accuracy and CPU time. For matrices whose Schur decomposition has large blocks with clustered eigenvalues, our method far outperforms the other. Since our method does not require the efficient computation of higher order derivatives, it has the additional advantage of being easily extended to other matrix functions (e.g., special functions).

Radial basis function-differential quadrature-based physics-informed neural network for steady incompressible flows

In this work, a radial basis function differential quadrature-based physics-informed neural network (RBFDQ-PINN) is proposed to simulate steady incompressible flows. The conventional physics-informed neural network (PINN) makes use of the physical equation as a constraint to ensure that the solution satisfies the physical law and the automatic differentiation (AD) method to calculate derivatives at collocation points. Although the AD-PINN is expedient in evaluating derivatives at arbitrary points, it is time-consuming with higher-order derivatives and may lead to nonphysical solutions with sparse samples. Alternatively, the finite difference (FD) method can facilitate the calculation of derivatives, but the FD-PINN will increase the computational cost when handling random point distributions, especially with higher-order discretization schemes. To address these issues, the radial basis function differential quadrature (RBFDQ) method is incorporated into the PINN to replace the AD method for the calculation of derivatives. The RBFDQ method equips with high efficiency in the calculation of high-order derivatives as compared with the AD method and great flexibility in the distribution of mesh points as compared with the FD method. As a result, the proposed RBFDQ-PINN is not only more efficient and accurate but also applicable to irregular geometries. To demonstrate its effectiveness, the RBFDQ-PINN is tested in sample problems such as the lid-driven cavity flow, the channel flow over a backward-facing step, and the flow around a circular cylinder. Numerical results reveal that the RBFDQ-PINN achieves satisfactory accuracy without any labeled collocation points, whereas the AD-PINN struggles to solve some cases, especially for high Reynolds number flows.

Extended Convergence for Two Sixth Order Methods under the Same Weak Conditions

High-convergence order iterative methods play a major role in scientific, computational and engineering mathematics, as they produce sequences that converge and thereby provide solutions to nonlinear equations. The convergence order is calculated using Taylor Series extensions, which require the existence and computation of high-order derivatives that do not occur in the methodology. These results cannot, therefore, ensure that the method converges in cases where there are no such high-order derivatives. However, the method could converge. In this paper, we are developing a process in which both the local and semi-local convergence analyses of two related methods of the sixth order are obtained exclusively from information provided by the operators in the method. Numeric applications supplement the theory.

A new class of higher-ordered/extended Boussinesq system for efficient numerical simulations by splitting operators

In this work, we numerically study the higher-ordered/extended Boussinesq system describing the propagation of water-waves over flat topography. A reformulation of the same order of precision that avoids the calculation of high order derivatives on the surface deformation is proposed. We show that this formulation enjoys an extended range of applicability while remaining stable. Moreover, a significant improvement in terms of linear dispersive properties in high frequency regime is made due to the suitable adjustment of a dispersion correction parameter. We develop a second order splitting scheme where the hyperbolic part of the system is treated with a high-order finite volume scheme and the dispersive part is treated with a finite difference approach. Numerical simulations are then performed under two main goals: validating the model and the numerical methods and assessing the potential need of such higher-order model. The applicability of the proposed model and numerical method in practical problems is illustrated by a comparison with experimental data.

A Simulink-based software solution using the Infinity Computer methodology for higher order differentiation

This paper is dedicated to numerical computation of higher order derivatives in Simulink. In this paper, a new module has been implemented to achieve this purpose within the Simulink-based Infinity Computer solution, recently introduced by the authors. This module offers several blocks to calculate higher order derivatives of a function given by the arithmetic operations and elementary functions. Traditionally, this can be done in Simulink using finite differences only, for which it is well-known that they can be characterized by instability and low accuracy. Moreover, the proposed module allows to calculate higher order Lie derivatives embedded in the numerical solution to Ordinary Differential Equations (ODEs). Traditionally, Simulink does not offer any practical solution for this case without using difficult external libraries and methodologies, which are domain-specific, not general-purpose and have their own limitations. The proposed differentiation module bridges this gap, is simple and does not require any additional knowledge or skills except basic knowledge of the Simulink programming language. Finally, the block for constructing the Taylor expansion of the differentiated function is also proposed, adding so another efficient numerical method for solving ODEs and for polynomial approximation of the functions. Numerical experiments on several classes of test problems confirm advantages of the proposed solution.

Determination of Solidification of Rigidity Point Temperature Using a New Method

This work aims to calculate the rigidity point temperature of aluminum alloys by three new methods and compare them with currently employed methods. The influence of major and minor alloying elements over the rigidity point temperature is also discussed. Until now it has been difficult to determine the exact temperature of the rigidity point, since small variations in the data obtained give variable results, making it difficult to automate the process with high accuracy. In this work we suggested three new mathematic methods based on the calculation of higher order derivatives of (dT/dt) with respect to time or temperature compared to those currently employed. A design of experiments based on the Taguchi method was employed to compare the effect of the major and minor alloying elements for the AlSi10Mg alloy, and to evaluate the accuracy of each developed method. Therefore, these systems will allow better automation of rigidity point temperature (RPT) determination, which is one of the most important solidification parameters for solidification simulators. The importance of the correct determination of this parameter lies in its relation to quality problems related to solidification, such as hot tearing. If the RPT presents very low-temperature values, the aluminum casting will be more sensitive to hot tearing, promoting the presence of cracks during the solidification process. This is why it is so important to correctly determine the temperature of the RPT. An adequate design of chemical composition by applying the methodology and the novel methods proposed in this work, and also the optimization of process parameters of the whole casting process with the help of the integrated computational modeling, will certainly help to decrease any internal defective by predicting one of the most important defects present in the aluminum industry.

Numerical Solution for the Extrapolation Problem of Analytic Functions

In this work, a numerical solution for the extrapolation problem of a discrete set of n values of an unknown analytic function is developed. The proposed method is based on a novel numerical scheme for the rapid calculation of higher order derivatives, exhibiting high accuracy, with error magnitude of O(10−100) or less. A variety of integrated radial basis functions are utilized for the solution, as well as variable precision arithmetic for the calculations. Multiple alterations in the function's direction, with no curvature or periodicity information specified, are efficiently foreseen. Interestingly, the proposed procedure can be extended in multiple dimensions. The attained extrapolation spans are greater than two times the given domain length. The significance of the approximation errors is comprehensively analyzed and reported, for 5832 test cases.

Statistical mechanics approach to the electric polarization and dielectric constant of band insulators

We develop a theory for the analytic computation of the free energy of band insulators in the presence of a uniform and constant electric field. The two key ingredients are a perturbation-like expression of the Wannier-Stark energy spectrum of electrons and a modified statistical mechanics approach involving a local chemical potential in order to deal with the unbounded spectrum and impose the physically relevant electronic filling. At first order in the field, we recover the result of King-Smith, Vanderbilt and Resta for the electric polarization in terms of a Zak phase - albeit at finite temperature - and, at second order, deduce a general formula for the electric susceptibility, or equivalently for the dielectric constant. Advantages of our method are the validity of the formalism both at zero and finite temperature and the easy computation of higher order derivatives of the free energy. We verify our findings on two different one-dimensional tight-binding models.

A generalized Taylor method of order three for the solution of initial value problems in standard and infinity floating-point arithmetic

A well-known drawback of algorithms based on Taylor series formulae is that the explicit calculation of higher order derivatives formally is an over-elaborate task. To avoid the analytical computation of the successive derivatives, numeric and automatic differentiation are usually used. A recent alternative to these techniques is based on the calculation of higher derivatives by using the Infinity Computer—a new computational device allowing one to work numerically with infinities and infinitesimals. Two variants of a one-step multi-point method closely related to the classical Taylor formula of order three are considered. It is shown that the new formula is order three accurate, though requiring only the first two derivatives of y(t) (rather than three if compared with the corresponding Taylor formula of order three). To get numerical evidence of the theoretical results, a few test problems are solved by means of the new methods and the obtained results are compared with the performance of Taylor methods of order up to four.

Observability of nonlinear power system dynamics using synchrophasor data

The increasing availability of high-resolution, synchronized measurements from phasor measurement units (PMUs) presents an opportunity to monitor power system dynamics in real time. However, the degree of observability of power system dynamics depends upon the power system states monitored by the PMUs. This paper presents a new algorithm to assess the observability level of rotor angle dynamics, based on PMU locations and monitored states. An observability index is proposed, which can be tracked along the trajectory of system states over time. To overcome the difficulty in calculation of high-order derivatives, an automatic differentiation technique is applied for observability calculation of large-scale systems. The proposed observability algorithms and techniques are validated on a 179-bus system. The proposed method is also helpful in identifying locations where PMUs should be installed for effective observability of nonlinear power system dynamics. This synchrophasor based observability monitoring method serves as a basis for control schemes that require knowledge of power system dynamics. Copyright © 2015 John Wiley & Sons, Ltd.

A new unfitted stabilized Nitsche’s finite element method for Stokes interface problems

In this paper, we introduce and analyze a new stabilized finite element method based on combining Nitsche’s method with a ghost penalty method for two-phase Stokes flows involving two different kinematic viscosities by using the lowest equal order velocity–pressure pairs. The interface between two-phase flows does not need to align with the mesh and interface conditions are imposed weakly using a Nitsche type approach. This method has some prominent features: parameter-free, avoiding calculation of higher order derivatives or edge data structures and stabilization being completely local behavior about pressure. We prove that the method is inf–sup stable and obtain optimal order a priori error estimates. We also show that the estimate for the condition number of the stiffness matrix is independent of the location of the interface. Finally, we present some numerical examples to support our theoretical results.

Efficient recursive algorithms for functionals based on higher order derivatives of the multivariate Gaussian density

Many developments in Mathematics involve the computation of higher order derivatives of Gaussian density functions. The analysis of univariate Gaussian random variables is a well-established field whereas the analysis of their multivariate counterparts consists of a body of results which are more dispersed. These latter results generally fall into two main categories: theoretical expressions which reveal the deep structure of the problem, or computational algorithms which can mask the connections with closely related problems. In this paper, we unify existing results and develop new results in a framework which is both conceptually cogent and computationally efficient. We focus on the underlying connections between higher order derivatives of Gaussian density functions, the expected value of products of quadratic forms in Gaussian random variables, and $$V$$ -statistics of degree two based on Gaussian density functions. These three sets of results are combined into an analysis of non-parametric data smoothers.

A simple algorithm for the fast calculation of higher order derivatives of the inverse function

The paper deals with the calculation of higher order derivatives of the inverse function. A simple and fast recursive procedure is presented and compared with other methods known in literature both with respect to the computation time and memory usage.

Multi-Integral Method for Solving the Forward Dynamics of Stiff Multibody Systems

The Hermite–Obreshkov–Padé (HOP) procedure is an implicit method for the numerical solution of a system of ordinary differential equations (ODEs) applicable to stiff dynamical systems. This procedure applies an Obreshkov condition to multiple derivatives of the system state vector, both at the start and end of a time step in the numerical solution. That condition is shown to be satisfied by the Hermite interpolating polynomial that matches the state vector and its derivatives, also at the start and end of a time step. The Hermite polynomial, in turn, can be specified in terms of the system state and its derivatives at the start of a step together with a collection of free parameters. Adjusting these free parameters to minimize magnitudes of the ODE residual and its derivatives at the end of a step serves as a proxy for matching the system state and its derivatives. A high-order Taylor expansion at the start of a time step interval models the residual and its derivatives over the entire interval. A variant of this procedure adjusts those parameters to match integrals of the system state over the duration of that interval. This is done by minimizing magnitudes of integrals of the ODE residual calculated from the extrapolating Taylor-series expansion, a process that avoids the need to determine integration constants for multiple integrals of the state. This alternative method eliminates the calculation of high-order derivatives of the system state and hence avoids loss in accuracy from floating-point round off. Numerical performance is evaluated on a dynamically unbalanced constant-velocity (CV) coupling having a high spring rate constraining shaft deflection.

Analytic rates of solutions to the Euler equations

The Cauchy problem of the Euler equations is considered with initial data with possibly less regularity. The time-local existence and the uniqueness of strong solutions were established by Pak-Park, when the initial velocity is in the Besov space $B^1_{\infty, 1}$. By treating non-decaying initial data, we are able to discuss the propagation of almost periodicity. It is also proved that if the initial data are real analytic, then the solutions become necessarily real analytic in space variables with an explicit convergence rate of the radius in Taylor's expansion. This result comes from the calculation of higher order derivatives, inductively.

An accelerated incremental algorithm to trace the nonlinear equilibrium path of structures

This paper deals with the convergence acceleration of iterative nonlinear methods. An effective iterative algorithm, named the three-point method, is applied to nonlinear analysis of structures. In terms of computational cost, each iteration of the three-point method requires three evaluations of the function. In this study the effective functions have been proposed to accelerate the convergence process. The proposed method has a convergence order of eight, and it is important to note that its implementation does not require the computation of higher order derivatives compared to most other methods of the same order. To trace the equilibrium path beyond the limit point, a normal flow algorithm is implemented into a computer program. The three-point method is applied as an inner step in the normal flow algorithm. The procedure can be used for structures with complex behavior, including: unloading, snap-through, elastic post-buckling and inelastic post-buckling analyses. Several numerical examples are given to illustrate the efficiency and performance of the new method. Results show that the new method is comparable with the well-known existing methods and gives better results in convergence speed.

A new stabilized finite element method for optimal control for a Ladyzhenskaya model for unsteady flows

AbstractThis article considers the time‐dependent optimal control problem of tracking the velocity for the viscous incompressible flows which is governed by a Ladyzhenskaya equations with distributed control. The existence of the optimal solution is shown and the first‐order optimality condition is established. The semidiscrete‐in‐time approximation of the optimal control problem is also given. The spatial discretization of the optimal control problem is accomplished by using a new stabilized finite element method which does not need a stabilization parameter or calculation of high order derivatives. Finally a gradient algorithm for the fully discrete optimal control problem is effectively proposed and implemented with some numerical examples. © 2010 Wiley Periodicals, Inc. Numer Methods Partial Differential Eq 28: 263–287, 2012

Free Vibration Analysis of a Circular Plate With Multiple Circular Holes by Using Indirect BIEM and Addition Theorem

In this paper, natural frequencies and natural modes of a circular plate with multiple circular holes are theoretically derived and numerically determined by using the indirect boundary integral formulation, the addition theorem, and the complex Fourier series. Owing to the addition theorem, all kernel functions are expanded into degenerate forms and further expressed in the same polar coordinates centered at one circle where the boundary conditions are specified. Not only the computation of the principal value is avoided but also the calculation of higher-order derivatives can be easily determined. By matching boundary conditions, a coupled infinite system of linear algebraic equations is derived as an analytical model for the free vibration of a circular plate with multiple circular holes. The direct-searching approach is utilized in the truncated finite system to determine the natural frequency through singular value decomposition. After determining the unknown Fourier coefficients, the corresponding mode shapes are obtained by using the indirect boundary integral formulations. Some numerical eigensolutions are presented and then utilized to explain some physical phenomenon such as the beating and the dynamic stress concentration. Good accuracy and fast rate of convergence are the main features of the present method, thanks to the analytical approach.

Eigensolutions of a circular flexural plate with multiple circular holes by using the direct BIEM and addition theorem

The purpose of this paper is to present an analytical formulation to describe the free vibration of a circular flexural plate with multiple circular holes by using the null field integral formulation, the addition theorem and complex Fourier series. Owing to the addition theorem, all kernel functions are represented in the degenerate form and further transformed into the same polar coordinates centered at one of circles, where the boundary conditions are specified. Thus, not only the computation of the principal value for integrals is avoided but also the calculation of higher-order derivatives in the flexural plate problem can be easily determined. By matching the specified boundary conditions, a coupled infinite system of simultaneous linear algebraic equations is derived as an analytical model for the title problem. According to the direct searching approach, natural frequencies are numerically determined through the singular value decomposition (SVD) in the truncated finite system. After determining the unknown Fourier coefficients, the corresponding mode shapes are obtained by using the direct boundary integral formulations for the domain points. Several numerical results are presented. In addition, the inherent problem of spurious eigenvalue using the integral formulation is investigated and the SVD updating technique is adopted to suppress the occurrence of spurious eigenvalues. Excellent accuracy, fast rate of convergence and high computational efficiency are advantages of the present method thanks to its analytical features.

Multithreaded derivative computation with generated libraries

The computation of derivatives via automatic differentiation is a valuable technique in many science and engineering applications. While the implementation of automatic differentiation via source transformation yields the highest efficiency results, particularly for gradient computations, the implementation via operator overloading remains a viable alternative for some application contexts. Examples include the computation of higher order derivatives or cases where C++ as the language of choice still proves to be too complicated for the currently available source transformation tools. In this paper, we utilize a code generator to create libraries that overload intrinsics for derivative computation, and discuss approaches to improve the efficiency of the generated code. We first discuss the use of limited loop unrolling, but the main focus of the paper is multithreaded derivative computation, in particular an asynchronous scheme for C++ and a synchronous scheme for Fortran. We present test results obtained with a proof-of-concept implementation.