• 산업경영시스템학회지
• /
• 제41권2호
• /
• pp.1-8
• /
• 2018

K-means algorithm is one of the most popular and widely used clustering method because it is easy to implement and very efficient. However, this method has the limitation to be used with fixed number of clusters because of only considering the intra-cluster distance to evaluate the data clustering solutions. Silhouette is useful and stable valid index to decide the data clustering solution with number of clusters to consider the intra and inter cluster distance for unsupervised data. However, this valid index has high computational burden because of considering quality measure for each data object. The objective of this paper is to propose the fast and simple speed-up method to overcome this limitation to use silhouette for the effective large-scale data clustering. In the first step, the proposed method calculates and saves the distance for each data once. In the second step, this distance matrix is used to calculate the relative distance rate ($V_j$) of each data j and this rate is used to choose the suitable number of clusters without much computation time. In the third step, the proposed efficient heuristic algorithm (Group search optimization, GSO, in this paper) can search the global optimum with saving computational capacity with good initial solutions using $V_j$ probabilistically for the data clustering. The performance of our proposed method is validated to save significantly computation time against the original silhouette only using Ruspini, Iris, Wine and Breast cancer in UCI machine learning repository datasets by experiment and analysis. Especially, the performance of our proposed method is much better than previous method for the larger size of data.

• 자원환경지질
• /
• 제53권6호
• /
• pp.703-713
• /
• 2020

수리지질학을 포함한 많은 지질학 분야에 있어 지하 매질의 불균질성을 규명하는 것은 큰 중요성을 가진다. 본 연구에서는 최근 Park(2020)에 의해 소개된 바 있는 대수층을 구성하는 매질의 수리물성을 예측할 수 있는 기법의 개념 및 이론을 간단히 소개하고, 가상의 대수층에 해당 기법을 적용하여 기법에 의한 결과들의 다양한 시사점을 도출하였다. 소개하는 기법은 공분산행렬 적응 진화전략이라는 광역최적화 기법을 사용하며, 개념적으로 대수층에 가해지는 수리적 스트레스에 의한 지하수위 변화 자료를 동화하여 대수층 불균질성을 특성화하는 방법론이다. 가상의 대수층의 수리전도도 예측에 개발 기법을 적용한 결과, 총 40000개 미지의 값을 매우 빠른 시간 내에 예측함을 확인하였다. 또한, 예측의 결과는 레퍼런스 수리전도도와 수치적 및 구조적으로 큰 유사성을 보여 예측의 질적 수준이 높음을 확인하였다. 본 연구에서는 매우 제한적인 케이스에 대하여 적용을 실시하였으나, 기법의 추가개발을 통하여 보다 다양한 케이스에의 적용이 가능할 것으로 예상되며 현재 이를 위한 추가 개발이 이루어지고 있는 상황이다. 개발 기법은 수리지질학 분야 뿐만 아니라 다양한 지질학 및 지구물리 분야에 적용될 수 있는 잠재성을 갖추고 있다.

• Structural Engineering and Mechanics
• /
• 제84권4호
• /
• pp.489-502
• /
• 2022

This paper is aimed at developing an optimization-based Finite Element model updating approach for structural damage identification and quantification. A modal flexibility-based error function is introduced, which uses modal assurance criterion to formulate the updating problem as an optimization problem. Because of the inexplicit input/output relationship between the candidate solutions and the error function's output, a robust and efficient optimization algorithm should be employed to evaluate the solution domain and find the global extremum with high speed and accuracy. This paper proposes a new multi-stage Selective Particle Swarm Optimization (SPSO) algorithm to solve the optimization problem. The proposed multi-stage strategy not only fixes the premature convergence of the original Particle Swarm Optimization (PSO) algorithm, but also increases the speed of the search stage and reduces the corresponding computational costs, without changing or adding extra terms to the algorithm's formulation. Solving the introduced objective function with the proposed multi-stage SPSO leads to a smart feedback-wise and self-adjusting damage detection method, which can effectively assess the health of the structural systems. The performance and precision of the proposed method are verified and benchmarked against the original PSO and some of its most popular variants, including SPSO, DPSO, APSO, and MSPSO. For this purpose, two numerical examples of complex civil engineering structures under different damage patterns are studied. Comparative studies are also carried out to evaluate the performance of the proposed method in the presence of measurement errors. Moreover, the robustness and accuracy of the method are validated by assessing the health of a six-story shear-type building structure tested on a shake table. The obtained results introduced the proposed method as an effective and robust damage detection method even if the first few vibration modes are utilized to form the objective function.

• 농촌지도와개발
• /
• 제27권3호
• /
• pp.125-134
• /
• 2020

The regional centrality plays a very important role in national and regional planning and it is measured by data such as people, goods, and information flows among regions. The inter-regional flows are usually considered by only direct flows, yet indirect flows, which are generated accordingly from direct flows, are not critically considered. Most centrality studies have also hardly reflected the indirect flow in the network analysis. This study demonstrates the significance of the indirect flows to enhance accuracy of the regional centrality. The nationwide dataset of inter-regional commuter traffic matrix is used in this study and analysed into two groups; one to consider only direct flow and the other both direct and indirect flows. The results indicate remarkable differences of centrality raking between two groups such as Yeongam of Jeonnam Province(+60th), Eumseong of Chungbuk Province(+57th), Gwacheon of Gyeonggi Province (-35th), and Nowon of Seoul (-32nd). It clearly shows the significant influence of indirect flow for regional centrality study. This also reveals regional centrality ranking in Korea by considering direct and indirect flows of commuters. Jung, Gangnam, and Jongno of Seoul are categorized in the highest rank group and Ulleung of Gyeongbuk, Ongjin of Incheon, and Jindo of Jeonnam are in the lowest group. The top group includes seven districts of Seoul, two of Busan, and one of Gyeonggi Province. The bottom group includes mostly island and costal areas. As this study shows an accurate method of centrality measurement, it has a significant implication to lead an effective regional planning.

• 한국소음진동공학회논문집
• /
• 제15권3호
• /
• pp.251-258
• /
• 2005

The free vibration of a spinning flexible disk-spindle system supported by hydro dynamic bearings (HDB) in an HDD is analyzed by FEM. The spinning flexible disk is described using Kirchhoff plate theory and von Karman non-linear strain, and its rigid body motion is also considered. It is discretized by annular sector element. The rotating spindle which includes the clamp, hub, permanent magnet and yoke, is modeled by Timoshenko beam including the gyroscopic effect. The flexible supporting structure with a complex shape which includes stator core, housing, base plate, sleeve and thrust pad is modeled by using a 4-node tetrahedron element with rotational degrees of freedom to satisfy the geometric compatibility. The dynamic coefficients of HDB are calculated from the HDB analysis program, which solves the perturbed Reynolds equation using FEM. Introducing the virtual nodes and the rigid link constraints defined in the center of HDB, beam elements of the shaft are connected to the solid elements of the sleeve and thrust pad through the spring and damper element. The global matrix equation obtained by assembling the finite element equations of each substructure is transformed to the state-space matrix-vector equation, and the associated eigen value problem is solved by using the restarted Arnoldi iteration method. The validity of this research is verified by comparing the numerical results of the natural frequencies with the experimental ones. Also the effect of supporting structures to the natural modes of the total HDD system is rigorously analyzed.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2003년도 추계학술대회논문집
• /
• pp.572-578
• /
• 2003

The free vibration of a spinning flexible disk-spindle system supported by hydro dynamic bearings in a HDD is analyzed by FEM. The spinning flexible disk is described using Kirchhoff plate theory and von Karman non-linear strain, and its rigid body motion is also considered. It is discretized by annular sector element. The rotating spindle which includes the clamp, hub, permanent magnet and yoke, is modeled by Timoshenko beam including the gyroscopic effect. The flexible supporting structure with a complex shape which includes stator core, housing, base plate, sleeve and thrust pad is modeled by using a 4-node tetrahedron element with rotational degrees of freedom to satisfy the geometric compatibility. The dynamic coefficients of HDB are calculated from the HDB analysis program, which solves the perturbed Raynolds equation using FEM. Introducing the virtual nodes and the rigid link constraints defined in the center of HDB, beam elements of the shaft are connected to the solid elements of the sleeve and thrust pad through the spring and damper element. The global matrix equation obtained by assembling the finite element equations of each substructure is transformed to the state-space matrix-vector equation, and the associated eigenvalue problem is solved by using the restarted Arnoldi iteration method. The validity of this research is verified by comparing the numerical results of the natural frequencies with the experimental ones. Also the effect of supporting structures to the natural modes of the total HDD system is rigorously analyzed.

• 한국지진공학회논문집
• /
• 제3권2호
• /
• pp.87-96
• /
• 1999

시간영역에서의 구조물의 해석은 직접적분법과 모드중첩법에 의하여 구해질 수 있다 그 중에서도 모드중첩법에 의한 해석방법은 몇가지 저차 진동모드를 사용하여 비교적 정확한 해를 구할 수 있기 때문에 동적해석에 널리 사용되고 있다, 그러나 비선형해석에서는 각 부재들의 상태에 따라 강성이 달라지므로 고유 진동모드를 정의할수 없거나 변화되는 강성에 따라 고유진동 모드를 지속적으로 다시 구하여야 하는 불편 있으므로 모드 중첩법을 이용한 비선형해석은 완전탄소성모델 등 극히 제한된 조건에서만 실행이 가능하였다 본논문에서는 강성행렬을 각 부재별로 분리시키고 비선형복원력과 초기선형복언력과 초기선형복원력의 차이를 하중항에 반영시킴으로써 모드중첩법을 이용하여 비선형 해석은완전탄소성모델 등 극히 제한된 조건에서만 실행이 가능하였다 본 논문에서는 강성행렬을 각 부재별로 분리시키고 비선형 복원력과 초기선형복원력의 차이를 하중항에 반영시킴으로써 모드중첩법을 이용하여 비선형해석이 가능한 방법을 제시하고자 한다. 특히 각 부재 강성을 각 부재의 상대변위의 함수로 나타냄으로써 연속적인 계산이 가능하게 하였다 본 논문에서 제시된 방법은 전단보모델에 적용하였으며 모드 개수를 변화시켜 지진하중에 의한 최대변위를 계산하여 이를 직접적분버에 의한 결과와 비교하였다.

• Journal of Animal Science and Technology
• /
• 제62권6호
• /
• pp.884-892
• /
• 2020

Corticosterone is known as a biological stress index in many species including birds. Feather corticosterone concentration (FCC) has increasingly been used as a measure for chronic stress status in broiler chickens. As sample preparation is the first step of analytical process, different techniques of feather matrix disruption need to be validated for obtaining better result in analysing corticosterone extraction. The current study was a validation of pulverizing the feather by bead beater (BB) and surgical scissors (SS) processing prior to corticosterone extraction in feather of broiler chickens. The type of feather processing prior to the hormone extraction may alter the final output. Thereby, finding a standard method according to laboratory facilities is pivotal. This study carried out to determine the effects of feather pulverization methods on the extraction amount of corticosterone in broiler chickens. Feathers were sampled from four weeks old Ross 308 broiler chickens (n = 12 birds). All broiler chickens were kept under the same environmental condition and had access to feed and water. Feather samples were assigned to one of the following processing methods 1) using a BB for pulverizing and 2) using a SS for chopping into tiny pieces. Each sample was duplicated into two wells during enzyme immunoassay (EIA) analysis to improve the accuracy of the obtained data. The results showed lower standard errors and constant output of FCC by using the BB method compared with the SS method. Overall comparison of FCC showed a significantly higher (p < 0.001) amount of the FCC in the BB compared with the SS. Overall, using the BB method is recommended over the SS method for feather processing due to the ability to homogenize a large number of samples simultaneously, ease of use and greater extraction of feather corticosterone.

• 한국융합학회논문지
• /
• 제9권6호
• /
• pp.45-52
• /
• 2018

본 논문은 단일 이미지에서 초해상도 영상 생성을 위해 저해상도 이미지에서 생성한 패치정보를 기반으로 선형보간하는 방법을 제안하였다. 기존의 초해상도 생성 방법인 전역 공간의 회귀 모델을 사용하면 특정 영역에 대해 참조할 정보가 부족하여 일반적으로 품질이 떨어지는 결과가 나타난다. 이러한 결과를 보완하기 위해 제안하는 방법은 초해상도 이미지 생성 과정에서 영상을 패치 단위로 지역을 분할하여 의미있는 정보를 수집하고, 수집된 정보를 기반으로 초해상도 이미지 생성을 위해 확장시킨 이미지 매트릭스 영역의 구성정보를 분석하여 선형 보간 과정을 거치고 패치정보를 대응시켜 탐색한 최적의 패치 정보를 기준으로 선형 보간하는 방법을 제안하였다. 실험을 위해 원본 이미지를 복원된 영상과 PSNR, SSIM으로 비교 평가하였다.

• 지능정보연구
• /
• 제13권4호
• /
• pp.35-44
• /
• 2007

본 논문에서는 인쇄체 숫자를 인식하기 위해 가중 원형 정합(weighted template matching) 방법을 제안한다. 원형 정합은 입력 영상 전체를 하나의 전역적인 특징으로 처리하는데 반해, 제안된 가중 원형 정합은 패턴의 특징이 나타나는 국부적인 영역에 해밍 거리(Hamming distance)의 가중치를 두어 패턴 특징을 강조하여 숫자 패턴의 인식률을 높인다. 또한 원형 정합은 영상을 이진화(binarization)할 때 발생되는 패턴의 윤곽선을 울퉁불퉁하게 만드는 랜덤 노이즈에 취약하다. 이러한 랜덤 노이즈를 제거하기 위해 본 논문에서는 체인 코드 트리밍(chain code trimming)방법을 제안한다. 체인 코드 트리밍은 패턴의 외부 윤곽선과 내부 윤곽선의 체인 코드 분석을 통해 특정 유형의 체인 코드를 제거, 교정함으로써 이루어진다. 실험에서는 기존의 원형 정합을 사용했을 때와 체인 코드 트리밍을 한 후 가중 원형 정합을 사용했을 때의 혼돈 행렬(confusion matrix)을 서로 비교한다. 실험 결과는 본 논문에서 제안한 방법에 의해 인쇄체 숫자의 인식률이 크게 향상된 것을 보인다.