• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.2
• /
• pp.179-184
• /
• 2006

An intuitive understanding of the dynamic pattern generation in asymmetric networks may be considered an essential component in developing models for the dynamic information processing. It has been reported that the neural network with cyclic connections generates multiple limit cycles. The dynamics of discrete time network with cyclic connections has been investigated intensively. However, the dynamics of a cyclic connection neural network in continuous-time has not been well-known due to the considerable complexity involved in its calculation. In this paper, the dynamic behavior of a continuous-time cyclic connection neural network, in which each neuron is connected only to its nearest neurons with binary synaptic weights of ${\pm}1$, has been investigated. Furthermore, the dynamics and stability of the network have been analyzed using a piece-wise linear approximation.

• Journal of the Institute of Convergence Signal Processing
• /
• v.1 no.2
• /
• pp.169-176
• /
• 2000

Deadbeat property is well established in digital control system design in time domain. But in continuous time system, deadbeat is impossible because of it's ripples between sampling points inspite of designs using the related digital control system design theory. But several researchers suggested delay elements. A delay element is made from the concept of finite Laplace Transform. From some specifications such as internal model stability, physical realizations as well as finite time settling, unknown coefficents and poles in error transfer functions with delay elements can be calulted so as to satisfy these specifications. For the application to the real system, robustness property can be added. In this paper, error transfer function is specified with 1 delay element and robustness condition is considered additionally. As the criterion of the robustness, a weighted sensitive function's $H_{infty}$ norm is used. For the minimum value of the criterion, error transfer function's poles are calculated optimally. In this sense, optimal design of the continuous time deadbeat controller is obtained.

• Proceedings of the KSME Conference
• /
• 2000.04a
• /
• pp.672-677
• /
• 2000

This paper compares the two kinds of friction model. The first model is classical stick/slip model. In the stick/slip model, the system is treated to have two different states, namely, stick state or slip state. The second one is continuous model developed by Dahl et. al, namely, Extended Dahl's model. Each model has unique properties, and can be best useful when it is applied on the appropriate system. In this paper, each model is applied on the simple two-block system and the complex automatic transmission system. And the simulation result including simulation accuracy and time required are compared.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.57 no.3
• /
• pp.41-54
• /
• 2015

시간-면적 기법을 이용해 유역의 수문과정을 묘사하는 분산형 (distributed) 연속 (continuous) 강우유출모형인 HYSTAR의 거동특성과 주요 매개변수에 대한 민감도를 분석하였다. 유역의 수문조건에 따른 모형거동의 변화를 분석하기 위해 연속되는 4개의 개별 강우사상에 대한 민감도를 조사하고 비교하였다. 또한, 매개변수의 상호작용이 민감도 분석결과에 미치는 영향을 파악하기 위해 두 가지 서로 다른 기법 (one-factor-at-a-time 과 all-factor-at-a-time 방법)을 이용하여 산정된 민감도를 비교하였다. 분석결과, 모형의 직접유출량, 첨두유량 및 도달시간 모의결과는 유출곡선번호 (curve number)에 가장 민감하게 반응하는 것으로 나타났으며, 토양의 깊이, van Genuchten 식의 매개변수, 작물계수에 큰 영향을 받았다. 한편, 모의된 기저유출량은 토양의 깊이를 비롯하여 van Genuchten 식의 매개변수, 작물계수 (crop coefficient), 이방성계수 (anisotropic coefficient), 유출곡선번호의 변화에 민감하였다. 매개변수에 대한 민감도는 분석에 이용된 강우사상에 따라 다르게 나타났으며, 유역의 토양수분조건에 따라 다르게 거동하는 모형의 중요한 특성을 잘 반영하였다. 두 가지 서로 다른 기법을 이용한 민감도 분석결과는 모의된 직접유출량 및 기저유출량의 변화가 매개변수의 상호작용에 의해 제한될 수 있음을 보여 주었다. 본 연구는 HYSTAR 모형의 매개변수에 대한 민감도 분석을 통해서 해당 모형의 거동을 정량적으로 보여주었고, 이를 통해 모형의 건전성 (soundness)을 입증할 수 있는 하나의 근거를 제시하였다. 본 연구결과는 향후 HYSTAR 모형을 이용한 수문분석 시 보정을 위한 매개변수 선정에 활용될 수 있을 것으로 사료된다. 또한, 본 연구결과에서 나타난 민감도의 수문조건 (또는 선정된 강우사상)에 대한 의존성은 연속유출 모형의 민감도 분석을 위한 강우 사상 선정 및 민감도 분석결과의 해석에 유용한 정보를 제공할 수 있을 것으로 기대된다.

• ETRI Journal
• /
• v.18 no.2
• /
• pp.87-106
• /
• 1996

This paper proposes a new model as a framework for forecasting demand and technological substitution, which can accommodate different patterns of technological change. This model, which we named, "Adaptive Diffusion Model", is formalized from a conceptual framework that incorporates several underlying factors determining the market demand for technological products. The formulation of this model is given in terms of a period analysis to improve its explanatory power for dynamic processes in the real world, and is described as a continuous form which approximates a discrete derivation of the model. In order to illustrate the applicability and generality of this model, time-series data of the diffusion rates for some typical products in electronics and telecommunications market have been empirically tested. The results show that the model has higher explanatory power than any other existing model for all the products tested in our study. It has been found that this model can provide a framework which is sufficiently robust in forecasting demand and innovation diffusion for various technological products.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.26 no.9
• /
• pp.1322-1331
• /
• 1989

This paper proposes a new adaptation algorithm with which a model reference adaptive control can give a local boundedness of the tracking error applied to a continuous-time linear time-invariant single-input single-output plant whose reduced-order model is of relative degree 1 and whose unmodeled dynamics may be represented in a sigular perturbation form. With the addition of an offset term and an extra adaptation structure, this algorithm is shown to have a robustness property in the sense that this gives zero residual tracking errors when the unmodeled dynamics are disappeared.

• Proceedings of the KIEE Conference
• /
• 1998.07b
• /
• pp.611-613
• /
• 1998

A structure of musculotendon model with a fatigue profile is investigated. The Hill-type musculotendon model can predicts the decline in muscle force for a given fatigue profile. It consists of nonlinear activation and contraction dynamics based on the physiological concepts. It is normalized for generalization to deal with the various muscles. Muscle force generated by continuous tetanic electrical monophasic pulsewidth modulation stimulation is decreased in time. A fatigue profile is expressed by a function of intramuscular acidification and applied to the relationship between muscle force and shortening velocity in contraction dynamics. The results of computer simulation are well matched with data in a literature which are isometrically performed for knee extension muscles. Also change in optimal fiber length has an effect only on muscle time, constant not on the steady-state tetanic force.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2002.05a
• /
• pp.378-384
• /
• 2002

In this paper, we propose a mixed integer optimization approach for solving the inventory problem with variable lead time, reorder point, crashing cost and price -quantity discount. Chang and Chang[15] study a continuous review inventory model in which lead time is a decision variable under price-quantity discount. However, their study cannot find the optimal solution due to the flaws in the modeling and the solution procedure. We present a complete procedure to find the optimal solution for the model. In addition to the above contribution, we also apply the minimax distribution free approach to the model to devise a practical procedure which can be used without specific information on demand distribution.

• Journal of Institute of Control, Robotics and Systems
• /
• v.7 no.7
• /
• pp.597-604
• /
• 2001

Recently, the study of the moving sliding mode in the variable structure control is in progress ac-tively. The conventional time-invariant sliding model control can\`t guarantee the sliding mode in the reaching phase, which is robust against the uncertainty. But with the time-varying method, the controller makes the states track the desired trajectories and keeps the sliding mode. Nevertheless, the piecewise continuous method of the past still has the reaching mode. Thus we propose the continuously moving sliding surface by the fuzzy algorithm. The proposed algorithm is made of the fuzzy rule considering both the error and the error velocity, and may apply to the entire phase plane without sacrificing sliding mode. Especially the proposed scheme can rotate tot he slope-decreasing direction, needless to say rotating to the slope-increasing direction. For showing that the proposed controller guarantees the sliding model and ensures the robustness, we apply the proposed method to the two-link robot manipulator simulation.

• Proceedings of the KIEE Conference
• /
• 1983.07a
• /
• pp.135-138
• /
• 1983

This paper studies a variable structure adaptive model following control system which can control a plant in which the parameters of the controlled plant can not be estimated because they vary with time and in which the controlled plant has noise. The values of the feedback gain matrices for given states are obtained the equivalent control law, and the adaptive controller has been designed using the adaptive mechanism which switches the matrices. The adaptive controller minimizes the state error vector, that is, the difference between the state vector of the model and the state vector of the controlled plant. A controlled plant which has time varying parameters, a controlled plant which has only noise, and a controlled plant which has both have been controlled by the designed adaptive controller. The continuous single input-output system has been analysed by computer. This control system may be used to control practical systems by the addition of a microcomputer.