JP3283951B2 - Production management method and system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for managing the production of various industrial products such as semiconductor parts such as LSIs, clothing and automobiles. Here, the management of production refers to scheduling of a product consisting of a plurality of varieties to determine a production order according to each processing step and urgency,
This includes monitoring the progress of manufacturing and acquiring processing data during manufacturing.

[0002]

2. Description of the Related Art A conventional production management system will be described below by taking an example of an LSI manufacturing process. FIG. 9 is a diagram illustrating an outline of a procedure for manufacturing an LSI. On a single crystal wafer, apply an insulating film such as an oxide film / nitride film, diffuse by heat treatment, form a pattern by photolithography, deposit a film by vapor deposition or sputtering, introduce impurities by ion implantation, scribe (divide) ), Mounting, bonding and other technologies are used at least once and many times dozens of times, while the functions are built on the semiconductor substrate,
An LSI is manufactured. Each technique is further subdivided into several processes. For example, as shown on the left side of FIG. 9, the photolithography technique includes (a) application of a resist as a photosensitive agent, (b) mask alignment / exposure, (c) pattern formation by development, and (d) processing by etching. , Etc., and such a processing unit is usually called a “step”. An LSI is manufactured by repeating these steps several hundred times. A list of procedures of this manufacturing process is a “process table” as shown in FIG.
Processing contents including at least a used device and processing conditions (recipe) are described.

[0003] This process chart is designated for each lot (processing unit: usually a cassette unit, and in single-wafer processing, one sheet unit).
LSIs can be manufactured by performing the processing in order according to the process chart. What is characteristic here is that, in this step, the same apparatus is used many times under different processing conditions. In the case of manufacturing different types of products, the same process is performed under different processing conditions. The processing conditions (recipe) here include not only the temperature for performing the heat treatment, the gas flow rate, the manufacturing conditions such as the mask used at the time of exposure, but also the processing conditions (recipe).
It also includes inspection conditions such as the position of dimension measurement. As described above, since a permutation combination of many elements is to be controlled, a complicated line is formed.

In recent years, not only LSIs but also diversified products and quick response to customer needs have been demanded. Along with this, the demand for the delivery date management and TAT control is also increasing for the LSI production management due to the improvement of the production efficiency by the automation of the conveyance. With the diversification of products, many types of products have been produced on one production line. In addition, these products not only have different production amounts, but also have different importance and processing priority for each product type. Here, the degree of importance depends on added value such as a design cost of a mask or the like and a manufacturing cost such as requiring a special process, and is usually determined when a lot is input. On the other hand, the priority indicates which should be processed first when processing is awaited by the same manufacturing apparatus, and a margin until the delivery date can be provided. For this reason, lots with low priority were initially expected to have sufficient margin until the delivery date, so it was necessary to raise the priority because the lot was late on schedule, In some cases, it is lowered.

[0005] Acquiring and monitoring the processing execution data not only prevents product failure beforehand, but also is useful in analyzing the cause of failure. Therefore, it is essential to acquire processing data including inspection data. Here, it is desirable that this processing data be automatically collected from the device,
It does not have to be automatic. When collecting the device data and inspection data, the same device is used many times in the process. Therefore, the data indicates which process of which lot (which process in the process chart) is the data. Key (word) information is added and stored in the database. It is convenient to use a processing schedule created by scheduling as the key information. That is, a method is used in which a lot is recognized by a barcode reader during lot processing, the processing schedule is received by network communication, and stored in a database together with processing data. Alternatively, there is a method in which a magnetic card or the like is transported together with a lot, the card is read, key information is read, and the key information is stored in a database together with processing data. Each of these has a feature that it can also serve as a lot progress management. FIG. 11 shows such a line management procedure. The schedule result is distributed to computers arranged in a line in the form of a schedule, and the worker executes processing based on work instructions according to the schedule. The execution record of the processing data, etc. is directly sent from the manufacturing / inspection equipment, and the progress is managed by feeding back to the scheduler by the online data collection that is directly transmitted from the manufacturing / inspection equipment or the offline progress notification that is made by the operator from the terminal. Reflect the situation.

In a mass production line in which products of the same type are mass-produced, old ones are sequentially processed from the accumulated lots waiting for processing in each apparatus (first in first out). In addition, a signboard method or the like in which the excess or deficiency of the arrival lot is visually displayed is often used as a production method. Information on which process of which lot is in which process is sequentially added to the data by key input or bar code reading on the spot, and the data is stored in a database. However, in this case, since the processing schedule of each process of each lot cannot be grasped, it is difficult to control the delivery date management and TAT (the number of days required to complete the lot).

[0007] On the other hand, in the production line of ASICLSI,
Lots of lots of different types having different processing priorities and importances exist. Among ASICs, an LSI using a full custom or a standard cell has a high importance because it is necessary to use a dedicated mask pattern from a substrate process. However, in the gate array LSI, since the substrate is common, the importance is low at the stage of preparing the substrate of the gate array, but the importance increases when the wiring process using a dedicated mask pattern is started. On the other hand, the priority indicates which should be processed first when processing is waited for by the same apparatus, and is determined by a margin until a delivery date. For this reason, at first, a lot with a low priority is expected because a sufficient margin is expected until the delivery date, and it is necessary to raise the priority because a lot is later than expected on the way. Conversely, there may be a margin on the way, and the priority may be lowered. Thus, the priority changes as the processing of the lot progresses.

[0008] The time (TAT) from the input of such a product to its completion varies depending on the number of lots input in the line. Further, when products having different priorities coexist, the processing of a low priority product is delayed because of a high priority product. For this reason, as is conventionally done,
For each lot, the exact completion date cannot be predicted only by predicting the progress based on the number of lots set in the apparatus or the processing time and determining whether or not the delivery date will be met. Therefore, in order to control the completion date of a lot having a high priority or importance (hereinafter referred to as a progress control lot) so as not to be delayed in the delivery date and to create key information for data acquisition, It is necessary to monitor the progress of the lot and execute regular scheduling about once a day and irregular scheduling accompanying a trouble of the apparatus or a delay in the progress of the lot to update the processing schedule sequentially. That is,
Monitor the progress of the lot and, based on the processing results,
Change the priority of which lot is processed first, and execute the scheduling after the processing is completed, taking into account the operation status of the equipment (information such as failures), maintenance status, or the operation status of the operator. I have to do it. In order to perform production control by such scheduling, all lots in a line are disassembled into processes, and it is determined whether or not a device to be used (or an operator in some cases) can be operated, and according to the priority of each lot. You must create a processing schedule. In particular, in LSI manufacturing, there are many processes that require continuity with the previous process (the process must be performed within a certain time). It is necessary to schedule the processing start schedule for each process of each lot in detail while checking.

[0009]

However, in preparing such a processing schedule, since there are a plurality of apparatuses of the same type, what kind of combinations of apparatuses and operators that can be processed and combinations in consideration of the continuity with the next process are required. There are also streets. Therefore, the memory capacity and the calculation time required to create a schedule by allocating each process of all lots while checking the operation state (vacant time zone) of the apparatus and the operator are determined by the number of lots and the number of apparatuses (considering the operator). In this case, the number of operators further increases with an increase in the number of operators. Particularly, when the number of lots is large, a large memory capacity and a large amount of calculation time are required.

[0010] An object of the present invention is to provide a system in which lots of high priority or importance whose progress is to be controlled are mixed with mass-produced products whose priority and importance are not high both, and An object of the present invention is to reduce the processing time and the memory used so that the progress management of a line and the completion date prediction of a lot having a high priority or importance can be accurately performed. In addition, the progress of a lot having a low priority can be managed. In particular, it is an object to enable adaptation to a production line having a large number of production lots and various types of products in the entire line.

[0011]

A first aspect of the present invention is a manufacturing process.
From the process chart showing the procedure of
The scheduled processing of each each manufacturing lot or the production apparatus,
Directly connected to the progress management / data collection system in the production line
A first step of creating a short-term processing schedule with the online scheduler;
Prepared by a second step of creating a long-term treatment plan offline scheduler based on the information, the first
A production management method for performing processing and storing data based on a processing schedule of one step, wherein the first step selects only a high-priority or important lot from the production lots. Steps, a process chart of the selected lot, a delivery date of the selected lot, a current priority of the selected lot, and a
Passing the processed final step of the lot to the second step; and determining the priority of the non-selected lot from the priority of the selected lot determined in the second step. And a step of creating a short-term processing schedule for the production lot , wherein the second step is the step of receiving the selection received from the first step.
Schedule of the selected lot and delivery of the selected lot
Period, current priority of the selected lot, and
A step of allocating the processing steps of the production lot to the apparatus based on the processed final step of the selected lot and creating a long-term processing schedule until the completion date to predict the completion date, wherein the completion date is later than the delivery date If it is, the priority is increased, the priority of the lot with a margin is reduced, the processing priority of the lot is determined, and the determined priority is passed to the first step. Features.

In a second aspect of the present invention, in addition to the first aspect, in the second step, a processing schedule is created by selecting lots extracted at regular intervals from the lots of low importance as important lots. The method further comprises the step of repairing. A third invention is a production management system, production
A process table showing the steps of the process, and an operation schedule of each manufacturing apparatus, for each manufacturing lot or each manufacturing apparatus, having a scheduler for calculating a scheduled processing time of the process to be processed,
In a production management system for performing a process and storing data based on a process schedule created by a scheduler, the scheduler includes a progress management / data collection system in a production line.
Online scheduler directly connected to the system and the online
An off-line scheduler capable of communicating with an in-scheduler , wherein the on- line scheduler includes at least a communication control unit for communicating with the off-line scheduler, and only a lot having a high priority or importance from among production lots. And a lot selection unit for transmitting a process chart, a delivery date, a processed final process, and a current priority of the corresponding lot to the communication control unit, and a priority by the offline scheduler. A schedule creation unit for setting the priority of all lots based on the schedule and creating a schedule for each lot, and the offline scheduler includes:
A communication control unit for communicating with the online scheduler process chart from the online scheduler, delivery,
It is composed of a schedule creation section that creates a processing schedule based on the data of the last processed step and the current priority, and a priority change section that checks the delivery delay from the processing schedule created by the schedule creation section and changes the priority. It is characterized by that.

[0013]

According to the production management system and method of the present invention, all the lots are not scheduled for all processes, and the offline scheduler targets only the lots with high importance. The memory used can be reduced. For this reason, a long-term schedule until the delivery date is possible. Further, the completed processing schedule reflects a long-term plan for lots whose progress must be controlled. On the other hand, in the online scheduler, a short-term work schedule is created in response to the priority change result in the offline scheduler. In this case, since the period is short, the calculation time can be reduced and the memory used can be reduced. Therefore, an accurate processing schedule can be created in consideration of the operation of the operator and the like.

[0014]

Embodiments of the present invention will be described below in detail with reference to the drawings. First, the configuration of the production management system of the present invention is shown in FIG. In this system, two schedulers are used. One is an online scheduler 1 directly connected to a progress management / data collection system 3 in the production line, and the other is an offline scheduler 2 having a role of predicting a completion date.

The online scheduler 1 manages work instructions (transmission of schedule information) and progress, and also provides a short-term work instruction to each terminal in the production line based on the priority change result from the offline scheduler. Create a schedule for processing. As described above, in the online scheduler 1, a processing schedule to be used as key information at the time of data storage is created, and thus, in principle, all lots are to be scheduled. However, when data storage is not required, it is also possible to use a pseudo process chart in which the lot is excluded from the schedule or the process is excluded from the schedule. The processing schedule of the online scheduler 1 is stored in the database 6 as key information at the time of data collection together with processing data (including manual input) from the apparatus.

The offline scheduler 2 receives the registration information and the progress result from the online scheduler 1, creates a long-term processing schedule thereafter, and predicts the completion date of the progress control lot (priority lot / important lot). Based on the result, the priority of each lot is changed and passed to the online scheduler 1. The long-term processing schedule and the lot processing result are displayed in the form of a lot progress chart 5 (Gantt chart) or the like.

FIG. 2 is a diagram showing the configuration of the online scheduler 1 and the offline scheduler 2. The online scheduler 1 has a database 10 (lot name, use process chart,
Delivery date, priority initial value or current priority, device information, etc.), a processing schedule for each lot, and a database 11 of processing results. Of these, the one scheduled to be processed is changed as appropriate. In addition, a lot selection unit 12 for selecting a lot with a high priority or a lot with a high importance from the registration information of each lot, a communication control unit 14 for transferring information of the selected lot to an offline scheduler, a schedule creation unit 13 Consists of The communication control unit 14 also receives the priority and the processing schedule of a part lot from the offline scheduler 2. Up to this point, the basic configuration of the online scheduler 1 according to the present invention has been described. In addition, a processing schedule extraction unit 15 can be provided to shorten the creation of the processing schedule. On the other hand, the offline scheduler 2 has a schedule creation unit 18 and a priority change unit 19 in addition to the communication control unit 17.

Next, a method of creating a processing schedule in the production management method of the present invention is shown in FIG. Steps S1 to S6
Is processing in the online scheduler 1, S7 to S12
Is the processing in the offline scheduler 2. First, the lot selection unit 12 in the online scheduler 1 checks whether each lot corresponds to a priority lot or an important lot while referring to the registration information, and checks a process chart of the lot, a current priority (or an initial value), The processed final step is passed to the communication control unit 14 (S1).

The communication control unit 14 transfers the data from the lot selection unit 12 to the communication control unit 17 of the offline scheduler 2 (S2). Here, among the lots to be selected,
In addition to the lots currently in the line, lots with high priorities and importance among lots to be introduced in the future are also included. The communication control unit 17 of the offline scheduler 2 that has received the data stores the data in the schedule creation unit 18.
(S7).

The schedule creating section 18 creates a processing schedule for these selected lots (S8). The schedule period includes at least a period until a newly input lot (the latest input lot) is completed, and usually extends for several months. Information on the estimated completion date, delivery date and priority of each lot calculated here is
It is sent to the priority changing unit 19.

The priority changing unit 19 checks whether each lot is completed by the delivery date.
The priority is changed, the schedule is executed again, and the completion date is checked (S9). If there is no late lot, the priority of each lot at that time (priority of which process is given priority when the lot overlaps the same device) and the process schedule are sent to the communication control unit 17 (S10).

On the other hand, if the delivery delay cannot be prevented even if the priority is changed and the rescheduling is performed several times, the calculation is stopped and an alarm is generated (S12). Communication control unit 17
Then, the received data is transferred to the communication control unit 14 of the online scheduler 1 (S11). Here, it is also possible to send only the priority and schedule all the lots again on the online scheduler side.

To change the priority, those with a large difference (margin) between the expected completion date and the delivery date are reduced, and those with a small margin are raised. Here, the expected completion date can be estimated by simply summing the processing time in each device as in the conventional method,
The calculation is not performed by adding the value of the statistical waiting time in each device, but is accurate because each process of each lot is assigned while checking the idle time of the device. Therefore, it is possible to accurately predict the delivery date and manage the progress based on the processing schedule. In the processing schedule creation of the offline scheduler 2, lots are selected and reduced. Therefore, if a schedule is created for the entire line scale, the number of devices is large for the number of lots, and the lots advance faster than the actual ones. In other words, the completion date is calculated faster than it actually is. In order to prevent this, when the devices in the line are multiplexed, the calculation is performed by reducing the number of the devices in proportion to the lot selection ratio. If the number of devices cannot be reduced, the operable time is limited to reduce the number of devices. Strictly speaking, when calculating a lot with a high priority, calculation is performed using a regular line configuration, and then, when calculating a lot with a low priority but important (such as a marker lot described later), the line scale is reduced. It is correct, but if the number of priority lots is small, the completion prediction does not deviate so much even if a schedule is created with pseudo lines reduced from the beginning. Further, the reduction of the line may be performed only for a few types of devices having a high operating rate that may become a bottleneck. This off-line schedule is independent of the online scheduler 1 that issues work instructions to the operator, so that re-scheduling does not hinder the actual work of the operator.

When the priority setting by the offline scheduler 2 is completed, the communication control unit 14 of the online scheduler 1 receives the result and passes the priority data to the schedule creation unit 13 of the online scheduler 1 (S3). The schedule creation unit 13 of the online scheduler 1 refers to this priority and resets the priority of all lots (S4). This reattachment is the same for lots with high priority. For lots that are not subject to the offline schedule, priorities are assigned with reference to the priorities assigned by the offline scheduler 2 for lots of the same type. (The preceding lots are ordered so as to be processed first.) Based on this priority, a short-term (usually several days) processing schedule for work instructions for all lots is created (S6).

In steps S3, S4 and S6, it is possible to re-schedule all lots again. However, since the processing schedule of lots with high priority is hardly changed, the following method is used for those with high priority. The time required to create a processing schedule can be shortened by using this function. That is, the processing schedule extracting unit 15 is provided in the online scheduler, and the result (processing plan) of the offline schedule is stored in the processing schedule extracting unit 15. FIG. 4 shows a method of creating a processing schedule in this case. FIG.
S1, S2 and S7 to S12 are omitted. When creating a schedule in the online scheduler 1,
The processing schedule of the lot in the designated period is extracted and set as the processing schedule. Other lots fill the processing schedule in this free time zone.

It is not always necessary to execute both the online schedule and the offline schedule at the same time. For example, the priority set by the offline scheduler 2 is stored, and if it is within about one week, the schedule is created by the online scheduler 1 using the value, and the offline scheduler 2 executes the schedule for one week. It is also possible to start the device about once.

In this method, since the online scheduler side has a short period and the offline scheduler side has a small number of target lots, the number of target steps to be calculated at a time is reduced. A production plan that takes delivery time into account can be accurately set without receiving the request. Therefore, it can be effectively applied to a line having a large number of lots. The off-line scheduler can also be used for simulation for making a lot input plan.

In the method shown in FIG. 3, the objects of the offline schedule are described as priority lots and important lots. In such a method, a processing schedule cannot be set for a general-purpose product lot with low priority and importance, so that it is difficult to perform production planning and lot progress management. To prevent this, there is a method of using a marker lot as a method of setting a processing schedule and reducing the time and memory required for the scheduling. The method is described below.

Lots having different priorities exist on the line. Among them, for the lot with the higher priority, the entire process schedule is implemented, the completion date is predicted, and the progress is controlled in order to manage the delivery date. However, if only high-priority lots are scheduled, the progress of low-priority lots is not known at all. For mass-produced products such as gate array substrates with low priority (those using the same process chart even if different masks are used can be regarded as mass-produced products), lots extracted at regular intervals from mass-produced products ( Marker lots: For example, every 10 lots) are treated as indices for performing progress management (delivery date management), and all processes are scheduled together with priority lots by an offline scheduler to predict progress,
Change the priority, submit the next lot, etc.

That is, as shown in FIG. 5, when mass production lots C _i, C _{i + 1,} ... _, C _{i + N} ,. A lot is created, and a processing schedule is created for this lot by an offline scheduler. In the creation of the processing schedule, by specifying the lot input timing or the processing start step of the day (the end step at a certain time may be notified), the processing according to the actual processing on the line is performed. A processing schedule can be set, and the progress of lots on the line can be managed. On the other hand, other lots are processed in the order of input lots. That is, when a plurality of lots are in a waiting state in the same process, if it is determined that the lot put in first should be processed first, the other lots between the two marker lots will be roughly Progress can be managed. In addition, data collection can be performed in the same manner as a lot that has been subjected to a normal schedule.

FIG. 6 shows the effect of using this marker lot. This figure shows that the number of lots that can be calculated by the method of scheduling all the normal lots and all processes (the number of lots whose delivery date can be simulated) is standardized. FIG. 6 is a diagram showing the ratio of the parameter as a parameter. For example, when the ratio of the priority lot is 0.01 (1%), the number of marker lots is 5
%, That is, the number of lots can be calculated to be 15 times that of the conventional lot as one lot per 20 lots. If there is no priority lot, the calculation amount decreases in proportion to the ratio of the marker lot. Considering a mass production line in which 100 lots are put in each day (30000 lots per year), if one lot among the general-purpose lots 100 put in on that day is a marker lot, the progress can be managed within one day. The amount of calculation in this case is 1/100 of that in a normal case, and a scheduler that can handle 300 lots may be used. If the accuracy of the progress management is about two days, a scheduler capable of handling a further half of the lot, ie, 150 lots, may be used. As described above, the marker lot method is an extremely efficient method in a mass production line.

In the online scheduler shown in FIG. 1, all processes of the progress control lot (including the marker lot) are scheduled by reflecting the priority of the offline schedule. For this reason, basically, both are scheduled to be the same processing. On the contrary,
As for the schedule of a general-purpose lot that is not a progress control lot (not a marker lot), the entire process can be scheduled as it is because the period is short. However, if the marker lot method is used, it is not always necessary to schedule all the processes for lots that are not progress control lots. That is, in addition to the progress prediction and delivery date management, the use purpose of the processing schedule includes use for key information at the time of data storage. In the latter case, data cannot be stored without schedule information (process schedule information). . For this reason, in a system that uses schedule information at the time of storing data, it is necessary to schedule all lots. However, since this scheduling is not for predicting progress, it is not necessary to be accurate in time. It is only necessary to create a processing schedule only for a device for which it is desired to know the processing time accurately or for a device for which data is to be obtained. In this case, the processing schedule may be set using the pseudo process chart. The pseudo process chart will be described below.

The simulated process chart is a process chart in which only a process for collecting data and a process required at least for progress management (a process for displaying in a lot progress chart) are targeted for a schedule. It is performed based on the work instructions according to the standard process chart. FIG. 7 is a diagram illustrating the concept of the pseudo process chart. For example, in the process chart, processes (K2, K5) such as inspection for surely acquiring data, a process (K4) for displaying on a lot progress chart and monitoring the progress, and other processes are mixed. For a progress control lot consisting of a lot with a high priority and a marker lot, the entire process is to be scheduled, so the original process table is used as it is. For a lot for which only data acquisition is performed and progress management is not performed, a pseudo process chart for data collection including only a data acquisition process is used. For example, it is a process chart only for dimension measurement, film thickness measurement, and the like. By using this pseudo process chart, it is possible to reduce memory and time during scheduling. From the viewpoint of the operator, data can be collected in the same procedure as in the case where there is a processing schedule based on a regular process chart, so that the risk of operation errors is reduced. (In the case of a command to each device,
During the marker lot, a command is issued by setting the processing schedule of these pseudo processes. At the time of online scheduling, short-term standard scheduling and pseudo-scheduling are simultaneously performed. First, a process of a lot with a high priority is reserved, and pseudo-scheduling is performed in a vacant time zone. Therefore, the contents of the standard scheduling are not changed by the pseudo scheduling. In the case of the pseudo-scheduling (non-progress control lot), the wafer process is separately performed by a standard process chart, and is performed by a production management method such as a capped method or a sign method between the progress control lots.

In the system shown in FIG. 1, a method of creating a scheduling incorporating a pseudo process chart will be described in more detail. A progress control lot for the purpose of TAT management (progress management such as prediction of completion date) requires long-term scheduling, but pseudo-scheduling for data storage only needs to be scheduled for several days. Therefore, long-term scheduling of only progress control lots and creation of a processing schedule for data storage of all lots (including non-progress control lots based on the pseudo process chart) (short term) cannot be scheduled at once. The latter schedule needs to be implemented by an online scheduler directly connected to the data collection system.
However, the former is implemented in the present system by an off-line scheduler, and the schedule of the progress control lot and the schedule of the non-progress control lot are separated into two schedulers. FIG. 8 shows the concept.

In the figure, a hatched portion indicates a progress control lot, and a lot having a high priority and a marker lot are scheduled for a long time by an offline scheduler according to a standard process chart. Based on this result, the priority of a lot that is likely to be late for delivery is increased, and the priority of a lot that has room before delivery is reduced. The online scheduler reflecting the change of the priority performs the short-term all lot scheduling of the shaded portion.
At this time, the standard scheduling and the pseudo scheduling for data storage of the non-progress control lot are performed simultaneously.
A process of a lot with a high priority is reserved, and pseudo scheduling is performed in the vacant time zone. Therefore, the online scheduling result of the priority lot is the same as the content of the offline scheduling. Typically, this schedule period extends over a period of two to three months because the offline scheduler calculates the delivery date. On the other hand, the online scheduler creates a processing schedule for several days.
Further, these schedule results are appropriately changed depending on the progress. Examples of the rescheduling timing include, for example, a regular schedule such as an online scheduler at a fixed time every day, an offline scheduler once a week, and a delay (for example, one day) during which a lot progresses from the first processing schedule. Some re-schedule.

[0036]

According to the production management system of the present invention, all lots are divided into an online scheduler and an offline scheduler without scheduling all processes. In the offline scheduler, only progress management lots (lots with a high priority or importance) are scheduled, so that a large amount of memory can be reduced when a schedule is created.

The processing schedule completed by the offline scheduler reflects a long-term plan for lots whose progress must be controlled. Therefore, the processing schedule of the entire line is created by the online scheduler in a short period of time based on the processing schedule. By executing the processing according to the processing schedule,
Precise delivery date control is possible.

Further, even for mass production lots that are not progress control lots, approximate progress can be managed by using marker lots. Therefore, it is effective to predict the delivery date by scheduling lines in which a large number of lots flow. Becomes possible.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a production management system of the present invention.

FIG. 2 is a diagram showing a configuration of an online scheduler and an offline scheduler of the production management system of the present invention.

FIG. 3 is a diagram showing a method of creating a processing schedule in an online scheduler and an offline scheduler.

FIG. 4 is a diagram illustrating a method of creating a processing schedule in an online scheduler when a processing schedule extracting unit is added.

FIG. 5 is a diagram showing a concept of performing progress management by a marker lot.

FIG. 6 is a diagram showing how the number of lots that can be calculated increases when a marker lot is used.

FIG. 7 is a diagram showing an example of a method for creating a pseudo process chart.

FIG. 8 is a diagram showing a concept of processing schedule creation in the production system of the present invention.

FIG. 9 is a diagram illustrating an outline of an LSI manufacturing procedure.

FIG. 10 is a diagram showing an example of a process chart.

FIG. 11 is a diagram showing a conventional production management method.

[Explanation of symbols]

1. Online scheduler 2. Offline scheduler 3. Progress management / data collection system 4.
Manufacturing / inspection equipment 5 ... Lot progress chart 6 ... Database 10 ... Registration information database 11 ... Processing result database 12 ... Lot selection unit 13,18 ... Schedule creation unit 14,17 ..Communication control unit 15 ・ ・ ・ Processing schedule extraction unit 16 ・ ・ ・ Processing schedule (short term) 19 ・ ・ ・ Priority change unit

Claims (3)

(57) [Claims] 1. A a process table showing the steps of the manufacturing process, from operation schedule of each manufacturing apparatus, each manufacturing lot or scheduled processing for each of the manufacturing apparatus, progress management / data in a production line
A first step of creating a scheduled processing of short <br/> period online scheduler directly connected to the data acquisition system, the on
Offline scheduling based on information from the line scheduler
A second step of creating a long-term processing schedule with a juler, and performing a process and storing data on the basis of the processing schedule of the first step . The step is a step of selecting only high-priority or important lots from the production lots, a process chart of the selected lots, a delivery date of the selected lots, and a current date of the selected lots. Priority and
The processed final step of the selected lot
Passing to the step of: determining the priority of a non-selected lot from the priority of the selected lot determined in the second step; and setting a short-term processing schedule of the production lot. is composed of a step of creating said second step, said selected lots received from the first step
Process chart, delivery date of the selected lot, and
The current priority of the selected lot and the
A step of allocating the processing steps of the production lot to the apparatus based on the processed final step, creating a long-term processing schedule until the completion date, and predicting the completion date; if the completion date is later than the delivery date, priority is given A step of increasing the priority, lowering the priority of a spare lot, and determining the processing priority of the lot; and transferring the determined priority to the first step. Method. 2. The method according to claim 2 , further comprising the step of re-creating a processing schedule as an important lot using lots extracted at regular intervals from the lots of low importance.
Production management method according to claim 1, characterized in that it has. 3. A scheduler for calculating a scheduled processing time of a process to be processed for each manufacturing lot or each of the manufacturing apparatuses from a process table indicating a procedure of the manufacturing process and an operation schedule of each manufacturing apparatus. In a production management system that performs processing progress and stores data based on a created processing schedule, the scheduler manages progress / data collection in a manufacturing line.
The online scheduler which is directly connected with the current system o
Offline schedule that can communicate with online scheduler
The online scheduler comprises at least a communication control unit that communicates with the offline scheduler, and selects only a lot having a high priority or importance from among the production lots, and processes the lot. Table, delivery date, processed
A final selection step, a lot selection unit for transmitting the current priority to the communication control unit, and setting the priority of all lots based on the priority by the offline scheduler, and creating a schedule for each lot. It consists of a scheduling unit, the off-line scheduler, a communication control unit for communicating with the online scheduler process chart from the online scheduler, delivery, treatment
It consists of a schedule creation unit that creates a processing schedule based on the data of the finished last process and the current priority, and a priority change unit that changes the priority by checking the delivery deadline from the processing schedule created by the schedule creation unit A production management system characterized by the following.