A derivative-free frame-based conjugate gradients algorithm is presented. Convergence is shown for $C^1$ functions, and this is verified in numerical trials. The algorithm is tested on a variety of low dimensional problems, some of which are ill-conditioned, and is also tested on problems of high dimension. Numerical results show that the algorithm is effective on both classes of problems. The results are compared with those from a discrete quasi-Newton method, showing that the conjugate gradients algorithm is competitive. The algorithm exhibits the conjugate gradients speed-up on problems for which the Hessian at the solution has repeated or clustered eigenvalues. The algorithm is easily parallelizable.

The extended system of nondegenerate simple bifurcation point of the Navier-Stokes equations is constructed in this paper. Due to its derivative has a block lower triangular form, we design a Newton-like method. Using the extended system and splitting iterative technique to compute transcritical nondegenerate simple bifurcation point, we not only reduce computational complexity, but also obtain quadratic convergence of algorithm.

In this paper, LCP is converted to an equivalent nonsmooth nonlinear equation system $H(x, y)=0$ by using the famous NCP function–Fischer-Burmeister function. Note that some equations in $H(x, y)=0$ are nonsmooth and nonlinear hence difficult to solve while the others are linear hence easy to solve. Then we further convert the nonlinear equation system $H(x, y)=0$ to an optimization problem with linear equality constraints. After that we study the conditions under which the K–T points of the optimization problem are the solutions of the original LCP and propose a method to solve the optimization problem. In this algorithm, the search direction is obtained by solving a strict convex programming at each iterative point. However, our algorithm is essentially different from traditional SQP method. The global convergence of the method is proved under mild conditions. In addition, we can prove that the algorithm is convergent superlinearly under the conditions: $M$ is $P_0$ matrix and the limit point is a strict complementarity solution of LCP. Preliminary numerical experiments are reported with this method.

This paper considers the asymptotic stability analysis of both exact and numerical solutions of the following neutral delay differential equation with pantograph delay.

where $B,C,D\in C^{d\times d},q\in (0,1)$,and $B$ is regular. After transforming the above equation to non-automatic neutral equation with constant delay, we determine sufficient conditions for the asymptotic stability of the zero solution. Furthermore, we focus on the asymptotic stability behavior of Runge-Kutta method with variable stepsize. It is proved that a L-stable Runge-Kutta method can preserve the above-mentioned stability properties.

By using Moore-Penrose generalized inverse and the general singular value decomposition of matrices, this paper establishes the necessary and sufficient conditions for the existence of and the expressions for the centrosymmetric solutions with a submatrix constraint of matrix inverse problem $AX = B$. In addition, in the solution set of corresponding problem, the expression of the optimal approximation solution to a given matrix is derived.

In the present paper, we investigate properties of lumped mass finite element method (LFEM hereinafter) eigenvalues of elliptic problems. We propose an equivalent formulation of LFEM and prove that LFEM eigenvalues are smaller than the standard finite element method (SFEM hereinafter) eigenvalues. It is shown, for model eigenvalue problems with uniform meshes, that LFEM eigenvalues are not greater than exact solutions and that they are increasing functions of the number of elements of the triangulation, and numerical examples show that this result equally holds for general problems.

This paper provides a convergence analysis of a fractional-step projection method for the controlled-source electromagnetic induction problems in heterogenous electrically conduting media by means of finite element approximations. Error estimates in finite time are given. And it is verified that provided the time step $\tau$ is sufficiently small, the proposed algorithm yields for finite time $T$ an error of $\mathcal{O}(h^s+\tau)$) in the $L^2$-norm for the magnetic field $\boldsymbol{H},$ where $h$ is the mesh size and $1/2 < s ≤ 1$.  

A number of new results on sufficient conditions for the solvability and numerical algorithms of the following general algebraic inverse eigenvalue problem are obtained: Given $n+1$ real $n\times n$ matrices $A=(a_{ij}),A_k=(a_{ij}^{(k)})(k=1,2,\cdots,n)$ and $n$ distinct real numbers $\lambda_1,\lambda_2,\cdots,\lambda_n,$ find $n$ real number $c_1,c_2,\cdots,c_n$ such that the matrix $A(c)=A+\sum\limits_{k=1}^{n}c_k A_k$ has eigenvalues $\lambda_1,\lambda_2,\cdots,\lambda_n.$

By following the geometric point of view in mechanics, a novel expression of the combined hybrid method for plate bending problems is introduced to clarify its intrinsic mechanism of enhancing coarse-mesh accuracy of conforming or nonconforming plate elements. By adjusting the combination parameter $\alpha \in (0,1)$ and adopting appropriate bending moments modes, reduction of energy error for the discretized displacement model leads to enhanced numerical accuracy. As an application, improvement of Adini's rectangle is discussed. Numerical experiments show that the combined hybrid counterpart of Adini's element is capable of attaining high accuracy at coarse meshes.  

The three-dimensional nonlinear Schrödinger equation with weakly damped that possesses a global attractor is considered. The dynamical properties of the discrete dynamical system which generate by a class of finite difference scheme are analysed. The existence of global attractor is proved for the discrete dynamical system.  

In this paper we study structured backward errors for some structured KKT systems. Normwise structured backward errors for structured KKT systems are defined, and computable formulae of the structured backward errors are obtained. Simple numerical examples show that the structured backward errors may be much larger than the unstructured ones in some cases.

The Hamiltonian and multi-symplectic formulations for RLW equation are considered in this paper. A new twelve-point difference scheme which is equivalent to multi-symplectic Preissmann integrator is derived based on the multi-symplectic formulation of RLW equation. And the numerical experiments on solitary waves are also given. Comparing the numerical results for RLW equation with those for KdV equation, the inelastic behavior of RLW equation is shown.

This paper attempts to develop kinetic flux vector splitting (KFVS) for the Euler equations with general pressure laws. It is well known that the gas distribution function for the local equilibrium state plays an important role in the construction of the gas–kinetic schemes. To recover the Euler equations with a general equation of state (EOS), a new local equilibrium distribution is introduced with two parameters of temperature approximation decided uniquely by macroscopic variables. Utilizing the well-known connection that the Euler equations of motion are the moments of the Boltzmann equation whenever the velocity distribution function is a local equilibrium state, a class of high resolution MUSCL–type KFVS schemes are presented to approximate the Euler equations of gas dynamics with a general EOS. The schemes are finally applied to several test problems for a general EOS. In comparison with the exact solutions, our schemes give correct location and more accurate resolution of discontinuities. The extension of our idea to multidimensional case is natural.