• Structural Engineering and Mechanics
• /
• 제42권2호
• /
• pp.175-189
• /
• 2012

This paper presents an algorithm for structural reliability with the response surface method. For this aim, an approach with three stages is proposed named as improved response surface method. In the algorithm, firstly, a quadratic approximate function is formed and design point is determined with First Order Reliability Method. Secondly, a point close to the exact limit state function is searched using the design point. Lastly, vector projected method is used to generate the sample points and Second Order Reliability Method is performed to obtain reliability index and probability of failure. Five numerical examples are selected to illustrate the proposed algorithm. The limit state functions of three examples (cantilever beam, highly nonlinear limit state function and dynamic response of an oscillator) are defined explicitly and the others (frame and truss structures) are defined implicitly. ANSYS finite element program is utilized to obtain the response of the structures which are needed in the reliability analysis of implicit limit state functions. The results (reliability index, probability of failure and limit state function evaluations) obtained from the improved response surface are compared with those of Monte Carlo Simulation, First Order Reliability Method, Second Order Reliability Method and Classical Response Surface Method. According to the results, proposed algorithm gives better results for both reliability index and limit state function evaluations.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.631-634
• /
• 2004

In this paper, the Bayesian recurrent neural network (BRNN) is proposed to predict time series data. Among the various traditional prediction methodologies, a neural network method is considered to be more effective in case of non-linear and non-stationary time series data. A neural network predictor requests proper learning strategy to adjust the network weights, and one need to prepare for non-linear and non-stationary evolution of network weights. The Bayesian neural network in this paper estimates not the single set of weights but the probability distributions of weights. In other words, we sets the weight vector as a state vector of state space method, and estimates its probability distributions in accordance with the Bayesian inference. This approach makes it possible to obtain more exact estimation of the weights. Moreover, in the aspect of network architecture, it is known that the recurrent feedback structure is superior to the feedforward structure for the problem of time series prediction. Therefore, the recurrent network with Bayesian inference, what we call BRNN, is expected to show higher performance than the normal neural network. To verify the performance of the proposed method, the time series data are numerically generated and a neural network predictor is applied on it. As a result, BRNN is proved to show better prediction result than common feedforward Bayesian neural network.

• Journal of Communications and Networks
• /
• 제8권1호
• /
• pp.59-69
• /
• 2006

Bluetooth is reputed as a wireless networking technology supplying ad-hoc networks between digital devices. In particular, Bluetooth scatternet is an essential part of dynamic ad-hoc networks. Yet, there have not been sufficient researches performed on scatternet environment. This paper proposes a scatternet reformation algorithm for ad-hoc networks for instances where some nodes enter or leave the scatternet. The proposed algorithm is a general algorithm that can be applied to many types of Bluetooth scatternet regardless of the topology. The proposed algorithm is made for two reformation cases, i.e., nodes leaving and nodes entering. For the reformation when nodes leave a scatternet, the recovery node vector (RNV) algorithm is proposed. It has short reformation setup delay because the process involves a single page process (not including inquiry process). For the reformation when nodes enter a scatternet, the entry node algorithm is proposed. This is a simple and easily implementable algorithm. In this paper, real hardware experiments are carried out to evaluate the algorithm's performance where the reformation setup delay, the reformation setup probability and the data transfer rate are measured. The proposed algorithm has shown improvement in the reformation setup delay and probability.

• 방송공학회논문지
• /
• 제10권3호
• /
• pp.348-356
• /
• 2005

본 논문에서는 H.264로 부호화 된 비트스트림을 이전보다 낮은 비트율을 갖는 비트스트림으로 변환시키기 위한 고속 비트율 감축 트랜스코딩을 제안한다. 제안된 트랜스코딩 기법은, 복호기에서 나온 블록 모드 정보와 움직임 벡터를 이용하여, 확률 기반의 블록 모드 결정 기법과 움직임 벡터 재사용 및 정제 기법을 사용하였다. 그 결과, 직렬 화소 영역 트랜스코딩과 비교하였을 때 약 0.1$\∼$0.3 dB 정도의 화질 저하가 있지만, 부호화 시간을 비교하였을 때 약 40배 빠른 결과를 얻었다.

• Journal of Information Processing Systems
• /
• 제11권3호
• /
• pp.421-437
• /
• 2015

This paper aims to present a supervised classification algorithm based on data fusion for the segmentation of the textured images. The feature extraction method we used is based on discrete wavelet transform (DWT). In the segmentation stage, the estimated feature vector of each pixel is sent to the support vector machine (SVM) classifier for initial labeling. To obtain a more accurate segmentation result, two strategies based on information fusion were used. We first integrated decision-level fusion strategies by combining decisions made by the SVM classifier within a sliding window. In the second strategy, the fuzzy set theory and rules based on probability theory were used to combine the scores obtained by SVM over a sliding window. Finally, the performance of the proposed segmentation algorithm was demonstrated on a variety of synthetic and real images and showed that the proposed data fusion method improved the classification accuracy compared to applying a SVM classifier. The results revealed that the overall accuracies of SVM classification of textured images is 88%, while our fusion methodology obtained an accuracy of up to 96%, depending on the size of the data base.

• Nuclear Engineering and Technology
• /
• 제39권1호
• /
• pp.63-74
• /
• 2007

In this paper, a reconfigurable controller consisting of a normal controller and a standby controller is designed to control the thermoelectric (TE) power in the SP-100 space reactor. The normal controller uses a model predictive control (MPC) method where the future TE power is predicted by using support vector regression. A genetic algorithm that can effectively accomplish multiple objectives is used to optimize the normal controller. The performance of the normal controller depends on the capability of predicting the future TE power. Therefore, if the prediction performance is degraded, the proportional-integral (PI) controller of the standby controller begins to work instead of the normal controller. Performance deterioration is detected by a sequential probability ratio test (SPRT). A lumped parameter simulation model of the SP-100 nuclear space reactor is used to verify the proposed reconfigurable controller. The results of numerical simulations to assess the performance of the proposed controller show that the TE generator power level controlled by the proposed reconfigurable controller could track the target power level effectively, satisfying all control constraints. Furthermore, the normal controller is automatically switched to the standby controller when the performance of the normal controller degrades.

• 대한산업공학회지
• /
• 제42권1호
• /
• pp.38-49
• /
• 2016

This paper proposes a mathematical model-based solution for sourcing decisions with an objective of minimizing the manufacturer's total cost in the two-echelon supply chain with supply capacity risk. The risk impact is represented by uniform, beta, and triangular distributions. For the mathematical model, the probability vector of normal, risk, and recovery statuses are developed by using the status transition probability matrix and the equations for estimating the supply capacity under risk and recovery statuses are derived for each of the three probability distributions. Those formulas derived are validated using the sampling method. The results of the simulation study on the test problem show that the sourcing decisions using the proposed solution reduce the total cost by 1.6~3.7%, compared with the ones without a consideration of supply capacity risk. The total cost reduction increases approximately in a linear fashion as the probability of risk occurrence or reduction rate of supply capacity due to risk events is increased.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권4호
• /
• pp.1263-1274
• /
• 2021

In online competitive social networks, each user can be influenced by different competing influencers and consequently chooses different products. But their interest may change over time and may have swings between different products. The existing influence spreading models seldom take into account the time-related shifts. This paper proposes a minimum cost influence maximization algorithm based on the competitive transition probability. In the model, we set a one-dimensional vector for each node to record the probability that the node chooses each different competing influencer. In the process of propagation, the influence maximization on Competitive Linear Threshold (IMCLT) spreading model is proposed. This model does not determine by which competing influencer the node is activated, but sets different weights for all competing influencers. In the process of spreading, we select the seed nodes according to the cost function of each node, and evaluate the final influence based on the competitive transition probability. Experiments on different datasets show that the proposed minimum cost competitive influence maximization algorithm based on IMCLT spreading model has excellent performance compared with other methods, and the computational performance of the method is also reasonable.

• 한국음향학회지
• /
• 제13권4호
• /
• pp.40-49
• /
• 1994

본 논문에서는 단어 인식 시스템의 성능 개선을 위하여 다음과 같은 두가지 방법을 제안한다. 첫번째 방법은 VQ 코드간의 천이를 안정화시키기 위하여 음성신호의 특징벡터 시퀀스에 관성을 적용하는 방법이고, 두번째 방법은 이산 HMM 모델에서 인접 프레임 간의 시간 상관성을 고려하기 위하여 VQ 코드의 천이행렬을 출력 심벌의 관측확률에 가중치로 이용하여 새로운 관측확률을 발생하는 방법이다. 특징벡터 시퀀스에 관성을 도입함으로서, SOFM상의 각 단어에 대한 반응경로에서 확률분포가 중첩되는 것을 억제하여 HMM의 상태천이를 안정화 시킬 수 있다. 기존의 이산 HMM에 VQ 코드의 천이행렬을 가중치로 적용함으로써, 특징벡터의 확률분포를 더욱 세분화하고, 특징분포를 적당한 영역으로 제한함으로써 인식시스템의 성능을 개선할 수 있다. 제안한 방법을 평가하기 위하여 50개의 DDD 지역명을 대상으로 인식 실험을 수행하였다. 실험 결과에 의하면, 제안된 방법이 기존의 HMM 모델에 비해 화자종속 실험에서는 $4.2\%$의 인식률 향상과 화자 독립 실험에서는 $12.45\%$의 인식률 향상을 얻을 수 있었다.

• 정보보호학회논문지
• /
• 제30권4호
• /
• pp.537-544
• /
• 2020

악성코드 기술 발전으로 변이, 난독화 등의 탐지 회피 방법이 고도화되고 있다. 이에 악성코드 탐지 기술에 있어 알려지지 않은 악성코드 탐지 기술이 중요하며, 배포된 악성코드를 통해 저자를 식별하여 알려지지 않은 악성코드를 탐지하는 악성코드 저자 식별 방법이 연구되고 있다. 본 논문에서는 바이너리 기반 저자 식별 방법에 대해 중요 정보인 컴파일러 정보를 추출하고자 하였으며, 연구 간에 특징 선택, 확률 및 비확률 모델, 최적화가 분류 효율성에 미치는 민감성(Sensitive)을 확인하고자 하였다. 실험에서 정보 이득을 통한 특징 선택 방법과 비확률 모델인 서포트 벡터 머신이 높은 효율성을 보였다. 최적화 연구 간에 제안하는 프레임워크를 통한 특징 선택 및 모델 최적화를 통해 높은 분류 정확도를 얻었으며, 최대 48%의 특징 감소 및 51배가량의 빠른 실행 속도라는 결과를 보였다. 본 연구를 통해 특징 선택 및 모델 최적화 방법이 분류 효율성에 미치는 민감성에 대해 확인할 수 있었다.