• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2015년도 학술발표회
• /
• pp.606-609
• /
• 2015

본 연구에서는 support vector regression (SVR) 및 매개변수 최적화 알고리즘을 이용한 하천수위 예측모델을 구축하고 이를 실제 유역에 적용하여 모델 효율성을 평가하였다. 여기서, SVR은 하천수위를 예측하기 위한 예측모델로서 채택되었으며, 커널함수 (Kernel function)로서는 radial basis function (RBF)을 선택하였다. 최적화 알고리즘은 SVR의 최적 매개변수 (C?, cost parameter or regularization parameter; ${\gamma}$, RBF parameter; ${\epsilon}$, insensitive loss function parameter)를 탐색하기 위하여 적용되었다. 매개변수 최적화 알고리즘으로는 grid search (GS), genetic algorithm (GA), particle swarm optimization (PSO), artificial bee colony (ABC) 알고리즘을 채택하였으며, 비교분석을 통해 최적화 알고리즘의 적용성을 평가하였다. 또한 SVR과 최적화 알고리즘을 결합한 모델 (SVR-GS, SVR-GA, SVR-PSO, SVR-ABC)은 기존에 수자원 분야에서 널리 적용되어온 신경망(Artificial neural network, ANN) 및 뉴로퍼지 (Adaptive neuro-fuzzy inference system, ANFIS) 모델과 비교하였다. 그 결과, 모델 효율성 측면에서 SVR-GS, SVR-GA, SVR-PSO 및 SVR-ABC는 ANN보다 우수한 결과를 나타내었으며, ANFIS와는 비슷한 결과를 나타내었다. 또한 SVR-GA, SVR-PSO 및 SVR-ABC는 SVR-GS보다 상대적으로 우수한 결과를 나타내었으며, 모델 효율성 측면에서 SVR-PSO 및 SVR-ABC는 가장 우수한 모델 성능을 나타내었다. 따라서 본 연구에서 적용한 매개변수 최적화 알고리즘은 SVR의 매개변수를 최적화하는데 효과적임을 확인할 수 있었다. SVR과 최적화 알고리즘을 이용한 하천수위 예측모델은 기존의 ANN 및 ANFIS 모델과 더불어 하천수위 예측을 위한 효과적인 도구로 사용될 수 있을 것으로 판단된다.

• 전자공학회논문지B
• /
• 제31B권1호
• /
• pp.81-87
• /
• 1994

This paper describes the recognition of the printed Hanguel(Korean Character) using Neural Network. In this study, Neural network is used in only specific classification. Hanguel is classified globally by using template matching. Neural network is learned using the segmented grapheme. The grapheme of Hanguel is segmented using the structural method. Neural network is constructed, which is corresponded to the kind and the shape of graphemes. Each neural network is multi layer perceptron. The learning algorithm is the modified error back propagation using descending epsilon method. With five test character sets, the recognition rate of 94.95% is obtained.

• 한국해양공학회지
• /
• 제13권4호통권35호
• /
• pp.124-131
• /
• 1999

In this paper, we show the numerical and experimental results of two-dimensional fluid motions inside a rectangular container with a vertical plate subjected to a background rotation added by a rotational oscillation. In the PIV experiment we apply a new algorithm, NTSS, to the velocity calculation. In the numerical computation, the linear Ekman-pumping model was used to take the bottom friction effect into account. It was found that it showed good agreement with the experimental results at low ${\epsilon}$ number.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제14권2호
• /
• pp.93-111
• /
• 2010

The purpose of this paper is to obtain new complexity results for a second-order cone optimization (SOCO) problem. We define a proximity function for the SOCO by a kernel function. Furthermore we formulate an algorithm for a large-update primal-dual interior-point method (IPM) for the SOCO by using the proximity function and give its complexity analysis, and then we show that the new worst-case iteration bound for the IPM is $O(q\sqrt{N}(logN)^{\frac{q+1}{q}}log{\frac{N}{\epsilon})$, where $q{\geqq}1$.

• 동력기계공학회지
• /
• 제3권4호
• /
• pp.16-21
• /
• 1999

Numerical simulation on two-dimensional turbine cascade flow has been performed using compressible Navier-Stokes equations. The flow equations are written in a cartesian coordinate system, then mapped into a generalized body-fitted ones. All direction of viscous terms are incoporated and turbulent effects are modeled using the extended ${\kappa}-{\epsilon}$ model. Equations are discretized using control volume SIMPLE algorithm on the nonstaggered grid sysetem. Applications are made at a VKI turbine cascade flow in atransonic wind-tunnel and compared to experimental data. Present numerical results are shown to be in good agreement with the experimental results and simulate the compressible viscous flow characteristics inside the turbine blade passage.

• 대한기계학회논문집
• /
• 제17권8호
• /
• pp.2110-2121
• /
• 1993

A numerical investigation has been carried out for the flow characteristics after exchange of some filters from the original layer to the new low pressure loss layer with equal filtering efficiency. The solution domain includes upper plenum, filter layer, clean space, access panels, and lower plenum. The concept of the distributed pressure resistance was applied to describe the momentum loss in filter layer and access panels. The evolution of the flow field is simulated using the low Reynolds number k-.epsilon. over bar turbulent model and SIMPLE algorithm based on the finite volume method. As a result, after the exchange of filter layer the power requirement can be reduced by 8-9 percent. The results also demonstrate that the perpendicularity of the flow near access panels may become worse at new filter layer than origianl filter layer. But the situation can be recovered by adjusting the jopening ratio of access panels.

• 유체기계공업학회:학술대회논문집
• /
• 유체기계공업학회 1998년도 유체기계 연구개발 발표회 논문집
• /
• pp.27-32
• /
• 1998

This paper presents the development of a two-dimensional model for investigating the fluid flow in water jet and calculating the velocity and pressure distributions. The mathematical formulation as a standard k-$\epsilon$ model was solved employing a general thermofluid-mechanics computer program, PHOENICS code, which is based on the Semi-Implicit Method Pressure Linked Equations(SIMPLE) algorithm. The developed code was applied to water jet design to determine the nozzle size, and investigated the effect of the change of nozzle location. Calculated results showed that the flow pattern is not changed as the change of nozzle location.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2007년도 춘계학술대회A
• /
• pp.506-511
• /
• 2007

Mechanical behavior of copper nanowire is investigated. An FCC nanowire model composed of 1,408 atoms is used for MD simulation. Simulations are performed within NVT ensemble setting without periodic boundary conditions. $Nos\acute{e}-Poincar\acute{e}$ MD algorithm is employed to guarantee preservation of Hamiltonian and temperature. Numerical tensile tests of Nanowire are carried out with constant strain rate. Additionally, temperature and strain rate effects are considered. Stress-strain curve is constructed from the calculated Cauchy stresses and specified strain values. In (22,4,4) Copper nanowire, non-linear behavior appears around ${\epsilon}\simeq0.09.$ At this instance, starting of structural reorientations are observed. At the onset of reorientation, the modulus characteristics are also investigated.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2004년도 춘계학술대회 논문집
• /
• pp.685-689
• /
• 2004

Numerical simulation is conducted to clarify the heat transfer and fluid flow characteristics of HVAC(Heating, Ventilating and Air-Conditioning) for double' deck train. The HVAC system is installed under the roof of carbody. In the lay-out of HVAC system, air duct must be installed to supply air to 1st and 2nd floor respectively. The standard k-$\epsilon$ and LES models for turbulence and SIMPLE algorithm for pressure equation hased on finite volume method are used to solve the physic a] HVAC model. To assure convergence, QUICK scheme for momentum equation and the 2nd order upwind scheme for turbulent equations arc used. From the results of simulation, the temperature and velocity magnitude are also distributed uniformly in the interior of double-deck passenger car.

• 호남수학학술지
• /
• 제14권1호
• /
• pp.107-121
• /
• 1992

In this paper, we consider an iterative algorithm for solving a new class of quasi complementarity problems of finding $u{\epsilon}R^{n}$ such that $g(u){\in}K(u)$, $Tu+A(u){\in}K^{*}(u)$, and < g(u), Tu + A(u) >=0, where T, A and g are continuous mappings from $R^{n}$ into itself and $K^{*}(u)$ is the polar cone of the convex cone K(u) in $R^{n}$. The algorithms considered in this paper are general and unifying ones, which include many existing algorithms as special cases for solving the complementarity problems. We also study the convergence criteria of the general algorithms.