JPH08137828A - Simulation device and automatic expert system generating device using same - Google Patents

Abstract

PURPOSE: To effectively generate a source program for the expert system by following up changes of external environment at all times. CONSTITUTION: A simulation model is generated by allocating loads, etc., to data required to produce a product under the conditions of a rule base 23, and simulation is carried out. This simulation result is evaluated under evaluation conditions 30 to change the conditions stored in the rule base 23, the history of the changed conditions is learnt, and based on the learnt change history of the conditions, the source program for the expert system is generated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a simulation device for executing an optimum simulation by obtaining the line efficiency of an automation line and an expert system automatic generation device using the simulation device.

[0002]

2. Description of the Related Art FIG. 20 is a block diagram of a simulation apparatus. The input device 1 inputs simulation conditions such as the type of product produced in an automated line, its size, and processing sequence, and sends it to the line efficiency calculation execution unit 2.

The line efficiency calculation executing section 2 calculates the line efficiency of the automation line based on the simulation conditions such as the inputted product type, its size, and the processing order, and displays this line efficiency on the display device 3. .

In such an apparatus, when the optimum line efficiency of the automated line and its condition are calculated, the simulation condition is changed and input from the input device 1 until the optimum line efficiency and its condition are obtained. Then, a series of operations for displaying the line efficiency calculated under this condition on the display device 3 is repeated.

However, in order to calculate the optimum line efficiency, it is necessary to repeatedly input the simulation conditions, execute the simulation, and evaluate the simulation result many times. It takes a very long time to obtain the conditions and is inefficient.

On the other hand, there is an expert system for supporting the design of the line. This expert system
There are two ways, one is to use an algorithm dedicated to the target system and the other is to use a general-purpose shell (engine).

When the algorithm dedicated to the target system is used, if the external environment such as the production scale or facility plan changes, the algorithm also needs to be changed accordingly, and software maintenance must always be performed.
In addition, it is difficult to change the algorithm even if it is diverted to another system.

When a general-purpose shell is used, since it has high versatility, the rule-based condition is used for processing. Also, the condition to be described must be obtained from an expert, but it is very difficult because the designer of the expert system conducts it in an interview format.

[0009]

As described above, in order to calculate the optimum line efficiency, it is necessary to repeat a series of operations such as inputting simulation conditions many times, which is very time-consuming and inefficient. Is.

On the other hand, in the expert system, maintenance and management in accordance with changes in the external environment are always important, which requires a lot of man-hours. Further, since acquisition of rule-based knowledge (processing / judgment conditions) is performed by a human system, it takes a lot of time and difficulty.

Therefore, an object of the present invention is to provide a simulation device which can efficiently obtain an optimum simulation result. It is another object of the present invention to provide an expert system automatic generation device capable of efficiently following the change of the external environment and efficiently generating the source program of the expert system.

[0012]

According to a first aspect of the present invention, in a simulation apparatus for evaluating a simulation result obtained by executing a simulation under preset simulation conditions based on the evaluation conditions, the simulation results are evaluated conditions. And a condition changing unit that changes at least the simulation condition and executes the simulation again if the simulation result does not satisfy the evaluation condition as a result of the determination by the determining unit. This is a simulation device that tries to achieve the purpose.

According to the second aspect of the present invention, the simulation is performed in which the data required for manufacturing the product stored in the manufacturing database is at least load-allocated according to the conditions stored in the rule base to create a simulation model and execute the simulation. The apparatus includes an evaluation condition database that stores evaluation conditions for simulation results, and an evaluation / condition changing unit that evaluates simulation results based on the evaluation conditions and changes the conditions stored in the rule base. This is a simulation device that aims to achieve

According to a third aspect of the present invention, a function of at least load-allocating data required for manufacturing a product stored in the manufacturing database according to the conditions stored in the rule base to create a simulation model and execute the simulation. In an expert system automatic generation device equipped with, an evaluation condition database for storing evaluation conditions for simulation results, and an evaluation / condition changing means for evaluating the simulation results based on the evaluation conditions and changing the conditions stored in the rule base. And learning means for learning the history of the conditions changed by the evaluation / condition changing means, and program generating means for generating a source program of the expert system based on the change history of the conditions learned by the learning means. Get ready for the above goals A expert system automatic generation system that.

According to the fourth aspect, the program generating means generates the source code of the condition change from the learned condition change history in the order of high priority, and the condition change source code is used as the variable algorithm in the source program of the expert system. It is added to the section.

[0016]

According to the first aspect, when the simulation is executed under the preset simulation condition, it is determined whether or not the simulation result satisfies the evaluation condition. If the simulation result does not satisfy the evaluation condition, at least Change the simulation conditions and run the simulation again. Thereby, the optimum simulation result can be efficiently obtained.

According to the second aspect of the present invention, the data required for manufacturing the product is allocated to the load according to the rule-based conditions to create a simulation model, and when the simulation is executed, the simulation result is based on the evaluation conditions. And change the conditions stored in the rule base. Then, the simulation model is created again and the simulation is executed. As a result, the source program of the expert system can be efficiently generated by always following changes in the external environment.

According to the third aspect of the present invention, the data required for manufacturing the product is allocated to the load in accordance with the rule-based conditions to create a simulation model, and when the simulation is executed, the simulation result is based on the evaluation conditions. And change the conditions stored in the rule base.

When the condition is changed as described above, the history of the changed condition is learned, and the source program of the expert system is generated based on the learned change history of the condition.

In this case, according to the fourth aspect, the program generating means generates the source code of the condition change from the learned condition change history in order of high priority, and the condition change source code is used as the source program of the expert system. Add to the variable algorithm part in.

[0021]

【Example】

(1) Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of the simulation apparatus. An input device 11 is connected to the optimization processing device 10.

The input device 11 inputs, as simulation conditions, for example, the type of product produced on an automated line and its lot size, processing order, flow order, and processing time in each processing.

The optimization processing device 10 has a function of calculating the optimum simulation condition and the line efficiency based on the simulation condition, and the line efficiency calculation executing unit 12, the operating rate determining unit 13, and the input condition changing unit 14 are provided. have.

The line efficiency calculation executing unit 12 has a function of calculating the line efficiency of the automated line based on the simulation conditions. The operating rate determination unit 13 calculates the operating rate of the automation line based on the simulation conditions, and compares it with the average average operating rate of all processes, which is set in advance as a determination criterion, for example, 80% or more, and changes the comparison result to the input condition. It has a function of sending to the section 14.

In this case, the operating rate K is obtained by calculating k = (each processing time / simulation) × 100 [%] (1).

The input condition changing unit 14 includes an operation rate determining unit 13
When the operation rate of the automated line does not meet the determination standard, the simulation condition set in the line efficiency calculation executing unit 12, for example, the lot size, the processing order, and the flow order are changed. ing.

When the operating rate of the automated line satisfies the criterion, the input condition changing unit 14 determines the type of the simulation condition and its lot size, the processing order, the flow order, the processing time in each processing, and the simulation result. Is displayed on the display device 15.

Next, the operation of the apparatus configured as described above will be described with reference to the optimization flowchart shown in FIG. First, from the input device 11, in step # 1, the line efficiency calculation executing unit 12 determines, as simulation conditions, the product type of the product produced on the automated line and its lot size, the processing order, the flow order, the processing time in each processing, and the like.
Input is set to.

Next, the operation rate determination unit 13 is instructed by step #
In 2, the average processing rate of all processes, for example, 80% or more is set as a criterion. When the simulation condition is set in this way, the simulation of the simulation model is executed under this simulation condition.

After execution of this simulation, the line efficiency calculation executing unit 12 calculates the line efficiency of the automated line based on the simulation result under the simulation condition in step # 3.

Next, the operation rate judging section 13 determines in step # 4.
In the above, the operation rate of the automated line is calculated based on the simulation results under the simulation conditions using the above formula (1).
It is calculated by calculating

The operation rate determining unit 13 also compares the calculated operation rate of the automated line with the average processing rate of 80% or more, which is a determination criterion, and sends the comparison result to the input condition changing unit 14.

The input condition changing unit 14 receives the determination result of the operation rate determining unit 13, and if the operation rate of the automated line does not satisfy the determination criterion, the process proceeds to step # 5 to the line efficiency calculation executing unit 12. Among the set simulation conditions, the lot size, processing order, and flow order are changed and set.

After changing and setting the simulation conditions, the simulation of the simulation model is executed again. After that, the line efficiency calculation executing unit 12 again calculates the line efficiency of the automated line based on the simulation result in step # 3.

At the same time, the operation rate determination unit 13 calculates the operation rate of the automation line based on the simulation result in step # 4. Again, the input condition changing unit 14
Receives the determination result of the operation rate determination unit 13, and if the operation rate of the automation line satisfies the determination criteria, the process proceeds to step # 6, and the type and lot size of the simulation condition, the processing order, the flow order, The processing time of the processing and the simulation result are displayed on the display device 15.

As described above, in the first embodiment,
The simulation is executed under preset simulation conditions, and it is determined whether or not this simulation result satisfies the evaluation condition judgment criterion. If this simulation result does not satisfy the evaluation condition, change the simulation condition and try again. Since the simulation is executed, the optimum simulation condition can be automatically obtained only by inputting the simulation condition once.

Therefore, a series of operations of inputting simulation conditions, executing simulations, and evaluating the simulation results are not repeated until the optimum simulation conditions are obtained, and the work can be reduced and the time can be shortened. It is possible to improve efficiency by connecting.

The first embodiment may be modified as follows. For example, the determination criterion is not limited to the operating rate, and may be, for example, the number of productions or the number of gimmicks,
Further, a plurality of settings may be set.

The first embodiment may be replaced with mechanical control software. (2) Next, a second embodiment of the present invention will be described. FIG. 3 is a block diagram of an expert system automatic generation device applied to a board mounting process.

The manufacturing data storage unit 20 comprises a manufacturing database 21 and a load amount database 22. Of these, the manufacturing database 21 includes parts configuration, production schedule,
Each data required for manufacturing a product such as parts information, facility information, line configuration, process route, etc. is stored.

Specifically, the manufacturing database 21 has a structure shown in FIG.
~ As shown in the schematic diagram of FIG. 6, ten types of data files, that is, a component configuration data file 21a, a production schedule data file 21b, and a component information data file 21.
c, PDG (Parts Data Code) data file 21
d, channel data file 21e, board information data file 21f, machine data file 21g, line data file 21h, process route data file 21i,
And the allocated parts data file 21j.

The component configuration data file 21a stores the component configuration data including the component name and the number of components to be mounted on the board. For example, a parent ordering code field is formed.

The production schedule data file 21b stores production schedule data indicating how many boards are to be mounted for each day. For example, fields such as a construction date and a parent order code are formed.

In the parts information data file 21c, parts information data which is specifications of parts are stored, and fields such as a child arrangement code and a product name number are formed, for example. The PDG data file 21d stores PDG data that is the shape of parts and the packing style at the time of arrival, and fields such as PDG, packing style, and tape material are formed.

In the channel data file 21e, channel data, which is the number of component supply channels of the mounting machine (machine), is stored, and fields such as PDG and machine name are formed.

The board information data file 21f stores board information data, which is board shape information, and includes fields such as a master arrangement code, width, and length.

In the machine data file 21g, machine data which is a capability value of the mounting machine is stored, and fields such as a machine name, loading time and mounting time are formed.

The line data file 21h stores line data which is a mounting machine configuration in the line,
For example, fields such as lines are formed. The process route data file 21i stores process route data indicating in which process order each substrate flows. For example, the parent arrangement code and each line # 1, # 2 ... The field is formed.

The allocated component data file 21j stores allocated component data indicating allocation of components handled by each mounting machine, and fields such as machine name and channel number are formed in the allocated component data file 21j.

The rule base 23 stores each rule of expert design know-how in designing a manufacturing line. FIG. 7 is a schematic diagram of component allocation to each mounting machine arranged on the manufacturing line in the rule base 23.

In the rule base 23 for this component allocation,
A priority order area 23a for component extraction, a conditional expression area 23b for describing the component extraction conditions, an allocation order area 23c for rearranging the extracted components, an allocation machine area 23d for describing from which mounting machine to allocate, a mounting machine area for each mounting machine. An allocation method area 23e is formed to make the allocation uniform.

For example, in the order "1" in the priority order area 23a, "first, the cell name 2
A part having a width of 125 and a tape width of 8 mm is extracted. Next, the extracted parts group is rearranged in the descending order of the number of parts. Next, the chips are alternately allocated to the two chip mounters (CP1 and CP2), and if the parts are not allocated yet and remain, they are allocated to the odd-shaped component mounting machine (IP2). The allocation method is to make the number of parts uniform. ".

The layout designing section 24 interactively arranges each equipment of the manufacturing line in the space given to arrange the manufacturing line from the equipment shape stored in the manufacturing database 21 on the display screen. It has the function of supporting the production line model creation.

The allocating unit 25 uses the rules of expert know-how stored in the rule base 23 to allocate the load to the manufacturing line by day / process-based load accumulation / mountain collapse.
It has the function of allocating auxiliary materials to equipment.

The allocation section 25 is a load allocation section 25.
a, a line allocating section 25b, and a component allocating section 25c. Of these, the load allocation unit 25a is
The load is allocated with respect to the elapsed time of operating the manufacturing line. As shown in FIG. 8, the load (production amount) is piled up and allocated so that the maximum processing capacity is obtained with respect to the elapsed time t. It has a function of leveling the load by breaking it down as indicated by an arrow so that the load becomes uniform with the processing capacity with respect to the elapsed time t.

The line allocator 25b allocates a load to each line. As shown in FIG. 9, a load is piled up and allocated to each line so as to obtain the maximum processing capacity, and then each line is allocated. On the other hand, it has a function of leveling the load by breaking it down so that the load is evened by the capacity.

The component allocating section 25c has a function of allocating a load to each mounting machine. Here, as shown in FIG. 10, the manufacturing line includes a solder printer 40, an adhesive applicator 41, chip component mounting machines 42 and 43, an IC component mounting machine 44, a variant component mounting machine 45, and a curing furnace 46. It is configured.

In this manufacturing line, the equipment for mounting components is four units, that is, each chip component mounting machine 42, 43, IC component mounting machine 44, and odd-shaped component mounting machine 45. Since these facilities 42 to 45 can mount a plurality of different components, they have a plurality of channels for supplying these components.

These channels are fixed channels in which parts are set so as to be compatible with different boards, variable channels that are exchanged each time the board changes, and external setup channels in which parts can be set even when the mounting machine is operating. There is.

Therefore, as shown in FIG. 11, the component allocating section 25c first stacks and allocates the components that the mounting machine is good at so as to maximize the performance of each mounting machine, and thereafter, As shown in FIG. 12, if the line balance between the mounting machines is poor, it has a function of leveling down the load to the other mounting machines to level the load.

The loading sequence creating unit 26 is a specialist stored in the load base 25a, the line allocating unit 25b, and the component allocating unit 25c in the allocating unit 25, and the load data obtained by the allocations and the rule base 23. It has a function to take in the rules of know-how and create an optimal order of loading under those conditions.

FIG. 13 and FIG. 14 show an example of creating the input sequence. In FIG. 13, the vertical axis shows the substrate and the horizontal axis shows the components set in the general-purpose channel. As shown in the figure, the loading sequence creating unit 26 has a function of loading the general-purpose channel components in order from the substrate in which the use is minimized in order to minimize the setup.

In FIG. 14, the vertical axis represents the substrate and the horizontal axis represents the takt time of the mounting machine. As shown in the figure, the loading sequence creation unit 26 has a function of loading from a board having a good line balance and operating the mounting machine in a state in which the operating rate is high.

The simulation model creating section 27 includes the process allocation of the manufacturing line model created by the layout designing section 24 and the load allocating section 25 in the allocating section 15.
a, the line allocating unit 25b, and the component allocating unit 25c, the respective process capacities obtained by the respective allocations, the optimum input order created by the input sequence creating unit 26, and the process arrangements, the process capacities, and the input orders. It has a function to automatically create a simulation model under the condition.

This simulation model uses each Petri net code based on the Petri net theory shown in FIG. That is, as shown in the figure, a place represents a state such as equipment, buffer, and transportation.

Transient represents connection between places (processes), and changes state (operation). The arc indicates the direction in which the work or information flows, and represents the relationship between the state and the motion.

The token indicates the work or information itself, and indicates the establishment of the state in the place. Therefore,
The simulation model is constructed by combining and connecting these place, transient, arc, and token Petri net codes.

The simulation unit 28 tracks changes in each behavior of equipment, objects, and people while changing the time under the conditions described in the simulation model created by the simulation model creation unit 27, and displays the simulation result. Simulation result database 29
It has a function to store in.

The simulation engine uses the Petri net theory as described above. The operating principle of this simulation is as follows. The simulation model is configured, for example, as shown in FIG. 16, in which three places are connected to the transition input side and two places are connected to the transition output side.

In the case of such a simulation model, 2 out of 3 places on the transition input side are used.
If tokens exist in one place and tokens exist in three places (all places) after this, firing is possible only when tokens do not exist in all places on the transition output side.

This firing means that the tokens disappear from all the places on the transition input side and the tokens occur on all the places on the transition output side. Further, the simulation engine has a function of giving a color to the token to identify its type and performing an extended Petri net capable of occupying the token in the place for a designated time.

On the other hand, the evaluation condition database 30 stores the evaluation condition data for the simulation result. 17 and 18 show the relationship between the evaluation of the simulation result and the change of the conditions. The evaluation condition is described in the <if then else> format, and the changing condition is described in each of the else and the else.

It should be noted that the input order is numbered in descending order of priority and indicates ascending ascending order A and descending descending order B. For example, N
o. In the example of 1, CP3 has more than 90
In addition, when the operation rate of CP3 is larger than 65%, the loading order is such that the boards are loaded in the order of the shortest mounting time, the largest number of components, and the smallest number of component types.

Evaluation / condition change unit (dispatching unit)
31 evaluates the simulation result stored in the simulation result database 29 based on the evaluation condition stored in the evaluation condition database 30, and changes the rule data stored in the rule base 23,
The change history is learned and the condition change history database 32
It has a function to store in.

The process design algorithm automatic generation unit (expert system algorithm generation unit) 33 executes the source program of the expert system in the process design field based on the condition change history of the rule base 23 stored in the condition change history database 32. It has a function to automatically generate.

That is, as shown in FIG. 18, the evaluation / condition changing unit 31 has a function of generating the source code of the changing condition from the changing condition history learned by the evaluation order in descending order of priority.

This condition changing source code is, for example, 1.
Implementation time and ascending order are described as SORT (code),
2. 2. The real number of parts and the descending order are described as SORT. S
The part type, ascending order, ... Are described as the ORT.

The condition-change source code generation is associated with each input condition, for example, the input sequence "1".
"1.SORT" corresponds to "1.SORT", and "4.SORT" corresponds to the input order "2".

By the way, the new engine for process design (source program of expert system) 34 is prepared in advance by being divided into a fixed algorithm part and a variable algorithm part.

Therefore, the process design algorithm automatic generation unit 33 adds the condition changing source code to the variable algorithm unit of the fixed algorithm unit and the variable algorithm unit of the new process design engine 34 to add the source program of the expert system. Has a function of automatically generating.

Next, the operation of the apparatus configured as described above will be described. In designing the manufacturing line in the board mounting process, the layout designing unit 24 interactively creates a model of the manufacturing line by arranging the equipment in a given space from the equipment shape stored in the manufacturing database 21. .

That is, on the display screen of the layout design section 24, each facility of the manufacturing line is arranged and a model of the manufacturing line is created. This manufacturing line includes, for example, a solder printer 40 and an adhesive applicator 4 as shown in FIG.
1, each chip component mounting machine 42, 43, IC component mounting machine 4
4, a modified component mounting machine 45, and a curing furnace 46.

In this manufacturing line, the equipment for mounting components is four units, that is, each chip component mounting machine 42, 43, IC component mounting machine 44, and odd-shaped component mounting machine 45. On the other hand, the allocating unit 25 reads out the expert know-how rule stored in the rule base 23, and uses this rule to load / distribute the load / manufacturing load for each day / process, load allocation to the manufacturing line, and equipment. Assign sub-materials.

That is, as shown in FIG. 8, the load allocator 25a piles up and allocates loads so as to obtain the maximum processing capacity with respect to the elapsed time t, and thereafter, the load is allocated with respect to the elapsed time t. The load is leveled so as to equalize the processing capacity.

As shown in FIG. 9, the line allocating section 25b piles up the loads so as to obtain the maximum processing capacity and allocates them to each line. The load is broken down so that it is leveled.

The component allocating section 25c allocates a load to each mounting machine. Here, the equipment for mounting the components on the manufacturing line is as described above for each chip component mounting machine 4
2, 43, IC component mounting machine 44, and odd-shaped component mounting machine 45.

Since these facilities 42 to 45 can mount a plurality of different parts, they have a plurality of channels for supplying these parts, and these channels are fixed channels in which the parts are set so as to be compatible with different boards. There are a variable channel that is exchanged each time the board changes, and an external setup channel that can set components even when the mounting machine is operating.

Therefore, as shown in FIG. 11, the component allocating section 25c firstly lays out the components that the mounting machine is good at and allocates them so as to maximize the performance of the mounting machine.

Thereafter, if the line balance between the mounting machines is poor as shown in FIG. 12, the component allocating section 25c breaks down the components to other mounting machines to level the load. The loading sequence creating unit 26 includes a load allocating unit 25a, a line allocating unit 25b, and a component allocating unit 25 in the allocating unit 25.
Each load data obtained by each allocation of c and the rule of the know-how of the expert stored in the rule base 23 are taken in, and the optimum loading sequence is created under the conditions.

For example, as shown in FIG. 13, the loading sequence creating unit 26 loads the general-purpose channel components in order from the substrate that minimizes the use of components in order to minimize the setup. or,
As shown in FIG. 14, the loading sequence creating unit 26 loads from a board having a good line balance, and operates the mounting machine in a state in which the operating rate is high.

In this way, the layout design unit 24 creates a manufacturing line model, and each process capability is obtained by the allocation of the load allocation unit 25a, the line allocation unit 25b, and the component allocation unit 25c, and the input sequence creation unit 26. When the optimum injection order is created, the simulation model creation unit 27 takes in the process layout of each manufacturing line model, each process capability, and the optimum injection order, and automatically creates a simulation model under these conditions.

FIG. 19 shows the simulation model creating section 2
7 shows an example of a simulation model created from 7. This simulation model is constructed by combining and connecting the places, transients, arcs, and tokens that are the Petri net codes shown in FIG. 15 according to the process arrangement, and describing the processing time, failure occurrence rate, etc. in each place.

The simulation unit 28 changes the time under the conditions described in the simulation model created by the simulation model creation unit 27, that is, the processing time and the failure occurrence ratio described in each place. Simulation is performed by tracking changes in the behavior of equipment, objects, and people.

That is, the simulation engine uses, for example, a token as a product, and makes this token exist in each place representing each facility for a time corresponding to the processing time.

The simulation engine fires in each transition only when there are tokens in all places on the transition input side and no tokens in all places on the transition output side. Tokens disappear from all places on the transition input side, and tokens occur on all places on the transition output side.

Further, the simulation engine causes the token to have a color for identifying its type, and occupies and moves the token in the place for a designated time.

In this way, the simulation engine executes the simulation for the simulation model of the manufacturing line by moving the token to each place according to the arc direction.

The result of this simulation execution is stored in the simulation result database 29. When the simulation is completed in this way, the evaluation / condition changing unit 31 reads the simulation result stored in the simulation result database 29, and evaluates the simulation result based on the evaluation condition stored in the evaluation condition database 30.

For example, the evaluation condition is No. 1 as shown in FIG. In the example of No. 1, "When the production amount of CP3 is more than 90 and the operation rate of CP3 is more than 65%, the loading order is in the order of shortest mounting time, high number of parts, and low number of parts. Substrate is loaded in order ”.

The evaluation / condition changing unit 31 makes an evaluation based on each evaluation condition in this way, changes the rule data stored in the rule base 23, learns the change history, and changes the condition change history database 32. Remember.

On the other hand, the process design algorithm automatic generation unit 33 automatically generates the source program of the expert system in the process design field from the learned condition change history of the rule base 23.

That is, the process design algorithm automatic generation unit 33, as shown in FIG.
The change condition history that has been learned by and stored in the condition change history database 32 is read out, and the source code of the change condition is generated from this change condition history in descending order of priority.

This condition change source code generation is associated with each input condition. For example, "1.SORT" for input sequence "1" and "4.SORT" for input sequence "2".
"SORT" corresponds.

Therefore, as the evaluation condition, for example, as shown in FIG. In the example of No. 1, "when the production volume of CP3 is more than 90 and the operating rate of CP3 is more than 65%", No. is input condition. 1, 3 and 4 are correspondingly interpreted as "loading boards in order of shortest mounting time, high number of components, and low number of component types".
With respect to such input conditions, the process design algorithm automatic generation unit 33 sets the input condition No. 1. Condition change source code “SOR
T (mounting time, ascending order) ”and input condition No.
The condition change source code “SORT (actual number of parts, descending order)” is generated for “3 in descending order of the number of parts” and the input condition No. 4. Generate the condition change source codes "SORT (component type, ascending order)" and "SORT (number of component type, ascending order)" for "Put boards in ascending order of component type".

When the condition change source code is generated in this way, the process design algorithm automatic generation unit 33 has the new process design engine 34 divided into the fixed algorithm unit and the variable algorithm unit. The source code of the expert system is automatically generated by adding the condition changing source code to the variable algorithm part.

As described above, in the second embodiment,
According to the conditions of the rule base 23, the data necessary for manufacturing the product is allocated by the allocation unit 25, the process layout is allocated by the layout design unit 24, and the input sequence creation unit 2
6, a simulation model is created by creating the input sequence, etc., the simulation of this simulation model is executed, the simulation result is evaluated based on the evaluation conditions, and the conditions stored in the rule base 23 are changed. Also, since the history of the changed conditions is learned and the source program of the expert system is generated based on the learned history of the changed conditions, it is linked to changes in the external environment such as production scale and facility planning. However, the aptitude value can be set by simulating load, line allocation, auxiliary material allocation, and the like.

Moreover, by learning the condition change history in the process of evaluating the simulation result, the expert system which can always follow the change of the external environment according to the situation and efficiently generate the source program of the expert system is created. Can be built.

The present invention is not limited to the second embodiment described above, but may be modified as follows. For example,
The invention is not limited to the board mounting process, but can be applied to other product manufacturing processes.

[0109]

As described above in detail, according to the present invention, it is possible to provide a simulation apparatus capable of efficiently obtaining an optimum simulation result. Further, according to the present invention, it is possible to provide an expert system automatic generation device capable of efficiently following the change of the external environment and efficiently generating the source program of the expert system.

[Brief description of drawings]

FIG. 1 is a block diagram of a simulation apparatus showing a first embodiment according to the present invention.

FIG. 2 is an optimization flowchart of the device.

FIG. 3 is a block diagram of an expert system automatic generation device showing a second embodiment according to the present invention.

FIG. 4 is a schematic diagram of a manufacturing database.

FIG. 5 is a schematic diagram of a manufacturing database.

FIG. 6 is a schematic diagram of a manufacturing database.

FIG. 7 is a schematic diagram of component allocation to each mounting machine in a rule base.

FIG. 8 is a diagram showing a mountain crushing of a load allocating portion.

FIG. 9 is a view showing a pile-up of a line allocating portion.

FIG. 10 is a view showing a manufacturing line.

FIG. 11 is a diagram showing stacking of component allocation parts.

FIG. 12 is a view showing a landslide of a component allocation portion.

FIG. 13 is a diagram showing an example of creating an input sequence.

FIG. 14 is a diagram showing an example of creating an input sequence.

FIG. 15 is a diagram of each Petri net code based on Petri net theory.

FIG. 16 is a diagram showing an operating principle of simulation.

FIG. 17 is a diagram showing a relationship between evaluation of simulation results and change of conditions.

FIG. 18 is a diagram showing a relationship between evaluation of simulation results and change of conditions.

FIG. 19 is a diagram showing an example of a simulation model.

FIG. 20 is a configuration diagram of a conventional simulation device.

[Explanation of symbols]

10 ... Optimization processing device, 11 ... Input device, 12 ... Line efficiency calculation execution unit, 13 ... Operating rate determination unit, 14 ... Input condition changing unit, 15 ... Display device, 20 ... Manufacturing data storage unit, 2
1 ... Manufacturing database, 22 ... Load amount database, 2
3 ... Rule base, 24 ... Layout design section, 25 ... Allocation section, 25a ... Load allocation section, 25b ... Line allocation section, 25c ... Component allocation section, 26 ... Loading sequence creation section, 2
7 ... Simulation model creation unit, 28 ... Simulation unit, 29 ... Simulation result database,
30 ... Evaluation condition database, 31 ... Evaluation / condition change unit, 32 ... Condition change history database, 33 ... Process design algorithm automatic generation unit, 34 ... Process design new engine.

Claims (4)

[Claims] 1. A simulation apparatus that evaluates a simulation result obtained by executing a simulation under preset simulation conditions based on the evaluation condition, and a determination unit that determines whether the simulation result satisfies the evaluation condition. And, if the simulation result does not satisfy the evaluation condition as a result of the determination by the determining unit, at least the condition changing unit that changes the simulation condition and executes the simulation again is provided. Simulation device. 2. A simulation device that creates a simulation model by performing at least load allocation according to conditions stored in a rule base for data required to manufacture a product stored in a manufacturing database, and executes a simulation result. And an evaluation condition database that stores the evaluation condition for, and an evaluation / condition changing unit that evaluates the simulation result based on the evaluation condition and changes the condition stored in the rule base. And a simulation device. 3. An expert system having a function of performing a simulation model by creating a simulation model by at least load-allocating data required for manufacturing a product stored in a manufacturing database according to a condition stored in a rule base. In the automatic generation device, an evaluation condition database that stores the evaluation condition for the simulation result, and an evaluation / condition changing unit that evaluates the simulation result based on the evaluation condition and changes the condition stored in the rule base. Learning means for learning the history of the condition changed by the evaluation / condition changing means, and program generating means for generating a source program of the expert system based on the change history of the condition learned by the learning means, Specially equipped Expert system automatic generation system according to. 4. The program generation means generates source code for condition change from the learned condition change history in order of high priority, and adds this condition change source code to a variable algorithm section in the source program of the expert system. The expert system automatic generation device according to claim 3, wherein