• 대한기계학회논문집A
• /
• 제39권5호
• /
• pp.535-541
• /
• 2015

본 연구에서는 벽걸이 모니터 브라켓 암의 다중목적 근사최적설계를 수행하였다. 이를 위해 브라켓 암의 자유도를 고려하여 평면내의 회전 각도를 선정해 응력과 처짐량이 크게 발생하는 경우에 대한 최적화 문제를 정식화 하였다. 직교배열표와 반응표면법을 사용하여 평균 및 파라미터 분석을 통해 성능지수에 대한 설계변수 민감도를 확인하였으며, 중심합성계획법과 D-최적 계획법을 사용하여 목적함수와 제한조건함수에 대하여 반응표면 근사모델을 생성하고 $R^2$ 값을 통해 정확도를 평가하였다. 이를 비지배 분류 유전알고리즘에 적용하여 최적화를 수행하고 유한요소해석을 통해 검증하였다. 또한, 중심합성 계획법과 D-최적 계획법을 이용한 최적해를 비교 분석하였다.

• 한국콘텐츠학회논문지
• /
• 제22권8호
• /
• pp.33-43
• /
• 2022

지식 그래프 및 단백질 상호 작용과 같은 실제 데이터에서 개체들과 개체들의 관계 및 구조를 나타내기 위해 레이블 그래프를 사용한다. IT의 급속한 발전과 데이터의 폭발적인 증가로 사용자에게 관심 있는 정보를 제공하기 위한 서브 그래프 매칭 기술이 필요하다. 본 논문은 레이블의 의미적 유사성과 그래프 구조 차이를 고려한 근사 Top-k 서브 그래프 매칭 기법을 제안한다. 제안하는 기법은 레이블 의미적 유사도를 고려하기 위하여 FastText을 활용한 학습 모델을 이용한다. 레이블 간 의미적 유사도를 미리 계산한 LSG(Label Similarity Graph)를 통해 처리 속도의 효율을 높인다. LSG를 통해 레이블이 정확하게 일치해야 확장이 가능한 기존 연구의 한계를 해결한다. 2-hop까지 탐색을 수행함으로써 질의 그래프에 대한 구조적 유사성을 지원한다. 매칭된 서브 그래프는 유사도 값 기반으로 Top-k 결과를 제공한다. 제안하는 기법의 우수성을 보이기 위하여 다양한 성능평가를 수행한다.

• Journal of applied mathematics & informatics
• /
• 제36권3_4호
• /
• pp.331-348
• /
• 2018

In this paper, an initial value method for solving a class of singularly perturbed delay differential equations with layer behavior is proposed. In this approach, first the given problem is modified in to an equivalent singularly perturbed problem by approximating the term containing the delay using Taylor series expansion. Then from the modified problem, two explicit Initial Value Problems which are independent of the perturbation parameter, ${\varepsilon}$, are produced: the reduced problem and boundary layer correction problem. Finally, these problems are solved analytically and combined to give an approximate asymptotic solution to the original problem. To demonstrate the efficiency and applicability of the proposed method three linear and one nonlinear test problems are considered. The effect of the delay on the layer behavior of the solution is also examined. It is observed that for very small ${\varepsilon}$ the present method approximates the exact solution very well.

• 대한수학회지
• /
• 제45권5호
• /
• pp.1275-1295
• /
• 2008

In this paper, we consider an impulsive nonlinear second order ordinary differential equation with nonlinear three-point boundary conditions and develop a monotone iteration scheme by relaxing the convexity assumption on the function involved in the differential equation and the concavity assumption on nonlinearities in the boundary conditions. In fact, we obtain monotone sequences of iterates (approximate solutions) converging quadratically to the unique solution of the impulsive three-point boundary value problem.

• Journal of Communications and Networks
• /
• 제10권1호
• /
• pp.63-70
• /
• 2008

Narrow-band interference (NBI) from the coexisting narrow-band services affects the performance of ultra wideband (UWB) systems considerably due to the high power of these narrow-band signals with respect to the UWB signals. Specifically, IEEE 802.11a systems which operate around 5 GHz and overlap the band of UWB signals may interfere with UWB systems significantly. In this paper, we suggest a novel NBI suppression technique based on singular value decomposition (SVD) algorithm in time hopping UWB (TH-UWB) systems. SVD is used to approximate the interference which then is subtracted from the received signals. The algorithm precision and closed-form bit error rate (BER) expression are derived in the wireless multipath channel. Comparing with the conventional suppression methods such as a notch filter and a RAKE receiver, the proposed method is simple and robust and especially suitable for UWB systems.

• 한국정보처리학회논문지
• /
• 제6권5호
• /
• pp.1303-1311
• /
• 1999

In order to learn in dynamic environments, reinforcement learning algorithms like Q-learning, TD(0)-learning, TD(λ)-learning have been proposed. however, most of them have a drawback of very slow learning because the reinforcement value is given when they reach their goal state. In this thesis, we have proposed a reinforcement learning method that can approximate fast to the goal state in maze environments. The proposed reinforcement learning method is separated into global learning and local learning, and then it executes learning. Global learning is a learning that uses the replacing eligibility trace method to search the goal state. In local learning, it propagates the goal state value that has been searched through global learning to neighboring sates, and then searches goal state in neighboring states. we can show through experiments that the reinforcement learning method proposed in this thesis can find out an optimal solution faster than other reinforcement learning methods like Q-learning, TD(o)learning and TD(λ)-learning.

• Kyungpook Mathematical Journal
• /
• 제56권1호
• /
• pp.221-233
• /
• 2016

The present paper proposes a new monotone iteration method for existence as well as approximation of the coupled solutions for a coupled periodic boundary value problem of first order ordinary nonlinear differential equations. A new hybrid coupled fixed point theorem involving the Dhage iteration principle is proved in a partially ordered normed linear space and applied to the coupled periodic boundary value problems for proving the main existence and approximation results of this paper. An algorithm for the coupled solutions is developed and it is shown that the sequences of successive approximations defined in a certain way converge monotonically to the coupled solutions of the related differential equations under some suitable mixed hybrid conditions. A numerical example is also indicated to illustrate the abstract theory developed in the paper.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1994년도 제11회 음성통신 및 신호처리 워크샵 논문집 (SCAS 11권 1호)
• /
• pp.223-228
• /
• 1994

We propose a novel method to incorporate temporal correlations into a speech recognition system based on the conventional hidden Markov model. With the proposed method using the extended logarithmic pool, we approximate a joint conditional PD by separate conditional PD's associated with respective components of conditions. We provide a constrained optimization algorithm with which we can find the optimal value for the pooling weights. The results in the experiments of speaker-independent continuous speech recognition with frame correlations show error reduction by 13.7% with the proposed methods as compared to that without frame correlations.

• 품질경영학회지
• /
• 제21권2호
• /
• pp.156-161
• /
• 1993

Limiting distributions of Studentized test statistics have been shown for testing the slope parameter in a simple linear structural model. Since the limiting distribution of Studentized one appears to yield inaccurate inference, this paper suggests adjustment of critical value and normalization of the Studentized one. As results, we can have procedures for refined inference based on our approximate distrbution instead of the limiting distribution.

• 한국경영과학회지
• /
• 제20권1호
• /
• pp.179-193
• /
• 1995

A multi-objective, single-stage scheduling problem with stochastic processing times is considered where the objective is to simultaneously minimize the expected value and the variance of total flowtime, and the mean probability of tardiness. In cases where processing times follow normal distributions, a method using pairwise interchange of two jobs(PITJ) is proposed to generate a set of the approximate efficient schedules. The efficient schedules are not dominated by the criterion vectors of any other permutation schdules in the feasible region. Numerical experiments performed to ascertain the effectiveness of PITJ algorithm are also reported in the results.