• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.04a
• /
• pp.216-219
• /
• 2011

이 기종의 컴퓨팅 자원들로 구성된 데이터센터에서 사용자가 제출한 클라우드 작업의 서비스 품질을 보장하기 위해서는 효율적인 작업할당 기법이 필요하다. 본 논문에서는 작업의 처리시간을 사용자의 QoS 로 정의하고 작업처리시간을 최소화하면서 동시에 모든 사용자에게 차별 없이 동일한 수준의 서비스를 제공할 수 있는 최적 작업할당 문제를 정의한다. 협력적 게임모형인 NBS(Nah Bargaining Solution)을 이용하여 정의한 문제를 해결하였고 이를 이용하여 각 자원에 할당되는 작업부하를 효율적으로 제어함으로써 파레토 최적을 유지할 수 있는 협력적 부하분산 정책을 도출하였다. 또한, 성능 평가를 통해 제안한 기법이 작업처리시간을 최소화하면서도 모든 자원에서의 작업처리시간이 동일함을 보였다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.2
• /
• pp.689-694
• /
• 2018

Recently, fuel consumption efficiency has become the most important issue in the vehicle development process due to the problem of environmental pollution. The air flow patterns of the vehicle body line and rear part are the most important elements affecting the fuel consumption efficiency. Especially, the airflow pattern of the vehicle rear part is the most important design factor to be considered in rear spoiler design. In this paper, the control factors affecting the airflow of the rear spoiler are determined, the airflow sensitivity of these control factors are tested and, then, the optimized control factors to reduce the airflow drag force are proposed. The model of optimized control factors is tested and the values of the optimized control factors are changed by analyzing the S/N ratio and mean value. Finally, the new modified model incorporating the optimized control factors is tested in an air flow tunnel and its ability to decrease the air drag and reduce the cost is verified.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.85-88
• /
• 2003

강화학습의 한가지 방법인 Q-learning은 최근에 Linear Quadratic Regulation(이하 LQR) 문제에 성공적으로 적용된 바 있다. 특히, 시스템 모델의 파라미터에 대한 구체적인 정보없이 적절한 입ㆍ출력만으로 학습을 통해 문제의 해결이 가능하므로 상황에 따라 매우 실용적인 방법이 될 수 있다. 뉴럴-큐 기법은 이러한 Q-learning의 Q-value를 MLP(multilayer perceptron) 신경망의 출력으로 대치시켜, 비선형 시스템의 최적제어 문제를 다룰 수 있게 한 방법이다. 그러나, 뉴럴-큐 기법은 신경망의 구조를 먼저 결정한 후 역전파 알고리즘을 이용해 학습하는 절차를 행하므로, 시행착오를 통해 신경망 구조를 결정해야 한다는 점, 역전파 알고리즘의 적용에 따라 신경망의 연결강도 값들이 지역적 최적해로 수렴한다는 점등의 문제점이 있다. 본 논문에서는 뉴럴-큐 학습의 도구로 KFD회귀를 이용하여 Q 함수의 근사 기법을 제안하고 관련 수식을 유도하였다. 그리고, 모의 실험을 통하여, 제안된 뉴럴-큐 방법의 적용 가능성을 알아보았다.

• Journal of Korean Association for Spatial Structures
• /
• v.4 no.3 s.13
• /
• pp.77-84
• /
• 2004

This study presents an effective stiffness-based optimal technique to control quantitatively lateral drift for tall frameworks using composit member subject to lateral loads. To this end, displacement sensitivity depending on behavior characteristics of tall frameworks is established and approximation concept that preserves the generality of the mathematical programming and can efficiently solve large scale problems is introduced. Specifically, under the 'constant-shape' assumption, resizing techniqe of composite member is developed. Two types of 50 story frameworks are presented to illustrate the features of the quantitative lateral drift control technique proposed in this study.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.37 no.11
• /
• pp.1104-1111
• /
• 2009

In the low altitude, mission of the aircraft is restricted by a variety of threats such as anti-air missiles and terrain obstacles. Especially, aircraft have always a risk of ground collision near terrain. In this study, to effectively solve this problem, we developed the flight path generation algorithm that is considered the terrain avoidance. In this flight path generation algorithm, waypoints that should be passed by the UAV are selected first. The waypoints are located in the middle of the terrain obstacles. Then, physically meaningful waypoints sets are classified by Dijkstra algorithm. The optimal waypoint guidance law based on the optimal control theory is applied to produce trajectory candidates. And finally the minimum control energy trajectory is determined.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.7
• /
• pp.923-929
• /
• 2021

TCP communication and packet communication require transmission control technology to ensure high quality and high reliability. However, in the case of real-time data transmission, an inefficient transmission problem occurs. In order to overcome this problem and transmit the packet reliability, in general, early congestion control using the buffer level as an index was used. Control of the congestion control point and the cancellation point is delayed because the point at which congestion is controlled is based on the buffer level. Therefore, in this paper, not only the buffer level indicator, but also the ideal buffer level, which determines the packet discard probability, is classified so that the transmission rate and buffer level that measure network congestion are close to the level above the optimal setting. As a result, it was shown that the average buffer level can be directly controlled by maintaining the average buffer level by the ideal buffer level set in the experiment to prove the proposed method.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.20 no.5
• /
• pp.661-669
• /
• 2007

Eyer since the Prandtl's experiment in 1934 and X-21 airjet test in 1950 both attempting to reduce drag, it was found that controlling the velocities of surface for extremely fast-moving object in the air through suction or injection was highly effective and active method. To obtain the right amount of suction or injection, however, repetitive trial-and error parameter test has been still used up to now. This study started from an attempt to decide optimal amount of suction and injection of incompressible Navier-Stokes by employing optimization techniques. However, optimization with traditional methods are very limited, especially when Reynolds number gets high and many unexpected variables emerges. In earlier study, we have proposed an algorithm to solve this problem by using step by step method in analysis and introducing SQP method in optimization. In this study, we propose more effective and robust algorithm and techniques in solving flow optimization problem.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.33 no.1
• /
• pp.11-19
• /
• 2005

In order to improve dynamic stall characteristics of an oscillating airfoil, optimal design has been performed for fixed nose droop and Gurney flap. Fixed nose droop is known to be very effective to improve pitching moment characteristics but may cause degeneration of aerodynamic lift at the same time. On the other hand, Gurney flap has the opposite characteristics. For fixed nose droop, location and angle are chosen as design variables, while length is defined as design variable for Gurney flap. Higher order response surface methodology and sensitivity based optimal design method are employed to handle highly nonlinear problem such as dynamic stall. Optimal design has been performed so that lift and pitching moment are simultaneously improved. The design results show that aerodynamic characteristics can be remarkably improved through present design approach and the present passive control method is as good as active control method which combines variable nose droop and Gurney flap.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.36 no.4
• /
• pp.301-304
• /
• 1987

The Optimal Control Problems for the singular system with the Generalized state space model are considered. It is shown that when the system is singular, the dimension can be reduced by coordinate transformation and the equivalent nonsingular system is got. After we have nonsingular system, the solution for the optimal control problem can be got by Riccati equation.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.33 no.12
• /
• pp.491-496
• /
• 1984

The optimal control and filtering problems for the systems with the generalized state space model are considered and the generalized Riccati equation is derived. Also the algorithm for the solution of the generalized algebraic Riccati equation is developed and it is shown that the algotithm can be applied to the case where the matrix R is singular or near singular.