• Proceedings of the KIEE Conference
• /
• 1997.11a
• /
• pp.182-186
• /
• 1997

This paper presents a mathematical approach to implementing distributed optimal power flow (OPF), wherein a regional decomposition technique is adopted to parallelize the OPF. Three mathematical decomposition coordination methods are introduced firs to implement the proposed distributed scheme: the Auxiliary Problem Principle (APP), the Predictor-Corrector Proximal Multiplier Method (PCPM), and the Alternating Direction Method (ADM). Then two alternative schemes for modeling distributed OPF are introduced; the Dummy Generator-Dummy Generator (DGDG) scheme and Dummy Generator-Dummy Load (DGDL) scheme. We present the mathematical analyses of the proposed approach, and demonstrate the approach on several test, systems, including IEEE Reliability Test Systems and parts of the ERCOT (Electric Reliability Council of Texas) system.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.11
• /
• pp.4556-4572
• /
• 2015

Dynamic texture (DT) recognition is a challenging problem in numerous applications. In this study, we propose a new algorithm for DT recognition based on group sparsity structure in conjunction with chaotic feature vector. Bag-of-words model is used to represent each video as a histogram of the chaotic feature vector, which is proposed to capture self-similarity property of the pixel intensity series. The recognition problem is then cast to a group sparsity model, which can be efficiently optimized through alternating direction method of multiplier algorithm. Experimental results show that the proposed method exhibited the best performance among several well-known DT modeling techniques.

• KIEE International Transactions on Electrophysics and Applications
• /
• v.5C no.5
• /
• pp.204-210
• /
• 2005

Alternating current (AC) losses of two Bi-2223 ([Bi, Pb]: Sr: Ca: Cu: O = 2:2:2:3) tapes [(Tape I, un-twist-pitch) and the other with a twist-pitch of 10 mm (Tape II)] were measured and compared. These samples, produced by the powder-in-(Ag) tube (PIT) method, are multi-filamentary. Also, it's produced by non-twist and different twist pitch (8, 10, 13, 30, 50 and 70 mm). The critical current measurement was carried out under the environment in liquid Nitrogen and in zero-field by 4-probe method. Susceptibility measurements were conducted while cooling in a magnetic field. Flux loss measurements were conducted as a function of ramping rate, frequency and field direction. The AC flux loss increases as the twist-pitch of the tapes decreased, in agreement with the Norris Equation. Neutron-diffraction measurements have been carried out investigate the crystal structure, magnetic structures, and magnetic phase transitions in Bi-2223([Bi, Pb]:Sr:Ca:Cu:O)

• Journal of the Korea Computer Graphics Society
• /
• v.22 no.3
• /
• pp.21-29
• /
• 2016

Cell segmentation is an important but time-consuming and laborious task in biological image analysis. An automated, robust, and fast method is required to overcome such burdensome processes. These needs are, however, challenging due to various cell shapes, intensity, and incomplete boundaries. A precise cell segmentation will allow to making a pathological diagnosis of tissue samples. A vast body of literature exists on cell segmentation in microscopy images [1]. The majority of existing work is based on input images and predefined feature models only - for example, using a deformable model to extract edge boundaries in the image. Only a handful of recent methods employ data-driven approaches, such as supervised learning. In this paper, we propose a novel data-driven cell segmentation algorithm for bright-field microscopy images. The proposed method minimizes an energy formula defined by two dictionaries - one is for input images and the other is for their manual segmentation results - and a common sparse code, which aims to find the pixel-level classification by deploying the learned dictionaries on new images. In contrast to deformable models, we do not need to know a prior knowledge of objects. We also employed convolutional sparse coding and Alternating Direction of Multiplier Method (ADMM) for fast dictionary learning and energy minimization. Unlike an existing method [1], our method trains both dictionaries concurrently, and is implemented using the GPU device for faster performance.

• International Union of Geodesy and Geophysics Korean Journal of Geophysical Research
• /
• v.26 no.1
• /
• pp.1-14
• /
• 1998

Various numerical methods for the two dimensional shallow water equations have been applied to the problems of flood routing, tidal circulation, storm surges, and atmospheric circulation. These methods are often based on the Alternating Direction Implicity(ADI) method. However, the ADI method results in inaccuracies for large time steps when dealing with a complex geometry or bathymetry. Since this method reduces the performance considerably, a fully implicit method developed by Wilders et al. (1998) is used to improve the accuracy for a large time step. Finite Difference Methods are defined on a rectangular grid. Two drawbacks of this type of grid are that grid refinement is not possibile locally and that the physical boundary is sometimes poorly represented by the numerical model boundary. Because of the second deficiency several purely numerical boundary effects can be involved. A boundary fitted curvilinear coordinate transformation is used to reduce these difficulties. It the curvilinear coordinate transformation is used to reduce these difficulties. If the coordinate transformation is orthogonal then the transformed shallow water equations are similar to the original equations. Therefore, an orthogonal coorinate transformation is used for defining coordinate system. A multigrid (MG) method is widely used to accelerate the convergence in the numerical methods. In this study, a technique using a MG method is proposed to reduce the computing time and to improve the accuracy for the orthogonal to reduce the computing time and to improve the accuracy for the orthogonal grid generation and the solutions of the shallow water equations.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.2
• /
• pp.72-78
• /
• 1996

This thesis deals with a method to solve a two-and-one-half-dimensional ($2\frac12$ D) problem, which means that the ocean environment is two-dimensional whereas the source is fully three-dimensionally propagating, including three-dimensional refraction phenomena and three-dimensional back-scattering, using two-dimensional two-way parabolic equation method combined with Fourier synthesis. Two dimensional two-way parabolic equation method uses Galerkin's method for depth and Crank-Nicolson method and alternating direction for range and provides a solution available to range-dependent problem with wave-field back-scattered from discontinuous interface. Since wavenumber, k, is the function of depth and vertical or horizontal range, we can reduce a dimension of three-dimensional Helmholtz equation by Fourier transforming in the range direction. Thus transformed two-dimensional Helmholtz equation is solved through two-way parabolic equation method. Finally, we can have the $2\frac12$ D solution by inverse Fourier transformation of the spectral solution gained from in the last step. Numerical simulation has been carried out for a canonical ocean environment with stair-step bottom in order to test its accuracy using the present analysis. With this spectral parabolic equation method, we have examined three-dimensional acoustic propagation properties in a specified site in the Korean Straits.

• Journal of Astronomy and Space Sciences
• /
• v.39 no.2
• /
• pp.67-77
• /
• 2022

The elements that impact the dynamics and collaborations of waves and particles in the magnetosphere of planets have been considered here. Saturn's internal magnetosphere is determined by substantiated instabilities and discovered to be an exceptional zone of wave activity. Interchanged instability is found to be one of the responsible events in view of temperature anisotropy and energization processes of magnetospheric species. The generated active ions alongside electrons that constitute the populations of highly magnetized planets like Saturn's ring electron current are taken into consideration in the current framework. The previous and similar method of characteristics and the perturbed distribution function have been used to derive dispersion relation. In incorporating this investigation, the characteristics of electromagnetic ion cyclotron wave (EMIC) waves are determined by the composition of ions in plasmas through which the waves propagate. The effect of ring distribution illustrates non-monotonous description on growth rate (GR) depending upon plasma parameters picked out. Observations made by Cassini found appropriate for modern study, have been applied to the Kronian magnetosphere. Using Maxwellian ring distribution function of ions and detailed mathematical formulation, an expression for dispersion relation as well as GR and real frequency (RF) are evaluated. Analysis of plasma parameters shows that, proliferating EMIC waves are not developed much when propagation is parallelly aligned with magnetosphere as compared to waves propagating in oblique direction. GR for the oblique case, is influenced by temperature anisotropy as well as by alternating current (AC) frequency, whereas it is much affected only by AC frequency for parallel propagating waves.

• ETRI Journal
• /
• v.42 no.6
• /
• pp.922-931
• /
• 2020

A hierarchical iterative algorithm for the canonical polyadic decomposition (CPD) of tensors is proposed by improving the traditional conjugate gradient least squares (CGLS) method. Methods based on algebraic operations are investigated with the objective of estimating the direction of arrival (DoA) and polarization parameters of signals impinging on an array with electromagnetic (EM) vector-sensors. The proposed algorithm adopts a hierarchical iterative strategy, which enables the algorithm to obtain a fast recovery for the highly collinear factor matrix. Moreover, considering the same accuracy threshold, the proposed algorithm can achieve faster convergence compared with the alternating least squares (ALS) algorithm wherein the highly collinear factor matrix is absent. The results reveal that the proposed algorithm can achieve better performance under the condition of fewer snapshots, compared with the ALS-based algorithm and the algorithm based on generalized eigenvalue decomposition (GEVD). Furthermore, with regard to an array with a small number of sensors, the observed advantage in estimating the DoA and polarization parameters of the signal is notable.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.18 no.1
• /
• pp.15-22
• /
• 1985

The study on computation time, accuracy, and convergency characteristic of the implicit finite difference method is presented with the variation of the space increment and time step in a two-dimensional transient heat conduction problem with a dirichlet boundary condition. Numerical analysis were conducted by the model having the conditions of the solution domain from 0 to 3m, thermal diffusivity of 1.26 $m^2/h$, initial condition of 272 K, and boundary condition of 255.4 K. The results obtained are summarized as follows : 1) The degree of influence with respect to the accuracy of the time step and space increment in the alternating-direction implicit method and Crank-Nicholson implicit method were relatively small, but in case of the fully implicit method showed opposite tendency. 2) To prescribe near the zero for the space increment and tine step in a two dimensional transient problem were good in a accuracy aspect but unreasonable in a computational time aspect. 3) The reasonable condition of the space increment and the time step considering accuracy and computation time could be generalized with the Fourier modulus increment, F, ana dimensionless space increment, X, irrespective of the solution domain.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2018.07a
• /
• pp.49-50
• /
• 2018

인공지능의 급격한 발전으로 빅 데이터의 활용이 증가되었지만 이로 인해 머신 러닝에서 일어나는 문제들 또한 해결해야할 과제이다. 본 논문에서는 이에 따라 초래되는 문제들 중 학습 데이터가 많아질 경우의 문제들을 방지하기 위해, 알고리즘의 수정 대신 병렬 처리 기반 시스템을 제안한다. 본 논문에서는 Alternating Direction Method of Multiplier(ADMM) 알고리즘을 소개하고 ADMM 기반의 최적화 기법을 적용하여 병렬 학습 시스템 최적화를 제안하였다.