• 한국지능시스템학회논문지
• /
• 제11권7호
• /
• pp.579-583
• /
• 2001

In the area of data clustering in high dimensional space, one of the difficulties is the time-consuming process for computing vector similarities. It becomes worse in the case of the agglomerative algorithm with the group-average link and mean centroid method, because the cluster similarity must be recomputed whenever the cluster center moves after the merging step. As a solution of this problem, we present an incremental method of similarity computation, which substitutes the scalar calculation for the time-consuming calculation of vector similarity with several measures such as the squared distance, inner product, cosine, and minimum variance. Experimental results show that it makes clustering speed significantly fast for very high dimensional data.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2008년도 학술대회
• /
• pp.334-336
• /
• 2008

A novel adaptive mesh refinement (AMR) strategy based on the Moment-of-Fluid (MOF) method for volume-tracking dynamic interface computation is presented. The Moment-of-Fluid method is a new interface reconstruction and volume advection method using volume fraction as well as material centroid. The mesh refinement is performed based on the error indicator, the deviation of the actual centroid obtained by interface reconstruction from the reference centroid given by moment advection process. Using the AMR-MOF method, the accuracy of volume-tracking computation with evolving interfaces is improved significantly compared to other published results.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2008년 추계학술대회논문집
• /
• pp.334-336
• /
• 2008

A novel adaptive mesh refinement (AMR) strategy based on the Moment-of-Fluid (MOF) method for volume-tracking dynamic interface computation is presented. The Moment-of-Fluid method is a new interface reconstruction and volume advection method using volume fraction as well as material centroid. The mesh refinement is performed based on the error indicator, the deviation of the actual centroid obtained by interface reconstruction from the reference centroid given by moment advection process. Using the AMR-MOF method, the accuracy of volume-tracking computation with evolving interfaces is improved significantly compared to other published results.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2003년도 춘계학술대회 논문집
• /
• pp.1545-1548
• /
• 2003

In rigid-plastic finite element method, there is a heavy computation time and convergence problem. In this study. static-explicit rigid-plastic finite element method will be introduced. This method is the way that restrict the convergence interval. In result, convergence problem and computation time due to large non-linearity in the existing numerical analysis method were no longer a critical problem. Also, we investigated the effect of punch stroke and the strain increment this method. It is expected that various results from the numerical analysis will give very useful information for the design of tools in sheet metal forming process.

• 한국전산유체공학회지
• /
• 제14권1호
• /
• pp.18-23
• /
• 2009

A novel adaptive mesh refinement(AMR) strategy based on the Moment-of-Fluid(MOF) method for volume-tracking dynamic interface computation is presented. The Moment-of-Fluid method is a new interface reconstruction and volume advection method using volume fraction as well as material centroid. The adaptive mesh refinement is performed based on the error indicator, the deviation of the actual centroid obtained by interface reconstruction from the reference centroids given by moment advection process. Using the AMR-MOF method, the accuracy of volume-tracking computation with evolving interfaces is improved significantly compared to other published results.

• 한국정밀공학회지
• /
• 제16권4호통권97호
• /
• pp.26-36
• /
• 1999

A general procedure for determining the optimum location of suspension hard points with respect to kinematic design parametes is presented. Suspensions are modeled as connection of rigid bodies by ideal kinematic joints. Constraint equations of the kinematic joints are expressed in terms of the generalized coordinates and hard points. By directly differentiating the constraint equations with respect to the hard points, kinematic sencitivity equations are obtained. In order to cope with algebraic complexity associated with the differentiation process, a symbolic computation technique is used. A performance index is defined in terms of static design parameters such as camber, caster, toe, ect.. Gradient of the performance index can be analytically computed from the kinematic sensitivity equations. Optimization results show the effectiveness and validity of the procedure, which is applicable to any type of suspension if its kinematic configurations are given.

• 한국시뮬레이션학회:학술대회논문집
• /
• 한국시뮬레이션학회 1999년도 추계학술대회 논문집
• /
• pp.75-80
• /
• 1999

This paper presents a new method to find a optimal solution for stochastic system. This method uses Genetic Algorithm(GA) and simulation. GA is used to search for new alternative and simulation is used to evaluate alternative. The stochastic system has one or more random variables as inputs. Random inputs lead to random outputs. Since the outputs are random, they can be considered only as estimates of the true characteristics of they system. These estimates could greatly differ from the corresponding real characteristics for the system. We need multiple replications to get reliable information on the system. And we have to analyze output data to get a optimal solution. It requires too much computation to be practical. We address the problem of reducing computation. The procedure on this paper use GA character, an iterative process, to reduce the number of replications. The same chromosomes could exit in post and present generation. Computation can be reduced by using the information of the same chromosomes which exist in post and present current generation.

• 한국정보통신학회논문지
• /
• 제8권4호
• /
• pp.955-959
• /
• 2004

공간 데이터 마이닝을 위한 기존의 클러스터링 기법들은 점 객체만을 대상으로 한다. 즉, 선이나 면 같은 다양한 공간 객체들을 지원하지 못한다. 이것은 클러스터링 과정에서 객체들 간의 거리 계산에 있어서, 점 객체는 용이하지만 선과 면인 경우에는 어렵기 때문이다. 본 논문에서는 이러한 문제점을 해결하기 위하여 균등 격자를 이용한 클러스터링 기법을 설계한다. 세부적으로 이 기법에서는 다각형 객체들 간의 거리 계산을 균등 격자를 이용하여 단순화시킴으로서 거리 계산에 따른 시간과 비용을 줄일 수 있다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1993년도 하계학술대회 논문집 A
• /
• pp.39-41
• /
• 1993

This paper presents the development of interactive graphical software for the educational purpose of power flow(PF) calculation. The developed software is specially designated to give a beginner the interest on PF problem as well as to increase the understanding of it with ease. the software developed in this paper is basically composed of the pull-down menu driver, in which various functions, such as Program Master, Data File Management, Case Study Option, PF Run and View Output, are prepared to handle the software easily and thus to be familiar with power flow calculation. A special design is also considered for interactive operation of the software, wi th which user can interrupt the computation process of PF to control the convergency of PF algorithm, With this function begineer can acquire the understanding of convergency characteristics and numerical sensitivity of PF algorithm as well as basic concept of its computation logic. Futhermore, various graphic illustrations is also provided to review and compare the computation results on monitor.

• Journal of Information Processing Systems
• /
• 제20권2호
• /
• pp.226-238
• /
• 2024

Recently, multi-access edge computing (MEC) has emerged as a promising technology to alleviate the computing burden of vehicular terminals and efficiently facilitate vehicular applications. The vehicle can improve the quality of experience of applications by offloading their tasks to MEC servers. However, channel conditions are time-varying due to channel interference among vehicles, and path loss is time-varying due to the mobility of vehicles. The task arrival of vehicles is also stochastic. Therefore, it is difficult to determine an optimal offloading with resource allocation decision in the dynamic MEC system because offloading is affected by wireless data transmission. In this paper, we study computation offloading with resource allocation in the dynamic MEC system. The objective is to minimize power consumption and maximize throughput while meeting the delay constraints of tasks. Therefore, it allocates resources for local execution and transmission power for offloading. We define the problem as a Markov decision process, and propose an offloading method using deep reinforcement learning named deep deterministic policy gradient. Simulation shows that, compared with existing methods, the proposed method outperforms in terms of throughput and satisfaction of delay constraints.