• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제10권2호
• /
• pp.146-151
• /
• 2010

This paper proposes a rank-based control method of mutation probability for improving the performances of genetic algorithms (GAs). In order to improve the performances of GAs, GAs should not fall into premature convergence phenomena and should also be able to easily get out of the phenomena when GAs fall into the phenomena without destroying good individuals. For this, it is important to keep diversity of individuals and to keep good individuals. If a method for keeping diversity, however, is not elaborately devised, then good individuals are also destroyed. We should devise a method that keeps diversity of individuals and also keeps good individuals at the same time. To achieve these two objectives, we introduce a rank-based control method of mutation probability in this paper. We set high mutation probabilities to lowly ranked individuals not to fall into premature convergence phenomena by keeping diversity and low mutation probabilities to highly ranked individuals not to destroy good individuals. We experimented our method with typical four function optimization problems in order to measure the performances of our method. It was found from extensive experiments that the proposed rank-based control method could accelerate the GAs considerably.

• 한국연소학회:학술대회논문집
• /
• 한국연소학회 2013년도 제46회 KOSCO SYMPOSIUM 초록집
• /
• pp.83-85
• /
• 2013

With an increase in the use of low rank coals in power plants, various operational issues were raised in the fuel storage and supply, combustion, boiler and flue gas treatment systems. In the fuel storage and supply system, the main issue is the self-heating propensity of low rank coals leading to spontaneous combustion in yard storage, transport and pulverization. This study evaluated the reactivity of various sub-bituminous and bituminous coals with oxygen at low temperatures by analyzing the temperature increase characteristics of coals under a constant flow rate of oxygen supply. The results were quantified to a self-heating index and the relation with the fuel properties were evaluated.

• 응용통계연구
• /
• 제21권6호
• /
• pp.933-945
• /
• 2008

발병연령과 이에 관련되었다고 의심되는 좌위와의 연관성이 실제로 존재하는 경우에 유전자형 정보에 따라 발병연령(age of onset) 분포의 추세가 뚜렷하게 나타난다. 그러나 연관성 검정에서 주로 채택하고 있는 발병연령의 상한연령(cutoff age)을 제한하는 표본추출은 유전자형에 따른 여러 군의 미완결 자료의 분포가 다름을 초래하게 되며, 이러한 미완결 분포 차이는 발병연령의 추세 검정에 있어 효율성을 낮추는 원인이 된다. 일반적으로 두 군의 경우에 그 대책으로써 윌콕슨(Wilcoxon) 통계량보다는 미완결 자료의 분포가 다름에 영향을 덜 받는다고 알려진 로그순위(log-rank) 통계량을 사용한다. 본 논문에서는 로그순위 통계량 적용을 유전자형에 따른 여러 군의 경우로 확장하여 Jones와 Browley (1989)에 언급된 일반화 로그순위 추세 검정통계량(generalized log-rank statistic for trend)을 제안하며, 연관성 연구에서 이 검정통계량과 여러 다른 추세 검정통계량의 효율성을 모의실험으로 알아본다.

• 한국컴퓨터정보학회논문지
• /
• 제25권4호
• /
• pp.1-9
• /
• 2020

본 논문에서는 데이터 압축을 통해 메모리 시스템의 에너지 효율 및 성능을 향상시키는 동적랭크 서브세팅 기법 (Dynamic Rank Subsetting, DRAS)을 제안한다. DRAS 기법은 하나의 메모리 랭크 (Rank)를 두 개의 서브랭크 (Sub-rank)로 동작되도록 하여, 데이터가 절반 크기로 압축될 경우 압축된 데이터를 하나의 서브랭크에만 저장한다. 이를 통해 DRAS 기법은 압축된 데이터에 대한 읽기 및 쓰기 동작의 메모리 대역폭을 두 배로 높일 수 있고, 동적 전력 소모도 절반으로 감소시킬 수 있다. 만약 데이터가 절반 크기로 압축되지 않는다면 기존 메모리 시스템에서와 같이 데이터를 두 서브랭크에 저장한다. 따라서 DRAS 기법은 데이터가 압축되지 않는 경우에 대해서는 기존 메모리 시스템 수준의 메모리 대역폭과 전력 효율성을 보장한다. 메모리 시뮬레이터를 사용한 실험 평가를 통해 DRAS 기법이 컴퓨터 시스템 성능을 평균 12% 향상시키고 메모리 시스템의 전력소모를 평균 24% 감소시킬 수 있음을 보인다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1997년도 한국자동제어학술회의논문집; 한국전력공사 서울연수원; 17-18 Oct. 1997
• /
• pp.105-108
• /
• 1997

This paper proposes a new approach of balanced model reduction using matrix pencil. The algorithm presented in this paper is to convert full-rank high-order system into rank-deficient system using perturbation made by matrix pencil method. Then the system can be truncated to a low-order system that we want via balanced realization. We discuss the comparison with other methods and the various observations by simulations.

• 호스피스학술지
• /
• 제5권2호
• /
• pp.1-14
• /
• 2005

The purpose of this study was to promote the popularization of hospice services by providing the information about the influences of hospice training on participants. We compared differences of attitudes toward death between those of pre-training and post-training by use of questionnaire. This study involved 137 volunteers participating in the hospice training held by one hospice center located in K-city. The questionnaire was composed of 46 items, 21 items of general information and 25 items of information about attitudes toward death. We applied Park's translated version(1992), originally distributed by Thorson and Powell(1988), in the assessment of participants' changed attitudes toward death. Participants were asked to fill out the questionnaire before and after the hospice training. The data were analged by frequency. ANOVA t-test, paired t-test. The results were as follows, 1.Pre-training items holding a high rank were 'The subject about the future life after death is distressing'(F=5.20), 'No worry about any occurrences around the body after a funeral'(F=4.75), 'Having little interest in being laid in a coffin after death'(F=4.56); items holding a low rank were 'Having little interest in any occurrences around the body after death'(F=2.22), 'Willing to deal with the aftermath of the death in detail'(F=1.94), 'No fear to have an idea to be cancerous'(F=1.72). 2.Post-training items holding a high rank were 'No fear to be dead as a result of sufferings from long-lasting illness'(F=3.18), 'No fear to have an idea to be cancerous'(F=3.16)', Having little interest in the future life after death'(F=3.09); items holding a low rank were 'Unpleasant to be not able to move after death(F=1.74), 'Distressing not to know about the future life after death'(F=1.61), 'The subject about the future life after death is distressing'(F=1.60). 3.There were significant difference in participants' attitudes toward death. Compared with pretraining(F=3.97), more positive were developed in post training(F=2.30). The results of this study indicate that hospice training renders positive regards toward death to the participants. Therefore every human being, who is able to speculate about the meaning of death, has to be exposed to the hospice training. In conclusion, hospice training should be popularized in the near future.

• Geomechanics and Engineering
• /
• 제6권6호
• /
• pp.613-624
• /
• 2014

The fractures in coal are the main migration and output channels of coalbed methane, directly influencing the permeability of the coal seams. It is of great significance to study the effect of fracture distribution in coals on the permeability of coal seam. The development rules of endogenetic and exogenetic fractures are different among various coal lithotypes. There is also difference in the fracture density for the same lithotype with different thicknesses. Through the observation and description of the macroscopic fractures in coal and the origin of fractures in coal, the effect of the coal lithotype and its thickness on fracture development in coal was discussed. It was found through the study that the density of fractures in vitrain band was the maximum for the same coal rank and thickness, followed by clarain band. There were few fractures developed in the durain band. However, the changes of fracture density in three types of bands presented different declining trends for low, medium and high coal rank. There were no fractures developed in the fusain. There were three variation patterns for the fracture densities at the same coal rank and coal lithotype: linear decrease, nonlinear decrease, and first decrease then remaining unchanged. However, the overall trend was that the fracture density decreased with the increase of thickness of coal band for the same coal rank and coal lithotype.

• 방송공학회논문지
• /
• 제26권2호
• /
• pp.125-131
• /
• 2021

최근 CNN(Convolutional Neural Network)은 영상 분류, 객체 인식, 화질 개선 등 다양한 비전 분야에서 우수한 성능을 보여주고 있다. 그러나 많은 메모리와 계산량이 요구되어 모바일 또는 IoT(Internet of Things) 장치와 같은 저전력 디바이스에 적용하기에는 제한이 따른다. 이에, CNN 모델의 임무 성능을 유지하면서 네트워크 모델을 압축하는 연구가 진행되고 있다. 본 논문에서는 행렬 분해 기술인 저계수 행렬 근사(Low-rank approximation)와 CP(Canonical Polyadic) 분해 기법을 결합한 CNN 모델 압축 기법을 제안한다. 제안기법은 하나의 행렬 분해 기법만을 적용하는 기존의 기법과 달리 CNN의 계층 유형에 따라 두 가지 분해 기법을 선택적으로 적용하여 압축 성능을 높인다. 제안기법의 성능 검증을 위하여 영상 분류 CNN 모델인 VGG-16, ResNet50, 그리고 MobileNetV2 모델을 압축하였고, 계층 유형에 따라 두 가지의 분해 기법을 선택적으로 적용함으로써 저계수 행렬 근사 기법만 적용한 경우 보다 1.5 ~ 12.1 배의 동일한 압축률에서 분류 성능이 향상됨을 확인하였다.

• 전기학회논문지
• /
• 제57권10호
• /
• pp.1708-1712
• /
• 2008

This paper describes a linear matrix inequality (LMI) approach to the design of robust low-order power system stabilizers (PSSs), which are used to damp out local-mode oscillations of synchronous generators. The performance of a PSS is expressed as the location of the closed-loop poles, and a single fixed-gain pole-placement controller is synthesized for a wide range of operating conditions. The synthesis results in simultaneous regional pole-placement stabilization. and is formulated as an LMI feasibility problem with a rank condition. A penalty method is applied to solve the rank-constrained LMI problem. Numerical experiments with a single-machine connected to an infinite bus system were performed to demonstrate the proposed method.

• International Journal of Contents
• /
• 제12권3호
• /
• pp.22-28
• /
• 2016

Tensor with missing or incomplete values is a ubiquitous problem in various fields such as biomedical signal processing, image processing, and social network analysis. In this paper, we considered how to reconstruct a dataset with missing values by using tensor form which is called tensor completion process. We applied Tucker factorization to solve tensor completion which was built base on optimization problem. We formulated the optimization objective function using components of Tucker model after decomposing. The weighted least square matric contained only known values of the tensor with low rank in its modes. A first order optimization method, namely Nonlinear Conjugated Gradient, was applied to solve the optimization problem. We demonstrated the effectiveness of the proposed method in EEG signals with about 70% missing entries compared to other algorithms. The relative error was proposed to compare the difference between original tensor and the process output.