• Korean Journal of Mathematics
• /
• v.13 no.1
• /
• pp.113-122
• /
• 2005

We consider the iterative schemes for the large sparse linear system to solve partial differential equations. Using spectral radius of iteration matrices, the optimal relaxation parameters and good parameters can be obtained. With those parameters we compare the effectiveness of the SOR and SSOR algorithms. Applying Crank-Nicolson approximation, we observe the error distribution according to domain decomposition. The number of processors due to domain decomposition affects time and error. Numerical experiments show that effectiveness of SOR and SSOR can be reversed as time size varies, which is not the usual case. Finally, these phenomena suggest conjectures about equilibrium time grid for SOR and SSOR.

• Communications of the Korean Mathematical Society
• /
• v.27 no.1
• /
• pp.167-174
• /
• 2012

In this paper, we introduce an asymptotically $\phi$-hemicontractive mapping with a $\phi$-normalized duality mapping and obtain some strongly convergent result of a kind of multi-step iteration schemes for asymptotically $\phi$-hemicontractive mappings.

• Proceedings of the Korea Society for Simulation Conference
• /
• 2001.05a
• /
• pp.169-169
• /
• 2001

With the aid of fast computing power, resampling techniques are being introduced for simulation output analysis (SOA). Autocorrelation among the output from discrete-event simulation prohibit the direct application of resampling schemes (Threshold bootstrap, Binary bootstrap, Stationary bootstrap, etc) extend its usage to time-series data such as simulation output. We present a new method for inference from a regenerative process, regenerative bootstrap, that equals or exceeds the performance of classical regenerative method and approximation regeneration techniques. Regenerative bootstrap saves computation time and overcomes the problem of scarce regeneration cycles. Computational results are provided using M/M/1 model.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.12
• /
• pp.1532-1540
• /
• 1988

Shape from shading problem, such as other most early visions, is ill-posed problems, which can be solved by the use of regularization methods. This paper proposes the three kinds of stabilizer for the regularization. These are integrability constraints and spline functionals. Parallel iterative schemes are derived in the form of the finite difference approximation. Experimental results, show that the average error in surface orientation is less than 5%.

• East Asian mathematical journal
• /
• v.30 no.3
• /
• pp.327-334
• /
• 2014

In this paper, we construct extrapolated expanded mixed finite element approximations to approximate the scalar unknown, its gradient and its flux of semilinear Sobolev equations. To avoid the difficulty of solving the system of nonlinear equations, we use an extrapolated technique in our construction of the approximations. Some numerical examples are used to show the efficiency of our schemes.

• Journal of Broadcast Engineering
• /
• v.23 no.2
• /
• pp.302-315
• /
• 2018

Highly non-linear electro-optical transfer function of the Perceptual Quantizer was approximated by a truncated Taylor series, resulting in a closed form solution for luma adjustment. This previous solution is fast and quite suitable for the hardware implementation of luma adjustment, but the approximation error becomes relatively large in the range of 600~3,900 cd/m2 linear light. In order to reduce such approximation error, we propose a new linear model, for which a correction is performed on the position and the slope of line based on the scope of approximation. In order to verify the approximation capability of the proposed linear model, a comparative study on the luma adjustment schemes was conducted using various high dynamic range test video sequences. Via the comparative study, we identified a significant performance enhancement over the previous fast luma adjustment scheme, where a 4.65dB of adjusted luma t-PSNR gain was obtained for a test sequence having a large portion of saturated color pixels.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.6
• /
• pp.505-514
• /
• 2020

Since a large amount of training data is typically needed to train Deep Neural Networks (DNNs), a parallel training approach is required to train the DNNs. The Stochastic Gradient Descent (SGD) algorithm is one of the most widely used methods to train the DNNs. However, since the SGD is an inherently sequential process, it requires some sort of approximation schemes to parallelize the SGD algorithm. In this paper, we review various efforts on parallelizing the SGD algorithm, and analyze the computational overhead, communication overhead, and the effects of the approximations.

• Computers and Concrete
• /
• v.23 no.2
• /
• pp.145-153
• /
• 2019

Adaptive schemes are constructed in this paper for modeling the effective chloride diffusion coefficient of cement-based materials (paste and concrete). Based on the polarization approximations for the effective conductivity of isotropic multicomponent materials, we develop some fitting procedures to include more information about the materials, to improve the accuracy of the scheme. The variable reference parameter of the approximation involves a few free scalars, which are determined through the available numerical or experimental values of the macroscopic chloride diffusion coefficient of cement paste or concrete at some volume proportions of the component materials. The various factors that affect the chloride diffusivity of cement-based material (porous material structure, uncertainty of value of the chloride diffusion coefficient in water-saturated pore spaces, etc.) may be accounted to make the predictions more accurate. Illustrations of applications are provided in a number of examples to show the usefulness of the approach.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.3
• /
• pp.1065-1085
• /
• 2020

With rapid growth of content demands, device-to-device (D2D) content sharing is exploited to effectively improve the service quality of users. Considering the limited storage space and various content demands of users, caching schemes are significant. However, most of them ignore the influence of the asynchronous content reuse and the selfishness of users. In this work, the user preferences are defined by exploiting the user-oriented content popularity and the current caching situation, and further, we propose the social-aware rate, which comprehensively reflects the achievable contents download rate affected by the social ties, the caching indicators, and the user preferences. Guided by this, we model the collaborative caching problem by making a trade-off between the redundancy of caching contents and the cache hit ratio, with the goal of maximizing the sum of social-aware rate over the constraint of limited storage space. Due to its intractability, it is computationally reduced to the maximization of a monotone submodular function, subject to a matroid constraint. Subsequently, two social-aware collaborative caching algorithms are designed by leveraging the standard and continuous greedy algorithms respectively, which are proved to achieve different approximation ratios in unequal polynomial-time. We present the simulation results to illustrate the performance of our schemes.

• Atmosphere
• /
• v.24 no.4
• /
• pp.555-563
• /
• 2014

The impacts of dynamic and thermodynamic schemes used in the Community Ice CodE (CICE), the Los Alamos sea ice model, on sea ice concentration, extent and thickness over the Arctic and Antarctic regions are evaluated. Using the six dynamic and thermodynamic schemes such as sea ice strength scheme, conductivity scheme, albedo type, advection scheme, shortwave radiation method, and sea ice thickness distribution approximation, the sensitivity experiments are conducted. It is compared with a control experiment, which is based on the fixed atmospheric and oceanic forcing. For sea ice concentration and extent, it is found that there are remarkable differences between each sensitivity experiment and the control run over the Arctic and Antarctic especially in summer. In contrast, there are little seasonal variations between the experiments for sea ice thickness. In summer, the change of the albedo type has the biggest influence on the Arctic sea ice concentration, and the Antarctic sea ice concentration has a greater sensitivity to not only the albedo type but also advection scheme. The Arctic sea ice thickness is significantly affected by the albedo type and shortwave radiation method, while the Antarctic sea ice thickness is more sensitive to sea ice strength scheme and advection scheme.