SLMQN is a subspace limited memory quasi-Newton algorithm for solving largescale bound constrained nonlinear programming problems. The algorithm is suitable to these large problems in which the Hessian matrix is diffi...
详细信息
SLMQN is a subspace limited memory quasi-Newton algorithm for solving largescale bound constrained nonlinear programming problems. The algorithm is suitable to these large problems in which the Hessian matrix is difficult to compute or is dense,or the number of variables is too large to store and compute an n x n matris. Due to less storage requirement, this algorithm can be used in PCs for solving medium-sized and large problems. The algorithm is implemented in Fortran 77.
In this paper the natural boundary reduction, suggested by Feng and Yu[1], is applied to deal with the three-dimensional problems. By expansion in spherical harmonics, we obtain the natural integral equations of harmo...
详细信息
In this paper the natural boundary reduction, suggested by Feng and Yu[1], is applied to deal with the three-dimensional problems. By expansion in spherical harmonics, we obtain the natural integral equations of harmonic problems over interior and exterior spherical domains. Meanwhile, we develop a numerical method for sloving these equations. Some numerical examples are also given to illustrate our method.
In using the methods given by [1] to compute the hypersingular integrals on interval,one should select the mesh carefully in such a way that singular point falls near the center of a subinterval. A numerical method gi...
详细信息
In using the methods given by [1] to compute the hypersingular integrals on interval,one should select the mesh carefully in such a way that singular point falls near the center of a subinterval. A numerical method given in this paper might solve this problem. This new method is very simple, easy to be implemented, and above all, notaffected by the location of singular point.
This paper deals with parallel implemeatation on distributed memory systemsof a pressure-correction projection scheme for the unsteady incompressible NavierStokes equations, the CNMT2 scheme (i.e., the Crank-Nickolson...
详细信息
This paper deals with parallel implemeatation on distributed memory systemsof a pressure-correction projection scheme for the unsteady incompressible NavierStokes equations, the CNMT2 scheme (i.e., the Crank-Nickolson Modified Temamscheme Ⅱ), presented in [1, 2]. The key point of this work is to study parallelizationof the fast Poisson solver [2] and analyse its parallel efficiency. Various techniques,such as pipelining and canon cyclic algorithm, were used to ensure good parallel performance and scalability of the algorithm. The algorithm has been implemented using MPI message passing environment and numerical tests have been carried out on various computers, including the home made Dawn-1000 MPP system and workstation clusters.
In this paper, based on the study of [1], the discretizations of the coupling of finite elemellt and boundary integral are presented to solve the initial boundary value problem of parabolic partial differential equati...
详细信息
In this paper, based on the study of [1], the discretizations of the coupling of finite elemellt and boundary integral are presented to solve the initial boundary value problem of parabolic partial differential equation defined on an unbounded *** semi-discrete scheme and fully discrete scheme are given, and stability theorem and error estimates, which correspond to discrete scheme respectively,are ***,the numerical example is provided,and numerical result shows that the method is feasible and effective.
This paper presents a class of high resolution KFVS (kinetic flux vector split-ting) finite volum methods for solving three-dimensional compressible Euler equa-tions with γ-gas law. The schemes are obtained based on ...
详细信息
This paper presents a class of high resolution KFVS (kinetic flux vector split-ting) finite volum methods for solving three-dimensional compressible Euler equa-tions with γ-gas law. The schemes are obtained based on the important connection between the Boltzmann equation and the Euler equations. According to the sign of the normal molecular velocity component at the surface of ally control volumes,one gives a splitting of the macroscopic flux vector, i.e. writes the macroscopic flux vector into the sum form of a positive flux and a negative flux. The initial reconstruction is applied to improve resolution of the schemes. Several numerical results are also presented to show the performance of our schemes.
This paper studies the three-term conjugate gradient method for unconstrained optimization. The method includes the classical (two-term) conjugate gradient method and the famous Beale-Powell restart algorithm as its s...
详细信息
This paper studies the three-term conjugate gradient method for unconstrained optimization. The method includes the classical (two-term) conjugate gradient method and the famous Beale-Powell restart algorithm as its special forms. Some mild conditions are given in this paper, which ensure the global convergence of general three-term conjugate gradient methods.
In this paper, we discuss the convergence properties of the memoryless quasi-Newton method proposed by Shanno (1978). In the two-dimensional quadratic case, we prove the global convergence of the method without any li...
详细信息
In this paper, we discuss the convergence properties of the memoryless quasi-Newton method proposed by Shanno (1978). In the two-dimensional quadratic case, we prove the global convergence of the method without any line search; if an exact line search is made at the first iteration, then the method gives the exact solution at most at the forth iteration. Numerical experiments further demonstrate these properties of the memoryless quasi-Newton method.
暂无评论