JPH0944208A - Simulation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To speedily perform a simulation processing again by model correction after retrogression to the simulation process at optional time halfway from a process. SOLUTION: When process results of processes 12 and 13 at the end of the processes are not good, model corrections are made by a process 14 and the simulation process is performed again for the corrected model after a retrogression process is performed by a process 15, but it is judged from a correction influence judgement table whether or not the state transition at last simulation process time is affected by the model correction before the simulation process is performed again, and when it is judged that no influence is exerted, the last simulation process is used as it is.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a discrete computer simulation method for predicting a production situation at a production site, and more particularly to a simulation method which is suitable when a model is corrected by trial and error. It is.

[0002]

2. Description of the Related Art Japanese Patent Application Laid-Open No. Sho 62-135958 discloses a conventional technique for storing a simulation process and outputting a simulation result using the process. Also,
Japanese Patent Application Laid-Open No. 61-237162 discloses a method of partially modifying the simulation model and performing the simulation again.In this case, at the time of the first simulation, all possible events are determined in advance. After that, when the simulation model is modified, the calculation is performed again only for the event that has been changed due to the model modification.

[0003]

However, in the case of the former simulation processing process storage method, the simulation processing is simply stored, and it is not used at all for the return of the simulation processing. I have. Further, in the latter simulation method, the simulation processing after the model correction (second time) is performed using the simulation processing result before the model correction (first time), but the effect of the model correction covers a wide range. In this case, more time is required for the calculation than in the simulation processing before the model correction. If it is determined in the simulation processing after the model correction that the calculation requires more time than that before the model correction, the simulation processing is stopped and the simulation processing is performed using the simulation method before the model correction. It had to continue.

In addition to these problems, when simulating the production situation at the production site, the facility capacity is not changed at any time during the simulation.
Simulation period (for example, Monday to Friday)
The facility capacity is often changed at a specific time (for example, early morning on Friday). In such a case, during the simulation process (for example,
It is necessary to return the simulation processing from the time) to the specific time, but until now, the simulation processing has been performed again from the beginning, and much time has been required.

An object of the present invention is to provide a simulation method which can return to the simulation process at an arbitrary time from the middle of the simulation process, and can reduce the processing time when the simulation process is performed again after the return and model correction. To serve.

[0006]

SUMMARY OF THE INVENTION The object of the present invention is to store, in parallel with a simulation process, a change in a model state that occurs in the course of a simulation process, and to execute the model process again when the simulation process is performed again after the model is corrected. This is achieved by determining whether or not the model state transition in the previously performed simulation processing can be used based on the modified influence determination data.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Prior to the detailed description of the present invention, an outline thereof will be described. In a simulation system, a state transition storage table for storing model state transitions as a simulation internal time and its transition state. Is provided for the previous and current simulations, and a modification influence determination table for determining which state transition in the state transition storage table becomes unusable at the time of re-simulation due to model modification is a state transition element (for example, product, (Equipment, waiting), and these tables are used to perform a return process and a re-simulation process.

That is, the simulation model stored in advance is sequentially shifted with the progress of the simulation process, and every time the model changes in such a process, the internal time and the shift at the time of the shift are changed. (For example, a product waits for a certain "wait" and processing starts at a certain facility) is stored in the state change storage table. In this way, the model is changed one after another, and the state change is stored. If it is necessary to correct the model and re-simulate in such a state, the repetition of such processing is stopped and the simulation processing is temporarily stopped. It is something to be done. After this, the simulation is retraced. In the retrace process, the content stored at the end of the state transition storage table is read out, and the time at which the transition state transits is replaced with the current internal time. . Thus, the internal time of the simulation is set back to the transition time of the read state transition. At the same time, the stored state transition (for example, waiting for a certain product at a certain “wait” and starting processing at a certain facility) is the opposite of this state transition (processing at a facility where a certain product is present). Stop and go back to the model that remembers the "wait"). As a result, the simulation can be returned by one state change, and by repeating such a return process, the simulation can be returned to a desired internal time.

When the re-simulation is performed after the model is modified, the internal time at which the state has changed during the previous simulation and the state change at that time are read one by one from the state change storage table corresponding to the previous simulation. The elements of the transition (for example, a product ends processing at one facility and then waits at the next “wait”) (the product, the facility, and the wait) are affected by the partial modification of the model performed this time. If not, that is,
If the element in the correction influence determination table is not affected, the same state change as in the previous time occurs, and the previous simulation process is used as it is. In other words, calculation processing of work time such as when and which product starts or ends processing in which equipment, and selection processing such as which product is processed and which equipment is used (these Is referred to as a normal simulation process). However, if there is at least one element affected by the model modification, the previous simulation process cannot be used, and in such a case, the normal simulation process is performed. As a result, when a part of the model is modified, the effect of the model modification is limited to only a part of the element, and most of the element is not affected. is there. This means that in the normal simulation process,
Many table comparisons have to be performed to find out which equipment is available, and a lot of time is required for the calculation. This is because the processing is completed in the time required for the comparison calculation.

[0010] However, the effect of the model correction is initially limited to only the elements related to the model correction, but the effects are extended to the elements not related to the model correction as the simulation progresses. For example, assuming that a model modification such as changing the processing capacity of a certain manufacturing facility is performed, a product passing through this facility is affected. If the affected product is further processed by another facility, the facility may be affected by the model modification. Each time this model modification is affected,
The corresponding equipment or product in the correction impact determination table is to be in an affected state.

Now, the present invention will be described in detail with reference to FIGS. First, a simulator according to the present invention will be described. FIG. 1 shows an entire system configuration in an example, and a case of simulating a production situation at a production site in a factory is assumed in the present example. It has become. As shown, the simulator 1 uses a CRT as a means for inputting a simulation model.
The display 2 and the keyboard 3 are provided, and the CRT display 2 and the printer 4 are provided as means for displaying the simulation result. Further, the simulator 1 has a model storage table 5 for storing a simulation model, a time at which a state changes in a process of a simulation performed before a partial modification of the model, an equipment name, a product name, and a change state. (State start, completion) state change storage table 6 for storing the effect of model correction, modification effect determination table 7 for storing the effect of model correction
Is provided. These tables are C
The simulation is advanced by the operation of the PU 8, and FIG. 2 shows an outline of the processing flow.

According to this processing flow, first, a model to be simulated is input in processing 10. The method of inputting the model is shown in FIG. As shown in the figure, the names of the products (A, B, C... In this example) to be put in, and furthermore, each product is any manufacturing equipment (Equipment 1 to in this example)
In 3), the number of minutes to be worked is input, and the number of work minutes is input in ST (Standard Time; standard work time). The order of operations on the product is set in a first step, a second step, a third step, and so on. For example, in the case of the product A, as shown in the drawing, firstly, after being worked in the facility 1 for 10 minutes, the work is performed in the facility 2 for 30 minutes. The simulation processing is performed by the processing 11 based on the model input as described above. FIG. 4 shows a processing flow of the simulation. In FIG. 4, five processes 23, 24, 29, 33, and 37 indicated by broken lines are used after the model is corrected, and details thereof will be described later.

As shown in FIG. 4, in the simulation process, first, at step 20, an equipment table (see FIG. 5), a waiting table (see FIG. 6) and a product table (see FIG. 7) as components of the model storage table are processed. It is to be initialized. Of these tables, the equipment table is the equipment name corresponding to the equipment and the equipment is used (in use) or not used (unused)
The state of all equipment extracted from the model storage table in the initialized state is all unused, The work completion time is set to zero. Also,
The waiting table includes a waiting name corresponding to the equipment and a product name waiting in the waiting state. In the initialization state, all the waiting product names are cleared. Further, the product table includes a product name, an equipment place where the product is currently being processed, and a state of the product at the place (“Waiting”, “Work”, “Completed”). In, the equipment of the first process of the product is read in the order of products in the model storage table, and if the read equipment is not used, the state of the corresponding equipment in the equipment table is set to “in use”, and the work completion time of the equipment table is The ST of the first step of the product in the storage table is substituted. Further, the equipment location of the product in the product table is set to the first process, and the state is set to "work".
Conversely, if the read equipment is “in use”, the product is registered in the waiting table corresponding to the first process in the model storage table of the product, and the equipment location of the product in the product table is further specified. The state is set to “waiting” in the first step. The above processing is performed for all products in the model storage table. After the processing is completed, the internal time of the simulator, that is, the time in the simulator is set to zero, whereby the pre-simulation processing is completed. ing.

In the process 21, it is determined whether or not the work has been completed for all the products. If the simulation is in progress, the process 22 determines the minimum non-zero work completion time in the equipment table. Things are selected,
This is set as the next internal time. By this process 22, the equipment whose work is to be completed next is selected, and the internal time is advanced to that time. Since the work is completed at the advanced internal time, the work is completed and, at the same time, the process 25 is stored in the state change storage table 26.
Thus, the state transition is stored in the format shown in FIG. This format is "Done"
The state transition memory that stores the state transition in such a format, including the internal time when the state transition occurs, the completed product name, the equipment name, and the state indicating the completed state transition. The table is used in the re-simulation described later, but the state transition memory table shows how the model has changed in the simulator, and the movement of the model can be reproduced later.

By the way, the product related to the work completion is sent to the next facility according to the information in the model table.
If it is determined in step 27 that the next equipment does not exist, the product is not considered to be processed thereafter because the work has been completed, and if the next equipment is vacant, The status of the facility is "in use" by process 28
After that, the work completion time is calculated by the process 30, and the start of the work on the product is additionally stored in the state change storage table 26 by the process 31. On the other hand, when the next equipment is “in use”, the product is registered at the end of the product waiting in the waiting table by the process 32, and then the status is changed to “waiting” by the process 34. The information is registered in the table 26. By these processes, the process for the product related to the work completion is completed at the internal time. However, since the equipment related to the completion of the work remains empty, a new work is started in the equipment. If there is a product waiting in front of the facility related to the completion of the work in the judgment by the process 35, the product registered at the top of the “wait” is to be worked. In the process, the state of the product table is set to "work" by the process 36, and the waiting product names in the waiting table are reduced one by one. Thereafter, the work completion time is calculated in the same manner as in the previous case, and the start of the work is additionally stored in the state change storage table 26 by the processing 38. Further, if there is no product waiting in front of the facility related to the work completion, the facility is opened by the processing 39. Note that various methods are known for taking out a product in a “waiting” state. For example, there is known a method of taking out in order of small ST or in order of small remaining man-hour. Depending on this take-out method, it takes a lot of time for the take-out process, and the productivity greatly changes, but in this example, the first-in first-out method is adopted.

With the above processing, the processing of one process for the product is completed, and at the same time, the transition information of the product is stored in the state transition storage table. By repeating the same processing, the simulation processing proceeds. It is something that has become.

By the way, if the processing result is displayed by the processing 12 when the simulation processing is completed, a result as shown in FIG. 9 can be obtained. In this example, the vertical direction indicates the equipment name, the horizontal direction indicates the internal time, and each rectangle indicates the start and the end of the work on the product described therein. In the process 13, the operator determines whether or not there is a production problem based on the displayed contents. If it is considered from the result that the processing of the product C in the first step is too long, the ST of the product C in the first step is 4 by the operator.
It is supposed to be changed from 0 minutes to 30 minutes. That is,
The model is partially modified.

By the way, it is useless to perform the simulation process again from the beginning with this slight model correction, and by using the state transition storage table storing the state transition in the previous simulation process,
The speed of the re-simulation process is increased. The re-simulation is performed at high speed by simplifying the process of removing the waiting product, the process of finding the work completion time, and the process of finding the minimum work completion time. This re-simulation will be described. It is as follows.

That is, if the result of the previous simulation processing is not satisfactory, the model is corrected in processing 14 and the corrected model is again subjected to simulation processing in processing 15, but prior to the processing. In order to determine whether or not the state change in the previous simulation processing is affected by the model correction, the presence or absence of the effect of the model correction is registered in the correction effect determination table. If it is determined from the correction influence determination table that the model is not affected by the model correction, the previous simulation process is used as it is. More specifically, all the impact items in the equipment modification impact determination table shown in FIG. 10 are "no impact", and all the impact items in the wait revision impact judgment table shown in FIG. 11 are "no impact". The influence item in the product modification influence judgment table shown in FIG.
The product C is set to “affected”. After the correction effect determination table is initialized as described above, resimulation is performed. The processing flow of the resimulation processing is substantially the same as that shown in FIG.
33 and 37 have increased. Explaining the processing especially increased for these re-simulation processings, in processing 29, 33, 37, it is determined whether or not the influence of the model modification has spread. That is, it is determined in a process 23 whether or not the previous simulation process can be used. If the process is not affected at all by the model correction, a high-speed process (the process flow is shown in FIG. 13) is performed in a process 24. However, if it is affected, a new simulation process is performed without using the previous simulation process.
However, the effects of model modification spread from products to equipment and waiting, and from equipment and waiting to products,
In these processes 29, 33, and 37, it is determined whether the equipment or waiting that has been affected by the model correction, the product has another new equipment or waiting, or the possibility that the model correction affects the product. Things. In process 29, if the completed product is affected by the model modification, the next equipment is affected, and if the next equipment is affected, the completed product is affected. Whether or not the completed product and the next equipment are affected is determined from the product correction influence determination table and the equipment correction influence determination table. If there is an effect, the product or equipment in each table is set to “affected”. Similarly, in the process 33, the correction influence determination table is updated by checking the influence relationship between the product for which the work is completed and the waiting, and in the process 37, the influence relation between the waited product and the completed facility is updated, and the model correction Whether or not the impact is spreading is being investigated at any time.

Here, the high-speed processing at the time of the re-simulation processing, that is, the algorithm of the processing 24 will be described with reference to FIG. 13. In this processing, the result of the current simulation processing is utilized by using the result of the previous simulation processing. Can be obtained at high speed. As shown in FIG. 13, first, among the work completion times at the equipment affected by the model correction, the minimum value is T, and the state transition stored at the top of the previous state transition storage table is Assuming that the time is F, T and F are compared by processing 41. It should be noted that the time required to search for T is smaller in the number of comparisons than in the normal case, so that the search can be performed earlier than that.

According to the comparison, it is known that the equipment will be completed in the nearest future from the present time. In the comparison, if T is smaller than F, it is the affected equipment, so the process returns to the normal simulation process. If T is larger than F, it is unaffected equipment, so that the process moves to the next high-speed processing. It has become.

The above process is a process of finding a facility to be completed next and advancing the internal time. After that, the state transition information stored at the top is read out from the state transition storage table and the information is read. The state transition is deleted from the table. In the process 42, it is determined whether or not the product and the location (equipment or waiting) in the read state change information are already affected by the model correction based on the correction effect determination table. If there is, the process returns to the normal simulation process, but if the product and location are not affected, the previous simulation process, that is, the read state change information can be used as it is. After that, the process branches to any of processes 44, 46, and 48 according to the item indicating the state in the information by the process 43, but basically, the equipment table, the waiting table, and the product table are updated. Are additionally stored in the state transition storage table for simulation this time. The table for this time is also used for storage in the state change storage table in the normal simulation processing.

After the re-simulation processing is performed as described above, the simulation processing result is displayed in the same manner as in the previous case. The simulation for is repeated.
By repeating such processing, production conditions considered to be optimal for production can be found.

The re-simulation process has been described above. Finally, the return process of the simulation will be described as follows.

That is, in an actual simulation, the simulation processing cannot be easily interrupted at a place where model correction or the like is to be performed. When the simulation process is stopped before the process to be interrupted, the simulation process is advanced little by little, so that the simulation process can be interrupted at a desired location and the model can be modified. However, when the vehicle has passed a desired location, there has been no other operation until now until the simulation is restarted from the beginning. It is the present invention that even when the simulation process has passed a desired location, the simulation process is easily returned to that location.

Now, the processing flow of the subsequent return processing is shown in FIG.
As shown in FIG. 4, when the simulation process is interrupted and the return process is performed, the process 1 is performed in the entire flow of the simulation process shown in FIG.
Backtracking processing is performed at 1 and 15. First, the state transition stored at the end of the state transition storage table is read by the processing 60, and the backward processing is performed based on the read state transition information. First, after the simulation internal time is returned until the time in the information read by the process 61, the subsequent processes are branched depending on the state of the state transition read by the process 62. If the status is "start", in other words, if the work has started at that time, the status of the product table corresponding to the product read by the process 63 is set to "waiting", and the top of the waiting table is set. Is registered, and the state of the equipment table corresponding to the equipment read by the process 64 is "unused". If the status is completed, the status of the product table is set to "work" by the process 65, the equipment table is set to "in use" by the process 66, and The internal time is set. Furthermore, if the status is waiting, the same processing is performed.
According to 67, the state of the product table is set to “work”, and the location is set to the previous process on the input model table. In the post-processing 68, the state of the equipment table corresponding to the equipment in the immediately preceding process of the read equipment is set to "in use", and the current internal time is set as the work completion time. ing. By repeating such processing based on the state transition storage table, the state of the simulation model can be returned.

The present invention has been described above. As can be seen from the above description, when the partial modification of the simulation model is small and the effect of the partial modification of the model does not extend over a wide range of the model, the calculation time of the simulation processing is reduced. Will be done. In particular, "waiting" which requires a long calculation time
By reducing the number of times that the product is taken out from the, the overall calculation time is greatly reduced. In addition, the calculation time for determining the next work completion time is considerably reduced, so that not only the simulation calculation time is reduced, but also the simulation can be repeated in a short time to determine the simulation result. It is also possible to save operator labor costs. Further, even when the influence of the model correction reaches a certain wide range, the determination as to whether or not the previous simulation process can be used is performed by using the correction influence determination table in 2
Since it is only necessary to refer to the times (product and equipment or product and waiting correction effect determination table), the increase in processing time due to this is almost negligible.

If the simulation can be easily moved backward for an arbitrary time, it is possible to improve the interactivity and responsiveness when the simulation is advanced too much.

[0029]

As described above, according to the present invention, after returning to the simulation processing at an arbitrary time from the middle of the simulation processing, when the simulation processing is performed again by the model correction after the returning, the processing is promptly performed. There is an effect that it is carried out.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a simulator according to the present invention.

FIG. 2 is a diagram showing a schematic flow of a simulation process according to the present invention;

FIG. 3 is a diagram for explaining a model input method;

FIG. 4 is a diagram showing a flow of a simulation process directly related to the present invention in FIG. 2;

FIG. 5 is a diagram showing an equipment table as an element constituting a model storage table.

FIG. 6 is a diagram showing a wait table as an element constituting a model storage table.

FIG. 7 is a diagram showing a product table as an element constituting a model storage table;

FIG. 8 is a diagram illustrating a state transition storage table;

FIG. 9 is a diagram illustrating a display example of a simulation processing result;

FIG. 10 is a diagram showing an equipment table as an element constituting a modification influence determination table;

FIG. 11 is a diagram showing a waiting table as an element constituting a modification influence determination table;

FIG. 12 is a diagram showing a product table as an element constituting a modification influence determination table;

FIG. 13 is a diagram showing a flow of high-speed processing at the time of resimulation processing;

FIG. 14 is a diagram depicting a flow of a post-simulation return process;

[Explanation of Codes] 1 ... Simulator, 5 ... Model storage table, 6 ... State transition storage table, 7 ... Correction impact determination table, 8 ... C
PU

Claims (1)

[Claims] 1. A discrete system simulation model to be simulated is stored as modifiable, and
This is a simulation method in which a future state is discretely created by sequentially shifting the model by a simulation process, and a transition state of the model as a result of the simulation process can be displayed. The transition of the model state that occurs in the process is stored as the simulation internal time and the transition state, and when the simulation processing is interrupted,
Each time a change in the stored model state is read out in the opposite direction to the storage order, the model at the time of the processing interruption is sequentially returned to the state before the change by one, and after the model is returned to the desired internal time. When the simulation is performed again in accordance with the model correction, it is determined whether or not the change in the model state in the previous simulation processing process can be used, using the data that has been permitted to be added as the correction influence determination data. A simulation method, characterized in that: