• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.15-17
• /
• 2024

최근 다양한 연구 분야에서는 CUDA 프레임워크를 이용하여 병렬 처리를 통해 연산 시간을 단축하는데 성공하고 있다. 이 중 숄레스키 분해는 양의 정부호 행렬을 하삼각행렬로 분해하는 과정에서 많은 행렬 곱셈이 요구되어 GPU 의 구조적 특징을 활용하면 상당한 가속화가 가능하다. 따라서 이 논문에서는 CUDA 코어에 연산을 할당할 때, 핵심 요소인 블록의 개수와 블록 당 쓰레드 개수를 조절할 수 있는 병렬 숄레스키 분해 연산 프로그램을 구현하였다. 서로 다른 세 종류의 행렬 크기에 대해 다양한 블록 수-쓰레드 수 환경을 설정하여 가속화 정도를 측정한 결과, 각 행렬 별 최적 환경에서 동일 그룹 내 최장 시간 대비, 1000x1000 행렬에서는 약 1.80 배, 2000x2000 행렬에서는 약 2.94 배의 추가적인 가속화를 달성하였다.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.40 no.3
• /
• pp.99-108
• /
• 2003

This paper is concerned with the problem of designing a guaranteed cost state feedback controller for singular systems with time-varying delays. The sufficient condition for the existence of guaranteed cost controller, the controller design method, and the optimization problem to get the upper bound of guaranteed cost function are proposed by LMI(linear matrix inequality), singular value decomposition, Schur complements, and change of variables. Since the obtained sufficient conditions can be changed to LMI form, all solutions including controller gain and the upper bound of guaranteed cost function can be obtained simultaneously. Moreover, the proposed controller design method can be extended to the problem of robust guaranteed cost controller design method for singular systems with parameter uncertainties and time-varying delays. The validity of the proposed design algorithm is investigated through a numerical example.

• Membrane Journal
• /
• v.28 no.5
• /
• pp.297-306
• /
• 2018

This review focused carbon-free hydrogen productions from ammonia decomposition including inorganic membranes, catalysts and the presently studied reactor configurations. It also contains general information about hydrogen productions from hydrocarbons as hydrogen carriers. A Pd-based membrane (e.g. a porous ceramic or porous metallic support with a thin selective layer of Pd alloy) shows its efficiency to produce the high purity hydrogen. Ru-based catalysts consisted of Ru, support, and promoter are the efficient catalysts for ammonia decomposition. Packed bed membrane reactor (PBMR), Fluidized bed membrane reactor (FBMR), and membrane micro-reactor have been studied mainly for the optimization and the improvement of mass transfer limitation. Various types of reactors, which contain various combinations of hydrogen-selective membranes (i.e. Pd-based membranes) and catalysts (i.e. Ru-based catalysts) including catalytic membrane reactor, have been studied for carbon-free hydrogen production to achieve high ammonia conversion and high hydrogen flux and purity.

• Journal of Korea Game Society
• /
• v.18 no.1
• /
• pp.125-134
• /
• 2018

This paper presents a practical method to efficiently extract the rotational part of a $3{\times}3$ matrix that changes continuously in time. This is the key technique in the corotational FEM and the shape matching deformation popular in physics-based dynamic deformation. Recently, in contrast to the traditional polar decomposition methods independent of time, an iterative method was proposed that formulates the rotation extraction in a physics-based way and exploits an incremental representation of rotation. We develop an optimization method that reduces the number of iterations under the assumption that the maximum magnitude of the incremental rotation vector is limited within ${\pi}/2$. Realistic simulation of dynamic deformation employs a sufficiently small time step, and thus this assumption is not problematic in practice. We demonstrate the efficiency and practicality of our method in various experiments.

• Smart Structures and Systems
• /
• v.19 no.1
• /
• pp.39-46
• /
• 2017

Finite element model updating is very effective procedure to determine the uncertainty parameters in structural model and minimize the differences between experimentally and numerically identified dynamic characteristics. This procedure can be practiced with manual and automatic model updating procedures. The manual model updating involves manual changes of geometry and analyses parameters by trial and error, guided by engineering judgement. Besides, the automated updating is performed by constructing a series of loops based on optimization procedures. This paper addresses the ambient vibration based finite element model updating of long span reinforced concrete highway bridges using manual model updating procedure. Birecik Highway Bridge located on the $81^{st}km$ of Şanliurfa-Gaziantep state highway over Firat River in Turkey is selected as a case study. The structural carrier system of the bridge consists of two main parts: Arch and Beam Compartments. In this part of the paper, the arch compartment is investigated. Three dimensional finite element model of the arch compartment of the bridge is constructed using SAP2000 software to determine the dynamic characteristics, numerically. Operational Modal Analysis method is used to extract dynamic characteristics using Enhanced Frequency Domain Decomposition method. Numerically and experimentally identified dynamic characteristics are compared with each other and finite element model of the arch compartment of the bridge is updated manually by changing some uncertain parameters such as section properties, damages, boundary conditions and material properties to reduce the difference between the results. It is demonstrated that the ambient vibration measurements are enough to identify the most significant modes of long span highway bridges. Maximum differences between the natural frequencies are reduced averagely from %49.1 to %0.6 by model updating. Also, a good harmony is found between mode shapes after finite element model updating.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.5
• /
• pp.1814-1828
• /
• 2021

Low-light image enhancement is a key technique to overcome the quality degradation of photos taken under scotopic vision illumination conditions. The degradation includes low brightness, low contrast, and outstanding noise, which would seriously affect the vision of the human eye recognition ability and subsequent image processing. In this paper, we propose an approach based on deep learning and Retinex theory to enhance the low-light image, which includes image decomposition, illumination prediction, image reconstruction, and image optimization. The first three parts can reconstruct the enhanced image that suffers from low-resolution. To reduce the noise of the enhanced image and improve the image quality, a super-resolution algorithm based on the Laplacian pyramid network is introduced to optimize the image. The Laplacian pyramid network can improve the resolution of the enhanced image through multiple feature extraction and deconvolution operations. Furthermore, a combination loss function is explored in the network training stage to improve the efficiency of the algorithm. Extensive experiments and comprehensive evaluations demonstrate the strength of the proposed method, the result is closer to the real-world scene in lightness, color, and details. Besides, experiments also demonstrate that the proposed method with the single low-light image can achieve the same effect as multi-exposure image fusion algorithm and no ghost is introduced.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.10
• /
• pp.1517-1524
• /
• 2022

In this paper, we consider a heterogeneous network (HetNet) consisting of one macro base station and multiple small base stations, and assume the coordinated multi-point transmission between the base stations. In addition, we assume that the channel between the base station and the user consists of path loss and Rayleigh fading. Under these assumptions, we present the energy efficiency (EE) achievable by the user for a given base station and we formulate an optimization problem of dynamic cell selection and transmit power allocation to maximize the total EE of the HetNet. In this paper, we propose an unsupervised deep learning method to solve the optimization problem. The proposed deep learning-based scheme can provide high EE while having low complexity compared to the conventional iterative convergence methods. Through the simulation, we show that the proposed dynamic cell selection scheme provides higher EE performance than the maximum signal-to-interference-plus-noise ratio scheme and the Lagrangian dual decomposition scheme, and the proposed transmit power allocation scheme provides the similar performance to the trust region interior point method which can achieve the maximum EE.

• Korean Journal of Environmental Agriculture
• /
• v.39 no.1
• /
• pp.58-64
• /
• 2020

BACKGROUND: Most of the researches on the dye removal using ozonation have been focused on the removal efficiency. However, the research on their removal characteristics and mechanism according to the reaction time has been still insufficient. METHODS AND RESULTS: In this study, the effects of initial pH and dye concentration with reaction time on the degradation characteristics of methyl orange (MO) and methylene blue (MB) by ozonation were evaluated. The degradation efficiency of MB by ozonation increased with increasing pH. On the other hand, the degradation efficiency of MO by ozonation did not show a significant difference with varing pH. The both MO and MB by ozonation were decomposed within 30 min irrespective of the dye concentration, but the decomposition rates of dyes were faster at lower initial dye concentration. The decomposition efficiency of total organic carbon (TOC) in each dye solution by ozonation was low, which was found to be effective for partial decomposition such as decolorization rather than complete degradation of the dye. CONCLUSION: Overall, ozonation was an effective method for removing nondegradable dyes. However, it is necessary to study the optimization of dye degradation under various environmental conditions for ozonation.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.6
• /
• pp.89-97
• /
• 2016

This paper addresses the problem of the construction planning of artillery positions, for which we present an optimization model and propose a heuristic algorithm to solve problems of practical size. The artillery position construction plan includes the assignment of engineers to support the artillery and the schedule of the support team construction sequence. Currently, in the army, managers construct the plan based on their experience. We formulate the problem as a mixed integer program and present a heuristic that utilizes the decomposition of the mixed integer model. We tested the efficacy of the proposed algorithm by conducting computational experiments on both small-size test problems and large-size practical problems. The average optimality gap in the small-size test problem was 6.44% in our experiments. Also, the average computation time to solve the large-size practical problems consisting of more than 200 artillery positions was 79.8 seconds on a personal computer. The result of our computational experiments shows that the proposed approach is a viable option to consider for practical use.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.9 no.1
• /
• pp.33-40
• /
• 1989

In the design of reinforced concrete framed structures, which consist of various design variables, the objective and the constraint functions are formulated in complicated forms. Usually iterative methods have been used to optimize the design variables. In this paper, multilevel formulation is adopted, and design variables are selected in reduced numbers at each level, to reduce the iterative cycle and to accelerate the convergence rate. At level 1, elastic analysis is performed to get the upper and lower bounds of the redistributed design moments due to inelastic behavior of the frame. Then the design moments are taken as design variables and optimized at level 2, and the sizing variables are optimized at level 3. The optimization of redistributed moments is performed using the design sensitivity obtained at the level 2, and force approximation technique is used to reflect the variation of design variables in the lower level to the upper level. The design variables are selected in reduced numbers at each level, and the optimization formulation is simplified effectively. A cost function is taken as the objective function, and the constraints of the stress of the structures are derived from BSI CP 110 following limit state theory. Numerical examples are included to prove the effectiveness of the developed algorithm.