• 디지털융복합연구
• /
• 제17권10호
• /
• pp.251-258
• /
• 2019

앞으로의 시대는 인공지능을 이용한 다양한 분야에 다양한 제품이2 생성될 것이다. 이러한 시대에 인공지능의 학습 방법의 동작 원리를 알고 이를 정확하게 활용하는 것은 상당히 중요한 문제이다. 이 논문은 지금까지 알려진 인공지능 학습 방법을 소개한다. 인공지능의 학습은 수학의 고정점 반복 방법(fixed point iteration method)을 기반으로 하고 있다. 이 방법을 기반으로 수렴 속도를 조절한 GD(Gradient Descent) 방법, 그리고 쌓여가는 양을 누적하는 Momentum 방법, 마지막으로 이러한 방법을 적절히 혼합한 Adam(Adaptive Moment Estimation) 방법 등이 있다. 이 논문에서는 각 방법의 장단점을 설명한다. 특히, Adam 방법은 조정 능력을 포함하고 있어 기계학습의 강도를 조정할 수 있다. 그리고 이러한 방법들이 디지털 신호에 어떠한 영향을 미치는 지에 대하여 분석한다. 이러한 디지털 신호의 학습과정에서의 변화는 앞으로 인공지능을 이용한 작업 및 연구를 수행함에 있어 정확한 활용과 정확한 판단의 기준이 될 것이다.

• 한국전산구조공학회:학술대회논문집
• /
• 한국전산구조공학회 2008년도 정기 학술대회
• /
• pp.20-25
• /
• 2008

The finite element analysis for porous media is severe job because constituents have different physical peoperties, and element's continuity and stability should be considered. Thus, we propose the new mixed finite element method in order to overcome the problems. In this method, multi time step, remeshing step, and sub iteration step are introduced. The multi time step and remeshing step make it possible to satisfy a stability and an accuracy during sub iteration in which global time is determined. Finally, the proposed method is compared with the ABAQUS(2007) software and exact solution(Schiffman 1967) through two dimensional consolidation model.

• 한국산업정보학회논문지
• /
• 제4권4호
• /
• pp.88-93
• /
• 1999

선형독립인 소수의 벡터들을 직교화 시키는 방법인 Cram-Schmidt 알고리즘, 몇 개의 극단적인 고유치를 구하는 방법인 란초스 알고리즘방법 등이, 한번의 반복동안 한번의 동기점만 가지도록 재구성된다. 즉 메시지 패싱 분산시스템에서 모든 프로세서들 사이에 한번의 통신만 요구되도록 알고리즘을 변화시킨다. 이러한 알고리즘들은 재구성되었다고 하며, 전통적인 방법에 비하여 더 나은 병렬성질을 가지게 된다.

• 전기학회논문지
• /
• 제67권2호
• /
• pp.277-284
• /
• 2018

In this paper, a new numerical method of finding the roots of a nonlinear system is proposed, which extends the conventional fixed point iterative method by relaxing the constraints on it. The proposed method determines the real valued roots and expands the convergence region by relaxing the constraints on the conventional fixed point iterative method, which transforms the diverging root searching iterations into the converging iterations by employing the metric induced by the geometrical characteristics of a polynomial. A metric is set to measure the distance between a point of a real-valued function and its corresponding image point of its inverse function. The proposed scheme provides the convenience in finding not only the real roots of polynomials but also the roots of the nonlinear systems in the various application areas of science and engineering.

• Structural Engineering and Mechanics
• /
• 제50권6호
• /
• pp.853-862
• /
• 2014

In this paper, two approximate analytical methods have been applied to forced nonlinear vibration problems to assess a high accurate analytical solution. Variational Iteration Method (VIM) and Perturbation Method (PM) are proposed and their applications are presented. The main objective of this paper is to introduce an alternative method, which do not require small parameters and avoid linearization and physically unrealistic assumptions. Some patterns are illustrated and compared with numerical solutions to show their accuracy. The results show the proposed methods are very efficient and simple and also very accurate for solving nonlinear vibration equations.

• Journal of applied mathematics & informatics
• /
• 제27권3_4호
• /
• pp.475-487
• /
• 2009

This paper studies the iteration method of solving a type of second-order three-point boundary value problem with non-linear term f, which depends on the first order derivative. By using the upper and lower method, we obtain the sufficient conditions of the existence and uniqueness of solutions. Furthermore, the monotone iterative sequences generated by the method contribute to the minimum solution and the maximum solution. And the error estimate formula is also given under the condition of unique solution. We apply the solving process to a special boundary value problem, and the result is interesting.

• Journal of applied mathematics & informatics
• /
• 제27권3_4호
• /
• pp.763-771
• /
• 2009

T.H. Kim, H.K. Xu, [Convergence of the modified Mann's iteration method for asymptotically strict pseudo-contractions, Nonlinear Anal.(2007),doi:l0.l016/j.na.2007.02.029.] proved the strong convergence for asymptotically strict pseudo-contractions by the classical CQ iterative method. In this paper, we apply the monotone CQ iterative method to modify the classical CQ iterative method of T.H. Kim, H.K. Xu, and to obtain the strong convergence theorems for asymptotically strict pseudo-contractions. In the proved process of this paper, Cauchy sequences method is used, so we complete the proof without using the demi-closedness principle, Opial's condition or others about weak topological technologies. In addition, we use a ingenious technology to avoid defining that F(T) is bounded. On the other hand, we relax the restriction on the control sequence of iterative scheme.

• Journal of Electrical Engineering and Technology
• /
• 제9권1호
• /
• pp.178-183
• /
• 2014

This paper presents a motor characteristics analysis method using an equivalent circuit. Motor characteristics analysis via equivalent circuit is very important for designing a high efficiency single phase induction motor. The accuracy of the motor characteristics depends on the accuracy of the parameters, especially saturation factor, which determines the cyclical relationship in the analysis process. Therefore, using the proposed method, the saturation factor was calculated using the iteration routine and numerical technique. The proposed method was verified by comparing the finite element method results and the dynamo test results of manufactured prototype model.

• 대한기계학회논문집A
• /
• 제25권9호
• /
• pp.1444-1451
• /
• 2001

The evolutionary structural optimization(ESO) method has been under continuous development since 1992. The bidirectional evolutionary structural optimization(BESO) method is made of additive and removal procedure. The BESO method is very useful to search the global optimum and to reduce the computational time. This paper presents the ranked bidirectional evolutionary structural optimization(R-BESO) method which adds elements based on a rank, and the performance indicator which can estimate a fully stressed model. The R-BESO method can obtain the optimum design using less iteration number than iteration number of the BESO.

• 대한수학회보
• /
• 제60권5호
• /
• pp.1181-1199
• /
• 2023

For solving complex symmetric positive definite linear systems, we propose a single step real-valued (SSR) iterative method, which does not involve the complex arithmetic. The upper bound on the spectral radius of the iteration matrix of the SSR method is given and its convergence properties are analyzed. In addition, the quasi-optimal parameter which minimizes the upper bound for the spectral radius of the proposed method is computed. Finally, numerical experiments are given to demonstrate the effectiveness and robustness of the propose methods.