Scheduling semiconductor wafer fabrication
TL;DR: In this paper, a variety of input control and sequencing rules are evaluated using a simulation model of a representative, but fictitious, semiconductor wafer fabrication, and the simulation results indicate that scheduling has a significant impact on average throughput time, with larger improvements coming from discretionary imput control than from lot sequencing.
Abstract: The impact that scheduling can have on the performance of semi-conductor wafer fabrication facilities is assessed. The performance measure considered is the mean throughput time (sometimes called cycle time, turnaround time or manufacturing interval) for a lot of wafers. A variety of input control and sequencing rules are evaluated using a simulation model of a representative, but fictitious, semiconductor wafer fabrication. Certain of these scheduling rules are derived by restricting attention to the sub-set of stations that are heavily utilized, and by using a Brownian network model, which approximates a multi-class queuing network model with dynamic control capability. Three versions of the wafer fabrication model, which differ only by the number of servers present at particular stations, are studied. The three versions have one, two, and four stations, respectively, that are heavily utilized (near 90% utilization). The simulation results indicate that scheduling has a significant impact on average throughput time, with larger improvements coming from discretionary imput control than from lot sequencing. The effects that specific sequencing rules have are highly dependent on both the type of input control used and the number of bottleneck stations in the fabrication. >
Citations
More filters
TL;DR: A review of research in this area to date, discuss the applicability of the various approaches and suggest directions for future research is presented in this article, where the authors describe the characteristics of the semiconductor manufacturing environment and review models related to performance evaluation and production planning.
Abstract: Although the national importance of the semiconductor industry is widely acknowledged, it is only recently that the production planning and scheduling problems encountered in this environment have begun to be addressed using industrial engineering and operations research.techniques. These problems have several features that make them difficult and challenging: random yields and rework, complex product flows, and rapidly changing products and technologies. Hence their solution will contribute considerably to die theory and practice of production planning and control. In a two-part project we present a review of research in this area to date, discuss the applicability of the various approaches and suggest directions for future research. In this paper, Part I, we describe the characteristics of the semiconductor manufacturing environment and review models related to performance evaluation and production planning. Part II will review research on shop-floor control in this industry to date.
599 citations
TL;DR: In this article, the problem of scheduling semiconductor burn-in operations is modeled as batch processing machines, where the processing time of a batch is equal to the largest processing time among all jobs in the batch.
Abstract: In this paper, we study the problem of scheduling semiconductor burn-in operations, where burn-in ovens are modeled as batch processing machines. A batch processing machine is one that can process up to B jobs simultaneously. The processing time of a batch is equal to the largest processing time among all jobs in the batch. We present efficient dynamic programming-based algorithms for minimizing a number of different performance measures on a single batch processing machine. We also present heuristics for a number of problems concerning parallel identical batch processing machines and we provide worst case error bounds.
433 citations
TL;DR: In this paper, the relationship between shop-floor control and production planning is discussed, and the relative advantages and disadvantages of the various approaches are discussed, as well as future research directions.
Abstract: In the first part of this review [62] we described the characteristics of semiconductor manufacturing environments and reviewed research on system performance evaluation and production planning. In this paper we focus on shop-floor control problems. We classify research to date by the solution techniques used, and discuss the relative advantages and disadvantages of the various approaches. We discuss the relationship between shop-floor control and production planning and suggest future research directions.
409 citations
TL;DR: In this article, a new class of scheduling policies, called fluctuation smoothing policies, were introduced to reduce the mean and variance of cycle time in semiconductor manufacturing plants, and they achieved the best performance in all configurations of plant models and release policies tested.
Abstract: The problem of reducing the mean and variance of cycle time in semiconductor manufacturing plants is addressed. Such plants feature a characteristic reentrant process flow, where lots repeatedly return at different stages of their production to the same service stations for further processing, consequently creating much competition for machines. We introduce a new class of scheduling policies, called Fluctuation Smoothing policies. Unanimously, our policies achieved the best mean cycle time and Standard Deviation of Cycle Time, in all the configurations of plant models and release policies tested. As an example, under the recommended Workload Regulation Release policy, for a heavily loaded Research and Development Fabrication Line model, our Fluctuation Smoothing policies achieved a reduction of 22.4% in the Mean Queueing Time, and a reduction of 52.0% in the Standard Deviation of Cycle Time, over the baseline FIFO policy. These conclusions are based on extensive simulations conducted on two models of semiconductor manufacturing plants. The first is a model of a Research and Development Fabrication Line. The second is an aggregate model intended to approximate a full scale production line. Statistical tests are used to corroborate our conclusions. >
401 citations
TL;DR: Typical scheduling problems found in semiconductor manufacturing systems are identified and important solution techniques used to solve these scheduling problems are presented by means of specific examples, and known implementations are reported.
Abstract: In this paper, we discuss scheduling problems in semiconductor manufacturing. Starting from describing the manufacturing process, we identify typical scheduling problems found in semiconductor manufacturing systems. We describe batch scheduling problems, parallel machine scheduling problems, job shop scheduling problems, scheduling problems with auxiliary resources, multiple orders per job scheduling problems, and scheduling problems related to cluster tools. We also present important solution techniques that are used to solve these scheduling problems by means of specific examples, and report on known implementations. Finally, we summarize some of the challenges in scheduling semiconductor manufacturing operations.
354 citations
Cites background from "Scheduling semiconductor wafer fabr..."
...But at the same time, scheduling approaches attracted researchers and people from industry working in semiconductor manufacturing for the last two decades (cf. Bitran and Tirupati 1988 and Wein 1988 for some older references)....
[...]
...However, this is a non-trivial task as many authors have discussed the difficulties of semiconductor manufacturing (cf. Wein 1988; Atherton and Atherton 1995; Sze 2001; Uzsoy et al. 1992; Sarin et al. 2011, among others)....
[...]
References
More filters
TL;DR: In this paper, it was shown that if the three means are finite and the corresponding stochastic processes strictly stationary, and if the arrival process is metrically transitive with nonzero mean, then L = λW.
Abstract: In a queuing process, let 1/λ be the mean time between the arrivals of two consecutive units, L be the mean number of units in the system, and W be the mean time spent by a unit in the system. It is shown that, if the three means are finite and the corresponding stochastic processes strictly stationary, and, if the arrival process is metrically transitive with nonzero mean, then L = λW.
2,536 citations
Book•
01 Jan 1979
TL;DR: This classic in stochastic network modelling broke new ground when it was published in 1979, and it remains a superb introduction to reversibility and its applications thanks to the author's clear and easy-to-read style.
Abstract: This classic in stochastic network modelling broke new ground when it was published in 1979, and it remains a superb introduction to reversibility and its applications. The book concerns behaviour in equilibrium of vector stochastic processes or stochastic networks. When a stochastic network is reversible its analysis is greatly simplified, and the first chapter is devoted to a discussion of the concept of reversibility. The rest of the book focuses on the various applications of reversibility and the extent to which the assumption of reversibility can be relaxed without destroying the associated tractability. Now back in print for a new generation, this book makes enjoyable reading for anyone interested in stochastic processes thanks to the author's clear and easy-to-read style. Elementary probability is the only prerequisite and exercises are interspersed throughout.
2,480 citations
TL;DR: This paper describes the Queueing Network Analyzer (QNA), a software package developed at Bell Laboratories to calculate approximate congestion measures for a network of queues and uses two parameters to characterize the arrival processes and service times.
Abstract: This paper describes the Queueing Network Analyzer (QNA), a software package developed at Bell Laboratories to calculate approximate congestion measures for a network of queues. The first version of QNA analyzes open networks of multiserver nodes with the first-come, first-served discipline and no capacity constraints. An important feature is that the external arrival processes need not be Poisson and the service-time distributions need not be exponential. Treating other kinds of variability is important. For example, with packet-switched communication networks we need to describe the congestion resulting from bursty traffic and the nearly constant service times of packets. The general approach in QNA is to approximately characterize the arrival processes by two or three parameters and then analyze the individual nodes separately. The first version of QNA uses two parameters to characterize the arrival processes and service times, one to describe the rate and the other to describe the variability. The nodes are then analyzed as standard GI/G/m queues partially characterized by the first two moments of the interarrival-time and service-time distributions. Congestion measures for the network as a whole are obtained by assuming as an approximation that the nodes are stochastically independent given the approximate flow parameters.
1,021 citations
TL;DR: The main emphases are on the difference between socially optimal and individually optimal (equilibrium) controls and on the use of dynamic-programming inductive analysis to show that an optimal control is monotonic or characterized by one or more "critical numbers".
Abstract: Congestion in a queueing system can sometimes be controlled by restricting arrivals, either by "closing a gate" or by charging an entrance fee or toll. We review both static (open-loop) and dynamic (closed-loop) models for control of admission to a queueing system. The main emphases are on the difference between socially optimal and individually optimal (equilibrium) controls and on the use of dynamic-programming inductive analysis to show that an optimal control is monotonic or characterized by one or more "critical numbers." We discuss the potential for use of these models in the analysis of computer/ communication systems and compare the results to certain others in the literature.
411 citations
01 Jan 1988
TL;DR: Consider an open queueing network with I single-server stations and K customer classes, where each customer class requires service at a specified station, and customers change class after service in a Markovian fashion.
Abstract: Consider an open queueing network with I single-server stations and K customer classes. Each customer class requires service at a specified station, and customers change class after service in a Markovian fashion. (With K allowed to be arbitrary, this routing structure is almost perfectly general.) There is a renewal input process and general service time distribution for each class. The correspondence between customer classes and service stations is in general many to one, and the service discipline (or scheduling protocol) at each station is left as a matter for dynamic decision making.
358 citations