中国科学院数学与系统科学研究院期刊网

数值计算与计算机应用 1997年 18卷

Please wait a minute...
选择: 显示/隐藏图片
1. C~k连续的保形插值2k+1次样条函数
王成伟
数值计算与计算机应用    1997, 18 (1): 1-7.   DOI: 10.12288/szjs.1997.1.1
摘要708)      PDF(pc) (235KB)(309)    收藏
In this paper, a method for shape-preserving interpolation is presented. By insertingsome new knots, a C~k (k> 2) shape-preserving interpolating spline function of degree 2k+1is obtained. The errors of interpolating spline functions are estimated. The approximationaccuracy is o(h3) (when k > 2) and o(h2) (when k = 2).
参考文献 | 相关文章 | 多维度评价
被引次数: CSCD(2)
2. 二维离散W变换的快速算法及其应用
曾泳泓,张小水
数值计算与计算机应用    1997, 18 (1): 8-14.   DOI: 10.12288/szjs.1997.1.8
摘要784)      PDF(pc) (254KB)(223)    收藏
In this paper, a 2D discrete W transform is turned to another 2D discrete transform.The kernel of the resulting transform is separable, thus it can be computed by the wellknown row-column algorithm. Therefore, a fast algorithm is obtained for 2D DWT witharbitrary length. Methods are also given in the paper for computing 2D cyclic convolutions,2D skew-cyclic convolutions and 2D generalized discrete Fourier transforms by 2D discreteW transform. Furthermore, running time of the algorithms on a kind of micro computeris given.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(5) CSCD(1)
3. 对称Toeplitz矩阵相乘的快速算法
沈光星
数值计算与计算机应用    1997, 18 (1): 15-19.   DOI: 10.12288/szjs.1997.1.15
摘要741)      PDF(pc) (181KB)(218)    收藏
In this paper, we give the fast algorithm for multiplication for two n-order symmetricToeplitz matrices, proving that the time complexity of this algoithm are n~2 (multiplication) and 2n~2 - 4n + 3 (addition), and that the space complexity of this algorithm isn2 + 4.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(20) CSCD(2)
4. 求解约束L_∞问题的新算法
贾国平
数值计算与计算机应用    1997, 18 (1): 20-29.   DOI: 10.12288/szjs.1997.1.20
摘要755)      PDF(pc) (394KB)(189)    收藏
This paper considers a kind of constrained L_∞ probleme, which involving linear functions only, and presents a new algorithm for the solution of this problem. In considerationof the special structure of L_∞ problems, we reduce our problem into some small-scaleproblems by means of the Benders decomposition method and avoid solving large-scaleproblem directly. Our approach is also applicable to the unconstrained case. We finallygive some numerical examples and compare our approach with other algorithms for L_∞problems. These results show that our method for soiving L_∞ problem is efficient.
参考文献 | 相关文章 | 多维度评价
5. EPE_k方法和可正定化矩阵
胡家赣,刘兴平
数值计算与计算机应用    1997, 18 (1): 30-39.   DOI: 10.12288/szjs.1997.1.30
摘要655)      PDF(pc) (381KB)(217)    收藏
In this paper, the EPE_k method is considered and the positive-definable matrix isdefined. The results of this paper can also be applied to other iterative method.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(27) CSCD(3)
6. 脉冲响应的一种有效数值解法
张智江,李岳生
数值计算与计算机应用    1997, 18 (1): 40-45.   DOI: 10.12288/szjs.1997.1.40
摘要689)      PDF(pc) (171KB)(195)    收藏
Impulse response is one of the absolutely necessary data in the modern scientific fields,such as resource exploration, directional design. In this paper, a effective numrical methodis obtained by the techniques of reducing the singularity and characteristics differences.Some examples are given.
参考文献 | 相关文章 | 多维度评价
7. 二维无结构三角形网格的高分辨率大粒子有限体积方法
宋松和,李荫藩
数值计算与计算机应用    1997, 18 (1): 46-52.   DOI: 10.12288/szjs.1997.1.46
摘要730)      PDF(pc) (264KB)(189)    收藏
Large-particle (FLIC) method, presented in 1960's, is a numerical method that be applied to solve unsteady flow. The computational scheme consists of two steps for each timemarch step: First, intermediate values are calculated for the velocities and energy, takinginto account the effects of acceleration caused by pressure gradients; Second transport effects are calculated. In this paper, we present a high resolution large-particle finite volumemethod for 2-D unstructured triangular mesh, the key idea of this method is monotonereconstruction of flow variables and solve "Riemann" problem in the first step. Finallythe result of the computation is satisfactory
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(5) CSCD(2)
8. 非定常不可压Navier-Stokes方程的高效和稳健的差分格式Ⅰ
伍亚丹
数值计算与计算机应用    1997, 18 (1): 53-63.   DOI: 10.12288/szjs.1997.1.53
摘要781)      PDF(pc) (473KB)(306)    收藏
An efficient and robust pressure correction projection method with the CNMT1 finitedifference scheme is presented in this paper for the numerical solution of the incompressible Navier-Stokes equations. It is proved that on fixed spatial grids the method is of secondorder global accuracy in time; this is confirmed with numerical experiment on an examplewith an exact solution. Then the method is used for numerical simulation of the drivencavity flow problems; the asymptotic periodic solution for Re=10000 is preseated.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(6) CSCD(2)
9. 非线性抛物组具并行本性的某些实用差分格式
周毓麟,沈隆钧,袁光伟
数值计算与计算机应用    1997, 18 (1): 64-73.   DOI: 10.12288/szjs.1997.1.64
摘要708)      PDF(pc) (305KB)(243)    收藏
In this paper some finite difference schemes with intrinsic parallelism for nonlinearparabolic system are constructed. For the nonlinear difference system with intrinsic parallelism, a mild restriction condition for the steplengths is derived.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(14) CSCD(4)
10. 求解对称三对角矩阵特征值的一种新的分而治之算法
罗晓广,李晓梅
数值计算与计算机应用    1997, 18 (2): 74-80.   DOI: 10.12288/szjs.1997.2.74
摘要937)      PDF(pc) (247KB)(229)    收藏

This paper presents a new divide-and-conquer algorithm for the eigenvalue problem ofsymmtric tridiagonal matrices. The new algorithm bases on bisection and secant iteration,which is different from Cuppen's method and Laguerre iteration. The results of theoreticalanalysis and numerical testing show that the convergent rate of our algorithm is obviouslyfaster than that of Laguerre iteration presented in [1]. When the problem scale is quite.large, with the same requirement of accuracy, more than 40% of the computing time canbe reduced by using this new algorithm. In the end, we parallelize this new algorithm andget satisfactory testing results.

参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(15) CSCD(1)
11. 可展Bézier曲面
叶正麟,孟雅琴,刘克轩
数值计算与计算机应用    1997, 18 (2): 81-86.   DOI: 10.12288/szjs.1997.2.81
摘要742)      PDF(pc) (217KB)(268)    收藏
Using Bezier curves of degree n + 1 as design curves XA(t) on one plane and Bezier curves of degree n + m + 1 as adjoint curves XB(t) on another parallel plane, the conditions of constructing developable surfaces of degree (n+ 1, n + m+ 1) are discussed.These conditions are determined by the control vextexes of the two Bezier curves and the matching functions. Furthermore, the methods for constructing developable surface of degree (n + 1, n + 2) and compositive surfaces are derived.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(58) CSCD(1)
12. 三维问题的区域分解法、软件实现及其在油藏数值模拟中的应用
戴民,石济民,林振宝,梁国平
数值计算与计算机应用    1997, 18 (2): 87-94.   DOI: 10.12288/szjs.1997.2.87
摘要744)      PDF(pc) (287KB)(263)    收藏
In this paper, a non-matching non-overlapping domain decomposition method for solving 3-dimensional problems is presented. Based on the method, we developed a powerful software which is suitable for solving practical problems. An application to numerical reservoir simulation is given to demonstrate the method.
参考文献 | 相关文章 | 多维度评价
13. 非线性热耦合和相变可动边界的BEM分析
冯振兴,唐少武,李正秀
数值计算与计算机应用    1997, 18 (2): 95-105.   DOI: 10.12288/szjs.1997.2.95
摘要796)      PDF(pc) (402KB)(213)    收藏

In this paper, a BEM model for transient heat transfer with phase-front as a mov-ing boundary is presented. A new time-marching scheme together with the analytical treatment of integrals for the diffusion fundamental solution and its numerical calcu-lation are formulated. Some examples show that the presented algorithm is of higher accuracy and stability. It could be used to two phase field case even with convection and other stronger nonlinear effects.

参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(1)
14. 解二维标量双曲型守恒律的一类满足极值原理的无结构三角形网格有限体积方法
宋松和,李荫藩
数值计算与计算机应用    1997, 18 (2): 106-113.   DOI: 10.12288/szjs.1997.2.106
摘要882)      PDF(pc) (288KB)(289)    收藏
A new finite volume scheme, based on first order monotone scheme and limited linear reconstruction, is constructed for scalar hyperbolic conservation laws in two dimension,the scheme satisfies the maximum principle and approximation the flux with second order accuracy. Numerical results for constant coefficient linear advection and Burgers'equation are presented.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(30) CSCD(6)
15. CAGD中的隐式曲线与曲面
徐国良
数值计算与计算机应用    1997, 18 (2): 114-124.   DOI: 10.12288/szjs.1997.2.114
摘要707)      PDF(pc) (551KB)(317)    收藏

In this paper, we review some recent developments on the study of implicitly defined curves and surfaces in the field of computer aided geometric design(CAGD), includ-ing mainly the research on the problems of parametrization, regularity and splines of algebraic curves and surfaces.

参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(43) CSCD(4)
16. 解高阶Hermitian矩阵特征值问题的并行块消去迭代法
孙家昶,邓健新,曹建文
数值计算与计算机应用    1997, 18 (2): 125-134.   DOI: 10.12288/szjs.1997.2.125
摘要756)      PDF(pc) (332KB)(326)    收藏

A parallel block elimination algorithm for solving Hermitian matrir large eigen-value problems was provided in this paper. The algorithm prossess crude grain parallel properties. The high-quality black-bos for solving matrix eigenvalue problems, multi-processors and it's local memory can be use effectively in this algorithm. It can be inplemented on high-performance distributed memory parallel computer.The convergencet error analysis of the algorithm, and parallel design are presented.A part of the numerical results are listed in this paper.

参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(13)
17. 优化星座聚类法和有机氟农药构效关系识别
陈曦,王丽君,胡上序,林学圃,吴军
数值计算与计算机应用    1997, 18 (2): 135-142.   DOI: 10.12288/szjs.1997.2.135
摘要854)      PDF(pc) (257KB)(333)    收藏
Since the study of organic fluorinated pesticides has been a noticeable field in the seaking of efficient pesticides, the research of their structure-activity relationships be-comes more and more important. If the structure features of the chemicals are consid-ered as pattern parameters, and the activities of pesticides are discretized into classes,then the quantitative structure-activity relationship (QSAR) problem can be treated by cluster analysis. Hence, the clustering method of constellation graph is employed to identify the QSAR of fluorinated organic pesticides. In order to obtain better clustering results, the weights of star-tracks are optimized by the Lagrange operator method and good results are achieved.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(18) CSCD(2)
18. 由谱数据和主子阵构造Jacobi矩阵
胡锡炎,张磊,黄贤通
数值计算与计算机应用    1997, 18 (2): 143-150.   DOI: 10.12288/szjs.1997.2.143
摘要721)      PDF(pc) (226KB)(176)    收藏
This paper considers the following problem: How to construct a Jacobian matrix from its spectrum and a submatrix. A new numerical method is given.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(43) CSCD(3)
19. 预处理2步非线性共轭梯度法的并行算法与数值试验
邓玲,李庆扬
数值计算与计算机应用    1997, 18 (2): 151-160.   DOI: 10.12288/szjs.1997.2.151
摘要802)      PDF(pc) (391KB)(275)    收藏
A 2-step nonlinear conjugate gradient (NCG) method has been introduced and an-alyzed in [2]. This method can be used to solve large sparse nonlinear equations with nonsymmetric Jacobin. In this paper, we presented a parallel algorithm of precohdi-tioned 2-step nonlinear conjugate gradient method and discuss how to choose the better precondition which is more suitable for shared memory multiprocessor. Also, we ana lyze how to implement efficiently polynomial preconditioning with step NCG method on shared memory parallel computer. Numerical result on solving nonlinear equations arising from the discretization of nonlinear elliptic partial equations have been given.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(1)
20. 基于快速汉克尔变换数值求解波方程
江少恩
数值计算与计算机应用    1997, 18 (3): 161-166.   DOI: 10.12288/szjs.1997.3.161
摘要864)      收藏
The fast Hankel transform (FHT) is introduced in this paper. Hankel transform is fast calculated using convolution property of FFT by variable transform. The sampling points problem is studied. The end correction of FHT is analysed in detail, and the results show that the accuracy is increased with end correction. So, FHT with end correction is a fast, right and efficient algorithm. FHT is applied to three-dimensional polar coordinate wave equations. It is confirmed that FHT is a fast and efficient numerical solution method to wave equations.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(2)
21. 基于P-中位模型的聚类分析的新的拉格朗日松弛算法
唐立新,杨自厚,王梦光
数值计算与计算机应用    1997, 18 (3): 167-172.   DOI: 10.12288/szjs.1997.3.167
摘要808)      收藏
Clusting problem based on P-Median may be formulated a as 0-1 interger programming problem. In this paper we develop a Lagrangean heuristic algorithms based on relaxation of the inequality constraint. The computational performance of the method is analysed for 10 problems. Computational results show that the average duality gap are within 2%.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(2)
22. SLMQN:一个解大规模边界约束非线性规划子空间有限储存拟牛顿方法的FORTRAN优化软件
倪勤,袁亚湘
数值计算与计算机应用    1997, 18 (3): 173-179.   DOI: 10.12288/szjs.1997.3.173
摘要800)      PDF(pc) (335KB)(251)    收藏
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.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(15) CSCD(2)
23. 一类分叉函数的向量算法──间接地址法
袁国兴,王丽华,杨朝霞
数值计算与计算机应用    1997, 18 (3): 180-186.   DOI: 10.12288/szjs.1997.3.180
摘要743)      PDF(pc) (264KB)(217)    收藏
This paper describes a vector algorithm for some branched functions-indirect address algorithm, and the results of numerical experiments on C3201 computer and YH-2 computer are also given.
参考文献 | 相关文章 | 多维度评价
24. 共享存储并行机的算法设计
刘兴平,张景琳
数值计算与计算机应用    1997, 18 (3): 187-198.   DOI: 10.12288/szjs.1997.3.187
摘要751)      PDF(pc) (391KB)(320)    收藏
In this paper we discuss algorithm design and programing of iterative methods of large sparse linear system on a vector computer and a shared memory parallel computer.Numerical examples indicate that the efficiency of the methods is high on the vector computer, but is low on the shared memory parallel computer.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(21) CSCD(1)
25. 一种改进的等参变换在边界元中的实现
胡圣荣,陈国华
数值计算与计算机应用    1997, 18 (3): 199-205.   DOI: 10.12288/szjs.1997.3.199
摘要833)      PDF(pc) (210KB)(231)    收藏
This paper implements an improved isoparametric transformation from FEM to BEM, which maps side nodes from global to local coordinate system with the same relative positions along the sides they reside in. An example shows this method can provide a reduction in error of up to 3-4 orders of magnitude compared with the standard transformation when side nodal placements are general.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(4)
26. 关于单纯形方法的若干新算法
朱明权,陈明飞
数值计算与计算机应用    1997, 18 (3): 206-217.   DOI: 10.12288/szjs.1997.3.206
摘要683)      PDF(pc) (321KB)(261)    收藏
In this paper, three new algorithms on the simplex method for linear programming are presented. Numerical examples show that these algorithms are efficient.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(10)
27. 并行多重网格计算:各向异性扩散问题
莫则尧,李晓梅
数值计算与计算机应用    1997, 18 (3): 218-229.   DOI: 10.12288/szjs.1997.3.218
摘要1155)      PDF(pc) (634KB)(298)    收藏
Parallel multigrid computations for anisotropic elliptic diffosion problems under distributed memory systems are discussed in this paper. Firstly, we analyse and overcome the effections to convergence factor due to the Jacobi components occured near the pseudo-boundary during line or plane relaxations by block based on domain decomposition. A new type of efficient communication structure is presented at the same time.Secondly, we organize the parallel multigrid computations for sir typical anisotropic diffusion problems based on parallel solutions of diagonial domillated tridiagonal linear algebric equations under the network of workstations (NOWs) with PVM software programming environment. Some useful conclusions are attained. The key factors which significantly decrease the parallel efficiency are also addressed.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(17) CSCD(2)
28. 高维正态概率积分中两个高斯型数值积分公式的比较
杨自强
数值计算与计算机应用    1997, 18 (3): 230-240.   DOI: 10.12288/szjs.1997.3.230
摘要886)      PDF(pc) (528KB)(299)    收藏
Both the classical Gauss-Hermite quadrature for dx and the littleknown Gaussian quadrature for given by Steen-Byrne-Gelbard (1969)given by Steen-Byrne-Gelbard (1969)can be used to evaluate the multivariate normal integrals. In the present paper, we compare the above quadratures for the multivariate normal integrals. The simulated results show that the efficiencies of two formulas have not the significant difference if the condition of integral is very good, however, when the dimension of integral is high or the condition of correlation matrix of the multivariate normal distribution is not good, Steen et.al. formula is more efficient. In appendis, an expanded table of Gaussian quadrature for Steen et.al. is given by the present author.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(2)
29. 线性模型的降维计算与病态分离
童恒庆
数值计算与计算机应用    1997, 18 (4): 241-245.   DOI: 10.12288/szjs.1997.4.241
摘要950)      收藏
In this paper an iterative algorithm of alternating projection between two convex sets is proposed to calculate regression coefficient in linear model. The descent computation and ill-condition seperating of regression coefficient are realized.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(7) CSCD(1)
30. 解实对称矩阵特征值问题的并行算法
邓健新
数值计算与计算机应用    1997, 18 (4): 246-258.   DOI: 10.12288/szjs.1997.4.246
摘要795)      PDF(pc) (472KB)(380)    收藏
Scientific research and engineering computations are becoming dependent the modern high performance parallel computers increasingly. The algorithms for solving matrir eigen-problems is an important part in more computing projects. A brief review of the parallel algorithms for solving symmetric matrix eigenproblems is given here. A major emphasis is to collect bisection / multisection algorithm, divide and conquer algorithm,homotopy continuation method, Jacobi-like algorithm and iterative method.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(23) CSCD(1)
31. 搜索图的全部可能回路的新算法
伊崇信,许克静,伊志伯
数值计算与计算机应用    1997, 18 (4): 259-265.   DOI: 10.12288/szjs.1997.4.259
摘要771)      PDF(pc) (215KB)(210)    收藏
This paper puts forward a new algorithm of searching out all cycles in graphs. Using this algorithm, we can accurately seek out all cycles in graphs without generating the basic cycles. The paper introduces the basic theory of the algorithm and the search method.
参考文献 | 相关文章 | 多维度评价
32. 对称逐步超松弛预处理共轭梯度法的改进迭代格式
林绍忠
数值计算与计算机应用    1997, 18 (4): 266-270.   DOI: 10.12288/szjs.1997.4.266
摘要858)      PDF(pc) (181KB)(327)    收藏
The symmetric successive over relaxation- preconditioned conjugated gradient method (SSOR-PCG) is a very efficient iterative method for solving large sparse linear equations.In this paper an improved iterative format 0f the SSORPCG method is pressented,which avoids the product operation of coefficient matrix and direction vector and thus saves computation work about 8%-50%.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(57) CSCD(18)
33. 分布式系统矩阵并行计算
迟学斌
数值计算与计算机应用    1997, 18 (4): 271-275.   DOI: 10.12288/szjs.1997.4.271
摘要771)      PDF(pc) (195KB)(259)    收藏
In this paper, parallel computation of matrix will be introduced, such as matrix multiplication, solving linear algebra equations, and solving eigenproblems on distributed memory systems. The numerical results have been done on Dawning-1000 MPP system, Workstation Cluster and PowerChallenge under PVM parallel environment. The actual performances on these distributed memory systems are very satisfactory.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(4)
34. 非线性RK方法求解微分代数方程
曹阳,李庆扬
数值计算与计算机应用    1997, 18 (4): 276-287.   DOI: 10.12288/szjs.1997.4.276
摘要937)      PDF(pc) (288KB)(354)    收藏
This paper considers nonlinear explicit Runge-Kutta methods for solving differential algebraic equations (DAEs), discusses the convrgence order for index-1 and index-2 DAEs. Numerical tests for two special formula show that global convergence order is the same as we propose.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(9) CSCD(1)
35. 范德蒙类矩阵之方程组的递进算法
徐仲,游兆永
数值计算与计算机应用    1997, 18 (4): 288-297.   DOI: 10.12288/szjs.1997.4.288
摘要754)      PDF(pc) (286KB)(182)    收藏
In this paper, we obtain fast progressive algorithms for the solution of the Vandermondelike systems involving polynomials that satisfy a three-term recurrence relation, in (n2) arithmetic operations and O(n) elements of storage.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(5)
36. 小系数二阶偏微分方程的组合稳定有限元方法
周磊,周天孝
数值计算与计算机应用    1997, 18 (4): 298-306.   DOI: 10.12288/szjs.1997.4.298
摘要806)      PDF(pc) (340KB)(255)    收藏
A second-order linear scalar differential equation including a zeroth order term is approximated by using a new stablized finite element method, which can attenuate the oscillations present using the standard Galerkin method with very small coefficient of the seconder-order term. By the helpness of additional variable which could be static condensated, we generalize the "unusual stabilized finite element method" developed in [1] from linear triangle element to bilinear quartdral element, and we establish a convergence study and obtain successful numerical simulation.
参考文献 | 相关文章 | 多维度评价
被引次数: Baidu(1)
37. 接触应力极小化外形设计方法
徐君开
数值计算与计算机应用    1997, 18 (4): 307-316.   DOI: 10.12288/szjs.1997.4.307
摘要753)      PDF(pc) (352KB)(342)    收藏
A method is proposed to contour design for contact stress minimization. It is indicated by computation of large-scale structural instances that the amount of computation will be reduced and the machine time can be saved. This method is a useful and significant quick algorithm.
参考文献 | 相关文章 | 多维度评价