• Proceedings of the Korea Water Resources Association Conference
• /
• 2005.09a
• /
• pp.330-332
• /
• 2005

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.10
• /
• pp.1590-1597
• /
• 2004

Three methods of design sensitivity analysis for structures such as numerical method, analytical method and semi-analytical method have been developed for the last three decades. Although analytical design sensitivity analysis can provide very exact result, it is difficult to implement into practical design problems. Therefore, numerical method such as finite difference method is widely used to simply obtain the design sensitivity in most cases. The numerical differentiation is sufficiently accurate and reliable fur most linear problems. However, it turns out that the numerical differentiation is inefficient and inaccurate in nonlinear design sensitivity analysis because its computational cost depends on the number of design variables and large numerical errors can be included. Thus the semi-analytical method is more suitable for complicated design problems. Moreover, semi-analytical method is easy to be performed in design procedure, which can be coupled with an analysis solver such as commercial finite element package. In this paper, implementation procedure fur the semi-analytical design sensitivity analysis outside of the commercial finite element package is studied and the computational technique is proposed for evaluating the partial differentiation of internal nodal force, so called pseudo-load. Numerical examples coupled with commercial finite element package are shown to verify usefulness of proposed semi-analytical sensitivity analysis procedure and computational technique for pseudo-load.

• Journal of applied mathematics & informatics
• /
• v.35 no.3_4
• /
• pp.277-302
• /
• 2017

In this paper, we have proposed a numerical method for Singularly Perturbed Boundary Value Problems (SPBVPs) of reaction-diffusion type of third order Ordinary Differential Equations (ODEs). The SPBVP is reduced into a weakly coupled system of one first order and one second order ODEs, one without the parameter and the other with the parameter ${\varepsilon}$ multiplying the highest derivative subject to suitable initial and boundary conditions, respectively. The numerical method combines boundary value technique, asymptotic expansion approximation, shooting method and finite difference scheme. The weakly coupled system is decoupled by replacing one of the unknowns by its zero-order asymptotic expansion. Finally the present numerical method is applied to the decoupled system. In order to get a numerical solution for the derivative of the solution, the domain is divided into three regions namely two inner regions and one outer region. The Shooting method is applied to two inner regions whereas for the outer region, standard finite difference (FD) scheme is applied. Necessary error estimates are derived for the method. Computational efficiency and accuracy are verified through numerical examples. The method is easy to implement and suitable for parallel computing. The main advantage of this method is that due to decoupling the system, the computation time is very much reduced.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.725-728
• /
• 2005

The scattering problem of the randomly rough surface is examined by the method of moments(MoM), small perturbation method (SPM), integral equation method (IEM) and the semi-empirical polarimetic model. To apply the numerical technique of the MoM to microwave scattering from a rough surface, at first, many independent randomly rough surfaces with a rms height and a correlation length are generated with Gaussian random deviate. Then, an efficient Monte Carlo simulation technique is applied to estimate the scattering coefficients of the surfaces.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2000.10a
• /
• pp.205-210
• /
• 2000

An optimum design algorithm of PSC box girder bridges using design sensitivity analysis is proposed in this paper. For the efficiency of the proposed algorithm, approximated reanalysis techniques using design sensitivity analysis are introduced. And also to save the numerical efforts, an efficient reanalysis technique through approximated structural responses is proposed. A design sensitivity analysis of structural response is executed by automatic differentiation(AD). The efficiency and robustness of the proposed algorithm, compared with conventional algorithm, is successfully demonstrated in the numerical example.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.43 no.6
• /
• pp.1020-1026
• /
• 1994

This paper presents a technique to optimize the shape of waveguide components in H-plane. The technique utilizes the numerical optimization process which employs the vector finite element method. In the optimization process, the sensitivity of an objective function with respect to design variables is computed by introducting adjoint variables, which makes the computation easy. The steepest descent method is then employed to update design variables. As a numerical example, an H-plane waveguide teejunction was considered to obtain optimized shape. Comparison between the initial and optimized shape was made.

• Journal of Advanced Marine Engineering and Technology
• /
• v.28 no.1
• /
• pp.124-135
• /
• 2004

An experimental study is performed turbulent flow behind a square cylinder by using 2-D PIV technique. The Reynolds number investigated are 10.000. 30.000 and 50,000. The mean velocity vector, time mean axial velocity turbulence intensity. kinetic energy and Reynolds shear stress behind the cylinder are measured, The numerical method used this study is a CFD code, STAR-CD. The numerical results are compared with these of experimental.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.22 no.1
• /
• pp.15-28
• /
• 2018

Recently Machine Learning algorithms are widely used to process Big Data in various applications and a lot of these applications are executed in run time. Therefore the speed of Machine Learning algorithms is a critical issue in these applications. However the most of modern iteration Machine Learning algorithms use a successive iteration technique well-known in Numerical Linear Algebra. But this technique has a very low convergence, needs a lot of iterations to get solution of considering problems and therefore a lot of time for processing even on modern multi-core computers and clusters. Tchebychev iteration technique is well-known in Numerical Linear Algebra as an attractive candidate to decrease the number of iterations in Machine Learning iteration algorithms and also to decrease the running time of these algorithms those is very important especially in run time applications. In this paper we consider the usage of Tchebychev iterations for acceleration of well-known K-Means and SVM (Support Vector Machine) clustering algorithms in Machine Leaning. Some examples of usage of our approach on modern multi-core computers under Apache Spark framework will be considered and discussed.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2001.05a
• /
• pp.827-832
• /
• 2001

A multi-level optimum design algorithm for prestressed concrete (PSC) box girders is proposed in this paper. To save the numerical efforts, a multi-level optimization technique using model coordination method that separately utilizes both tendon profile design and section design is incorporated. And also, a reduced basis technique for the efficient tendon profile optimization is proposed in this paper. From the numerical investigations, it may be positively stated that the optimum design of PSC box girder based on the new approach proposed in this study will lead to more rational and economical design compared with the currently available designs.

• Industrial Engineering and Management Systems
• /
• v.7 no.2
• /
• pp.143-149
• /
• 2008

There have been some approximation analysis methods for a GI/G/1 queueing system. As one of them, an approximation technique for the steady-state probability in the GI/G/1 queueing system based on the iteration numerical calculation has been proposed. As another one, an approximation formula of the average queue length in the GI/G/1 queueing system by using the diffusion approximation or the heuristics extended diffusion approximation has been developed. In this article, an approximation technique in order to analyze the GI/G/1 queueing system is considered and then the formulae of both the steady-state probability and the average queue length in the GI/G/1 queueing system are proposed. Through some numerical examples by the proposed technique, the existing approximation methods, and the Monte Carlo simulation, the effectiveness of the proposed approximation technique is verified.