JPH05174030A - Job shop scheduling expert system - Google Patents

Abstract

PURPOSE:To satisfy the time limit of delivery instructed for production and also to improve the working ratio of each machine so as to simplify the allocation of a job shop line. CONSTITUTION:A parts production instruction is decomposed into the jobs and these jobs are allocated to each job based on a calendar in a job shop scheduling expert system. This system contains the master data 2-5 which include the proper information on the parts, the processes, the machines, the jigs, etc., the working type calendar data 6 which includes the information on the working time type of each machine, the limitation data 7 which includes the limitation information on allocation, and a scheduling means 1 which decomposes a production instruction into jobs based on each of the preceding data and then allocates the jobs to the machines respectively with application of a rule having the priority on the time limit of delivery and a rule having the priority on the working ratio. The means 1 divides the allocation of jobs into the general machines, the M/C and the outside order processes to apply the rule having the priority on the time limit of delivery to the general machines and to apply the rule having the priority on the working ratio to the M/C respectively. Otherwise an excessive time-limit-of-delivery margin delay rule is applied together with a time-limit- of-delivery margin limit-priority rule, a daily machined product-priority rule, a multiprocess parts-priority rule, and a continuous machining-priority rule.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a job shop scheduling expert system for scheduling empirical knowledge of scheduling in a program according to a work procedure performed at a production site.

[0002]

2. Description of the Related Art In recent years, scheduling of mechanical equipment has become an important issue in the manufacturing industry. Scheduling of machining in production management means MRP (Matter
Receive a part order to be processed from a higher-level production management system such as ial requirement planning), and specify a schedule that clearly indicates the parts to be machined, the quantity, the planned start date, the planned completion date, etc. for each machine equipment. This is a procedure to create.

There is a job shop line as a line form for machining. In contrast to a flow shop (so-called flow work), where materials are put into a line where machines are lined up and products are made as the materials flow in the order in which the machines are lined up, the job shop line is the order in which machines line up Regardless of the order, it is a line that requires allocation of machines to be processed according to orders, that is, scheduling is required. In such a job shop line, it is difficult to find out what allocation should be performed to raise the operating rate as a whole and to secure the required delivery date.

Generally, there are several hundreds to several thousands of orders for dozens of machines, and the combination of jobs in that case is enormous. Therefore, it is not realistic to examine all of these combinations to find the optimum combination and determine the allocation because it takes a long processing time even if a computer is used.

Therefore, at many production sites, it is difficult to create a detailed schedule in advance, so daily study and adjustment are performed based on a rough schedule created manually to make the schedule more detailed and the work schedule fixed. I am trying.

[0006] Further, conventionally, several methods for allocating job shop lines have been studied and proposed. One of them is a method called an OR method, which is solved by a mathematical model, and the priority of each order is obtained and assigned by using a linear expression weighting delivery date, inventory, operating rate, cost, etc. as an evaluation expression. Is to do.

Another method is a method by simulation. In this case, for example, an evaluation scale for prioritizing a delivery delay as small as possible is given and held as a movable parameter. .. Therefore, in this method, a simulation is performed with another fixed combination of parameters, and the parameters are gradually changed to obtain a solution.

[0008]

However, in the conventional method using the mathematical model, the job shop line cannot be well controlled unless proper weighting is given, and it is difficult to give weighting. There is. In addition, the situation changes dynamically as the allocation is made, but if the weighting is fixed, there is a problem that the change cannot be dealt with.

Further, if various situations occur in the process of allocation, there is a best schedule according to the situation, and an expert on site can derive the best schedule. In the method using a function, scheduling is always performed by the same evaluation method, so that it is not possible to deal flexibly in each situation.

The present invention is to solve the above-mentioned problems, and a job shop scheduling expert system capable of easily allocating a job shop line satisfying the delivery date of a manufacturing instruction and increasing the operation rate of a machine. It is intended to provide.

[0011]

To this end, the present invention is a job shop scheduling expert system that decomposes parts manufacturing instructions into jobs and allocates the decomposed jobs to each machine according to a calendar. , Master data with unique information about jigs, operation type calendar data with information about operation time type for each machine, constraint data with allocation constraint information, and manufacturing instruction by referring to each data The present invention is characterized by comprising a scheduling means for disassembling and applying the delivery date priority rule and the operation rate priority rule to allocate each job to each machine. Further, the scheduling means assigns jobs to the general machine and M /
C (Machining Center) and outsourced process, apply the delivery date priority rule to general machines, the operation rate priority rule to M / C, or the excess delivery time delay delay rule, delivery time allowance limit priority rule, daily processed products It is characterized by applying the priority rule, the multi-process item priority rule, and the continuous processing priority rule.

[0012]

The job shop scheduling expert system of the present invention has master data, operation type calendar data, and constraint data, and the scheduling means refers to these to decompose the manufacturing instructions into jobs, and the delivery date priority rule and operation rate priority rule. , Overdue margin delay rule, Deadline margin limit priority rule, Daily processed goods priority rule,
Applying the multi-process item priority rule and continuous processing priority rule,
Since each job is assigned to each machine by dividing it into a general machine, an M / C, and an outsourcing process, it is possible to perform scheduling with an increased operating rate of the machine while satisfying the delivery date of the manufacturing instruction.

[0013]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing an embodiment of a job shop scheduling expert system of the present invention.

In FIG. 1, the scheduling unit 1
Depending on M / C (machining center), general machines other than M / C, and outsourced processes, delivery priority rule, operation rate priority rule, delivery allowance margin rule, excess delivery allowance delay rule, daily processed goods allocation rule, multi-process parts priority By applying rules, continuous processing allocation rules, etc., jobs that are subject to parts order are allocated to the calendar for each machine,
This allocation is performed by referring to the master data composed of the component master 2, the process master 3, the machine master 4, and the jig master 5 based on the input data.

The input data includes a part order (manufacturing instruction) 9, a work record (machining record) 10, a calendar file 6 storing an operation calendar and an operation type calendar, and a constraint file 7, and stores a scheduling result described later. Do the scheduling result file 8
Is.

The manufacturing instruction has information on an order number, a part number, a processed quantity, a material arrival date, and a delivery date. The processing record has information such as an order number, a part number, a record number of completed records, a start schedule, and a completion schedule. The operation calendar has information on date, day of week, and operation flag. The operation type calendar has information on a date, a machine symbol, an operation possible time type, and a setup possible time type. The constraint file has information on the machine operation standard time, assembly parts, multi-process parts, delivery time margin days, excess delivery time days, and daily processed product reference days.

The component master 2 has information on a component number, a component name, and the number of processes. The process master 3 has information on a part number, a work order, a machine symbol, a processing time, and a setup time. Machine master 4 is a machine symbol, machine name, type (M / C, general, outsourced), operating time type,
It has setup time type information. And
The jig master 5 has information on the jig name, the number, the part number, the work order, and the number of processed pieces.

Therefore, in a parts order, first, a part is specified by the part number, and a manufacturing instruction is issued to process the part by what time. Therefore, we use the part number as a key to find out what kind of part it is.
Information about parts and processes is stored in the part master 2 and the process master 3. In the part master 2, the name of the part and how many steps there are are known.
You can see what process is used in what order, what kind of machine is used, how long it takes to process, and how long it takes to set up. Furthermore, if you use the machine as a key, you can see the machine master 4, and you can see the name of the machine, standard operating patterns, and the time zones when setup is possible. Some machines require a jig, and M / C is applicable, but in this case, machining is not possible even if the machine is empty without a jig. For that purpose, it has a jig master 5, the name of the jig, how many such jigs exist, and what step of which part to process.
It has information on how many pieces can be processed at one time. Furthermore, whether or not the machine and jig are available depends on whether or not allocation is possible. I am looking at 5.

Allocation (scheduling) in the scheduling unit 1 is an operation of placing jobs on a calendar. In principle, the job has a calendar for a fixed scheduling period of one month, three months, etc. This is to repeat the operation of embedding those jobs in a calendar for each machine. In this allocation, on the one hand, there is a demand to increase the operating rate of the machine as much as possible, that is, to increase the operating time of the machine as much as possible, and on the other hand, to keep the delivery date and finish the processing quickly. In general, the two are in a relationship that if the operating rate of the machine is high, the machining is completed early, but there is a case where the processing is not necessarily so. That is, when it is desired to finish the machining of a certain component quickly, there arises a problem that the operating rate of the machine cannot be increased due to the machining of the certain component. Therefore, it is necessary to properly use the operation rate priority rule and the delivery date priority rule.

In the present invention, the respective priority rules are applied to each of the M / C, the general machine other than the M / C, and the three equipment resources having different evaluation methods for the outsourcing process. For example, the delivery date priority rule is applied to general machines in principle, and the operating rate priority rule is applied to M / C in principle. With regard to outsourced processing, we consider that the capacity is infinite, and assign it indefinitely. Also, M
/ C is allocated separately at night and day when it operates in unmanned operation.

As described above, since the operation rate is important for M / C, a job with a short machining time, which requires frequent setup change during the day, is assigned. In other words, since people are on during the daytime, the setup change is carried out rapidly, and the one with a short processing time is done. And those with long processing time are allocated at night. Therefore, sort according to the length of processing time, and assign from the shortest daytime to the longest nighttime.

The delivery date priority rule gives priority to the job having the smallest delivery date margin among the conflicting jobs. In the case of this delivery date priority rule, first, the material arrival date has passed in the first process, the processing from the first process to the previous process has been completed after the second process, and the assembly must be performed. For items that cannot be processed, the delivery margins for each of the allocation targets are calculated and sorted, provided that the assembly process has been completed. In this case, when the delivery time margin exceeds a certain value and there is too much delivery time margin, an alarm is given and the allocation is excluded from the allocation target (excess delivery time allowance delay rule). This is for avoiding such a delivery delay because a product with a long leg may cause a delivery delay. For example, in job A, where there is not enough delivery time,
At the time of T1, the allocation could not be made due to the relationship with the previous process, but at the time of T2, the allocation was made possible because the previous process was completed. In this case, if the job B having an excessive delivery time is allocated, the job A cannot be allocated and the allocation is delayed until the time T3, resulting in a delay in the delivery. Therefore, in order to avoid such a problem, with respect to the job B having an excessive delivery time margin, the allocation is delayed until the delivery time becomes less than a certain delivery time margin.

If, for example, the operating rate priority rule is applied, the operating rate will increase, but some will not meet the delivery date. In this way, when the allocation rate is delayed or the deadline margin is reduced to a certain value or less by applying the operation rate priority rule or the excess deadline margin delay rule, an alarm is generated and the priority is increased and allocation is performed according to the deadline margin limit priority rule. This is to relieve the delayed delivery of items that were postponed. In the present invention, delay in delivery is avoided while increasing the operating rate in this way.

The delivery time allowance is the number of days to express how much time is left until the delivery time in the process of allocating jobs for the remaining processes by occupying the machine, and delivery time allowance = delivery time-assignment date−Σ (remaining time It is calculated by each processing time of the processing step / standard processing time of one day) -1. Therefore, if the standard machining time of M / C is increased, the order of priority of the orders including many M / C steps is lowered. This is because if the standard machining time of M / C is increased, the machining capability will be greatly evaluated, and therefore the delivery margin will be increased. On the other hand, the allocation priority of orders that hardly include M / C processes goes up.

In the case of M / C, if the number of parts to be processed is large, it is necessary to devise a jig that requires no setup change or labor. Therefore, in such a case, it is more efficient to process a fixed number every day, so calculate the number of processed items per day from the number of processed items for each order and process a fixed number each day. Allocation is given priority over processed products (daily processed product rule). In addition, for example, multi-process parts (spindle) having more than a dozen processes are time-consuming and prone to delays in delivery, so allocation is given top priority (multi-process part rule).
In a process where the same M / C is continuous, reworking and work-in-progress will be reduced when processing is performed by combining continuous processes. Therefore, like the current combination of continuous processes, work in continuous processes is prioritized and assigned. (Continuous processing rule).

Next, the operation procedure will be described. 2 is a diagram for explaining the overall operation flow, FIG. 3 is a diagram showing an example of a schedule parameter setting screen, FIG. 4 is a diagram showing an example of a schedule evaluation screen, and FIG. 5 is an example of an order number order result display screen. , FIG. 6 is a diagram showing an example of display of results by machine,
FIG. 7 is a diagram showing a result display example by M / C, and FIG. 8 is a diagram showing a machine operation status graph display example.

As shown in FIG. 2, first, a component master 2, a process master 3, a machine master 4, and a jig master 5
Are maintained (step S1).

The part master 2, the process master 3, and the jig master 4 have a mechanism for searching by using the part number as a key. Therefore, when correcting or adding the part number,
Be sure to edit the component master 2 first, and edit the other items whenever there is a change. Then, whether the machine designated as the process master 3 is registered in the machine master 4, the jig corresponding to the M / C process of the process master is registered in the jig master 5, or the number of processes of the component master 2 is Check the consistency of the master such as whether it matches the number of processes of the process master.

In the part master editing, new parts are registered, old parts are deleted, part numbers are changed, and model names, part names, and specifications are changed. The number of processes is automatically input by inputting the process master. If the part number is corrected and the master file is saved, all the common parts of the part master, the process master, and the old part number of the jig master are rewritten to the new part number.

In the process master editing, the process order, machine symbol,
The processing time, setup time A (time for switching lots), and B (time for changing workpieces) are added, deleted, and changed.

In machine master editing, a machine symbol, machine name, type, number of pallets (M / C of multi-sided pallet), operating time type (operating time zone), setup time type are added, deleted, and changed.

In the jig master editing, the jig name, the number, the part number, the work order, and the number of processes are added, deleted, and changed. In the operation master calendar edit, set the factory operation date. Operation type of machines other than M / C (other machines),
The setup operation type of M / C is "0" except the operation day.
Set to (closed).

Next, the operation type calendar is edited (step S2). In this edit ,. 8:00 to 17: 0
0 ,. 8: 00-18: 00 ,. 8: 00-19:
00 ……. A machine operation time zone (operation type) and a setup time zone are set every day such as 8:00 to 6:00.

Then, an order is input and edited (step S3). In order input, for example, a part manufacturing order created by an MRP (material requirement planning) system is received by a floppy disk and input. Create an order file by combining the uncompleted orders of the last confirmed orders and the orders read this time.

Then, by setting the inference condition, the scheduling parameter (reference value used for scheduling),
Substitute machine settings and daily processed product settings (step S
4). Normally, scheduling is performed with default values, so the scheduling parameters are input only when changing.

As shown in FIG. 3, the scheduling parameters include, for example, a marginal number of days for delivery due to be assigned with the highest priority, an excess number of days for extra delivery to make the allocation order the lowest, and a number of days for calculating the number of processed pieces per day for each processed product , Standard processing time per day for delivery time allowance calculation, M for delivery time allowance calculation
There is a standard processing time such as / C per day.

Next, for example, the scheduling period is set to 90.
Scheduling is executed for a certain period of time (about 3 months) (step S5). During the execution of this scheduling, it is classified according to the number of jobs to be allocated, such as all-day stop, non-retention operation, slightly retention, retention, and significant retention, for example, color-coded to display the operating status of each machine and the retention status of jobs. .. In the case of M / C and general machines, the number of jobs left unassigned for each machine is totalized and displayed at the end of the one-day assignment. In the case of the outsourcing process, the job retention is not detected because the allocation is performed assuming that the capacity is infinite.

Next, the scheduling result is examined (step S6). As for the scheduling result, "Schedule evaluation" in which the entire schedule as shown in Fig. 4 is evaluated with statistical values, "Delivery due to delayed order order result display", "Order number order result display", "Machine result display" , “Result display by M / C”, and “Machine operation status graph display” screen that displays the planned operation status of the machine in a graph.

In the bar chart display, either the result of scheduling just before (latest data) or the result of holding last time (holding data) is selected and displayed, and the data automatically confirmed after the work instruction is confirmed. indicate.

As described above, the evaluation of the entire schedule is performed by using statistical values. For example, if the delivery date is delayed, the operating rate of the general machine is 30 days from the scheduling start date due to the difference between the delivery date and the scheduled completion date. The sum of the scheduled operating time is calculated for the target, and this is divided by the sum of the available operating time for 30 days. The M / C operating rate is calculated based on the operating rate calculated for 30 days. Of the job is evaluated by the time the job is waiting for the machine to be vacant, and the M / C is evaluated by the total time of waiting the machine to be vacant.

In the display of the result of the delayed delivery order, the bar charts corresponding to the machines are displayed in order from the order with the largest delivery delay, and a machine symbol is added to the bar chart. In the order number order result display, as shown in FIG. 5, a bar chart corresponding to the machines is displayed in order number order. In the result display by machine, the operation schedule for each machine is displayed in a bar chart corresponding to a job (lot unit) as shown in FIG. In the result display by M / C, the operation schedule of M / C is displayed in a bar chart corresponding to the work, as shown in FIG.
In the machine operation status graph display, the operation status of each machine and the job retention status are displayed in a graph as shown in FIG. 8, the daily operation time of the selected machine is displayed in a bar graph, and the job retention number is displayed in a line graph. .. In these cases, the bar charts of fixed orders and the bar charts of other orders are displayed in different colors so that they can be distinguished from each other. Detailed data that cannot be represented by a bar chart are displayed in a pop-up window as in the bottom right of the screen shown in the figure. This pop-up window opens by clicking the area such as order number, machine code, M / C name, etc., and "Close" in the pop-up window.
Close by clicking on the area.

When a satisfactory schedule is obtained as described above, the schedule is fixed, work instructions are issued, and machining procedures are issued (step S7).

If necessary, the entire schedule evaluation, machine-specific schedule evaluation, delivery delay order, order schedule, order number order, order schedule,
A management form such as a machine-specific schedule and an M / C-specific schedule is output (step S8).

In the schedule evaluation, the delivery delay average, maximum, minimum, machine (excluding M / C) operation rate average, maximum, minimum,
M / C operating rate average, maximum, minimum, machine (excluding M / C) waiting time average, maximum, minimum, M / C waiting time average, maximum, minimum are output. The machine-specific schedule evaluation summarizes the operating status and job (or work) retention status for each machine. The total available operating time for 30 days from the scheduling start date, the available operating time for the entire scheduling period (90 days) Total of 30 from the start date of scheduling
Total scheduled operating time for the day, total scheduled operating time for all scheduling periods, 30 from the start date of scheduling
Daily operating rate (total scheduled operating time / total available time), operating rate for all scheduling periods, total time waiting for machine (lot) waiting for machine to be empty, M / C
The item of the total time the work is waiting for the machine to be vacant (including the time set on the pallet and waiting) is output.

Then, processing is performed according to the work instruction sheet,
Fill in the work instruction sheet with the actual number of processed pieces and return it. This actual number of processed pieces is input (step S9) and used as basic data for the next scheduling. The order in which all the steps are completed is output (step S10) and input to the MRP system.

Further, the scheduling execution of the present invention will be described in detail. 9 is a diagram showing an allocation flow of processes other than M / C, FIG. 10 is a diagram showing a daytime allocation flow of the M / C process, and FIG. 11 is a diagram showing a nighttime allocation flow of the M / C process.

Scheduling is performed by a scheduler that incorporates the empirical knowledge of experts, and the scheduler performs scheduling according to the set inference conditions. As shown in FIG. 2, when the result of scheduling using the default value of the inference condition does not satisfy the request, rescheduling is performed by changing the inference condition, changing the machine capacity, changing the order, or the like.

In the scheduling, the jobs are forwardly assigned on a daily basis in order from the first day of the given scheduling period. The jobs are assigned on a lot basis to a machine other than M / C and on a M / C basis. The C step is performed in units of work (units of parts set in a jig during processing).

The premise of scheduling is that processing cannot be performed over a plurality of processes on the same day, provided that the same M
It can be processed in the process where / C continues. In addition, it is possible to perform setup regardless of the ability of the worker within the time zone during which setup is possible. The job allocation priority order is determined as follows.

As shown in FIG. 9, jobs using machines other than M / C are applied in principle to the due date priority rule and sorted in ascending order of the due date margin. Apply and assign the margin limit priority rule.

In order to increase the operating rate of M / C, jobs using M / C are sorted by processing time, and in principle, operating rate priority rules, multi-process parts rules, excessive delivery delay rules, delivery allowance limits. Applying the priority rule, the daily processed product allocation rule, and the continuous processed allocation rule, as shown in FIG. 10, they are allocated in the ascending order of the processing time at daytime and in the ascending order of the processing time at night as shown in FIG.

When the schedule plan obtained as a result of scheduling does not satisfy the request, the scheduling conditions are changed as follows and rescheduling is performed.

Change scheduling parameters.
If it is judged that the machine capacity is insufficient, we will deal with it by increasing the operating time and setup possible time. If it is necessary to equalize the load on the machine, consider changing the material arrival date, using alternative machines, dividing lots, dividing orders (creating new orders), etc.

As described above, in the general machine, the allocation is performed in the ascending order of delivery time. In the M / C, in order to increase the operation rate of the machine, the allocation is performed in the order of shorter processing time in the day and in the order of longer processing time at night. However, if an order occurs that cannot meet the delivery date, it is necessary to improve the schedule to meet the delivery date of all orders within a range in which the machine operating rate of the M / C does not significantly decrease. Therefore, the deadline margin days are set as a threshold value for raising the job allocation order that is likely to not meet the deadlines due to insufficient deadlines, and during the process of scheduling, the deadline margins will be greater than the deadline margin days value. Allocate the job that has run out with the highest priority.

For example, when the number of machining days calculation is set to 25, the number of machining per day in the order of 48 total machining is 48/25 + 1 = 2. Therefore, if the number of days for calculating the number of processed pieces is increased, the number of processed pieces per day is reduced, and as a result, the number of processed pieces per day is increased. On the contrary, if the number of days for calculating the number of processed pieces is reduced, the number of processed pieces per day increases and the number of processed pieces per day decreases per day. Therefore, by adjusting the number of days for calculating the number of processed pieces, it is possible to balance the allocation of the daily processed products and the ordered processed products.

As described above, in the job shop scheduling of the present invention, when a job generated from an order is temporarily stored or a job assigned in the middle of the scheduling is temporarily stored, it is possible to directly and directly from the program. A database with each job as a basic unit is created in a storage area that can be written and referenced at high speed.
This is called working memory. This database function can be realized by using the logic language Prolog, and unlike conventional programming, there is no need to clearly distinguish between data and programs, and all are described in basic units called Horn clauses. By adding / deleting this horn clause, data can be added / deleted and programs can be added / deleted.

In the job shop scheduling of the present invention, there are various restrictions that must be satisfied when jobs are assigned, and they can be classified into the following three types.

Constraints received from orders …… Start date, work in process / quantity, delivery date Constraints received from equipment resources …… Operating status, available operating hours, setup changeable hours, jig usage, pallet usage allocation Restrictions received from assigned jobs ........ Pre-process completion, assembly, immediately preceding work, daily processing number of processed products Limitation To check whether these constraints are satisfied, first of all, information on jobs to be allocated is set as various master data. And the input data, and then the information about the individual constraints of is searched from the various masters and the input data to compare them. Moreover, in job shop scheduling that requires allocation of a large number of jobs, it is necessary to perform data search and comparison at high speed.

Therefore, data is searched for and compared by finding a combination of data satisfying the given constraint from the entire working memory. By utilizing this function for checking constraints when assigning jobs, scheduling can be performed in a short time. This function corresponds to, for example, a pattern matching function called unification provided in the well-known Prolog.

Further, equipment resources such as machines and jigs include machines and jigs that can be replaced, and when the machines and jigs that should be originally used cannot be used, these alternative resources are used.
At this time, first check whether the constraint is satisfied for the machine that should be originally used, if it is not satisfied, check whether the constraint is satisfied similarly for the alternative equipment, and if the constraint is still not satisfied, The next alternative equipment will be examined in order. In the present invention, such repetition is not described in the conventional explicit loop, but the constraint is automatically satisfied and the alternative facility resource is searched by declaratively describing the constraint which the facility resource should satisfy. It is realized by using a mechanism to perform. Prolog has such a function. Prolog
If pattern matching fails during execution, it will automatically go back and restart execution from the place where there is another candidate for pattern matching. Prolo
By utilizing the backtracking function of g, constraint checking of equipment resources including alternative equipment can be easily described. Further, similarly, there are a plurality of job allocation rules, and if one allocation rule is applied and the constraint condition is not satisfied, the next allocation rule is applied, and this is repeated until the application rule is established. This process can also be easily described by taking advantage of Prolog's backtracking function as well as alternative equipment.

In the present invention, the job shop scheduling uses the method of the expert system which utilizes the empirical knowledge acquired from the expert. When an expert performs job shop scheduling, various items such as "margin until delivery date", "machine operation rate", "ease of setup change", and "some work in progress" are assigned to each job to be assigned. Consider the conditions and decide the allocation priority to allocate to the machines. The empirical knowledge of such experts is the "assignment condition section" that evaluates various conditions and selects one job.
It can be expressed by an "allocation rule" composed of a pair of "allocation work units" that indicate how to allocate it. You can add these allocation rules in the middle of development,
Frequently deleting or changing the content of rules.

Although it is possible to describe such rules in a conventional procedural language represented by FORTRAN, it becomes very difficult to add / delete / change. On the other hand, Pr
If rules are described using a language structure called an olog clause, Prologs are independent in terms of clauses, so additions, deletions, and changes can be easily performed.

There is a large amount of data that increases or decreases momentarily during the scheduling process, such as job allocation candidates and jigs on M / C pallets. A function to handle such variable length data is required,
Traditional languages do not have the ability to handle variable length data. Pro
Variable length data can be easily handled by using the log list processing function.

[0065]

As described above, according to the present invention, in principle, the operation rate priority rule and the delivery date priority rule are separately used for scheduling, so that the machine equipment can be installed while satisfying the delivery date of each manufacturing instruction. It is possible to formulate a schedule that maximizes the utilization rate of.
In addition, even a skilled person can perform a scheduling work that has required one day or more by hand in just a few minutes, and can perform a quick scheduling. It can also be used for simulations for process changes and introduction of new mechanical equipment.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of a job shop scheduling expert system of the present invention.

FIG. 2 is a diagram for explaining an overall operation flow.

FIG. 3 is a diagram showing an example of a schedule parameter setting screen.

FIG. 4 is a diagram showing an example of a schedule evaluation screen.

FIG. 5 is a diagram showing an example of an order number order result display screen.

FIG. 6 is a diagram showing an example of display of results by machine.

FIG. 7 is a diagram showing an example of displaying results by M / C.

FIG. 8 is a diagram showing a display example of a machine operation status graph.

FIG. 9 is a diagram showing an allocation flow of processes other than M / C.

FIG. 10 is a diagram showing a daytime allocation flow of the M / C process.

FIG. 11 is a diagram showing a nighttime allocation flow of the M / C process.

[Explanation of symbols]

1 ... Scheduling unit, 2 ... Parts master, 3 ... Process master, 4 ... Machine master, 5 ... Jig master, 6 ...
Calendar file, 7 ... Constraint file, 8 ... Scheduling result file, 9 ... Parts order (manufacturing instruction), 10 ... Work record (machining record)

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yuichi Tanioka 1-3-2 Shibaura, Minato-ku, Tokyo Shimizu Construction Co., Ltd. (72) Inventor Shuro Omi 1-3-2 Shibaura, Minato-ku, Tokyo Shimizu Construction company (72) Inventor Tomoyuki Aratake 1-2-3 Shibaura, Minato-ku, Tokyo Shimizu Construction company

Claims (11)

[Claims] 1. A job shop scheduling expert system that decomposes parts manufacturing instructions into jobs and allocates the decomposed jobs to each machine according to a calendar, and has unique information about parts, processes, machines, jigs, etc. Master data, operation type calendar data that has information about the operating time type for each machine, constraint data that has allocation constraint information, and manufacturing instructions are decomposed into jobs by referring to each of the above data, and delivery date priority rule and operation rate priority rule A job shop scheduling expert system characterized by comprising a scheduling means for applying each of the jobs to allocate each job to each machine. 2. The scheduling means divides the job allocation into a general machine, M / C, and an outsourcing process, applies the delivery date priority rule to the general machine, and applies the operation rate priority rule to the M / C. 2. The job shop scheduling expert system according to claim 1. 3. The job shop scheduling expert system according to claim 1, wherein an excessive deadline delay delay rule that excludes the jobs from the allocation target is applied to a job having an excessive deadline exceeding the fixed value. 4. The job shop scheduling expert system according to claim 1, wherein the allocation priority is raised for jobs whose delivery deadlines are less than a certain value. 5. The job shop scheduling expert system according to claim 1, wherein a daily processed product rule for processing a fixed number of parts every day is applied to a part having a large number of processed parts. 6. The job shop scheduling expert according to claim 1, wherein a continuous processing rule is applied to a part in which the processing steps are continuous so that the started part is continuously processed until the final step of its M / C. system. 7. The job shop scheduling expert system according to claim 1, wherein a multi-step parts priority rule for increasing the priority is applied to parts having a large number of steps. 8. The scheduling result is examined, and if there is a defect, the date of arrival of material, delivery date, operation pattern of machinery and equipment,
Rescheduling is performed by changing the scheduling conditions. 9. The job shop scheduling expert system according to claim 1, wherein the scheduling means displays the scheduling result in a bar chart. 10. The job shop scheduling expert system according to claim 9, wherein the scheduling result of the late delivery order, order, machine, and M / C is displayed in a bar chart. 11. The job shop scheduling expert system according to claim 9, wherein detailed data on the schedule result display screen is displayed on a pop-up window.