JP2001005509A - Method for deciding path of plural working heads and its device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make reducible the developing cost, to make shortenable the developing period, and to make automatically decidable the path not manually by making it unnecessary to use any complicated theory such as a 'neural network' or 'genetic algorithm'. SOLUTION: In this method for deciding the paths of plural working heads, working map data including plural positioning patterns indicating the moving time of working heads between adjacent two working positions are loaded, and the total time of the moving time of the working heads from the working start position of each working head to the working final position decided according to the combination of positioning patterns normalized by the working map data within a limited time in the priority order of the working heads is made the shortest within the limited time, and the combination of the positioning patterns covering all the working positions is decided so that path array can be prepared. This processing is repeated while the limited time is updated, and the path array in the shortest time is decided as the path of the plural working heads.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for determining a path of a plurality of processing heads, and more particularly to a method and an apparatus for determining a path for minimizing a moving time when a plurality of processing locations are processed by a plurality of processing heads. .

[0002]

2. Description of the Related Art For example, in the production of automobile parts such as radiators, there are many processed parts in the automobile parts. As an example, in the case of a radiator, a process of caulking a large number of aluminum caulking portions to upper and lower plastic tanks made of resin is required. This processing is performed while moving a plurality of processing heads to a processing location using a robot or the like. In such a case, it is necessary to minimize the moving time of the plurality of processing heads.

[0003] As a solution to the "Travel Salesman Problem" for finding a route with the shortest travel distance by visiting a plurality of places only once, a method using a "neural network" or a "genetic algorithm" is known. Have been. In some cases, the path of the processing head is determined by human trial and error.

[0004]

However, when developing a method for determining the path of a plurality of processing heads using a conventional "neural network" or "genetic algorithm", a plurality of processing heads are required to operate at the same time. It is necessary to avoid processing one processing location, and to prevent each processing head from visiting the same processing location even at different times, and to find a route with the shortest moving time. Become. In order to satisfy such a condition, it is expected that algorithms of a “neural network” and a “genetic algorithm” for each processing head will be complicated. Therefore, there is a problem that the development cost of the program for that purpose becomes high and the development period becomes long.
This goes against the recent demand for low cost, short term equipment development. Therefore, it is practically difficult to implement a method for determining the path of a plurality of machining heads by adding a method using the above-mentioned "neural network" or "genetic algorithm" to the automation equipment.

Further, according to the method of determining the path of the machining head by trial and error by a person, in addition to the problem that the work load by the person is large and the problem that the time required for the path determination is too long, As a result, even a target device of the same model may follow a different machining path, and there is a problem that the production capacity is influenced by the person who created the machining path.

SUMMARY OF THE INVENTION In view of the above problems in the prior art, an object of the present invention is to eliminate the necessity of using a complicated theory such as a "neural network" or a "genetic algorithm", to reduce the development cost and to shorten the development period. It is an object of the present invention to provide a path determination method and apparatus for a plurality of processing heads, which is short and enables a path to be determined automatically without manual operation.

[0007]

In order to achieve the above object, what is provided by the present invention is to provide a plurality of processing heads in a predetermined priority order from the processing start position of each processing head. Start, determine the combination of positioning patterns that cover all of the machining positions within the time limit, create a path array, and update the number of array creations if the path position covers all of the processing positions At the same time, the step of creating a path array by subtracting a predetermined value from the time limit is repeated, and if all of the machining positions are not covered by the path array and the number of array creations is not the initial value, the path array is created immediately before And a device for determining a path of a processing head, wherein the path arrangement is determined as a path of a plurality of processing heads.

Starting from the machining start position, a combination of positioning patterns defined by the machining map data is determined within a time limit, and the shortest time at which the combination of the positioning patterns satisfies a predetermined condition while reducing the time limit. Total travel time, so there is no need to use complex theories such as "neural networks" and "genetic algorithms", which reduces development costs, shortens the development period, and reduces Route can be determined automatically without the need.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an example of a system in which a route determination device according to the present invention is used. In the drawing, reference numeral 1 denotes a path determination device according to the present invention, 2 denotes an equipment control device that controls equipment (for example, a processing head) using the path determination device 1, 3 denotes a servo controller that controls the positioning of the processing head at an arbitrary position, 4 Is a processing head. In the following description, a processing head is used as equipment for which a route is determined by the route determining device 1, but the present invention is not limited to this.

The route determining device 1 and the equipment control device 2 are connected by a data communication bus 5. The equipment control device 2 notifies the path determination device 1 via the data communication bus 5 of the model and / or product number of the target device to be processed, and in response thereto, the path data of each processing head, which will be described in detail later, is transmitted to the path determination device. 1 through the data communication path 5 to the equipment control device 2
Will be notified.

Equipment control device 2 and positioning controller 3
Are connected by a data communication bus 6. The positioning controller 3 and each of the processing heads 7 are connected by a data communication bus 7. FIG. 2 is a block diagram schematically showing the configuration of the route determination device 1 in the device shown in FIG. In the figure, a route determination device 1 includes an input unit 21 including a keyboard and a mouse, and a display unit 2 including a CRT and the like.
2 and a control unit 23.

The control unit 23 includes a central processing unit (CPU) 2
31, a memory 232, a data registration unit 233 composed of a hard disk (HD) or the like, a communication control unit 234 which is an interface composed of an RS232C intelligent unit or the like, and an interface bus 235. The control unit 23 can be realized by installing a program for a method of determining a path of a plurality of machining heads according to the present invention on a commercially available personal computer.

The data registration unit 233 stores a processing map data file, a data file for each time layer, and an operation constant data file, which will be described in detail later. The processing map data file stores the “processing position (point number)”, “positioning pattern between processing positions”, and “processing start position (point number) of each processing head” for each model of the target device. This is the file to be executed.

The time stratified data file is a file for storing the "positioning time" for the "positioning pattern between each processing position". The calculation constant data file is a file for storing initial values and constants necessary for the route determination calculation processing. FIG. 3 shows the CPU 2 in the control unit 23 shown in FIG.
It is a functional block diagram centering on 31. In the figure, the data registration unit 233 includes a path array creating unit 31, a determining unit 32, a number counting unit 33, a time limit updating unit 34, and a count value determining unit 35.

FIG. 4 is a flow chart for explaining the operation of the route determination device shown in FIGS. According to FIG.
A method of determining a path of a processing head when processing each of a plurality of processing positions of the target device by any of the plurality of processing heads according to the embodiment of the present invention will be described.

First, in step S41, processing map data corresponding to the model of the target device is loaded into the memory 232. As described above, the processing map data includes a point number indicating a processing position, a plurality of positioning patterns indicating a moving time of the processing head between two adjacent processing positions, and a movement start point number of each processing head. It is composed of In this step S41, the time stratified data file and the operation data file are also loaded.

FIG. 6 is a diagram showing an example of the processing map data. In the example shown in FIG. 6, the point numbers indicating the processing positions are 1, 2, 4, 6, 8, 9, 10, 11, 12, and 13, and there are two types of positioning patterns, A and B. , Pattern A has a moving time TA of the processing head, and pattern B has a moving time TB of the processing head.
The number of processing heads is two, the processing start position of the first processing head is point number 1, and the processing start position of the second processing head is point number 13.

The processing of the target device by the processing head is performed only at the position of the point number registered in the processing map data, and the processing by the processing head is not performed at the position of the point number not registered in the processing map data. T
The meaning of A <TB means that the movement time of the processing head in the processing of pattern A is short, for example, because there are few obstacles, whereas the movement of the processing head in the processing of pattern B is an obstacle. This means that the travel time is longer.

Referring again to FIG. 4, in step S42, the number of array creations is initialized to 0, and the time limit is set to an initial value t1. Next, in step S43, based on t1, TA, and TB, the processing start position of each processing head, and the time limit, the travel time by the combination of the number of times TA and TB is within the time limit. When
A route pattern table including the above combinations and the corresponding positioning times is created. Details of the method of creating the route pattern table will be described later with reference to FIG.

Next, in step S44, the priority is 1
A path array is created in order from the first processing head. This is performed by finding a combination of path patterns that minimizes the moving time of the processing head in the path pattern table created in step S43. Next, in step S45, the path arrangement of the processing head having the second priority is also found in the same manner as described above. In step S46, the path array of the processing head having the lowest priority is found. Details of steps S44 to S46 will also be described later with reference to FIG. The priority order of the processing heads is predetermined. The order of priority may be arbitrary, but may be, for example, the order of ease of processing.

Next, in step S47, it is determined whether all the machining points have been stored in any of the path arrays. That is, all of the point numbers of the positions to be processed are
It is determined whether any of the route numbers is included in the point number. If the result of this determination is YES, step S4
Proceeding to step S8, the number of array creations is incremented, the time limit is updated by subtracting a predetermined value from the time limit t1 in step S49, and step S43 and subsequent steps are repeated.

In this way, even in the second and subsequent arrangements, a combination of path patterns that minimizes the moving time of the processing head within the reduced time limit is found in the same manner as described above. When steps S43 to S46 are repeated, eventually, at step S47, at least one of the point numbers of the positions to be processed is not included in any of the point numbers in the path array.

At that time, the process proceeds to step S50, where it is determined whether the number of array creations is not zero. If the number of array creations is not 0, the process ends normally in step S51, and the path array found in the previous cycle is adopted as the path array when each of the machining heads moves. Step S
If the number of array creations is determined to be 0 in the determination of 50, the process ends abnormally in step S52. In this case, it is considered that the initial value of the time limit is inappropriate, so the initial value is changed and the processing from step S41 is performed again.

By the above operation, the path arrangement of the processing head is determined so as to minimize the longest time in the total of the moving times of the plurality of processing heads. FIG. 5 is a flowchart illustrating details of the operations in steps S44 to S46 in FIG. In the figure, step S5
At 3, the count value of the positioning pattern counter is initialized to zero. The positioning pattern counter counts the number of times of the pattern for each type of pattern. If there are two moving patterns, for example, the pattern TA and the pattern TB on the processing map, the count A which is the number of the pattern TA and the pattern A There is a count B which is the number of TB. These positioning counters are provided in the path array creating means 31 of FIG. Hereinafter, for the sake of simplicity, the movement pattern in which the processing head moves between the point numbers is assumed to be two types of a pattern TA and a pattern TB.

Next, in step S54, the next point number which can be reached in the shortest time is assigned to the variable P, and whether the movement pattern for moving to the point number is TA or TB is stored. For example, at the position of point number 1 which is the processing head start position in the processing map of FIG. 6, the next point number 2 is substituted for the variable P, and the pattern TA is stored.

In step S55, 1 is added to count A or count B according to the stored movement pattern. That is, when moving in the pattern TA, 1 is added to the count A, and when moving in the pattern TB, the count B is added.
Is added to. Next, in step S56, count A
Is smaller than the maximum positioning count Amax based on the pattern TA, and whether the count B is smaller than the maximum positioning count Bmax based on the pattern TB. This determination is for determining whether or not the combination is within the range of the combination in the route pattern table described in detail later with reference to FIG. If YES, the value of the point number P substituted in step S54 is stored in the route array in step S57, and step S5
Return to 3. If NO, Amax in step S58
Alternatively, it is determined whether Bmax can be updated. Updating of Amax or Bmax means that the maximum value of count A or the maximum value of count B can be increased within the time limit. If Yes, Amax and / or Bmax are updated in step S59.
The method of updating is as follows.

In the case of TA> TB, 1 is added to Amax to obtain a new Amax. t1-
The maximum number of times of positioning by the pattern TB within the time of TAXAmax is set as a new Bmax. In the case of TA <TB, 1 is added to Bmax to obtain a new Bmax.

The maximum number of times of positioning by the pattern TA within the time of t1-TBXBmax is set as a new Amax.
If the determination in step S58 is NO, the maximum positioning count within the time limit can no longer be updated, so the processing of creating the path array of the machining head ends, and the point number stored in the path array immediately before this processing is completed. The row is a path array of the processing head.

The operation of creating the route pattern table will be specifically described with reference to the examples shown in FIGS. 6 and 7 to 9. FIG.
In the processing 1 shown in FIG. 7, the movement time TA between the point numbers by the pattern TA is 4 seconds, the movement time TB between the point numbers by the pattern TB is 7 seconds, and the initial value t1 of the time limit is 25 seconds. is there. A route pattern is created based on these values as shown in the figure. That is, when the possible combinations of the patterns TA and TB in the time limit of 25 seconds or less are listed, first, since TA <TB, the maximum number of times of positioning by the pattern TB is incremented in order from 0. Since the maximum value Bmax of the number of times of the pattern TB is 3 (7 × 3 = 21 ≦ 25, 7 × 4 = 28> 25), the maximum number of times of positioning Bmax by the pattern TB can be from 0 to 3 times.

Maximum number of positioning times Bm based on pattern TB
The maximum value Amax of the maximum number of times of positioning based on the pattern TA when ax is 0 is 6 (7 × 0 + 4 × 6 = 24).
≦ 25,7 × 0 + 4 × 7 = 28> 25). Therefore,
In the first combination, pattern TB is 0 and pattern TA is 6.
Times, and the total maximum positioning time at that time is 24 seconds.

Pattern TA when pattern TB is one time
The maximum value Amax of the number of times is 4 (7 × 1 + 4 × 4 =
23 ≦ 25,7 × 1 + 4 × 5 = 27> 25). Thus, the next combination has one TB and four TAs, for a total maximum positioning time of 23 seconds. TB is 2
The maximum value of the number of times of TA is 2 (7 × 2 + 4
× 2 = 22 ≦ 25,7 × 2 + 4 × 3 = 26> 25). Therefore, the third combination has two TBs and two TAs, and the total maximum positioning time at that time is 22 seconds.

When the number of TBs is three, the maximum value of the number of TAs is one (7 × 3 + 4 × 1 ≦ 25, 7 × 3 + 4 × 2 = 2).
9> 25). Accordingly, the fourth combination has three TBs and one TA, and the total maximum positioning time at that time is 25 seconds. In the above example, the pattern TB is incremented from 0 and the maximum value of the number of corresponding patterns TA is obtained. However, the pattern TA may be incremented from 0 and the maximum value of the number of corresponding patterns TB may be obtained. . In this case, there is a drawback that the processing time becomes longer because the number of combinations is larger than in the case described above.

Referring back to FIG. 4, when the route pattern table has been created as described above, steps S44 to S44 are executed.
At 46, starting from the processing start position of each processing head in the order of priority of the plurality of processing heads, the maximum positioning of each pattern in the combination of the positioning patterns defined by the processing map data within the time limit The route array is created by determining the point numbers corresponding to the number of times. The process of creating the route array is performed according to the flowchart of FIG.

In the example of FIG. 7, since the first processing head has the highest priority, the path arrangement of the first processing head is created first. The counts A and B are initialized to 0 in step S53 in FIG. When the maximum value Bmax of the count B within the time limit of 25 seconds is 0, the maximum value Amax of the count A
Is 6. When the maximum value Bmax of the count B is 3,
The maximum value Amax of the count A is 1. Next, in step S54, a point number that can be reached in the shortest time is substituted for a variable P. This corresponds to storing the point number 2 of the next processing position by the first processing head in the variable P when the priority of the first processing head is high. Next, in step S55, the count A of the number of times of positioning of the pattern TA and the count B of the number of times of positioning of the pattern TB by the first processing head within 25 seconds, which is the time limit, are incremented as follows.

(1) When TA is moved once and TB is moved zero times to point number 2 (A = 1, B = 0, moving time of the processing head is 4 × 1 + 7 × 0 = 4 ≦ 25). (2) Point number 2 and 4 with 1 TA and 1 TB
(A = 1, B = 1, the moving time of the processing head is 4 × 1 + 7 × 1 = 11 ≦ 25). (3) Once TA and twice TB, point numbers 2, 4,
6 (A = 1, B = 2, the moving time of the processing head is 4 × 1 + 7 × 2 = 18 ≦ 25).

(4) When TA is moved to point numbers 2, 4, 6, and 8 once with TA and three times with TB (A = 1, B = 3,
The moving time of the processing head is 4 × 1 + 7 × 3 = 25 ≦ 2
5). Other combinations cannot be adopted because the time limit is exceeded. Thus, the first processing head is determined in the above (4). Therefore, point numbers 9, 10, 11,
Step 12 needs to be performed by the second processing head. In this case, the moving time of the second processing head is 4 times TA and 0 times TB, and the total moving time is 4 × 4 + 7 × 0 = 16 ≦
Since it is 25, it is within the time limit and can be adopted.

In this case, the moving time of the processing head is determined by the longer one of the moving times of the first processing head and the second processing head. That is, in the processing 1 of FIG.
5 seconds. Since all the processing points have been stored in the path array, 1 is added to the number of array creations in step S48 in FIG. 4, and a value 21 obtained by subtracting a predetermined value 4 from the time limit t1 = 25 is set as a new time limit. Then, step S43 and subsequent steps are repeated again.

Process 2 shown in FIG. 8 shows a second array creation process. Similarly to the processing 1 in the path pattern table within the time limit of 21 seconds, the maximum number of times of positioning by the pattern TB can be from 0 to 3 times. Pattern T
Pattern TA when the maximum number of positioning by B is 0
The maximum value of the maximum number of times of positioning is 5 (7 × 0 +
4 × 5 = 20 ≦ 21, 7 × 0 + 4 × 6 = 24> 21).
Therefore, in the first combination, the pattern TB is 0 times and the pattern TA is 5 times, and the total maximum positioning time at that time is 20 seconds.

Pattern TA when pattern TB is one time
Is 3 (7 × 1 + 4 × 3 = 19 ≦ 2)
5,7 × 1 + 4 × 4 = 23> 21). Thus, the next combination has one TB and three TAs, for a total maximum positioning time of 19 seconds. When the number of TBs is two, the maximum value of the number of TAs is one (7 × 2 + 4 × 1 = 1).
8 ≦ 21, 7 × 2 + 4 × 2 = 22> 21). Thus, the third combination has two TBs and one TA,
The total maximum positioning time at that time is 18 seconds.

When the number of TB is three, the maximum value of the number of TAs is 0 (7 × 3 + 4 × 0 ≦ 21, 7 × 3 + 4 × 1 = 2).
5> 21). Accordingly, the fourth combination has three TBs and zero TAs, and the total maximum positioning time at that time is 21 seconds. When the second route pattern table is created as described above, step S4 in FIG.
In steps S4 to S46, the path arrangement of the first and second processing heads is created in the same manner as in the processing 1 according to the flowchart of FIG. 5 and the processing map of FIG.

The possible combinations of the count A of the number of times the pattern TA is positioned by the first processing head and the count B of the number of times the pattern TB is positioned by the first processing head within the limited time of 21 seconds after the update are as follows. . (1) When TA is moved once and TB is moved 0 times to point number 2 (A = 1, B = 0, the moving time of the processing head is 4
× 1 + 7 × 0 = 4 ≦ 25).

(2) When TA is moved once and TB is moved once to point numbers 2 and 4 (A = 1, B = 1, the moving time of the processing head is 4 × 1 + 7 × 1 = 11 ≦ 25) . (3) Once TA and twice TB, point numbers 2, 4,
6 (A = 1, B = 2, the moving time of the processing head is 4 × 1 + 7 × 2 = 18 ≦ 25). Other combinations cannot be adopted because the time limit is exceeded. Thus, considering the processing map data of FIG. 6, the path arrangement of the first processing head is determined in the above (3).

Therefore, the point numbers 8, 9, 10,
Steps 11 and 12 need to be performed by the second processing head.
In this case, the moving time of the second processing head is 5 times for TA and 0 times for TB, and the total moving time is 4 × 5 + 7 × 0 =
Since 20 ≦ 21, it is within the time limit and can be adopted. The movement time of the processing head is the first processing head and the second
Of the processing head movement time. That is,
In the process 2 of FIG. 8, the movement time is 20 seconds. Since all the processing points have been stored in the path array, 1 is added to the number of array creations in step S48 in FIG. 4, and a value 17 obtained by subtracting a predetermined value 4 from the time limit t1 = 21 is set as a new time limit. Then, step S43 and subsequent steps are repeated again.

Process 3 shown in FIG. 9 shows a third array creation process. In the route pattern table within the time limit of 17 seconds, the maximum number of times of positioning by the pattern TB is 0.
There can be up to two times. The maximum value of the maximum number of times of positioning by the pattern TA when the maximum number of times of positioning by the pattern TB is 0 is 4 (7 × 0 + 4 × 4 = 1).
6 ≦ 17,7 × 0 + 4 × 5 = 20> 17). Therefore, in the first combination, the pattern TB is 0 and the pattern TA
Is four times, and the total maximum positioning time at that time is 17
Seconds.

Pattern TA when pattern TB is one time
Is 2 (7 × 1 + 4 × 2 = 15 ≦ 1)
7,7 × 1 + 4 × 3 = 18> 17). Thus, the next combination has one TB and two TAs, for a total maximum positioning time of 15 seconds. When the number of TBs is two, the maximum value of the number of TAs is 0 (7 × 2 + 4 × 0 = 1).
4 ≦ 17,7 × 2 + 4 × 1 = 18> 17). Thus, the third combination has two TBs and zero TAs,
The total maximum positioning time at that time is 14 seconds.

When the third route pattern table is created as described above, steps S44 to S44 in FIG.
In S46, a path array of the first and second processing heads is created in the same manner as in the processing 1 according to the flowchart of FIG. 5 and the processing map of FIG. Pattern TA by first processing head within 17 seconds, which is the time limit after updating
The possibility of the combination of and the pattern TB is as follows.

(1) When TA is moved once and TB is moved zero times to point number 2 (movement time of the processing head is 4 × 1
+ 7 × 0 = 4 ≦ 17). (2) Point number 2 and 4 with 1 TA and 1 TB
When moving to (the moving time of the processing head is 4 × 1 + 7 ×
1 = 11 ≦ 17). Other combinations cannot be adopted because the time limit is exceeded. Thus, the path arrangement of the first processing head is determined in the above (2).

Therefore, the point numbers 6, 8, 9, 1
0, 11, and 12 must be performed by the second processing head. In this case, the moving time of the second processing head is 5 for TA.
The number of times is TB once, and the total travel time is 4 × 5 + 7 ×
Since 1 = 27> 17, it cannot be adopted because it exceeds the time limit. After all, there is no pattern combination that can be adopted within the time limit after updating.

Since the number of array creations in process 3 is not 0, the process ends normally in step S51 of FIG. 4, and the path array created in process 2 (the first processing head moves the point number 2 in the pattern TA, The point numbers 4 and 6 are moved by the pattern TB, and the second processing head is moved by the point numbers 12, 11, 10, 9, and 8 by the pattern TA.) The movement path of the processing head in the shortest time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of a system in which a route determination device according to the present invention is used.

FIG. 2 is a block diagram showing an outline of a configuration of a route determination device 1 in the device of FIG.

FIG. 3 is a functional block diagram mainly showing a CPU 231 in a control unit 23 shown in FIG.

FIG. 4 is a flowchart illustrating an operation of the route determination device shown in FIGS. 2 and 3;

FIG. 5 is a flowchart illustrating details of operations in steps S44 to S46 in FIG.

FIG. 6 is a diagram showing an example of processing map data.

FIG. 7 is a diagram specifically illustrating an example 1 of a creation operation of a route pattern table;

FIG. 8 is a diagram specifically illustrating a second example of the creation operation of the route pattern table.

FIG. 9 is a diagram specifically illustrating a third example of the operation of creating a route pattern table;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Route determination device 23 ... Control part 31 ... Route array preparation means 32 ... Determination means 33 ... Number counting means 34 ... Limited time updating means 35 ... Count value determination means

Claims (2)

[Claims] 1. A method for determining a path of a processing head when each of a plurality of processing positions of a target device is processed by any one of a plurality of processing heads, wherein the processing position is determined according to a model of the target device; Loading from a data file processing map data comprising a processing start position of each of the processing heads, and a moving time of the processing head between two adjacent processing positions represented by one of a plurality of positioning patterns. Initializing the number of array creations, setting a time limit to an initial value, and in a predetermined priority order of the plurality of processing heads,
Starting from the processing start position of each processing head, determining a combination of positioning patterns covering all of the processing positions within the time limit, and creating a path array; and Updating all of the array creation times, subtracting a predetermined value from the time limit, and repeating the step of creating the path array, if all are covered, A path determination method for a processing head, wherein the path array created immediately before is determined as the paths of the plurality of processing heads when not all of the paths are covered and the number of array generations is not the initial value. . 2. A path determining device for a processing head when processing each of a plurality of processing positions of a target device by any of a plurality of processing heads, wherein, in accordance with a model of the target device, the processing position; A data registration unit that includes a processing start position for each of the processing heads, and stores processing map data that represents a moving time of the processing head between two adjacent processing positions by one of a plurality of positioning patterns. In the order of the predetermined priority of the plurality of processing heads,
Starting from the processing start position of each processing head, determining a combination of positioning patterns defined by the processing map data within the time limit, and generating a path array by generating a path array; Determining means for determining whether all of the processing positions are covered; number-of-times counting means for counting the number of array creation times; and count value determining means for determining whether the count value of the number-of-times counting means is an initial value, A time limit updating unit that updates the time limit by subtracting a predetermined value from the time limit, and when all of the machining positions are covered by the path array, the adding unit calculates the number of array creation times. Updating and updating the time limit by the updating means and then creating a route array by creating the route array, If all of the machining positions are not covered by the path array and the number of array creations by the number counting means is not an initial value, the path array created immediately before is determined as a path of the plurality of machining heads. A path determining device for a processing head.