JPH05233590A - Production system simulator device using petri net - Google Patents

Abstract

PURPOSE:To immediately and efficiently change a product to be produced, trouble like a fault of a machine, and addition/reduction of buffers and to perform simulation in the time direction opposite to the production progress from arrangement and the time of a certain token. CONSTITUTION:A data input part 1 takes in data for modeling of a production system and outputs data of each product data, machine data, buffer data, etc., and a model constituting part 2 generates a Petri net model for each of these data and combines Petri net models for respective data to constitute a Petri net model of the production system. Therefore, the Petri net for data corresponding to the change is only replaced to change the model of the production system in the case of the change of the product, the machine, or buffers.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a Petri net-based production system simulator for analyzing the flow of information in a production system, state analysis, etc. The present invention relates to a Petri net-based production system simulator device that enables flexible planning and countermeasures for future production in the event of an accident.

[0002]

2. Description of the Related Art As one of methods for analyzing information flow, state analysis, etc. in a discrete event system such as a production system by modeling, there is a modeling method by Petri net (reference document: "Introduction to Petri net"). Author; J.
L. Pitasun. Publisher; Kyoritsu Publishing Co., Ltd.).

In the Petri net, as shown in FIG. 25, a condition is represented by "O" called a place, and an event is represented by "|" called a transition. Therefore, the Petri net has two kinds of nodes, that is, a place PL and a transition TR, and these nodes are connected by an arrow called an arc indicating the direction of state transition by its direction. The arcs are coupled from the place PL toward the transition TR, or vice versa, from the transition TR toward the place PL.

As described above, the static relationship between the condition and the event of the production system is simply expressed by the graph structure of the Petri net, and the dynamic behavior of the production system is called a token at each place in the Petri net. "Is placed and is represented by the movement of the token accompanying the occurrence and completion of the event.

The token is moved according to the initial marking indicating the initial state of the system, that is, the tokens assigned in advance to each place, according to the firing rules shown in the following, and the tokens move in the Petri net according to this firing rule. By moving, it represents the dynamic behavior of the system.

[0006] Tokens move within the Petri Net by firing a Petri Net transition.

[0007]. In order to fire a transition, the transition must be fired, that is, there must be tokens in all the input places of the transition (meaning that the condition is met and the event occurs).

[0008] If the transition fires, move the token from the input place to the output place, that is, remove the token from the input place, generate a new token, and put the token in the output place.

As described above, the Petri net is most suitable for modeling a production system that undergoes state transitions due to the occurrence and completion of discrete events. The production system, according to Petri Net, is
By expressing the system that is hierarchically integrated with the machine including the parts, the work cell, and the factory by the place and the transition of the Petri net, it is possible to express the hierarchy from the lower net to the upper net.

[0010]

However, such a conventional production system simulator device using a Petri net has the following problems.

[0011]. Since all the data of the entire production system was modeled together, the model is immediately changed when the type of product to be manufactured and the required number are changed, an accident such as a machine failure, or the addition or reduction of buffers occurs. Can not. For this reason, it was not possible to swiftly respond to or deal with the production plan changes in the variable-variety / variable-volume production system and the above-mentioned accidents / changes.

[0012]. A method of introducing the concept of time delay into the firing rule of Petri net has been proposed, but how to actually incorporate the concept of time delay into the Petri net and build a model is a procedural and automated method. It has not been proposed, and it has been difficult to accurately and easily model motions such as work accompanied by changes over time in complex problems.

.. In the conventional method, when an accident such as a machine failure and the movement of personnel occurs during operation, it is not possible to immediately predict a situation that may occur in the future due to the accident.

.. In the conventional method, the progress of the production, that is, the state transition is only simulated in the forward direction in time. Therefore, if the production is carried out based on the given production plan, the demand such as delivery date is satisfied. We were able to analyze the progress of production, but in order to satisfy it,
It was not possible to examine the situation in the opposite direction of the production progress such as by what time the parts should be prepared, considering the time.

Therefore, the present invention has been made in view of such a problem, and changes promptly and efficiently in response to changes in products to be produced, accidents such as machine failure, and addition / reduction of buffers. An object of the present invention is to provide a production system simulator device by a Petri net that can perform a simulation in the reverse time direction of production progress from the arrangement and time of a certain token.

[0016]

In order to achieve the above object, the invention according to claim 1 constructs a model of a production system by a Petri net and simulates the production system by the model. The simulator device incorporates data for modeling the production system, product data indicating the type of product and the work required for each product, machine data indicating the type and number of machines and the arrangement of machines, and the type of buffer And a petri net model for each data such as product data, machine data, buffer data, etc. divided by the above input part, and a petri net model for each data. A petri net model of the production system is constructed by combining the net models. Characterized by comprising a model construction section which, a.

According to the second aspect of the present invention, data for modeling a production system in a production system simulator device by a Petri net which constructs a model of the production system by the Petri net and simulates the production system by the model. Based on the data input by the input unit and the data input by the input unit, the work required for each product that causes a state transition is determined by a unit time transition having a predetermined unit time of ignition duration corresponding to the work time. Provided in one or a plurality of series, each with a zero time transition indicating the start and end of the work and having no ignition duration at the front and the end of the unit time transition provided in one or more, respectively, Place a place between each adjacent transition While showing the transition between the touching transition and the place by an arc indicating the direction of the state transition, the machine and the buffer etc. that do not cause the state transition are shown by the place, and the zero time transition of the work required for each product A model construction unit for procedurally and automatically constructing a Petri net model of a production system by connecting the machine and a place indicating a buffer with the arc, and a Petri net model of the production system constructed by the model construction unit In the above, the token is moved from the unit time transition to the place after the firing duration of the unit time transition has elapsed after the firing at the unit time transition, while the token is moved from the zero time transition to the place. It is performed by firing in a zero-hour transition, and is equipped with a simulation unit that indicates the state of product production at any time, the state variable values such as machine free space and buffer remaining capacity by the number of tokens in the place. And

In a third aspect of the present invention, data for modeling the production system in a production system simulator device by the Petri net that constructs a model of the production system by the Petri net and simulates the production system by the model. Incorporating the above, the product data indicating the product type and required work for each product, the machine data indicating the machine type and number of machines and the machine layout, and the buffer data indicating the buffer type and capacity etc. In the case of manufacturing data including necessary work for each product that causes a state transition, an input section and a Petri net model for each data such as product data, machine data and buffer data divided by the input section , A unit time transition with a predetermined unit time of ignition duration to the working time Therefore, a zero-hour transition that indicates the start and end of the work and has no ignition duration is provided at the forefront and the end of the unit time transition provided in one or more series. Machine data and buffer data that do not cause a state transition, while placing a place between each adjacent transition and connecting each adjacent transition and place with an arc that indicates the direction of the state transition In the case of etc., a machine, a buffer, etc. are indicated by a place, and a petri net model of a production system is constructed by connecting the zero-time transition of the Petri net of the product data and the place indicating the machine and the buffer by the arc. The model building section and the model In the Petri net model of the production system constructed by the construction unit, the token movement from the unit time transition to the place is performed after the firing duration time of the unit time transition has elapsed from the firing in the unit time transition, while the zero time has passed. The token is moved from the transition to the place by firing at the zero-time transition, and the state variable value such as the product production status at any time, the machine's free space, the remaining capacity of the buffer, etc. is indicated by the number of tokens in the place. And a simulation unit.

In a fourth aspect of the present invention, data for modeling a production system in a production system simulator device by a Petri net which constructs a model of the production system by the Petri net and simulates the production system by the model. Based on the data input by the input unit and the data input by the input unit, the work required for each product that causes a state transition is determined by a unit time transition having a predetermined unit time of ignition duration corresponding to the work time. Provided in one or a plurality of series, each with a zero time transition indicating the start and end of the work and having no ignition duration at the front and the end of the unit time transition provided in one or more, respectively, Place a place between each adjacent transition While showing the transition between the touching transition and the place by an arc indicating the direction of the state transition, the machine and the buffer etc. that do not cause the state transition are shown by the place, and the zero time transition of the work required for each product By constructing the Petri net model of the production system for forward simulation by connecting the machine and the place showing the buffer by the arc, the unit time transition and zero in the Petri net model of the work required for each product. Reversing the direction of the arc connecting the time transition and the place, and connecting the zero time transition of the work required for each product and the place indicating the machine and the buffer by the arc. By the opposite direction When performing a forward simulation in each of the model building unit that builds the Petri net model of the production system for simulation, and the Petri net model of the production system for the forward and backward simulations built by the model building unit While setting the token placement and time in the initial state, set the target token placement and time in the final state when performing a simulation in the reverse direction, and move the token from the unit time transition to the place above. Is performed after the firing duration of the unit time transition has elapsed since the firing in the unit time transition, while the token is moved from the zero time transition to the place by firing in the zero time transition, and the product is produced at any time. The present invention is characterized by comprising a simulation unit which indicates a state variable value such as a production state, a machine free space, and a remaining capacity of a buffer by the number of tokens in the place.

[0020]

According to the first aspect of the invention, the input unit takes in data for modeling the production system and divides it into data such as product data, machine data and buffer data, and the model construction unit produces the product data. , A Petri net model is created for each data such as machine data and buffer data, and a Petri net model for the production system is constructed by combining the Petri net model for each data. Therefore, when there is a change in product production, a machine, or a buffer, the model can be changed simply by replacing the sheet corresponding to the change.

According to the second aspect of the present invention, the input section takes in data for modeling the production system, and the model construction section receives the data and performs necessary work for each product causing a state transition, A unit time transition having an ignition duration of a predetermined unit time is provided in one or a plurality of series corresponding to the work time, and the work is performed at the front and the end of the unit time transition provided in the one or a plurality of series. Zero transitions that indicate the start and end of each transition and have no firing duration are provided, and a place is placed between each adjacent transitions, and adjacent transitions are connected by an arc indicating the direction of the state transition. By showing.
In addition, machines and buffers that do not cause state transitions are indicated by places, and zero-time transitions of work and places indicating machines and buffers are connected by arcs to construct a Petri net model of a production system.

In the Petri net model of the production system constructed in this way, the simulation unit ignites the unit time transition and the zero time transition based on the ignition duration, and the state of product production at any time. In addition, state variables such as machine free space and buffer remaining capacity are indicated by the number of tokens in each place.

According to the third aspect of the present invention, the input unit takes in data for modeling the production system and divides the data into product data, machine data, buffer data, etc., and the model construction unit produces the product data. , Create a Petri net model considering the time of state transition for each data such as machine data and buffer data.

That is, in the case of manufacturing data including work required for each product that causes a state transition, one or a plurality of unit time transitions having a firing duration of a predetermined unit time are associated with the work time. Provided in series, zero time transitions indicating the start and end of the work and having no ignition duration are provided at the front and the end of the unit time transitions provided in series or in series, and adjacent to each other. This is shown by arranging places between transitions and connecting adjacent transitions and places with an arc indicating the direction of state transition. On the other hand, in the case of machine data and buffer data that do not cause a state transition, the machine and the buffer are indicated by a place, and the zero time transition of the Petri net of the product data and the place indicating the machine and the buffer are connected by an arc. Build a Petri net model of the production system.

Then, in the Petri net model of the production system constructed in this way, the simulation unit ignites a unit time transition and a zero time transition based on the ignition duration, and the state of product production at an arbitrary time. In addition, state variables such as machine free space and buffer remaining capacity are indicated by the number of tokens in each place.

According to another aspect of the invention, the input unit takes in data for modeling the production system, and the model construction unit constructs a Petri net model of the production system for forward simulation based on the data. In the Petri net model of the work required for each product, the direction of the arc connecting the unit time transition and the zero time transition and the place is reversed, and the zerot time transition of the work required for each product and the machine. We construct a Petri net model of the production system for backward simulation by connecting and and the place showing the buffer with an arc.

Then, in each of the Petri net models of the production system for the forward and backward simulation, the simulation unit sets the initial token arrangement and time when performing the forward simulation, while When performing a simulation, set the target token placement and time in the final state, and fire based on the firing duration of the unit time transition and zero-hour transition to determine the product production state at any time and the machine's State variable values such as vacancy and remaining capacity of the buffer are indicated by the number of tokens in the place.

[0028]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a production system simulator device using a Petri net according to the present invention will be described below with reference to the drawings.

FIG. 1 briefly shows the structure of the production process simulator, data for modeling the production system, and the flow thereof.

This production process simulator has an input unit 1,
It is composed of a model construction unit 2, a simulation unit 3, an environment setting unit 4, and an output unit 5.

The input unit 1 inputs various data for constructing a Petri net model of a production system to produce a product,
It is configured such that the elements of the production activity in the factory such as machines and buffers, that is, the elements on the production system are divided and output as product, machine and buffer data 1a to 1c. That is, the input unit 1 includes a product name (number) for each product, a work required for each product and their predecessor work, a work name (number) for each job, required work and their predecessor work, and each predecessor work. Product data 1a in which information such as work name (number), machine used in each work, work time, etc. is described, information such as machine name (number) and its number, machine layout (dependency), etc. The machine data 1b that has been stored, the buffer name (number) of the buffer that is the storage location, the buffer data 1c that describes information such as the capacity and available semi-finished products (work), and the like are divided for each related data. Output.

The model construction unit 2 is a product from the input unit 1,
It is configured to construct a Petri net model 2a for forward and backward simulations based on the machine and the buffer data 1a to 1c as described later.

The simulation unit 3 executes the simulation of the Petri net model 2a for the forward and backward simulations constructed by the model construction unit 2 under the simulation environment set by the environment setting unit 4 to make a transition. When the ignition conflicts occur, the ignition conflict is resolved based on the work priority rule stored in the rule base 3a, and the simulation result such as the start time of each work is output as the schedule 3b.

The environment setting unit 4 increases or decreases the number of required products, increases the number of required products of a certain product to the simulation unit 3 based on the simulation environment data 3a that influences the simulation environment such as machine failure and stop, or It is configured to set up a simulation environment such as stopping a machine.

The output unit 5, together with the schedule 3b generated and output by the simulation unit 3, completes various evaluation criteria such as the completion time calculated based on the data on the schedule 3b, delivery delay, machine operation rate, etc. as a final result. 5a is configured to be output via a display, a printer or the like.

FIG. 2 shows in detail the processing procedure for modeling in the model building unit 2, that is, the Petri net model building. Hereinafter, the modeling process in the model construction unit 2 will be described based on FIG. 2 and with reference to FIGS.

First, in the model construction unit 2, the product, the machine and buffer data 1a to c divided from the input unit 1 for each element of the production system such as the machine and the buffer are input, That is, a Petri net model is created for each element, and each model is output in units of sheets 1A to 1C.

That is, for the product data 1a, a Petri net (see FIG. 3) for each work is created based on the work time and the like described in the product data 1a (step 10).
0), and then create a Petri net for each job (see FIG. 4) from each Petri net for each work based on the preceding relationship between the jobs (step 110), and further for each job based on the preceding relationship between the jobs. The Petri net for each product in the product data 1a, that is, the product sheet 1A for each product (see FIGS. 5 and 6) is created (step 120).
For the machine data 1b, the machine data 1b
Based on information such as machine name and its layout described in
Each machine is shown by a place, and each place is arranged according to the arrangement of the machines to create a Petri net of the machine data 1b, that is, the machine sheet 1B (see FIG. 7) (step 200), while the buffer data 1c is ,
Each buffer is indicated by a place based on information such as the buffer name described in the buffer data and its capacity, and the Petri net of the buffer data 1c, that is, the buffer sheet 1C.
(See FIG. 8) is created (step 300).

After the product sheet 1A, the machine sheet 1B and the buffer sheet 1C are prepared, the respective sheets 1A-
Based on C, a Petri net model for forward and backward simulation (hereinafter referred to as forward model and backward model) is constructed.

When constructing the forward direction model, first, the product sheet 1A and the machine sheet 1B are joined by an arc in a work unit as shown in FIG. 9 (step 400), and then the product sheet 1A and the buffer sheet C. And are united in work units by arcs as shown in FIG. 10 (step 41).
0). This builds the forward model.

On the other hand, when constructing the backward model,
First, all the arcs of the product sheet 1A are changed to the opposite direction (step 500, see FIG. 11), and then the product sheet 1A and the machine sheet 1B are operated in the unit of work as in the case of the model construction for the forward direction. And are joined by an arc as shown in FIG. 12 (step 510), and further, the product sheet 1A and the buffer sheet 1C are joined by an arc as shown in FIG. 13 (step 520). This builds a model for reverse simulation.

Therefore, in the present embodiment, not all the data is collected at once and the production system is modeled, but as described above, each data indicating the elements on the production system such as products, machines and buffers. Petri nets, that is, sheets are created, and a model of the production system is constructed by combining these sheets, so when changing the type of product or the required number, only the product sheets are replaced, and when a machine failure occurs Replaces only the machine sheet, and replaces only the buffer sheet to add / reduce buffers.It is only necessary to replace the sheet to which the changed data belongs. Partial modification of the model can be performed immediately and efficiently compared to the conventional model that was given.

As a result, high-speed production planning, production plan change, future production volume prediction / countermeasure planning at the time of an accident, etc. in a production system for performing variable product quantity / variable volume production
High performance.

3 (a) and 3 (b) show Petri net model PMw of each work.

In (a), a manufacturing operation required for each product is performed by a unit time transition (hereinafter, 1 hour) having a firing duration of a predetermined unit time (for example, 1 hour in this embodiment).
-Transition) TR1 is provided in series or in series corresponding to the work time, the start and end of the work are shown at the foremost part and the last part of the 1-transition TR1, and there is no ignition duration. Zero time transition (hereinafter referred to as 0-transition) TR
s and TRe are provided respectively, and a place PL is arranged between adjacent transitions to make each transition TR
And each place PL are connected by an arc so as to indicate the direction of state transition.

In this figure, since the work requires a work time of τ time, 1-transition TR1 requires τ set, and the work is added to the front and the end of the 1-transition connected in series. 0-transition TRs and TRe indicating the start and the end of the transition are respectively provided, and a place PL is placed between the transitions TR.
Are provided and are connected by an arc.

In (b), the Petri net model PMw of the work shown in (a) is simplified by a τ time transition TRτ that has an ignition duration of τ time. Therefore, the Petri net model shown in (a) and the Petri net model shown in (b) are equivalent.

4 (a) and 4 (b) show Petri net models PMw 'of each work.

Here, the relationship between work and work for manufacturing a product is premised on that one work is composed of a plurality of works.

(A) shows a Petri net model PMw 'of a certain work. This work consists of five works, and five works are simplified and expressed as shown in FIG. 3 (b). The transitions TR τ1 to 5 each showing the Petri net model of each work are expressed by connecting them by arcs via the place PL based on the preceding relation between the works constituting the work. In the figure, τ1-5
Indicates the duration of each transition TR τ1 to 5, that is, the working time of each work.

(B) is a simplified representation of the work Petri net model PMw 'shown in (a). In (a), the work start transition TR of the work having no preceding work is shown.
τ1 and the work end TRτ5 of the work having no succeeding work are respectively defined as the start and end of the work, and the work start and end transitions TRτ1 and TRτ5 are represented by enclosing them with a broken line. As in the case of FIG. 3, the Petri net model shown in (a) and the Petri net model shown in (b) are equivalent.

FIGS. 5A and 5B show Petri net models PMg of each product.

Here, it is assumed that a plurality of products are produced in the production system handled in this embodiment, and one product is composed of a plurality of jobs.

(A) shows that a certain product is manufactured through five jobs. That is, the Petri net model PMg of the product is a Petri net model P of five jobs shown in a simplified manner as shown in FIG.
Mw'1-5 are connected by an arc through a place PL based on the preceding relation between the jobs, and a place P indicating the start and end of the product Petri net, respectively.
Place Ls and PLe, and set the place PLs indicating the start and the work Petri nets PMw'1 and PMw'2 without preceding work to 0.
-While coupling through the transition TRs and the place PL, the place PLe indicating the end and the work Petri net PMw'5 which does not have a succeeding work are 0-transition TR.
Represented by combining via e and place PL.

(B) shows the Petri net model PMg of the product shown in (a) in a simplified manner by surrounding the place PLs indicating the start and the place PLe indicating the end thereof with solid lines. Petri net model P shown in a)
Equivalent to Mg. The Petri net PMg constructed for each product is the product sheet 1A.

FIG. 6 shows the product sheet 1A1 of each product 1-N.
~ N is shown.

In this way, product sheets 1A1 to N are created for each product 1 to N, and each product sheet 1A1 to N is based on the product data for each product 1 to N as shown in FIGS. A petri net of work / work is constructed in Fig. 1 to show the types of products 1 to N and the order of work / work.

FIG. 7 shows the machine sheet 1B.

The machine seat 1B is provided with as many places PLM1 to m corresponding to the machines M1 to m, and the arrangement of the machines M1 to m is indicated by the arrangement of the places PL. There is. In addition, the place P of the machines M1-3
Places connected by chain lines are 1 like LM1-3
Will show the line.

FIG. 8 shows the buffer sheet 1C.

In the buffer sheet 1C, a plurality of buffers (storage locations) B 1 to b are indicated by capacity places PLB1 to b, respectively, and the capacity is set for each capacity place PLB1 to b. Buffer B 1 ~
The type of b and the capacity of each buffer B1-b will be indicated. The number of tokens in each capacity place PLB1 to PLB indicates the remaining capacity of each buffer, that is, the free capacity.

FIG. 9 shows the combination of a work Petri net model PMw in the forward model product sheet 1A and the machine place PLM on the machine sheet 1B.

As shown in this figure, 0-transitions TRs and TRe indicating the start and end of the working Petri net model PMw and the place PLM of the machine sheet are connected by an arc. As a general rule, 1
0-transitions TRs, TRe indicating the start and end of one working Petri net model PMw are coupled to the machine place PLM on the same machine sheet 1B.

FIG. 10 shows the forward model product sheet 1A.
Work Petri net models PMw1 and PMw2 above,
The coupling with the capacitive place PLB on the buffer sheet 1C is shown.

This combination is 0-transition TRe1 indicating the end of the work Petri net PMw1 and 0-transition indicating the start of the work Petri net model PMw2 which is the succeeding work.
The same buffer sheet 1C as transition TRs2
In principle, it is coupled to the capacity place PLB shown in the upper buffer, and the direction of the arc is from 0-transition TRs2 indicating the start of the subsequent working Petri net model PMw2 to the capacity place PLB showing the buffer, and from the capacity place PLB to the preceding place. Work Petri net PMw1 0
-Go to transition TRe1. The number of tokens in the place PL at the end of the work Petri net PMw1 represents the number of product stocks in the work, and the number of tokens in the capacity place PLB represents the free capacity of the buffer.

FIG. 11 shows a working Petri net model PMw.
Shows a change to the reverse model of.

That is, this figure shows the arc of the work Petri net PMw shown in FIG. 3 in the direction opposite to the production proceeding direction, that is, the left-right opposite direction in the figure. A reverse model is created by reversing the directions of all arcs on 1A.

FIG. 12 shows the reverse direction product sheet 1A.
The work Petri net model PMw above and the machine place PLM on the machine sheet 1B are shown to be coupled.
In the backward model, 0-transitions TRs and TRe indicating the start and end of each work Petri net PMw,
The machine place PLM is coupled by an arc in the opposite direction to the direction of the arc of the Petri net shown in FIG.

FIG. 13 shows the reverse model product sheet 1A.
The coupling between the working Petri net models PMw1 and PMw2 in FIG. 3 and the capacity place PLB of the buffer sheet 1C is shown.

In the backward model, the directions of all arcs in the respective work Petri net models PMw1 and PMw2 are opposite to those shown in FIG.
The direction of the arc connecting the work petri net models PMw1 and PMw2 is also opposite to the case shown in FIG. 10, that is, 0-transition TRe1 indicating the start of the immediately preceding work petri net model PMw1 to the capacity place PLB. 0 from the subsequent work Petri net PMw2 from PLB
Suitable for transition TRs2.

Next, the simulation of the forward and backward models of the production system constructed by the model constructing unit 2 in this way will be described.

FIG. 14 shows the simulation processing procedure in the simulation section 3 in detail.

The simulation unit 3 first inputs the simulation environment data 4a, selects the simulation direction under the simulation environment set by the environment setting unit 4, and sets one of the forward direction and the reverse direction ( Step 600).

When the forward model is selected, the initial state of marking and time is set in the forward model (step 610), while when the backward model is selected, marking and time are set in the backward model. The final state which is the target of is set (step 615), and in both cases, the 0-transition capable of firing is extracted (step 620).

At this time, if there is an ignitable 0-transition and there is a conflict, a 0-transition with a high work priority is selected based on the rule base, and an ignitable 0-transition exists and there is a conflict. If not, select the 0-transition (step 63).
0), the token is moved from the input place to the output place of the 0-transition associated with the firing of the 0-transition, that is, the marking is changed (step 6).
40), followed by extraction of the 1-transition that can be fired (step 650). On the other hand, when there is no ignitable 0-transition when extracting the ignitable 0-transition, the ignitable 1-transition is immediately extracted without selecting and igniting the 0-transition (step 650).

Then, in the extraction of the ignitable 1-transition, if there is the ignitable 1-transition,
That is, in the middle of working on a certain product (step 6
50 "yes"), the marking of the input place PL is changed with the firing of the 1-transition, that is, the token is deleted from the input place PL (step 660), and the time is 1-transition firing duration "1". Only (step 670), and then change the marking of the output place PL due to the firing of the 1-transition, that is, set a token in the output place PL (step 68).
0), and subsequently, it is determined whether or not to interrupt the simulation (step 690).

If it is determined that the simulation should be continued (step 690 "No"
)), While repeating the above-described process of extracting the ignitable 0-transition (step 620) and thereafter, while determining that the simulation is to be interrupted (step 690 “Yes”), this time in the normal simulation direction. In the selection (step 600), the backward model is selected, and the marking and time target final states are set in the backward model (step 615), and the above-described ignitable 0-transition extraction process (step 6).
20) The subsequent processing is repeated. At that time, 0- which can be fired
In the selection of transitions, the conflict detection method is based on the forward direction, and the rule for solving it is prepared with its own work priority.

On the other hand, when there is no ignitable 1-transition at the time of extracting the ignitable 1-transition in step 650, that is, in the state where the work with the product is completed (step 650 “absent”), first of all, The simulation based on certain environmental data set in
A feasible schedule assuming that the system is actually executed is determined (step 700), and then the feasible schedule is evaluated (step 710).

The expression format of the executable schedule includes, for example, a method of sequentially displaying the work whose schedule (allocation) is confirmed in the Gantt chart format, a method of displaying the movement and history of tokens on the Petri net model, and the like.

When the evaluation result is "good" (step 710 "OK"), the feasible schedule is output as a suitable schedule (step 720), and the simulation process is terminated. If the evaluation result is not “good” (step 710 “No”), the process returns to the beginning and restarts the simulation process, that is, inputs environmental data and sets the simulation direction (step 600) and thereafter. The process of is executed.

Therefore, in the present embodiment, the simulation in the forward direction model as well as the simulation in the backward direction model is enabled, so that the final state is given and traced back from the relevant state even when the conditions relating to the production process are complicated. It is possible to calculate the preparation conditions of raw materials, parts, etc. for satisfying the given requirements by simulation.

Further, since the simulation interruption processing in step 690 allows the simulation to be temporarily stopped at any time to change the simulation environment, change the marking, replace the rule, etc.,
For example, in case of accidents such as machine breakdowns and personnel movements,
At any time during the progress of the plan, the simulation environment can be changed and the forward and backward simulations can be performed immediately, so that future prediction, countermeasure planning, lead time calculation, etc. at the time of an accident can be executed efficiently.

In this evaluation, the result of the forward simulation and the result of the backward simulation are compared and compared to determine "good". It is also possible to perform a plurality of simulations in the forward direction and the reverse direction to obtain a plurality of feasible schedules in both directions, and then select the most suitable schedule among them.

Next, as an example of the Petri net model constructed according to the present invention, a Petri net model of a tank painting process in a production line of a certain motorcycle factory will be described.

FIG. 15 is a block diagram showing the process of tank coating.

In this tank painting process, two types of tanks, a monotone color tank and a two-tone color tank, are painted. First, the welded tank from the previous process is made into a monotone by the tank painting monotone process S1 with paint 1. After the coating, the painted and welded tank is stocked in the storage location (buffer) 1, and then the painted tank is taped by the masking taping process S2 and stocked in the storage location 2.

In the case of a monotone color tank, the tank painting step is completed through the tank painting clear step S3 with the next paint 2, while in the case of a two-tone color tank, the tank with the next paint 3 is used. It is configured such that the painting two-tone process S4, the stock at the storage location 3 and the feedback to the masking taping process S2 again, and then the stock at the storage location 2 and the tank coating clear process S3 with the paint 2 finish the tank coating process. ing.

Next, the product data necessary for modeling such a tank painting process with a Petri net,
Machine data and buffer data will be described.

In the case of this example, since there are two types of products, a monotone color tank and a two-tone color tank, two types of product sheets are created, and the product data of each product sheet is divided into the following Tables 1 and 2, respectively. It is set as shown in. In this example, of all the processes of the motorcycle factory, only the tank painting process is focused on, and the number of jobs in the product data is only one tank painting (job 1).

[0090]

[Table 1]

[Table 2] Table 1 shows the product data of the monotone color tank. Explaining Table 1, the work required for tank painting of a monotone color tank (Product 1) is tank painting monotone S1 (work 1), masking taping S2.
(Work 2) and tank coating clear S3 (Work 3) are three types of work. Each of Work 1 to 3 is used with Machines 1 to 3, and the processing time (work time) of 10, 4 and 8 is used. It is required and is processed in the order of operations 1 to 3 (according to the column of preceding operation).

Table 2 shows the product data of the two-tone color tank. Explaining Table 2, as in the case of Table 1, the work required for tank painting of the two-tone color tank (Product 2) is tank painting monotone S1 (Work 1), masking taping S2 (Work 2), Two-tone tank painting S4 (work 3), masking taping S2
(Work 4) and tank painting clear S3 (Work 5) are five types of work, and each work 1 to 5 is a machine 1, 2, 4, 2,
3 is used, processing time of 10, 4, 10, 4, 8 is required, and processing is performed in the order of operations 1 to 5.

Further, in this example, as shown in FIG. 15, four machines are used for each work, that is, each of the steps S1 to S4, and the machine data are shown in the following Table 3 and The machine layout, or line configuration, is shown in FIG.

[0093]

[Table 3] Referring to Table 3 and FIG. 16, in this example, four machines, that is, the painting machine 1 (machine 1), the masking machine (machine 2), the painting clear machine (machine 3), and the painting machine 2 (machine 4) are used. One machine is provided for each machine, and the machines 1 to 4 are arranged around the machine 2 as shown in FIG.

Further, in this example, as shown in FIG. 15, three storage locations, that is, buffers are used, and buffer data is shown in Table 4 below.

[0095]

[Table 4] Explaining Table 4, in this example, storage locations, that is, storage location 1 (buffer 1), storage location 2 (buffer 2), and storage location 3 (buffer 3) are provided at three locations, each of which has a capacity. 3 shows the work that can be used with each of the buffers 1 to 3 as a condition.

That is, as shown in FIG.
In the case of, the masking taping process S2 (work 2) for tank coating (job 1) of the monotone color tank (product 1) and the masking taping process S2 (tank 2) for tank coating (job 1) of the two-tone color tank (product 2) Work 2) is available. In the case of the buffer 2, the tank coating clearing step S3 (work 3) of the tank coating (work 1) of the monotone color tank (product 1), the tank coating of the two-tone color tank (product 2) (work 1)
The tank coating two-tone process S4 (work 3) and the two-tone color tank (product 2) tank coating (work 1) tank coating clear process S3 (work 5) can be used. In the case of the buffer 3, the tank coating clear step S3 (operation 5) of the tank coating (job 1) of the two-tone color tank (product 2) can be used.

Next, creation of Petri net models for each data constructed based on the above product data, machine data, and buffer data, that is, product sheet, machine sheet, and buffer sheet will be described.

FIGS. 17 (a) to 17 (c) show Petri net models of the respective operations for the monotone color tank (Product 1).

That is, (a) shows the Petri net of the tank painting monotone process S1 (work 1) of the tank painting (work 1) of the monotone color tank (product 1),
This Petri net is provided with ten 1-transition TR1s in series according to the processing time 10 (see Table 1) of the work, and the start and end of the work are indicated at 0 at the forefront and the end. -Transition TRs
, TRe are provided and are connected to each place PL provided in the transition TR by an arc.

Similarly, (b) is the Petri net of the masking taping step S2 (work 2) of the tank coating (work 1) of the monotone color tank (product 1), (c).
Shows the Petri net of the tank painting clear process S3 (Work 3) of the tank painting (Job 1) of the monotone color tank (Product 1), and each Petri net depends on the processing time of each work (see Table 1). As a result, 4 to 8 1-transition TR1s are provided in series.

FIGS. 18 (a) to 18 (e) show Petri net models of respective operations for the two-tone color tank (product 2).

That is, (a) is a two-tone color tank (product 2) tank painting (job 1) tank painting monotone step S1 (work 1) Petri net, (b) is a two-tone color tank (product 2) Petri net for masking taping process S2 (work 2) for tank coating (work 1), tank painting for two-tone color tank (product 2) (job 1) two-tone process S3 (work 3) petri net for tank painting (work 1) , (D) Petri net of masking taping process S2 (work 4) for tank coating (job 1) of two-tone color tank (product 2), (e) tank coating of two-tone color tank (product 2) (job 1) Shows the Petri net of the tank painting clear step S3 (work 5) of FIG. Transition TR1 is 10,
4, 10, 4, 4 are provided in series.

19 (a) and 19 (b) show Petri net models of tank coating (work 1) for a monotone color tank (product 1), respectively.

That is, the Petri net model of product 1-job 1 shown in (a) shows a Petri net constructed by arranging the Petri nets shown in FIGS. 17 (a) to 17 (c) based on the order of work. As described with reference to FIG. 4, it is equivalent to the model of (b).

FIGS. 20 (a) and 20 (b) show tank coating (job 1) for a two-tone color tank (product 2).
Shows a Petri net model of.

That is, (a) shows the Petri net model of product 2-work 1, and shows the Petri net in which the Petri nets shown in FIGS. 18 (a) to 18 (e) are arranged based on the order of the works. And is equivalent to the model in (b).

FIGS. 21A and 21B show Petri net models of the monotone color tank (Product 1).

In this example, as the work of the monotone color tank (product 1), only one work of tank painting (work 1) is taken into consideration. Therefore, the Petri net model of the monotone color tank (product 1) is as shown in FIG. It is shown as (a) using the simplified model shown in 19 (b).
5B shows a product sheet showing the Petri net model shown in FIG. 5A, that is, a Petri net of a monotone color tank (Product 1) as described in FIG. 5, and is equivalent to (a).

22 (a) and 22 (b) show Petri net models of two-tone color tanks (product 2), respectively.

In the case of the two-tone color tank (product 2), as in the case of the monotone color tank (product 1), only the tank painting (work 1) is taken into consideration as the work, so that the two-tone color tank (product 2) is used. The Petri net model is shown as in (a) using the simplified model shown in FIG. 20 (b). (B) shows a product sheet showing the Petri net model shown in (a), that is, a Petri net of a two-tone color tank (Product 2),
It is equivalent to (a).

Next, the combination of each product sheet, machine sheet and buffer sheet will be described. Although the three types of sheets are originally joined together, since the figure becomes complicated, FIG. 23 shows the joining of the product sheet and the machine sheet, and FIG. 24 shows the joining of the product sheet and the buffer sheet. I will show them separately.

FIG. 23 shows the combination of the product sheets 1A1 and 1A2 and the machine sheet 1B.

That is, this figure shows each work 1 to 3 on the product sheet 1A1 of the monotone color tank (product 1).
And each operation 1 to 5 on the product sheet of the two-tone color tank (product 2) and machines 1 to 4 on the machine sheet
And FIG. 16 shows the connection by arc based on the process drawing of the tank coating process shown in FIG.

Referring to the drawing, in the machine place PLM1 showing the painting machine 1 performing the tank coating monotone process S1, the 0-transition TRe indicating the end of the tank coating monotone process (work 1) in the product sheet 1A1,
And 0-transition TRe indicating the end of the tank coating monotone process (work 1) on product sheet 1A2
From the machine place PLM1 to the 0-transition TRs indicating the start of work 1 of the product sheet 1A1 and the 0-transition TRs indicating start of work 1 of the product sheet 1A2 from the machine place PLM1. Connected. This is when both the monotone color tank (product 1) and the two-tone color tank (product 2) tank coating monotone process (work 1) are completed, both tank coating monotone process (work 1).
Expresses that it can be restarted.

Further, the machine place PLM2 showing the masking machine for performing the masking taping step S2 is shown in FIG.
The arc goes from 0-transition TRe which indicates the end of the masking taping process (work 2) on the product sheet 1A2 based on 5 and 0-transition TRe which indicates the end of the masking taping process (work 4) on the product sheet 1A2. While connected
From the machine place PLM2, 0-transition TRs indicating the start of masking taping (work 2) on product sheet 1A1, 0-transition TRs indicating the start of masking taping (work 2) on product sheet 1A2, and masking on product sheet 1A2. An arc is connected to the 0-transition TRs indicating the start of the taping process (operation 4). This is because when the two masking taping steps (work 2, work 4) in the two-tone color tank (product 2) are completed, the masking taping (work 2) of the monotone color tank (product 1) and the two-tone color tank (work 2) are performed. Masking taping for product 2) (work 2, work 4)
Can be restarted.

The other machines 3 and 4 can be similarly described.

FIG. 24 shows the combination of the product sheets 1A1 and 1A2 and the buffer sheet 1C.

That is, this figure shows each work 1 to 3 on the product sheet 1A1 of the monotone color tank (product 1).
And product sheet 1 of two-tone color tank (Product 2)
Each work 1 to 5 in A2 and the capacity 3 in the buffer sheet 1C, that is, each buffer 1 to 3 capable of holding a maximum of three tokens are connected by an arc based on the process drawing of the tank coating process shown in FIG. Is shown.

To explain the figure, the capacity place PLB1 of the buffer 1 has a 0-transition T indicating the start of masking taping (work 2) in the product sheet 1A1.
Rs and 0-transition TR indicating the start of masking taping (work 2) on product sheet 1A2
While the arc is connected from s, from the capacity place PLB1, 0-transition TRe indicating the end of the tank coating monotone process (work 1) on the product sheet 1A1 and the tank coating monotone process (work 1) on the product sheet 1A2. The arc is connected to the 0-transition TRe, which indicates the end of).

This is because the start of masking taping (work 2) of the monotone color tank (product 1) and the start of masking taping (work 2) of the two-tone color tank (product 2) put the token into the capacity place of buffer 1. 0-transition indicating the end of tank coating monotone (work 1) of the monotone color tank (product 1) and 0-transition indicating completion of tank coating monotone of the two-tone color tank (product 2) (work 1). Since the token is removed by the firing of TRe, the number of the tokens indicates the numerical value of the semi-finished product for which the previous work has been completed, and also represents the remaining capacity of the capacity place.

The same can be said for the capacity place PLB2 of the buffer 2 and the capacity place PLB3 of the buffer 3.

As described above, in the Petri net model of the production system constructed in this way, the simulation unit 3 is shown in FIG.
Simulation is performed based on the flowchart shown in FIG. The Petri net model described with reference to FIGS. 17 to 24 is for quasi-direction simulation. Therefore, if simulation is performed with this model, each manufacturing lead in the case where monotone and two-tone color tanks are flowed in the same line is used. You can know the time. Further, by recording the ignition history of the transition TR, it becomes a reference when making a production plan based on the Gantt chart.

Although the reverse model of the tank coating process has not been described, the product sheet, the machine sheet and the buffer sheet are subjected to the processing described in steps 500 to 520 of FIG. A reverse model can be constructed by connecting them in reverse, and by designating the delivery date of the product, which is the target final state in the backward model, the time to start making each product is determined retroactively from the delivery date. be able to.

In the above embodiment, when the Petri net model for each data is created, the data is divided into three types of data which are the elements on the production system: product data, machine data and buffer data. In the present invention, when a Petri net is created for each such data, it is not necessary to be limited to these three types of division, and a Petri net is created with machine data and buffer data and these two data are combined into one. One sheet and construct a Petri net model of the production system by combining two types of this sheet and a product sheet that is a Petri net model of product data for each product, or work by dividing the product sheet a little more You can build a Petri net model of the production system in units of sheets for each work, or even a production system. Product data for built, the machine is divided in addition elements other than the buffer, the Petri net model for each said element may be constructed Petri net model of the production system.

[0125]

As described above, according to the invention of claim 1, the data for modeling the production system are
The data is divided into product data, machine data, buffer data, etc. to create a Petri net model for each data, and the Petri net model for each data is combined to build the Petri net model of the production system. If there is a change or accident such as a product variation / variation product, machine failure, buffer addition / reduction, etc., simply change the Petri net of the data corresponding to the change, etc. It is possible to respond to such issues quickly and efficiently.

According to the second aspect of the present invention, the data for modeling the production system is converted into unit time transitions having a predetermined unit time of ignition duration for the work required for each product causing the state transition. Is provided corresponding to the work time, a zero-time transition indicating the start and end of the work is provided and expressed, and machines and buffers that do not cause state transitions are expressed by places, which are required for each product. A zero-time transition of work and a place showing a machine and a buffer are connected by an arc to build a Petri net model of a production system, and in such a Petri net model, a transition is fired based on its firing duration, Status changes such as machine free space and buffer remaining capacity at any time Since the value was shown by the number of tokens in each place, by detecting the history of the token, the state of the production system at an arbitrary time can be recognized accurately and readily.

According to the invention described in claim 3, a Petri net model considering the time of state transition is created for each data such as product data, machine data and buffer data as in the invention described in claim 2. The model of the production system is constructed by combining the models of and the simulation is performed on the Petri net model, so that the state of the production system at any time can be accurately and easily recognized, and the product change or Even if a machine accident or the like occurs, simply changing the Petri net of the data corresponding to the change can promptly and efficiently respond to the change.

According to the invention described in claim 4, in consideration of the time of state transition, a Petri net model for forward simulation and a Petri net model for backward simulation are constructed in the same manner as the invention described in claim 2. , In each of the forward and backward Petri net models, the token placement and time are set to the initial state when performing the forward simulation, while the token placement and time are set when the backward simulation is performed. Since the simulation is performed by setting the target final state, the simulation can be performed based on the given production plan, and at the same time, the target final state or the planned delivery date is given and the time is traced back from the delivery date. What time should I prepare the raw materials, parts, etc. to meet the delivery date? You can determine the conditions under which gave the final state, such as whether start making product from any stage in consideration of the time.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration and the like of a production system simulator device using a Petri net according to the present invention.

FIG. 2 is a flowchart showing in detail a modeling procedure in a model building unit.

FIG. 3A and FIG. 3B are explanatory views showing Petri net models for each work.

FIG. 4A and FIG. 4B are explanatory diagrams showing Petri net models of respective jobs.

5A and 5B are explanatory views showing Petri net models of respective products.

FIG. 6 is an explanatory diagram showing product sheets of products 1 to N.

FIG. 7 is an explanatory view showing a machine sheet.

FIG. 8 is an explanatory diagram showing a buffer sheet.

FIG. 9 is an explanatory view showing the combination of the product sheet and the machine sheet in the forward direction model.

FIG. 10 is an explanatory view showing the combination of the product sheet and the buffer sheet in the forward model.

FIG. 11 is an explanatory diagram showing a change of a work Petri net model to a backward model.

FIG. 12 is an explanatory view showing the combination of the product sheet and the machine sheet in the backward model.

FIG. 13 is an explanatory diagram showing the combination of the product sheet and the buffer sheet in the backward model.

FIG. 14 is a flowchart showing in detail a processing procedure of simulation.

FIG. 15 is a block diagram showing a tank coating process in a production line of a certain motorcycle factory.

FIG. 16 is an explanatory diagram showing a line configuration of each machine.

17 (a) to 17 (c) are explanatory views each showing a Petri net model of each operation for the monotone color tank (Product 1).

18 (a) to 18 (e) are explanatory diagrams each showing a Petri net model of each operation for a two-tone color tank (product 2).

19 (a) and 19 (b) are explanatory diagrams showing Petri net models of tank coating (work 1) for a monotone color tank (product 1).

20 (a) and 20 (b) are explanatory views each showing a Petri net model of tank coating (work 1) for a two-tone color tank (product 2).

21 (a) and 21 (b) are explanatory views each showing a Petri net model of a monotone color tank (Product 1).

22A and 22B are explanatory views each showing a Petri net model of a two-tone color tank (product 2).

FIG. 23 is an explanatory view showing a combination of each product sheet and a machine sheet.

FIG. 24 is an explanatory view showing a combination of a product sheet and a buffer sheet.

FIG. 25 is an explanatory diagram showing a conventional Petri net model and dynamic behavior in the model.

[Explanation of symbols]

 1 Input Section 1a Product Data 1b Machine Data 1c Buffer Data 1A Product Sheet 1B Machine Sheet 1C Buffer Sheet 2 Model Building Section 3 Simulation Section 3a Rule Base 3b Schedule 4 Environment Setting Section 4a Simulation Environment Data 5 Output Section TR Transition PL Place

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Akimasa Horizawa, Inventor, 10th Hanazono Dodocho, Ukyo-ku, Kyoto Omron Co., Ltd. (72) Inventor Norio Yoshikawa, 10th Hanazono Todocho, Ukyo-ku, Kyoto

Claims (4)

[Claims] 1. A production system simulator using a Petri net, which constructs a model of a production system by a Petri net, and simulates the production system by the model, incorporates data for modeling the production system,
Product data indicating the type of product and the work required for each product,
An input unit that divides each data such as machine data that indicates the type and number of machines and the machine layout, buffer data that indicates the type and capacity of the buffer, and product data, machine data and buffers that are divided by the input unit. A Petri net production characterized by comprising a model building unit that creates a Petri net model for each data such as data and combines the Petri net model for each data to build the Petri net model of the production system. System simulator device. 2. A production system simulator apparatus using a Petri net, which constructs a model of a production system by a Petri net, and simulates the production system by the model, and an input unit for receiving data for modeling the production system, Based on the data taken in by the input unit, a work required for each product that causes a state transition is provided with one or more unit time transitions having a predetermined unit time of ignition duration in series corresponding to the work time. , A zero-time transition indicating the start and end of the work and having no ignition duration is provided at the front and the end of the unit time transition provided in series or in series, and between adjacent transitions. Placed adjacent transitions The place and place are shown by connecting them with an arc that indicates the direction of the state transition, while the machine and the buffer etc. that do not cause the state transition are shown by the place, and the zero time transition of the work required for each product and the machine and In the model construction unit that constructs the Petri net model of the production system by connecting the place indicating the buffer with the arc, and in the Petri net model of the production system constructed by the model construction unit, from the unit time transition to the place Tokens are moved after the firing duration of the unit time transition has elapsed after the firing duration of the unit time transition has elapsed, while the tokens are moved from the zero time transition to the place by firing at the zero time transition. And a simulation unit that shows the state of product production at any given time, state variables such as machine vacancy, buffer remaining capacity, etc., by the number of tokens in the above place, and production by Petri nets. System simulator device. 3. A production system model is constructed by a Petri net, and a production system simulator device by the Petri net for simulating the production system by the model incorporates data for modeling the production system,
Product data indicating the type of product and the work required for each product,
An input unit that divides each data such as machine data that indicates the type and number of machines and the machine layout, buffer data that indicates the type and capacity of the buffer, and product data, machine data and buffers that are divided by the input unit. In the case of manufacturing data including work required for each product that causes a state transition, a Petri net model for each data such as data corresponds to a unit time transition that has a predetermined unit time of ignition duration One or more in series, and each of the unit time transitions provided in the one or more series shows the start and end of the work at the forefront and the end of the unit time transition and has a zero-time transition with no ignition duration, respectively. Place and place a place between each adjacent transition, as well as provide Between each other by an arc that indicates the direction of the state transition, while in the case of machine data and buffer data that do not cause a state transition, the machine and buffer are indicated by a place, and the Petri net of the above product data is zero. In the Petri net model of the production system constructed by the model construction unit, which constructs the Petri net model of the production system by connecting the time transition and the place indicating the machine and the buffer by the arc, The token is moved from the unit time transition to the place after the firing duration of the unit time transition has elapsed after the firing at the unit time transition, while the token is moved from the zero time transition to the place at the zero. It is performed by firing in a time transition, and it has a simulation unit that indicates the state of product production at any time, the state variable values such as machine vacancy, remaining capacity of the buffer, etc. by the number of tokens in the above place. Production system simulator device by Petri Net. 4. A production system simulator apparatus using a Petri net, which constructs a model of a production system by a Petri net, and simulates the production system by the model, and an input unit for receiving data for modeling the production system, Based on the data taken in by the input unit, a work required for each product that causes a state transition is provided with one or more unit time transitions having a predetermined unit time of ignition duration in series corresponding to the work time. , A zero-time transition indicating the start and end of the work and having no ignition duration is provided at the front and the end of the unit time transition provided in series or in series, and between adjacent transitions. Placed adjacent transitions The place and place are shown by connecting them with an arc that indicates the direction of the state transition, while the machine and the buffer etc. that do not cause the state transition are shown by the place, and the zero time transition of the work required for each product and the machine and By constructing a Petri net model of the production system for forward simulation by connecting the place showing the buffer with the arc, the unit time transition and the zero time transition in the Petri net model of the work required for each product By reversing the direction of the arc connecting the places, the zero direction transition of the work required for each product and the place showing the machine and the buffer are connected by the arc to reverse the direction. Raw for simulation In each of the model building unit that builds the Petri net model of the system and the Petri net model of the production system for the forward and backward simulations built by the model building unit, the initial state when performing the forward simulation While setting the token placement and time, when performing the reverse simulation, set the target token placement and time in the final state, and move the token from the unit time transition to the place to the unit time transition. While after the ignition duration of the unit time transition has elapsed from the ignition in, the token movement from the zero time transition to the place is performed by the ignition in the zero time transition, the product production state at any time, and the machine Sky , The production system simulator apparatus according to Petri net, characterized by comprising a simulation section showing a state variable value, such as the remaining capacity of the buffer by the number of tokens in the places, the.