• 한국경영과학회지
• /
• 제20권2호
• /
• pp.1-10
• /
• 1995

In this study, a threshold policy is considered for the M/M/2 queueing system with server vacations. The probability generating function for the number of customers present in the system is derived using an embedded Markov chain approach. Then, assuming a linear cost structure, an efficient procedure to find an optimal threshold policy is presented. The Laplace-Stieltjes transofrm for th waiting time of an arbitrary customer under a "FIFO" discipline is also derived.o derived.

• 산업경영시스템학회지
• /
• 제39권3호
• /
• pp.147-155
• /
• 2016

Industry 4.0's goal is the 'Smart Factory' that integrates and controls production process, procurement, distribution and service based on the fundamental technology such as internet of the things, cyber physical system, sensor, etc. Basic requirement for successful promotion of this Industry 4.0 is the large supply of semiconductor. However, company I who produces dicing blades has difficulty to meet the increasing demand and has hard time to increase revenue because its raw material includes high price diamond, and requires very complex and sensitive process for production. Therefore, this study is focused on understanding the problems and presenting optimal plan to increase productivity of dicing blade manufacturing processes. We carried out a study as follows to accomplish the above purposes. First, previous researches were investigated. Second, the bottlenecks in manufacturing processes were identified using simulation tool (Arena 14.3). Third, we calculate investment amount according to added equipments purchase and perform economic analysis according to cost and sales increase. Finally, we derive optimum plan for productivity improvement and analyze its expected effect. To summarize these results as follows : First, daily average blade production volume can be increased two times from 60 ea. to 120 ea. by performing mixing job in the day before. Second, work flow can be smoother due to reduced waiting time if more machines are added to improve setting process. It was found that average waiting time of 23 minutes can be reduced to around 9 minutes from current process. Third, it was found through simulation that the whole processing line can compose smoother production line by performing mixing process in advance, and add setting and sintering machines. In the course of this study, it was found that adding more machines to reduce waiting time is not the best alternative.

• 한국정보처리학회논문지
• /
• 제6권6호
• /
• pp.1502-1510
• /
• 1999

다중처리기 시스템을 위한 적응적 작업할당 방법에서는 작업의 대기시간을 줄이기 위하여 사용 가능한 서브큐브에 맞도록 작업의 크기를 반으로 줄인다. 이 방법에서는 작업을 수행하기에 적합한 서브큐브가 존재하지 않을 경우 작업크기를 축소하여 할당함으로써 대기시간을 줄일 수 있지만, 작업크기 축소로 인해 늘어나게 되는 작업수행시간 때문에 전체작업수행 비용이 오히려 증가될 수도 있다. 본 논문에서는 기존의 적응적 작업할당 방법을 개선한 Estimate-fold 작업할당 방법을 제안한다. 이 방법에서는 작업을 수행하기에 적합한 서브큐브가 존재하지 않으면 적합한 서브큐브가 형성될 때까지 대기하는 경우와 작업크기를 축소하여 할당하는 경우에 대한 작업수행비용을 예측하여 그 중에서 유리한 쪽을 선택한다. 시뮬레이션으로 Estimate-fold 작업할당 방법과 기존의 적응적 작업할당, 그리고 버디, 그레이코드 작업할당 방법들의 평균작업수행비용을 구하여 비교 평가함으로써 제안하는 Estimate-fold 작업할당 방법이 가장 효율적임을 보인다.

• 한국정보통신학회논문지
• /
• 제22권4호
• /
• pp.575-580
• /
• 2018

본 논문에서는 병원에서 환자의 대기시간을 실시간으로 모니터링하기 위한 오픈소스 기반의 시스템을 제안한다. 환자의 위치 데이터를 실시간으로 분석, 처리하기 위한 고성능 스트림 처리 시스템을 비용 효율적으로 구축하기 위해서는 오픈소스 프로젝트를 활용하는 것이 필요하다. 빅데이터를 처리하기 위한 오픈 소스 시스템으로 다양한 하둡 서브프로젝트들로 구성된 하둡 에코시스템이 있다. 본 논문에서는 먼저 시스템 요구사항을 정의하고 하둡 에코시스템에서 이를 만족시키기 위한 몇 가지 오픈소스 프로젝트들을 선정한다. 그리고, 선정된 아파치 스파크, 아파치 카프카 등을 이용한 시스템 구조 설계 및 상세 모듈 설계를 제안한다. 제안된 시스템은 기존 시스템과의 연계 및 오픈소스 프로젝트를 통해 구축비용을 절감할 수 있으며, 또한 분산 스트림 처리를 통해 고성능과 안정성을 확보할 수 있다.

• 대한조선학회논문집
• /
• 제47권2호
• /
• pp.258-264
• /
• 2010

Various kinds of ballast water treatment systems (BWTS) have been developed corresponding to reinforced IMO rules for marine environment. Some of them got the certificate of IMO and others are waiting for it. Selection of optimum BWTS is very important. Optimum BWTS means not only functional requirements but also economic efficiency. This paper presents economic analysis model for optimum BWTS according to ship type and size. In this study 10 kinds of BWTS whose initial installation cost and maintenance cost are known are analyzed based on present worth method. It is assumed that all BWTS satisfy minimum functional requirements and we need to consider different economical efficiency. Through the economic analysis we could select optimum BWTS.

• 대한산업공학회지
• /
• 제21권3호
• /
• pp.345-356
• /
• 1995

The aim of this paper is to analyze the batch service queueing model M/M(a, b ; ${\mu}_k/1$) under general bulk service rule with mean service rate ${\mu}_k$ for a batch of k units, where $a{\leq}k{\leq}b$. This queueing model consists of the two-dimensional state space so that it is characterized by two-dimensional state Markov process. The steady-state solution and performane measure of this process are derived by using Matrix Geometric method. Meanwhile, a new approach is suggested to calculate the two-dimensional traffic density R which is used to obtain the steady-state solution. In addition, to determine the optimal service initiation threshold a, a decision model of this queueing system is developed evaluating cost of service per batch and cost of waiting per customer. In a job order production system, the decision-making procedure presented in this paper can be applicable to determining when production should be started.

• 품질경영학회지
• /
• 제17권2호
• /
• pp.142-148
• /
• 1989

In order to determine the level of service which minimizes the total of expected cost of service and the expected cost of waiting for that service, the important considerations are to evaluate the distance traveled to and from a service facility (D) and the expected number of mechanics in queueing system (L). The travel-time models are very useful when the servers must travel to the customer from the service facility. Thus, in this paper we studied on the determination of the optimal service level by the travel-time models. In order to decide the optimal service level, (D) has been introduced as a uniform distribution and (L) has been introduced as M/M/S model of queueing theory.

• 대한교통학회지
• /
• 제14권3호
• /
• pp.27-44
• /
• 1996

The purpose of the thesis is providing a new formulation for the transit assignment problem. The existing models dealing with the transit assignment problem don't consider the congestion effects due to the insufficient capacity of transit vehicles. Besides, these models don't provide solutions satisfying the Wardrop's user equilibrium conditions. The congestion effects are considered to be concentrated at the transit stops. For the transit lines, the waiting times at the transit stops are dependent on the passenger flows. The new model suggests the route section cost function analogous to the link performance function of the auto assignment to reflect the congestion effects in congested transit network. With the asymmetric cost function, the variational inequality programming is used to obtain the solutions satisfying Wardrop's condition. The diagonalization algorithm is introduced to solve this model. Finally, the results are compared with those of EMME/2.

• 대한산업공학회지
• /
• 제26권2호
• /
• pp.81-87
• /
• 2000

As for M/G/1 queueing system, we use various control policies, with which we can optimize the system. Up to now the most widely adopted policies are N-Policy, T-Policy, D-Policy, and so on. The existing researches are largely concerned to find an optimal operation condition or to optimize the system under single policy in M/G/1 system. There are, however, few literatures dealing with multiple control policies at once to enhance the flexibility of the model. In this study, we consider M/G/1 system adopting N-Policy and T-Policy simultaneously. If one of two conditions is satisfied, then, the server starts the service. We call this Min(N,T)-Policy. We find the probability distribution of the number of customers and mean waiting time in steady state and derive a cost function. Next, we seek the $N^*$, optimal threshold under various N values. Finally, we reveal the characteristics of cost function.

• 한국정보통신학회논문지
• /
• 제24권8호
• /
• pp.1109-1112
• /
• 2020

This paper presents a low cost hardware message passing engine of enhanced atomic pipelined broadcast based on processing node status. In this algorithm, the previous atomic pipelined broadcast algorithm is modified to reduce the waiting time until next broadcast communication. For this, the processor change the transmission order of processing nodes based on the nodes' communication channel. Also, the hardware message passing engine architecture of the proposed algorithm is modified to be adopted to multi-core processor. The synthesized logic area of the proposed hardware message passing engine was reduced by about 16%, compared by the pre-existing hardware message passing engine.