• 대한기계학회논문집A
• /
• 제28권4호
• /
• pp.475-480
• /
• 2004

Function approximation is one of the most important and active research fields in design optimization. Accurate function approximations can reduce the repetitive computational effort fur system analysis. So this study presents an enhanced two-point diagonal quadratic approximation method. The proposed method is based on the Two-point Diagonal Quadratic Approximation method. But unlike TDQA, the suggested method has two quadratic terms, the diagonal term and the correction term. Therefore this method overcomes the disadvantage of TDQA when the derivatives of two design points are same signed values. And in the proposed method, both the approximate function and derivative values at two design points are equal to the exact counterparts whether the signs of derivatives at two design points are the same or not. Several numerical examples are presented to show the merits of the proposed method compared to the other forms used in the literature.

• 대한기계학회논문집A
• /
• 제35권3호
• /
• pp.259-266
• /
• 2011

본 논문에서는 SD-TDQAO (Sequential Dual - Two-point Diagonal Quadratic Approximate Optimization)라는 쌍대기법을 이용한 순차적 최적설계 알고리즘을 제안한다. 이 방법은 비선형 목적함수와 제한조건이 포함되어 있는 공학적인 문제를 효과적으로 풀 수 있도록 하는데 목적이 있다. 기존의 볼록성과 분리성이 만족되지 않는 eTDQA2 방법을 이용하여 쌍대기법에 이용할 수 있도록 이차 근사함수의 헤시언 대각요소에 이를 적용하여 쉽게 볼록성과 분리성을 보장할 수 있도록 하였다. 또한 이를 수학적 예제와 위상 최적설계문제를 통해 기존의 쌍대기법 알고리즘인 MMA 와의 비교로 그 성능을 입증하였다.

• 한국콘크리트학회:학술대회논문집
• /
• 한국콘크리트학회 2010년도 춘계 학술대회 제22권1호
• /
• pp.3-4
• /
• 2010

대각선 철근이 있는 PC 벽체의 강도를 단면해석을 이용해 구하고자 한다. 대각선 철근의 변형률은 PC 벽체의 주철근의 변형률을 이용해 구할 수 있으며 이러한 변형은 벽체에 추가적인 인장력을 발생시킨다. 대각선 철근에 의한 추가 인장력은 주철근 간격의 1/3지점에서 발생한다고 가정한다.

• Journal of Mechanical Science and Technology
• /
• 제15권9호
• /
• pp.1257-1267
• /
• 2001

For efficient mechanical system optimization, a new two-point approximation method is presented. Unlike the conventional two-point approximation methods such as TPEA, TANA, TANA-1, TANA-2 and TANA-3, this introduces the shifting level into each exponential intervening variable to avoid the lack of definition of the conventional exponential intervening variables due to zero-or negative-valued design variables. Then a new quadratic approximation whose Hessian matrix has only diagonal elements of different values is proposed in terms of these shifted exponential intervening variables. These diagonal elements are determined in a closed form that corrects the typical error in the approximate gradient of the TANA series due to the lack of definition of exponential type intervening variables and their incomplete second-order terms. Also, a correction coefficient is multiplied to the pre-determined quadratic term to match the value of approximate function with that of the previous point. Finally, in order to show the numerical performance of the proposed method, a sequential approximate optimizer is developed and applied to solve six typical design problems. These optimization results are compared with those of TANA-3. These comparisons show that the proposed method gives more efficient and reliable results than TANA-3.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2001년도 춘계학술대회논문집C
• /
• pp.386-391
• /
• 2001

This paper presents a new two-point approximation method based on the exponential intervening variable. To avoid the lack of definition of the conventional exponential intervening variables due to zero- or negative-valued design variables the shifting level into each exponential intervening variable is introduced. Then a new quadratic approximation, whose Hessian matrix has only diagonal elements of different values, is proposed in terms of these intervening variables. These diagonal elements are computed in a closed form, which correct the typical error in the approximate gradient of the TANA series due to the lack of definition of exponential type intervening variables and their incomplete second-order terms. Also, a correction coefficient is multiplied to the pre-determined quadratic term to match the value of approximate function with that of the original function at the previous point. Finally, the authors developed a sequential approximate optimizer, solved several typical design problems used in the literature and compared these optimization results with those of TANA-3. These comparisons show that the proposed method gives more efficient and reliable results than TANA-3.

• 대한기계학회논문집A
• /
• 제25권9호
• /
• pp.1423-1431
• /
• 2001

Based on the exponential intervening variable, a new two-point approximation method is presented. This introduces the shifting level into each exponential intervening variable to avoid the lack of def inition of the conventional exponential intervening variables due to zero-or negative-valued design variables. Then a new quadratic approximation whose Hessian matrix has only diagonal elements of different values is proposed in terms of these intervening variables. These diagonal elements are determined in a closed form that corrects the typical error in the approximate gradient of the TANA series due to the lack of definition of exponential type intervening variables and their incomplete second-order terms. Also, a correction coefficient is multiplied to the pre-determined quadratic term to match the value of approximate function with that of the previous point. Finally, in order to show the numerical performance of the proposed method, a sequential approximate optimizer is developed and applied to solve six typical design problems. These optimization results are compared with those of TANA-3. These comparisons show that the proposed method gives more efficient and reliable results than TANA-3.

• 한국해양공학회지
• /
• 제21권6호
• /
• pp.7-15
• /
• 2007

A mechanical model was developed to predict the behavior of point-loaded RC slender beams (a/d > 2.5) without stirrups. It is commonly accepted by most researchers that a diagonal tension crack plays a predominant role in the failure mode of these beams, but the failure mechanism of these members is still debatable. In this paper, it was assumed that diagonal tension failure was triggered by the concrete cover splitting due to the dowel action at the initial location of diagonal tension cracks, which propagate from flexural cracks. When concrete cover splitting occurred, the shape of a diagonal tension crack was simultaneously developed, which can be determined from the principal tensile stress trajectory. This fictitious crack rotates onto the crack tip with load increase. During the rotation, all forces acting on the crack (i.e, dowel force of longitudinal bars, vertical component of concrete tensile force, shear force by aggregate interlock, shear force in compression zone) were calculated by considering the kinematical conditions such as crack width or sliding. These forces except for the shear force in the compression zone were uncoupled with respect to crack width and sliding by the proposed constitutive relations for friction along the crack. Uncoupling the shear forces along the crack was aimed at distinguishing each force from the total shear force and clarifying the failure mechanism of RC slender beams without stirrups. In addition, a proposed method deriving the dowel force of longitudinal bars made it possible to predict the secondary shear failure. The proposed model can be used to predict not only the entire behavior of point-loaded RC slender shear beams, but also the ultimate shear strength. The experiments used to validate the proposed model are reported in a companion paper.

• 한국정밀공학회지
• /
• 제25권2호
• /
• pp.43-48
• /
• 2008

This paper proposes a new method for measuring the power of spherical and cylindrical lens with 6 points light source, which is composed of a LED and six holes, and magnification ellipse fitting algorithm. Each measured diagonal length of 6 points light source is determined by the target lens power. After finding the center position of each light point with threshold method, 3 axis-diagonal lengths were calculated. The long axis and short axis power of cylindrical lens can be calculated by using magnification ellipse fitting algorithm with the magnification relationships between the initial diagonal lengths and the measured diagonal lengths changed by lens power.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권7호
• /
• pp.1841-1857
• /
• 2023

CenterNet, a novel object detection algorithm without anchor based on key points, regards the object as a single center point for prediction and directly regresses the object's height and width. However, because the objects have different sizes, directly regressing their height and width will make the model difficult to converge and lose the intrinsic relationship between object's width and height, thereby reducing the stability of the model and the consistency of prediction accuracy. For this problem, we proposed an algorithm based on the regression of the diagonal half-length and the center angle, which significantly compresses the solution space of the regression components and enhances the intrinsic relationship between the decoded components. First, encode the object's width and height into the diagonal half-length and the center angle, where the center angle is the angle between the diagonal and the vertical centreline. Secondly, the predicted diagonal half-length and center angle are decoded into two length components. Finally, the position of the object bounding box can be accurately obtained by combining the corresponding center point coordinates. Experiments show that, when using CenterNet as the improved baseline and resnet50 as the Backbone, the improved model achieved 81.6% and 79.7% mAP on the VOC 2007 and 2012 test sets, respectively. When using Hourglass-104 as the Backbone, the improved model achieved 43.3% mAP on the COCO 2017 test sets. Compared with CenterNet, the improved model has a faster convergence rate and significantly improved the stability and prediction accuracy.

• 대한기계학회논문집A
• /
• 제26권12호
• /
• pp.2483-2491
• /
• 2002

This study presents an efficient method for robust optimal design. In order to avoid the excessive evaluations of the exact performance functions, two-point diagonal quadratic approximation method is employed for approximating them during optimization process. This approximation method is one of the two point approximation methods. Therefore, the second order sensitivity information of the approximated performance functions are calculated by an analytical method. As a result, this enables one to avoid the expensive evaluations of the exact $2^{nd}$ derivatives of the performance functions unlike the conventional robust optimal design methods based on the gradient information. Finally, in order to show the numerical performance of the proposed method, one mathematical problem and two mechanical design problems are solved and their results are compared with those of the conventional methods.