• 정보과학회 컴퓨팅의 실제 논문지
• /
• 제23권6호
• /
• pp.379-385
• /
• 2017

디스크 기반 RAID와 SSD를 결합한 하이브리드 스토리지가 기업 환경에서 널리 이용되고 있다. RAID 상단의 SSD는 주로 데이터 캐시로 사용된다. 최근 SSD를 사용하여 패리티 갱신 비용을 감소시키는 LeavO 캐시 기법이 제안되었으며, 본 논문에서는 데이터 캐시와 LeavO 캐시를 결합한 혼합 캐시기법을 제안한다. 특히 SSD 공간을 데이터 캐시와 LeavO 캐시, 두 영역으로 최적으로 분할하기 위해 비용 모델을 도출하고, 워크로드에 따라 두 영역의 크기를 동적으로 조절하는 적응형 혼합 캐시 기법을 개발하였다. 실험 결과에 따르면 적응형 혼합 캐시 기법은 기존 데이터 캐시 기법에 비해 좋은 성능을 보이며, 오프라인 최적 기법과 대등한 성능을 보인다.

• 응용통계연구
• /
• 제13권2호
• /
• pp.343-352
• /
• 2000

K-평균 군집화(K-means clustering)는 비계층적 군집화 방법이 하나로서 큰 자료에서 개체 군집화에 효율적인 것으로 알려져 있다. 그러나 종종 비교적 균일한 대군집의 일부를 소군집에 떼어주는 오류를 범하기도 한다. 이 연구에서는 그러한 현상을 정확히 인지하고 이에 대한 대책으로서 ‘이중 K-평균 군집화(double K-means clustering)’방법을 제시한다. 또한 실증적 사례에 새 방법론을 적용해보고 토의한다.

• 한국CDE학회논문집
• /
• 제7권2호
• /
• pp.131-139
• /
• 2002

Custom-tailored products always vary their sizes and shapes to satisfy the customers' tastes and requirements but they have to be fabricated as fast as possible when ordered. One way to solve this problem is to prepare several representative products in advance and each custom-tailored product is generated by machining the closest representative product. To realize this approach, it would be necessary to be able to group the products into several groups each of which has a representative product. Once the similar products are identified to be grouped, the representative shape can be generated such that all the products in the group can be made by machining the representative product. The custom-tailored products considered in this work have similar shapes but different sizes. Since these products have free surfaces, that are hard to be compared, their convex hulls are used for the grouping. Among all the products to be grouped, one product is chosen as a base shape. The shape and overall similarity values between the base shape and the remaining shapes are calculated as their convex hulls are rotated virtually. By calculating these similarity values at each rotation, the optimal alignment of the reference shape with respect to the base shape is determined. Overall similarity value at this optimal alignment is used as a measure for grouping. A prototype system based on the proposed methodology has been implemented and used to group the shoe-lasts for custom-tailored shoes.

• 복식
• /
• 제63권3호
• /
• pp.95-109
• /
• 2013

This thesis was conducted with the purpose of proposing a systematic and comprehensive system for women's upper garment sizes so that the satisfaction level of women purchasing the upper garment products in Internet shopping malls is enhanced. To achieve this, this study first conducted a survey of women from the ages of 18 to 39 and attempted to discover consumer satisfaction levels and preferences of the clothing product sizing system of Internet shopping malls. While keeping track of the global distribution environment, an optimal clothing sizing system for Korean women that fit recent changes in their body shapes was proposed. The results of this study are as follows. First, A result of studying the satisfaction levels and preferences of consumer's purchase experience and the sizing system showed that 48.6% of the total respondents were dissatisfied with the current sizing system. Second, based on the research of the size classification system of domestic and foreign upper garment for women, unlike domestic Internet shopping malls, overseas generally offer several size classifications. Third, results of studies 1 and 2 was used to propose an optimal clothing products sizing system method. Also, the body and product sizes and the measurement methods should be offered together. In summary of all these results, by establishing globally compatible sizing system, consumers are able to recognize their sizes on their own and by doing this, it will lower perceived risk of the consumers at the time of a Internet shopping mall purchase, and this will raise their level of satisfaction while making purchases.

• 대한교통학회지
• /
• 제18권3호
• /
• pp.55-76
• /
• 2000

실시간 통행시간관련자료의 집계시간간격은 보다 신뢰성있는 통행시간정보제공과 교통정보센터의 효율적인 운영을 위해 매우 중요한 요소이다. 그러나 대부분의 기존 VDS 및 TCS교통정보 데이터는 통계학적·공학적 차원에서의 합리적인 연구나 검증없이 경험적 간격으로 집계되고 있다. 본 연구의 목적은 링크 및 교통축(Corridor) 통행시간 산정 및 예측시의 최적 집계 시간간격을 결정할 수 있는 통계학적 모형을 개발하고 실제 도로망에서 수집되는 통행시간자료에 적용하는 것이다 첫째로, 본 연구는 링크 및 교통축 통행시간 산정 및 예측으로 인한 오차를 계량화하는 통계학적 모형을 제시하고, 제시된 모형의 의미를 교통류이론 측면과 통행시간정보 이용자측면에서 살펴보았다. 둘째로, 미국 Texas, Houston의 도시고속도로에서 AVI시스템을 통해 수집된 통행시간자료를 제시된 모형에 적용하였다. 적용결과 링크통행시간 산정을 위한 최적 집계시간간격보다 링크통행시간예측을 위한 최적 집계시간간격이 큰 것으로 나타났으며, 교통축 통행시간 산정 및 예측을 위한 최적 집계시간간격은 교통축을 구성하는 링크간의 상관관계 (Correlation)에 큰 영향을 받는 것으로 분석되었다.

• 한국식품과학회지
• /
• 제36권5호
• /
• pp.733-739
• /
• 2004

과실의 연화 및 변색이 가장 적게 발생되는 과실 크기별 봉지 규격을 구명하고 이와 더불어 연화 및 변색과 상관관계를 가지는 봉지 내 산소와 이산화탄소 농도와의 관계를 이용하여, 최적 기체 농도가 이루어지도록 PE 필름 봉지의 적정 규격을 모델링하고자 하였다. 다섯 등급으로 선과된 과실(LL:261g, L:217g, M:188g, S:168g, SS:154g)을 작업성과 미관을 고려하여 과실과 봉지 사이의 간격을 1-1.5cm 되게 PE 필름 봉지를 적용시켰을 때, LL은 150mm, L은 140mm, M과 S는 130mm, SS는 120mm의 봉지 폭에 각각 376mm, 357mm, 344mm, 333mm, 318mm의 봉지 길이가 적절하였다. 연화 및 흑변과 갈변 발생을 최소화하는 봉지 내 산소와 이산화탄소 농도는 각각 0.5-1.0%와 6.0-8.0%였으며, 이를 유지할 수 있는 과실 크기별 PE 필름 봉지의 최적두께는 LL은 $45{\mu}m$, L은 $50{\mu}m$, M은 $55{\mu}m$, S는 $60{\mu}m$, SS는 $65{\mu}m$로 조사되었다. 주어진 과실 크기에 맞는 봉지의 면적$(mm^2)$은 과실의 중량(g)에 대한 2차 단순회귀$(Y=-4055.707+627.993X_1-0.701{X_1}^2)$가 매우 적합하였고, 이 때 PE 필름 봉지 내에 원하는 수준의 산소나 이산화탄소 농도(Y, %)를 유지할 수 있는 PE 필름의 두께를 구하는 회귀식은 과실 중량$(X_1,\;g)$과 봉지 두께$(X_2,\; {\mu}m)$의 선형 중회귀$(O_2:Y=5.798-0.0109X_1-0.0491X_2,\;CO_2:Y=-2.427+0.01927X_l+0.09646X_2)$가 가장 적합하였다.

• 대한산업공학회지
• /
• 제12권1호
• /
• pp.31-41
• /
• 1986

In this paper, an integrated production inventory model is developed for two products on a single facility in which the raw materials are fed to the production system as required at the facility. The aim is to determine simultaneously the optimal production schedule for two products and the optimal procurement policies for the raw materials in a way that minimizes the total variable cost of the production system. The production schedule for the more frequently set-up product is formulated by the method of equal or unequal lot sizes. For the model developed, we present a computational scheme of finding the optimal policies and carry out sensitivity analysis through example problem.

• 한국멀티미디어학회논문지
• /
• 제13권6호
• /
• pp.817-824
• /
• 2010

Linear Discriminant Analysis (LDA) has been successfully applied for dimension reduction in face recognition. However, LDA requires the transformation of a face image to a one-dimensional vector and this process can cause the correlation information among neighboring pixels to be disregarded. On the other hand, 2D-LDA uses 2D images directly without a transformation process and it has been shown to be superior to the traditional LDA. Nevertheless, there are some problems in 2D-LDA. First, it is difficult to determine the optimal number of feature vectors in a reduced dimensional space. Second, the size of rectangular windows used in 2D-LDA makes strong impacts on classification accuracies but there is no reliable way to determine an optimal window size. In this paper, we propose a new algorithm to overcome those problems in 2D-LDA. We adopt an ensemble approach which combines several classifiers obtained by utilizing various window sizes. And a practical method to determine the number of feature vectors is also presented. Experimental results demonstrate that the proposed method can overcome the difficulties with choosing an optimal window size and the number of feature vectors.

• Steel and Composite Structures
• /
• 제29권1호
• /
• pp.39-51
• /
• 2018

Position of a circular or elliptical cutout within an orthotropic plate has great influence on its buckling behavior. This paper aims at finding the optimal position (both location and orientation) of a single circular/elliptical cutout, within an orthotropic rectangular plate, that maximizes the critical buckling load. We consider linear buckling of simply supported orthotropic plates under uniaxial edge loads. To obtain the optimal positions of the cutouts, we have employed a MATLAB optimization routine coupled with buckling computation in ANSYS. Our results show that the position of the cutout that maximizes the buckling load has great dependence on the material properties, laminate configurations, and the geometrical parameters of the plate. These optimal results, for a number of plate geometries and cutout sizes, are reported in this paper. These results will be useful in the design of perforated orthotropic plates against buckling failure.

• Advances in Energy Research
• /
• 제4권2호
• /
• pp.163-176
• /
• 2016

Rapid depletion of fossil based oil, coal and gas reserves and its greater demand day by day necessitates the search for other alternatives. Severe environmental impacts caused by the fossil fire based power plants and the escalating fuel costs are the major challenges faced by the electricity supply industry. Integration of Distributed Generators (DG) especially, wind and solar systems to the grid has been steadily increasing due to the concern of clean environment. This paper focuses on a new simple and fast load flow algorithm named Backward Forward Sweep Algorithm (BFSA) for finding the voltage profile and power losses with the integration of various sizes of DG at different locations. Genetic Algorithm (GA) based BFSA is adopted in finding the optimal location and sizing of DG to attain an improved voltage profile and considerable reduced power loss. Simulation results show that the proposed algorithm is more efficient in finding the optimal location and sizing of DG in 15-bus radial distribution system (RDS).The authenticity of the placement of optimized DG is assured with other DG placement techniques.