• 대한산업공학회지
• /
• 제40권2호
• /
• pp.215-222
• /
• 2014

This paper provides a real-time loading strategy for batch processing machines in which a number of jobs are simultaneously processed as a batch. The batch processing machines can be seen in both manufacturing industries (e.g., semiconductor, automobile and metal working) and service industries (transportation vehicles, mail shipment and theme park). This paper focuses on batch processing machines in semiconductor manufacturing. We present a look-ahead loading strategy for tardiness minimization where future arrivals and due dates are taken into consideration. Simulation tests are performed on the presented strategy and some existing loading heuristics under various production settings with different traffic intensities and forecasting errors. Experimental results show that our strategy provides the performance of good quality.

• 한국경영과학회지
• /
• 제25권1호
• /
• pp.37-50
• /
• 2000

We are concerned with estimating the average performance of a re-entrant line with single-job machines and batch machines. The system has multiclass jobs, which will be processed in predetermined routes. An analytical approach may be intractable since the system would not be modeled by product form queueing networks due to the inclusion of batch machines and the consideraton of multiclass jobs which have different processing times. We propose an approximation method based on the Mean Value Analysis(MVA). Our method obtains the mean walting time in each buffer of a workstation and the mean cycle time using the MVA and heuristics. numerical experiments show that the errors of our method are within 5% compared with simulation.

• Journal of the Korean Data and Information Science Society
• /
• 제24권5호
• /
• pp.989-998
• /
• 2013

최근 산/학계에서 주목받고 있는 빅 데이터는 정의상 한꺼번에 자료를 메모리에 올려 분석할 수 없기 때문에 기존의 데이터마이닝 시대에 개발된 일괄처리 (batch processing) 방식의 알고리즘을 적용할 수 없게 된다. 따라서 가장 시급히 해결해야 하는 문제는 기존의 여러 가지 기계학습방법을 빅 데이터에 적용할 수 있도록 분산처리 (distributed processing)를 수행하는 적절한 알고리즘을 개발하는 것이라 볼 수 있다. 본 논문에서는 분류문제에서 각광받는 지지벡터기계 (support vector machines)의 여러 알고리즘을 살펴보고자 한다. 특히 빅 데이터 분류문제에 유용할 것으로 예상되는 온라인 타입 알고리즘과 병렬처리 알고리즘에 대하여 소개하고, 이러한 알고리즘들의 성능 및 장단점을 선형분류에 대한 모의실험을 통해서 살펴본다.

• Management Science and Financial Engineering
• /
• 제13권2호
• /
• pp.47-77
• /
• 2007

This paper considers a scheduling problem where a customer orders multiple products(jobs) from a production facility. The objective is to minimize the sum of the order(batch) completion times. While a machine can process only one job at a time, multiple machines can simultaneously process jobs in a batch. Although each job has a unique processing time, we consider the case where batch processing times are identical. This simplification allows us to develop heuristics with improved performance bounds. This problem was motivated by a real world problem encountered by foreign electronics manufacturers. We first establish the complexity of the problem. For the two parallel machine case, we introduce two simple but intuitive heuristics, and find their worst case relative error bounds. One bound is tight and the other bound goes to 1 as the number of orders goes to infinity. However, neither heuristic is superior for all instances. We extend one of the heuristics to an arbitrary number of parallel machines. For a fixed number of parallel machines, we find a worst case bound which goes to 1 as the number of orders goes to infinity. Then, a tighter bound is found for the three parallel machine case. Finally, the heuristics are empirically evaluated.

• 대한산업공학회지
• /
• 제33권2호
• /
• pp.246-253
• /
• 2007

All of the machines in a production line can be classified into bottleneck and non-bottleneck machines. A bottleneck is a resource whose capacity limits the throughput of the whole production facility. This paper addresses a batch sizing problem at the bottleneck machine. Traditionally, most batch sizing decisions have been made based on the EOQ (economic order quantity) model where setup and inventory costs are considered while throughput rate is assumed to be given. However, since batch size affects the capacity of the bottleneck machine, the throughput rate may not be constant. As the batch size increases, the frequency of the setup decreases. The saved setup time can be transferred to processing time, which results in higher throughput. But, the larger batch size may also result in longer lead time and larger WIP inventory level. This paper presents an alternative method to determine batch size at the bottleneck machine in a manufacturing line. A linear search algorithm is introduced to find optimal throughput rate and batch size at the same time. Numerical examples are provided to see how the proposed method works and to investigate the effects of some parameters.

• Management Science and Financial Engineering
• /
• 제9권2호
• /
• pp.47-68
• /
• 2003

We consider the customer order scheduling problem with job capacity restriction where the number of jobs in the shop at the same time is fixed. In the customer order scheduling problem, each job is part of some batch (customer order) and the composition of the jobs (product) in the batch is pre-specified. The objective function is associated with the completion time of the batches instead of the completion time of the jobs. We first summarize the known results for the general customer order scheduling problems. Then, we establish some new properties for the problems with job capacity restriction. For the case of unit processing time with the objective of minimizing makespan, we develop a polynomial-time optimal procedure for the two machine case. For the same problem with a variation of no batch alternation, we also develop a polynomial-time optimal procedure. Then, we show that the problems with the objectives of minimizing makespan and minimizing average batch completion time become NP-hard when there exist arbitrary number of machines. Finally, We propose optimal solution procedures for some special cases.

• 한국경영과학회지
• /
• 제25권2호
• /
• pp.33-46
• /
• 2000

The problem of scheduling n independent jobs on serial stages with identical parallel machines at each stage is considered. Each job lot is allocated evenly to all machines at each stage for processing and moved in transfer batches between states., This scheduling strategy is called an identical production pattern. The objective is to find a permutation schedule that minimizes makespan. A branch and bound algorithm is suggested to find an optimal permutation schedule for a transformed problem A number examples is presented to illustrate the branch and bound algorithm, Computational results for 640 problems generated randomly show that within a resonable time the suggested algorithm can be used for transfer batch scheduling in a flexible flowshop.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 1999년도 추계학술대회 및 정기총회 : 정보통신기술의 활용과 21세기 전자상거래
• /
• pp.13-13
• /
• 1999

• Management Science and Financial Engineering
• /
• 제11권2호
• /
• pp.19-43
• /
• 2005

This paper considers a variation of the customer order scheduling problem, and the variation is the case where the machine-job assignment is fixed. We examine the parallel machine environment, and the objective is to minimize the sum of the completion times of the batches. While a machine can process only one job at a time, different machines can simultaneously process different jobs in a batch. The recognition version of this problem is known to be NP-complete in the strong sense even if there exist only two parallel machines. When there are an arbitrary number of parallel machines, we establish three lower bounds and develop a dynamic programming (DP) algorithm which runs in exponential time on the number of batches. We present two simple but intuitive heuristics, SB and GR, and find some special cases where SB and GR generate an optimal schedule. We also find worst case upper bounds on the relative error. For the case of the two parallel machines, we show that GR generates an optimal schedule when processing times of all batches are equal. Finally, the heuristics and the lower bounds are empirically evaluated.

• 경영과학
• /
• 제17권2호
• /
• pp.67-73
• /
• 2000

This paper describes a scheduling support system for a plant where the machine environment may be modeled as non-identical parallel machine lines (NPML). That is, there are a number of stages in series with various different-capability-machines at each stage. Arriving continuously are jobs with their specific due dates, processing times and candidate processing machines. We’ve developed a real-time scheduling module in conjunction with a supporting production information system which supplies necessary data to the module. This scheduling module is one of the 9 modules in this system, and is composed of both a scheduling interface and a production monitoring interface. This module allows users to generate many candidate schedules by selecting their business policies. The selective arguments which are available consist of allocation costs, batch sizes and machine selection intervals. They are now being implemented at a powder metallurgy plant.