Due-date assignment in wafer fabrication using artificial neural networks

Due-date assignment (DDA) is the first important task of shop floor control in wafer fabrication. Due-date related performance is impacted by the quality of the DDA rules. Assigning order due dates and timely delivering the goods to the customer will enhance customer service and competitive advantage. A new methodology for lead-time prediction, artificial neural network (ANN) prediction is considered in this work. An ANN-based DDA rule combined with simulation technology and statistical analysis is developed. Besides, regression-based DDA rules for wafer fabrication are modelled as benchmarking. Whether neural networks can outperform conventional and regression-based DDA rules taken from the literature is examined. From the simulation and statistical results, ANN-based DDA rules perform a better job in due-date prediction. ANN-based DDA rules have a lower tardiness rate than the other rules. ANN-based DDA rules have better sensitivity and variance than the other rules. Therefore, if the wafer fab information is not difficult to obtain, the ANN-based DDA rule can perform better due-date prediction. The SFM_sep and JIQ in regression-based and conventional rules are better than the others.     

Minimizing total weighted tardiness on heterogeneous batch processing machines with incompatible job families

This paper addresses a specific class of scheduling parallel batching problem, which is observed in steel casting industries. The focus of this research is to minimize the total weighted tardiness on heterogeneous batch processing machines under conditions of dynamic job arrivals, incompatible job families and non-identical job sizes. This type of parallel batching problem arises in a number of different settings, including diffusion in wafer fabrication, heat treatment operations in aircraft industries, and metal working. The problem is viewed as a three stage-decision-problem: the first stage involves selecting a machine from the heterogeneous batch processing machines for scheduling; the second stage involves the selection of a job family from the available incompatible job families; and the third stage involves the selection of a set of jobs to create a batch from the selected job family based on the capacity of the selected batch-processing machine. Since the problem is NP-hard, a few greedy heuristics are proposed. The computational experiments show that the proposed greedy heuristic algorithms are capable of consistently obtaining near-optimal solutions (statistically estimated) in very reasonable computational time on a Pentium III 650 Mz with 128 MB RAM.     

Minimizing total weighted tardiness on heterogeneous batch processing machines with incompatible job families

This paper addresses a specific class of scheduling parallel batching problem, which is observed in steel casting industries. The focus of this research is to minimize the total weighted tardiness on heterogeneous batch processing machines under conditions of dynamic job arrivals, incompatible job families and non-identical job sizes. This type of parallel batching problem arises in a number of different settings, including diffusion in wafer fabrication, heat treatment operations in aircraft industries, and metal working. The problem is viewed as a three stage-decision-problem: the first stage involves selecting a machine from the heterogeneous batch processing machines for scheduling; the second stage involves the selection of a job family from the available incompatible job families; and the third stage involves the selection of a set of jobs to create a batch from the selected job family based on the capacity of the selected batch-processing machine. Since the problem is NP-hard, a few greedy heuristics are proposed. The computational experiments show that the proposed greedy heuristic algorithms are capable of consistently obtaining near-optimal solutions (statistically estimated) in very reasonable computational time on a Pentium III 650 Mz with 128 MB RAM.     

半导体生产线动态调度方法研究

为了满足半导体生产线多目标优化要求,提出了多目标优化动态调度规则。它由正常状态调度规则、瓶颈设备较低在制品水平的调度规则、非瓶颈设备较高在制品水平的调度规则、多批加工设备调度规则以及紧急工件调度规则五种类型的调度规则组成。使用具有半导体生产线本质特征的简化模型,将多目标优化动态调度规则与先入先出法、最早交货期法、临界值法进行了比较,结果表明,多目标优化动态调度规则能够改善半导体生产线的整体性能,更好地优化模型的生产率、加工周期与在制品水平,最终提高准时交货率。     

Performance evaluation of priority dispatching rules in multi-level assembly job shops with jobs having weights for flowtime and tardiness

The problem of scheduling in dynamic shops is an important operational problem in view of its complexity and significance in terms of associated costs of scheduling. While a number of research studies have investigated the problem of scheduling in flow shops and job shops, only some attempts have been done to study the problem of scheduling in assembly job shops that manufacture multi-level jobs. The problem of scheduling in dynamic assembly job shops with jobs having weights for holding and tardiness of jobs deserves due attention. In this study an attempt has been made to propose new priority dispatching rules that minimize the performance measures related to weighted flowtime and weighted tardiness of jobs. The existing unweighted dispatching rules have been modified in view of the consideration of weights for flowtime and tardiness of jobs. The performances of the (modified) existing dispatching rules and the proposed dispatching rules are compared through exhaustive simulation experiments with the consideration of a number of different experimental settings involving due-date setting, utilization levels and types of job structures. The proposed dispatching rules are found to perform better than the existing ones in most experimental settings and with respect to a number of measures of performance.     

Scheduling of Wet Etch and Furnace Operations with Next Arrival Control Heuristic

In wafer fabrication facilities, it is critical to schedule batching operations efficiently. Most previous studies in scheduling these operations have the objective to efficiently create the batches with the given jobs at the batching operation and decide whether to start a partial batch or to wait for the next arrival to make a full batch. Typical batch operations in wafer fab can be found in the diffusion area which is mainly composed of pre-cleaning (wet etch) and furnace. Glassey and Weng (Dynamic batching heuristic for simultaneous processing. IEEE Trans Semicond Manuf 4(2), May 1991) first introduced the concept of using future arrival information and called it look-ahead. Fowler et al. (Real-time control of multiproduct bulk-service semiconductor manufacturing processes. IEEE Trans Semicond Manuf 1992) expanded the concept to multiple products and created the heuristic method called next arrival control heuristic (NACH). NACH considers only the next arrival and makes the decision whether to start the batch now or to wait for the next arrival. The new proposed scheduling system called NACH+ tries to control the incoming inventory into the batch operation. The preliminary batches are scheduled to arrive at the batch operation just in time so that any unnecessary waiting time does not happen. In this paper, NACH+ is compared to the minimum batching size rule.     

单机成组调度问题的约束满足建模与求解方法

为适应多品种小批量生产需求,企业普遍采用基于成组技术的混流生产,由此产生的成组调度需要平衡安装时间减少与满足交期之间的冲突关系。在分析安装时间是否依赖工件组排序、工件组能否分割加工等成组特征的基础上,以最小化加权流程时间与加权拖期为目标,构建了单机成组调度问题的约束满足模型,提出了以变量排序启发式搜索和前向约束传播相结合的求解方法。典型生产数据的实证分析表明,所提出的方法建模能力强,解的适应性好。     

Vehicle dispatching for highly loaded semiconductor production considering bottleneck machines first

Semiconductor wafer fabrication lines can be characterized by re-entrant product flow, long production lead-time, large variety of production processes, and large capital investment. These distinctive characteristics make the flow control in the fab very complicated. Throughput rate and lead-time are among the most important performance measures. The throughput rate is usually determined by a bottleneck resource, and the lead-time depends on the machine utilization level and the amount of variability in the system. Due to the high efficiency of material handling and reduced particles, automated material handling systems such as automatic guided vehicles (AGVs), overhead hoist transporters (OHTs), and overhead shuttles (OHSs) are being widely used in wafer fabrication lines (wafer fabs) instead of human operators. Although a material handling system itself is seldom a bottleneck of production in a fab, it is important for that to effectively support the bottleneck machines to maximize the throughput and reduce production lead-time. This paper presents a vehicle dispatching procedure based on the concept of theory of constraints, in which vehicle dispatching decisions are made to utilize the bottleneck machines at the maximum level. Simulation experiments have been performed to compare the proposed vehicle dispatching procedure with existing ones under different levels of machine utilization, vehicle utilization, and local buffer capacity.     

Minimizing total weighted tardiness on a single machine with sequence-dependent setup and future ready time

This paper minimizes total weighted tardiness on a single machine with sequence-dependent setup time and future ready time. The sequence-dependent setup is classified into two categories: continuous sequence-dependent setup and separable sequence-dependent setup. The former is the conventional type, which needs a part on a machine to setup, while the latter does not. The weighted shortest processing time (WSPT) term, the slack term, and the ready time term of the apparent tardiness cost (ATC)-based rules are analyzed. Two dispatching rules, ATC with ready time and continuous setup (ATCRCS) and ATC with ready time and separable setup (ATCRSS), are proposed for the problems which setup are the continuous type and the separable type, respectively. These new rules are evaluated by the best-of-the-best measure and the territory measure of the performance.     

A Look-Ahead Routing Procedure for Machine Selection in a Highly Informative Manufacturing System

Many dispatching rules have been developed for the on-line control of product flow in a job shop. The introduction of a highly informative manufacturing system (HIMS) has added a new requirement to a classical job-shop control problem: the selection of machines by parts of different types. An HIMS can keep a great deal of information on the status of the system, such as information on what is scheduled in the near future with great accuracy, which can be used for shop floor control. For example, the knowledge of the time when the next parts arrive at the machines can be used for better routing. This article tests the effect of the use of this knowledge for part routing on the part's flow time and tardiness under a look-ahead routing procedure (LARP). LARP assigns a new part to a machine so that the assignment minimizes the flow time or tardiness of the current part and the next N parts arriving after the current part. A test shows that the reduction of part flow time is up to 11% and the reduction of tardiness is up to 21% for the cases with this procedure.     

基于蚁群算法的单台批加工设备调度

从半导体生产线上批加工设备的实际情况着手,考虑了工件动态时间到达和菜单间整定时间等问题,利用蚁群算法,实现了半导体生产线上单台批加工设备的优化调度。基于实际生产线模型仿真验证的结果表明,能够在合理的时间内取得满意解。     

A hybrid look-ahead SOM-FBPN and FIR system for wafer-lot-output time prediction and achievability evaluation

A hybrid system is constructed in this study for wafer-lot-output time prediction and achievability evaluation, which are critical tasks for a wafer fab (fabrication plant). In the first part of the hybrid system, a look-ahead self-organization map fuzzy-back-propagation network (SOM-FBPN) is constructed to predict the output time of a wafer lot. Compared with traditional approaches in this field, the look-ahead SOM-FBPN has three advanced features: incorporating the future release plan, classifying wafer lots, and incorporating expert opinions. According to experimental results, the prediction accuracy and efficiency of the look-ahead SOM-FBPN were significantly better than those of many existing approaches. In the second part of the hybrid system, a set of fuzzy inference rules (FIRs) are established to evaluate the achievability of an output time forecast, which is defined as the possibility that the fabrication on the wafer lot can be finished in time before the output time forecast. Achievability is as important as accuracy and efficiency but has been ignored in traditional studies. With the proposed methodology, both output time prediction and achievability evaluation can be concurrently accomplished.     

Analysis of dispatching rules in a stochastic dynamic job shop manufacturing system with sequence-dependent setup times

Stochastic dynamic job shop scheduling problem with consideration of sequence-dependent setup times are among the most difficult classes of scheduling problems. This paper assesses the performance of nine dispatching rules in such shop from makespan, mean flow time, maximum flow time, mean tardiness, maximum tardiness, number of tardy jobs, total setups and mean setup time performance measures viewpoint. A discrete event simulation model of a stochastic dynamic job shop manufacturing system is developed for investigation purpose. Nine dispatching rules identified from literature are incorporated in the simulation model. The simulation experiments are conducted under due date tightness factor of 3, shop utilization percentage of 90 % and setup times less than processing times. Results indicate that shortest setup time (SIMSET) rule provides the best performance for mean flow time and number of tardy jobs measures. The job with similar setup and modified earliest due date (JMEDD) rule provides the best performance for makespan, maximum flow time, mean tardiness, maximum tardiness, total setups and mean setup time measures.     

Single machine scheduling with multiple objectives in semiconductor manufacturing

To deal with today’s stiff competition in the logistics of semiconductor manufacturing, the development of an efficient scheduling approach for the complex semiconductor back-end testing operations is very essential, where more than one objective, such as cycle time, machine utilization and due date accuracy are kept in the focus to various degrees simultaneously. In this paper, the problem of scheduling N independent jobs on a single testing machine with due dates and sequence-dependent setup times is addressed, where the multiple objectives are to minimize average cycle time and average tardiness and to maximize machine utilization. A near optimal solution, which is not inferior to any other feasible solution in terms of all objectives, is generated combining the analytically optimal and conjunctive simulated scheduling approach. First, the machine-scheduling problem is modeled using the discrete event simulation approach and the problem is divided into simulation clock-based lot selection sub-problems. Then, at each decision instance in simulated time, a Pareto optimal lot is selected using the compromise programming technique for multi-objective optimization. With the help of a broad experimental design, this developed solution is then compared with common heuristic-dispatching rules used in industry such as the shortest processing time (SPT) and earliest due date (EDD). The developed scheduling method shows better results for all the objectives over a wide range of problems. It shows approximately 16.7% reduction in average cycle time, 25.6% reduction in average tardiness, and 21.6% improvement in machine utilization over the common dispatching rules, SPT and EDD.     

Real-time OHT Dispatching Mechanism for the Interbay Automated Material Handling System with Shortcuts and Bypasses

As a key to improve the performance of the interbay automated material handling system (AMHS) in 300 mm semiconductor wafer fabrication system, the real-time overhead hoist transport (OHT) dispatching problem has received much attention. This problem is first formulated as a special form of assignment problem and it is proved that more than one solution will be obtained by Hungarian algorithm simultaneously. Through proposing and strictly proving two propositions related to the characteristics of these solutions, a modified Hungarian algorithm is designed to distinguish these solutions. Finally, a new real-time OHT dispatching method is carefully designed by implementing the solution obtained by the modified Hungarian algorithm. The experimental results of discrete event simulations show that, compared with conventional Hungarian algorithm dispatching method, the proposed dispatching method that chooses the solution with the maximum variance respectively reduces on average 4 s of the average waiting time and average lead time of wafer lots, and its performance is rather stable in multiple different scenarios of the interbay AMHS with different quantities of shortcuts. This research provides an efficient real-time OHT dispatching mechanism for the interbay AMHS with shortcuts and bypasses.     

Comparing the effect of different strategies of continuous improvement programmes on repair time to reduce lead time

To cope with uncertainty in semiconductor wafer fabrication facilities (fabs), scheduling methods are required to produce quick real-time responses. They should be well tuned to track the changes of a production environment to obtain good operational performance. This paper presents an efficient adaptive dispatching method (ADM) with parameters determined dynamically by real-time state information of fabs. ADM is composed of a dispatching rule considering both batch and non-batch processing machines to obtain improved fab-wide performance, several feature selection methods to determine key scheduling-related real-time state information, and a linear regression model to find the relations between the weighting parameters of the dispatching rule and the determined real-time state information. A real fab simulation model is used to demonstrate the proposed method. The simulation results show that ADM is adaptive to changing environment with better performance than a number of commonly used rules (such as FIFO, EDD, CR, LPT, LS, SRPT, and SPT) and an adaptive dispatching rule that considers only real-time ratio of hot jobs to the number of all jobs in a fab and the ratio of jobs with one third of photo steps left to the number of all jobs.     

复合优先级控制策略研究及其在半导体生产线调度中的应用

复合优先级控制策略从提高制造系统综合性能的角度出发,利用复合优先级系数反映系统内的局部信息及全局信息对待加工任务优先级的影响;通过计算比较待加工任务的复合优先级系数,可确定哪个工件先进行加工,还可确定新工件投入生产线进行加工的时间。仿真结果表明,所提出的复合优先级控制策略具有综合优化系统性能指标的能力,可同时降低平均加工周期及其方差,减少在制品WIP数,提高生产率,缩短工件的等待时间。     

Scheduling on unrelated parallel machines with sequence- and machine-dependent setup times and due-date constraints

We consider the unrelated parallel-machine scheduling problem with sequence- and machine-dependent setup times and due-date constraints. There are N jobs, each having a due date and requiring a single operation on one of the M machines. A setup is required if there is a switch from processing one type of job to another. Due to the characteristics of machines, the processing time depends upon the job and machine on which the job is processed, and the setup time is sequence and machine dependent. In addition, certain jobs have strict due-date constraints. An effective heuristic based on a modified apparent-tardiness-cost-with-setup procedure, the simulated annealing method, and designed improvement procedures is proposed to minimize the total tardiness of this scheduling problem. Computational characteristics of the proposed heuristic are evaluated through an extensive experiment using a newly created data set. Computational results show that the proposed heuristic is able to effectively improve the initial solutions, obtained by a modified apparent-tardiness-cost-with-setup procedure, and obtains better results than a random descent heuristic.     

考虑设备维修与维护的装配作业车间动态生产仿真研究

装配作业车间生产调度的关键是动态协同产品各零件的生产进度以满足产品的齐套需求。设备的随机故障将会扰乱已有的生产计划,给各零件之间的进度协同带来更高的挑战。预防性维护能够提高设备的可靠度以减少故障的发生,但过度的维护又会减少生产加工的时间。从调度规则构成属性的角度分析设备故障将带来的影响,并提出一种考虑设备故障的改进型调度规则。通过仿真试验分析了企业常见的3种预防性维护策略对车间性能的影响和7种代表性调度规则的性能。试验结果表明:预防性维护策略对车间有明显的影响;且考虑了设备故障的调度规则有一定的改进效果。     

A hybrid neural network and selective allowance approach for internal due date assignment in a wafer fabrication plant

A hybrid neural network and selective allowance approach is proposed in this study for internal due date assignment in a wafer fabrication plant (wafer fab). In the first part, a look-ahead self-organization map-fuzzy back propagation network (SOM-FBPN) is constructed to estimate the completion time of a job. Subsequently, a selective allowance policy is established to determine the allowance that will be added to the estimated job completion time. Compared with traditional approaches, the look-ahead SOM-FBPN has many features (including incorporating the future release plan, classifying jobs, and incorporating expert opinions), and the selective allowance policy is novel by considering two different production control issues to avoid missing the due date. According to experimental results, the prediction accuracy of the look-ahead SOM-FBPN was significantly better than those of many existing approaches. Besides, the selective allowance policy considerably improved the due date related performance after adding some extra allowance.