JP6324163B2 - Process classification apparatus, process classification program, and storage medium - Google Patents

Description

  The present invention relates to a process classification apparatus, a process classification program, and a storage medium that can suitably classify processes included in industrial product development processes and simulation models.

  When developing industrial products such as vehicles, electronic devices, and software, those development processes include a number of processes. In addition, many processes are complicatedly intertwined. Furthermore, many engineers are involved in the design of each process, and it is not easy to grasp the whole picture.

  When a development process includes a large number of manufacturing processes, it is preferable to group highly related processes into a group. This is because it is easy for engineers to understand the entire process. That is, it leads to process design support. As a method of grouping, for example, there is a Newman method using modularity (refer to Formula (4) on page 2 of Non-Patent Document 1). The Newman method is a very general method for grouping information on a certain network by paying attention to the relevance of the information. Further, Patent Document 1 discloses an apparatus that supports component design by grouping castings in a manufacturing process using a Newman method.

JP 2012-125895 A

Aaron Clauset, M.E.J.Newman, and Cristopher Moore, Finding community structure in very large networks, Physical Review E 70, 066111 (2004)

  By the way, in the industrial product development process, the test result of the prototype may be fed back to the design and redesigned. Thus, it is common that there is a “return” that reflects the result of the subsequent process in the previous process. In such a case, the previous process is repeatedly performed based on the reflected result. If there are repeated steps, the development period will be enlarged. Therefore, it is important to manage “return” appropriately.

  However, in the Newman method, it is difficult to appropriately handle “return”. This is because the Newman formula does not include a term for evaluating “return”. For example, by using the Newman method, it is assumed that grouping is performed such that the downstream process group (A2) is performed after the upstream process group (A1) is performed. In this case, a “return” that feeds back from the process of the downstream process group (A2) to the process of the upstream process group (A1) generally occurs.

  Thus, if there is a “return” across groups, the overall picture of the development process becomes complicated. In addition, the development group in charge of the upstream process group (A1) changes the design and development many times, leading to a prolonged overall development period. Therefore, it is necessary to perform grouping considering “return”.

  By the way, in an industrial product development process, for example, a simulation model such as SIMULLINK (registered trademark) may be used. The simulation model has a plurality of processing elements. Then, after the upstream processing element is executed, the downstream processing element is executed. Also in the simulation model, there is a “return” as in the industrial product development process.

  The present invention has been made to solve the above-described problems of the prior art. That is, the problem is to provide a process classification apparatus, a process classification program, and a storage medium in consideration of the relationship between processes and the “return” between processes. Here, the “process” includes one process included in the industrial product development process, one processing content included in the program, and one processing element included in the simulation model.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
適応関数ｆ１（Ｑｍ）を用いてプロセスを分類する。 The process classification apparatus according to the first aspect includes a process information acquisition unit that acquires a plurality of processes, processing order information indicating the processing order of the plurality of processes, and the plurality of processes as one of a plurality of groups. And a process classification unit for classifying. The process classification part is expressed by the following formula:

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
a, b: positive constant
c1: Positive constant
c2: classify processes using positive constant adaptation function f1 (Qm).

  This process classification apparatus classifies processes included in the development process of industrial products and simulation models. The larger the adaptive function f1 (Qm), the more suitable classification can be performed. The adaptation function f1 (Qm) depends on the modularity Q and also on the feedback number fb1. Therefore, it is possible to classify closely related processes into the same group, and to reduce “return” across different groups as much as possible.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
Ｇｒ：グループ数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
ｃ３：正の定数
ｎ：２以上の自然数
適応関数ｆ２（Ｑｍ）を用いて、グループ数の増減とプロセスの分類とを同時に行う。 The process classification apparatus according to the second aspect includes a process information acquisition unit that acquires a plurality of processes and processing order information indicating the processing order of the plurality of processes, and sets the plurality of processes to one of a plurality of groups. And a process classification unit for classifying. The process classification part is expressed by the following formula:

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: Positive constant
n: Increase / decrease the number of groups and classify processes simultaneously using a natural number adaptation function f2 (Qm) of 2 or more.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
Ｇｒ：グループ数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
ｃ３：正の定数
適応関数ｆ３（Ｑｍ）を用いて、グループ数の増減とプロセスの分類とを同時に行う。 The process classification apparatus according to the third aspect includes a process information acquisition unit that acquires a plurality of processes, processing order information indicating the processing order of the plurality of processes, and the plurality of processes as one of a plurality of groups. And a process classification unit for classifying. The process classification part is expressed by the following formula:

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: The increase / decrease of the number of groups and the classification of processes are simultaneously performed using the positive constant adaptation function f3 (Qm).

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
適応関数ｆ１（Ｑｍ）を用いてプロセスを分類する。 The process classification program according to the fourth aspect includes a process information acquisition function for acquiring a plurality of processes and processing order information indicating the processing order of the plurality of processes, and the plurality of processes as one of a plurality of groups. This is to allow a computer to realize the process classification function to classify. The process classification function is expressed by the following formula:

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
a, b: positive constant
c1: Positive constant
c2: classify processes using positive constant adaptation function f1 (Qm).

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
Ｇｒ：グループ数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
ｃ３：正の定数
ｎ：２以上の自然数
適応関数ｆ２（Ｑｍ）を用いて、グループ数の増減とプロセスの分類とを同時に行う。 The process classification program according to the fifth aspect includes a process information acquisition function for acquiring a plurality of processes, processing order information indicating the processing order of the plurality of processes, and the plurality of processes in any of a plurality of groups. This is to allow a computer to realize the process classification function to classify. The process classification function is expressed by the following formula:

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: Positive constant
n: Increase / decrease the number of groups and classify processes simultaneously using a natural number adaptation function f2 (Qm) of 2 or more.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
Ｇｒ：グループ数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
ｃ３：正の定数
適応関数ｆ３（Ｑｍ）を用いて、グループ数の増減とプロセスの分類とを同時に行う。 The process classification program according to the sixth aspect includes a process information acquisition function that acquires a plurality of processes, processing order information that indicates a processing order of the plurality of processes, and the plurality of processes as one of a plurality of groups. This is to allow a computer to realize the process classification function to classify. The process classification function is expressed by the following formula:

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: The increase / decrease of the number of groups and the classification of processes are simultaneously performed using the positive constant adaptation function f3 (Qm).

  In the process classification program according to the seventh aspect, the plurality of processes are processing elements in the simulation model.

  In the process classification program according to the eighth aspect, the plurality of processes are steps in an industrial product development process.

  A storage medium according to the ninth aspect stores the above-described process classification program and is readable by a computer.

  In the present invention, a process classification apparatus, a process classification program, and a storage medium are provided in consideration of the relationship between processes and “return” between processes.

It is a figure which illustrates the block diagram handled with the process classification apparatus which concerns on 1st Embodiment. It is a schematic block diagram which shows the structure of the hardware of the process classification apparatus which concerns on 1st Embodiment. It is a block diagram for demonstrating the structure of the control part of the process classification apparatus which concerns on 1st Embodiment. It is a flowchart which shows the flow of the process classification apparatus which concerns on 1st Embodiment. It is a figure which shows the block diagram after grouping the block diagram of FIG. 1 by the process classification apparatus which concerns on 1st Embodiment. FIG. 6 is a first expanded view of a part of the block diagram of FIG. 5. FIG. 6 is a developed view (part 2) of a part of the block diagram of FIG. It is a figure which shows the case where grouping by the Newman method is implemented. It is a block diagram which shows the case where it groups using the 2nd adaptive function. It is the figure which expanded a part of block diagram of FIG.

  Embodiments according to the present invention will be described below with reference to the accompanying drawings. In the following embodiments, a process classification apparatus, a process classification program, and a storage medium that group a large number of processes will be described. In the following embodiments, the process is a concept including one process included in the industrial product development process, one processing content included in the program, and one processing element included in the simulation model. From the first embodiment to the third embodiment, a case will be described in which the processes of the robot arm control software are classified. In the fourth embodiment, a case in which each process of the industrial product development process is classified will be described as an example.

(First embodiment)
1. Block Diagram FIG. 1 is a block diagram showing processing elements of robot arm control software classified by the process classification apparatus of this embodiment. On the left side of FIG. 1, an overall view ST1 of the control software for the robot arm is shown. The overall view ST1 has a lower hierarchy ST2. The lower hierarchy ST2 further has lower hierarchy ST3, ST4. Thus, the control software has a hierarchical structure.

  As shown in FIG. 1, the block diagram includes a processing element surrounded by a frame and a connection line connecting one processing element and another processing element. Here, the processing element indicates a process. The connection indicates relationship information between processing elements. In FIG. 1, the processing element E1 and the connection K1 are illustrated. Connections are indicated by arrows. This arrow indicates an execution procedure in which the processing element at the tip of the arrow is executed after the processing element at the root of the arrow is executed. In the block diagram shown in FIG. 1, processing elements are generally executed from the left side to the right side in FIG. That is, the upstream process is arranged on the left side in FIG. 1, and the downstream process is arranged on the right side. Then, based on the arrow, data is transferred from the upstream side to the downstream side, and the process proceeds.

  Moreover, the connection F1 is shown by FIG. The connection line F1 exits from the right processing element in FIG. 1 and enters the left element. This means that the upstream process uses the result of the downstream process. Thus, the reuse of the result of the downstream process by the upstream process is referred to as “return”.

  This return processing must be appropriately managed from the viewpoint of readability and group-by-group processing. For example, assume that the robot arm control software is divided into a plurality of programs. In that case, the program cannot be executed normally unless data is exchanged between a plurality of programs. Therefore, it is difficult to test the control software at the development stage. Therefore, it is preferable to reduce the return from one group to the other group.

2. Process classification apparatus The process classification apparatus 1 of this embodiment is for implementing grouping according to the degree of the depth of the relationship between processes about the system which has many processes. In that case, it is possible to classify the processes in consideration of (1) the degree of the depth of the relationship between processes and (2) few “returns” to return from one group to another group. Here, the process of this embodiment includes a software subsystem and module.

  FIG. 2 shows the hardware configuration of the process classification apparatus 1 of the present embodiment. The process classification device 1 includes a processing device 10, an input unit 20, and a display unit 30. The processing device 10 specifically performs various processes of the process classification device 1. The input unit 20 is for receiving input from the user. Alternatively, an output result from another computer may be accepted. An example is a keyboard. The display unit 30 is for displaying the results of executing various processes by the processing apparatus 10. For example, a display is mentioned.

  The processing device 10 includes a control unit 100 and a nonvolatile memory 200. The control unit 100 performs various controls. The control unit 100 includes a CPU (C1) and a memory M1. The control unit 100 performs various controls by using the CPUC1 and the memory M1. The non-volatile memory 200 stores a program P1. As will be described later, the program P1 is a process classification program to be realized by a computer. As a result, the computer functions as the process classification device 1.

3. Control Unit FIG. 3 is a block diagram for explaining each unit that performs various controls in the control unit 100. Each unit of the control unit 100 functions by using at least a part of the CPUC1 and at least a part of the memory M1.

  The control unit 100 includes a process information acquisition unit 110, a process classification unit 120, an input reception unit 130, and other control units 140.

  The process information acquisition unit 110 has a relationship information acquisition function for acquiring information related to the block diagram. That is, a plurality of processes, dependency relationship information between a dependency source process and a dependency destination process, and a hierarchical structure in which a part of the plurality of processes constitute one or more lower layers, get. Other information is also acquired.

  The process classification unit 120 classifies processes using an adaptive function described later. The higher the value of the adaptive function, the better the grouping is performed. The adaptive function will be described later.

  The input receiving unit 130 has an input receiving function for receiving input from the user. The input receiving unit 130 includes a constant input receiving unit 131 and a group number input receiving unit 132. These are for receiving input from the user for the parameters of the adaptive function described later. Note that the group number input reception unit 132 is not used in the present embodiment. Used in the second and subsequent embodiments.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
この第１の適応関数ｆ１（Ｑｍ）を用いることにより、プロセスを分類することができる。 4). Adaptive function (fitness function)
4-1. First Adaptive Function In the present embodiment, the first adaptive function is given by the following equation (1).

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
a, b: positive constant
c1: Positive constant
c2: Positive constant By using this first adaptive function f1 (Qm), the process can be classified.

  Here, the modularity Q is defined by Equation (4) of Non-Patent Document 1 (Aaron Clauset, MEJ Newman, and Cristopher Moore, Finding community structure in very large networks, Physical Review E 70, 066111 (2004)). Modularity. The feedback number fb1 represents the number of feedback returning from the downstream group process to the upstream group process. That is, the number of feedback across groups. Therefore, “return” returning from a process of a group to the process of the same group, that is, “return” within the same group is not counted. Here, the feedback number fb1 was originally devised and introduced by the present inventors. The constants c1 and c2 are positive real numbers of 0.5 or more and 2.0 or less. The constants a and b are real numbers given by the user. These will be described later.

  The modularity Q often takes a value from 0 to 1. However, it may take a negative value for calculation. Therefore, in the present embodiment, the modified modularity Qm defined by Qm = Q + c1 is used. Further, the feedback number fb1 may be zero. In that case, in Formula (1), it will divide by zero. In order to avoid this, the modified feedback number fbm defined by fbm = fb1 + c2 is used. The first adaptive function f1 (Qm) does not diverge infinitely. That is, the first adaptive function f1 (Qm) takes a maximum value under certain conditions.

4-2. Calculation Method In order to classify the processes, the first adaptive function f1 (Qm) of Equation (1) is calculated. Then, iterative calculation is performed so that the value of the first adaptive function f1 (Qm) becomes large. Then, the iterative calculation is repeated until the value of the first adaptive function f1 (Qm) converges. Whether or not the value of the first adaptive function f1 (Qm) has converged is determined when the increase width of the first adaptive function f1 (Qm) is equal to or less than a predetermined threshold value. It is determined that the value of (Qm) has converged. Otherwise, iterative calculation may be continued.

  Thus, although the maximization of Expression (1) is performed, the maximum value of the first adaptive function f1 (Qm) is not necessarily obtained in the calculation. A value of 90% or more of the maximum value of the first adaptive function f1 (Qm) may be sufficient.

4-3. Feature Point of First Adaptive Function The right side of the equation (1) has a first term (a · Qm) and a second term ((b / fbm) · Qm). The first term (a · Qm) is a term for evaluating the relationship between processes. The second term ((b / fbm) · Qm) is a term for evaluating “return”. The first term (a · Qm) is a value that increases as the highly related processes are grouped together. On the other hand, the second term ((b / fbm) · Qm) has a smaller value as the number of “returns” is larger, and a larger value as the number of “returns” is smaller.

  Thus, the first adaptive function f1 (Qm) is represented by the sum of the first term (a · Qm) and the second term ((b / fbm) · Qm). Therefore, even when the first adaptive function f1 (Qm) is maximized, the first term is large and the second term is relatively small, and the first term is relatively small and the second term is large. There is. In some cases, the size of the first term is approximately equal to the size of the second term. When the first term is large, the relevance between the processes in the group is high. When the second term is large, the number of “returns” across the group is small. When the first term and the second term are substantially equal, the relationship between the processes in the group is moderately high, and the number of “returns” across the group is moderately small. In addition, when the user selects and inputs the constants a and b, it is possible to select whether to perform the grouping with priority on the modularity Q or to perform the grouping with priority on the reduction of “return”.

4-4. Here, the case where priority is given to the modularity Q will be described. Modularity Q may be given priority when priority is given to increasing the relevance between processes classified into groups rather than reducing “returns” across groups. In this case, in formula (1), the constant a is sufficiently larger than the constant b. The user may input a constant a that is larger than the constant b to the process classification apparatus 1.

4-5. Here, a case where priority is given to a reduction in the number of “returns” between groups will be described. In the case where priority is given to reducing “returns” across groups rather than increasing the relevance between processes classified into groups, reduction of the number of “returns” may be given priority. In this case, in equation (1), the constant b is sufficiently larger than the constant a. The user may input a constant b that is larger than the constant a into the process classification device 1.

5. Process Classification Method The process classification method in the present embodiment will be described with reference to the flowchart of FIG. First, the process classification apparatus 1 reads a software program (S101). At this time, the process information acquisition unit 110 reads a plurality of processes and relationship information between the processes.

  Usually, a program has a plurality of subroutines. The subroutine corresponds to the group in the present embodiment. Therefore, it has already been divided into groups at the stage of reading. The process classification apparatus 1 of the present embodiment can use this read state as an initial state. Alternatively, an initial state other than the read state can be used.

  Next, the constant input receiving unit 131 of the input receiving unit 130 receives an input of a constant used for calculating the adaptive function (S102). For example, the user directly inputs the values of constants a and b into the input unit 20. Alternatively, a file in which the constants a and b are input may be prepared in advance, and the values of the constants a and b may be read from the file. Alternatively, an input of a constant other than the constants a and b used for calculating the adaptive function may be accepted. For example, constants c1, c2, etc.

  Next, the process classification unit 120 calculates an adaptive function (S103). If the value of the adaptive function has converged (S104: Yes), the process proceeds to S105. If the value of the adaptive function has not converged (S104: No), the process returns to S103 and the adaptive function is calculated again. When the adaptive function has converged (S104: Yes), the grouped result is displayed (S105). The grouped result may be stored in the memory M1 or the like.

  In calculating the adaptive function, a plurality of initial conditions are prepared. For example, 10 initial conditions. Next, an adaptive function is calculated for each of the ten initial conditions. A plurality of conditions close to the initial condition of the value of the largest adaptive function among the 10 adaptive functions are prepared. For example, five secondary conditions. Then, an adaptive function is calculated for the five secondary conditions. And the largest adaptive function is employ | adopted among five adaptive functions. The group when the adaptation function is maximized is the most preferable group. This most suitable group varies depending on input values such as a and b. The above is only an example, and other numbers may be used as the number of prepared initial conditions and secondary conditions. Here, the initial condition and the secondary condition are variations for designating which group each process belongs to. Further, iterative calculation may be performed using a cubic condition or higher.

6). Execution of grouping (specific example)
6-1. State after Grouping FIG. 5 is a diagram illustrating a state after the robot arm control software program is classified by the process classification device 1 of the present embodiment. That is, the process classification device 1 groups, for example, from the state of FIG. 1 to the state of FIG. As shown in FIG. 5, the process classification device 1 is classified into nine parts from the subsystem S1 to the subsystem S9.

  Here, the subsystem S1 is a trajectory function creation group. The subsystem S2 is a moving coordinate acquisition calculation group. The subsystem S3 is an angle / coordinate system conversion group. The subsystem S4 is a Y coordinate position adjustment group. The subsystem S5 is an X coordinate position adjustment group. Subsystem S6 is a group of inverse kinematic calculations. Subsystem S7 is an inverse kinematics calculation 2 group. Subsystem S8 is an inverse kinematics calculation 3 group. The subsystem S9 is an angle radian transform / result output group. Thus, each subsystem corresponds to each group. Inverse kinematic calculations are divided into three groups.

  FIG. 6 shows details of the subsystem S1. As shown in FIG. 6, there are returns represented by arrows F2, F3, F4, F5, and F6 within the subsystem S1. As described above, the total number of returns in the overall view ST1 does not change before and after grouping. Therefore, instead of confining the return inside the subsystem, it is possible to prevent the return from crossing the subsystem.

  FIG. 7 shows the inside of the subsystems S3, S4, and S5. No return is included in the subsystems S3, S4, and S5.

6-2. Effect of Grouping FIG. 5 shows the result when priority is given to “return”. Therefore, in FIG. 5, there is no “return” for returning from one subsystem to another. Note that the total number of returns in the overall view ST1 does not change before and after grouping. Instead, as shown in FIG. 6, for example, there is a “return” to return from one process in the subsystem S1 to another process. That is, the “return” is confined within the group, and there is no “return” across the group.

  By the grouping of the present embodiment, it was possible to carry out a simple and clear grouping as described above. Also, there is almost no “return” across groups. In other words, the subsystem is highly independent. Therefore, when developing this software, any of the subsystems S1 to S9 may be developed first. Of course, separate development groups may perform development separately. In addition, the subsystems S1 to S9 can be implemented independently when the program test is performed. Also, it is not necessary to test the upstream subsystem after testing the downstream subsystem.

7). Grouping by Newman Method FIG. 8 shows the result of grouping by the conventional Newman method. As shown in FIG. 8, highly related processes could be classified into a plurality of groups. However, as shown by arrows F7 and F8 in FIG. 8, a “return” across the group occurs to some extent. Therefore, it is not easy to test the subsystem alone. Also, the assignment of each person in charge is not as easy as this embodiment.

8). Effects of this embodiment Generally, a program becomes larger as development progresses. Also, as development progresses, parameters that are different from the parameters introduced in the initial stage are often introduced. Therefore, by carrying out the grouping of the present embodiment, a highly related program group among the programs is handled as one group. Thereby, the readability of the program is improved. In addition, when one project is shared by multiple people, it is easy to decide the person in charge. For example, one person in charge may develop one grouped group. In addition, when testing the entire program, the tests can be sequentially performed for each group. Therefore, debugging work can be made more efficient.

9. Modification 9-1. Storage Medium In the present embodiment, the process classification device 1 and the process classification program P1 have been described. However, the present invention can also be applied to a storage medium in which the process classification program P1 is recorded. This storage medium is a computer-readable storage medium that stores the process classification program P1. Examples of the storage medium include CDROM and DVDROM. Of course, it may be other than this.

9-2. Dependency Source and Dependency Destination In this embodiment, dependency relationship information having a dependency source and a dependency destination is used. However, even a bidirectional relationship that is not divided into a dependency source and a dependency destination can be similarly applied.

10. Summary of the present embodiment As described in detail above, the process classification device 1 of the present embodiment is a device that groups a number of processes using the first adaptive function f1 (Qm). As a result, the grouping can be performed in consideration of both the relationship between the processes and the number of “returns”. This grouping can improve the readability of many processes and programs. In addition, it is possible to suitably divide processes and programs. Therefore, the development of the program can be suitably distributed to each person in charge. Tests can also be performed in order.

  In addition, a present Example is only a mere illustration. Therefore, naturally, various improvements and modifications can be made without departing from the scope of the invention.

(Second Embodiment)
A second embodiment will be described. In the present embodiment, an adaptation function different from the adaptation function of the first embodiment is used. The other points are the same as in the first embodiment. Therefore, the adaptive function will be described.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
Ｇｒ：グループ数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
ｃ３：正の定数（自然数）
ｎ：２以上の自然数
この第２の適応関数ｆ２（Ｑｍ）を用いることにより、グループ数の増減とプロセスの分類とを同時に行うことができる。 1. Adaptive function (fitness function)
1-1. Second Adaptive Function In this embodiment, the following second adaptive function is used.

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: Positive constant (natural number)
n: Natural number equal to or greater than 2 By using the second adaptive function f2 (Qm), increase / decrease of the number of groups and classification of processes can be performed simultaneously.

1-2. Number of groups (second adaptive function)
Equation (3) has a group function Grm. The group function Grm has a constant c3. The constant c3 is input from the user. The closer the variable Gr is to the constant c3, the larger the group function Grm. Therefore, as the iterative calculation is repeated, the variable Gr becomes a natural number close to the constant c3.

  For example, when the user wants to set the number of groups to about “5”, “5” is input to the constant c3. The group number input receiving unit 132 receives an input of “5” by the user. Then, the group number input receiving unit 132 inputs “5” to the constant c3.

  As the iterative calculation is performed using the second adaptive function f2 (Qm), the variable Gr takes values such as “4”, “5”, “6”, and the like. Because of the influence of the modularity Q and the feedback number fb1, the value of the variable Gr does not always remain “5”. Considering the function form of the group function Grm, the value of the variable Gr tends to be a natural number close to “5”.

  In this way, the process classification unit 120 performs grouping while increasing or decreasing the number of groups to which a plurality of processes belong. That is, the process classification unit 120 has a group number increasing / decreasing function for increasing / decreasing the number of groups.

2. Execution of grouping (specific example)
FIG. 9 shows the results using the second adaptive function. FIG. 9 shows the result when the value of the constant c3 is set to 5 in Equation (3). As shown in FIG. 9, six subsystems are created.

  FIG. 10 is a diagram showing a state where the inverse kinematic calculation, the angle radian transform, and the result output group are expanded in the subsystem.

  The second adaptive function f2 (Qm) is different in function form from the first adaptive function f1 (Qm). However, grouping considering both the modularity Q and the feedback number fb1 can be performed. Also, by using the second adaptive function f2 (Qm), the user can roughly specify the number of groups.

(Third embodiment)
A third embodiment will be described. In the present embodiment, an adaptation function different from the adaptation function of the first embodiment is used. The other points are the same as in the first embodiment. Therefore, the adaptive function will be described.

Ｑ：Ｎｅｗｍａｎ法におけるモジュラリティ
ｆｂ１：下流のグループのプロセスから上流のグループのプロセスに戻るフィードバックの個数
Ｇｒ：グループ数
ａ，ｂ：正の定数
ｃ１：正の定数
ｃ２：正の定数
ｃ３：正の定数（自然数）
この第３の適応関数ｆ３（Ｑｍ）を用いることにより、グループ数の増減とプロセスの分類とを同時に行うことができる。 1. Adaptive function (fitness function)
1-1. Third Adaptive Function In the present embodiment, the following third adaptive function is used.

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: Positive constant (natural number)
By using the third adaptive function f3 (Qm), the increase / decrease of the number of groups and the classification of processes can be performed simultaneously.

1-2. Number of groups (third adaptive function)
Equation (5) has a group function Grm. The group function Grm has a constant c3. The constant c3 is input from the user. The closer the variable Gr is to the constant c3, the larger the group function Grm. Therefore, as the iterative calculation is repeated, the variable Gr becomes a natural number close to the constant c3. As described above, the point affected by the group function Grm is the same as in the second embodiment.

  The third adaptive function f3 (Qm) is different in function form from the second adaptive function f2 (Qm). However, grouping considering both the modularity Q and the feedback number fb1 can be performed. The user can also specify a rough number of groups.

(Fourth embodiment)
A fourth embodiment will be described. From the first embodiment to the third embodiment, the robot arm control software programs are grouped. However, the industrial product development process may be grouped.

  For example, consider grouping the development process of a refrigerator. In this development process, a post process is performed after the pre process. The pre-process and the post-process are connected by a connection represented by an arrow. In addition, there is a “return” that feeds back the result of the subsequent process to the previous process. Therefore, the first adaptive function f1 (Qm), the second adaptive function f2 (Qm), or the third adaptive function f3 (Qm) described in the first to third embodiments is used. Can be grouped together.

  For example, by performing such grouping, it is assumed that there is no return between groups and the result is that the number of groups is “7”. Then, it suffices to divide the development personnel into seven groups and develop each group. In this way, by feeding back the result of one group to another group, it is possible to avoid following the same process again. However, the same process may be tried many times within the same group. Even if there are some returns that feed back the results of one group to other groups, the effect of shortening the development period can be obtained if the number of returns can be reduced.

  In this way, it is possible to suitably manage the return in which the result of the post-process is fed back to the pre-process. In addition, the development period can be shortened. As described above, a design process for designing an industrial product development process can be suitably performed.

  Similarly, the above grouping can be applied to a simulation model. This is because there are many processes that run from upstream to downstream, and there are returns.

P1 ... Program 1 ... Process classification device 10 ... Processing device 20 ... Input unit 30 ... Display unit C1 ... CPU
M1 ... Memory 100 ... Control unit 110 ... Process information acquisition unit 120 ... Process classification unit 130 ... Input reception unit 200 ... Non-volatile memory

Claims (9)

A process information acquisition unit for acquiring a plurality of processes, and processing order information representing a processing order of the plurality of processes;
And processes the classification section for classifying the plurality of process to one of multiple groups,
Have
The process classification unit
It is expressed by the following formula

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
a, b: positive constant
c1: Positive constant
c2: A process classification apparatus for classifying processes using a positive constant adaptive function f1 (Qm). A process information acquisition unit for acquiring a plurality of processes, and processing order information representing a processing order of the plurality of processes;
And processes the classification section for classifying the plurality of process to one of multiple groups,
Have
The process classification unit
It is expressed by the following formula

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: Positive constant
n: A process classification apparatus that performs increase / decrease in the number of groups and classification of the processes at the same time using a natural number adaptation function f2 (Qm) of 2 or more. A process information acquisition unit for acquiring a plurality of processes, and processing order information representing a processing order of the plurality of processes;
And processes the classification section for classifying the plurality of process to one of multiple groups,
Have
The process classification unit
It is expressed by the following formula

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: A process classification apparatus that performs increase / decrease of the number of groups and classification of the processes at the same time using a positive constant adaptation function f3 (Qm). A process information acquisition function for acquiring a plurality of processes and processing order information representing the processing order of the plurality of processes;
A process classification function for classifying the plurality of processes into one of a plurality of groups;
In the process classification program for realizing
The process classification function is
It is expressed by the following formula

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
a, b: positive constant
c1: Positive constant
c2: A process classification program for classifying processes using a positive constant adaptation function f1 (Qm). A process information acquisition function for acquiring a plurality of processes and processing order information representing the processing order of the plurality of processes;
A process classification function for classifying the plurality of processes into one of a plurality of groups;
In the process classification program for realizing
The process classification function is
It is expressed by the following formula

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: Positive constant
n: A process classification program for performing increase / decrease in the number of groups and classification of the processes at the same time using a natural number adaptation function f2 (Qm) of 2 or more. A process information acquisition function for acquiring a plurality of processes and processing order information representing the processing order of the plurality of processes;
A process classification function for classifying the plurality of processes into one of a plurality of groups;
In the process classification program for realizing
The process classification function is
It is expressed by the following formula

Q: Modularity in the Newman method
fb1: Number of feedbacks returning from the downstream group process to the upstream group process
Gr: Number of groups
a, b: positive constant
c1: Positive constant
c2: positive constant
c3: A process classification program for performing increase / decrease of the number of groups and classification of the processes at the same time by using a positive constant adaptation function f3 (Qm). In the process classification program according to any one of claims 4 to 6,
The plurality of processes are:
A process classification program characterized by being a processing element in a simulation model. In the process classification program according to any one of claims 4 to 6,
The plurality of processes are:
A process classification program characterized by being a process in the development process of industrial products. A storage medium in which the process classification program according to any one of claims 4 to 8 is stored and readable by a computer.

Images