Device for machine learning, PSP generation device and method for machine learning

Abstract

The task is to provide a technology capable of generating a suitable multi-purpose PSP. A machine learning device comprises a receiver and a (self-)learning control unit. The receiving part receives initial information regarding a WBS work package of a standard, initial information regarding a WBS work package of a project conforming to the standard, and final information regarding a process diagnostic result of a project conforming to the standard. Based on this initial, final, and final information, the learning control part performs a learning process regarding a WBS work package, thereby generating association information that links the initial, final, and final information.

Description

Technical field

[0001] The present disclosure relates to a machine learning device, a PSP generation device and a machine learning method. Technological background

[0002] A product development project is an undertaking that creates a product which meets a consumer's quality, price, and deadline requirements. However, particularly in software development, it is difficult to determine whether the project has been completed as planned until the product is integrated into a system. For software development projects that present this challenge, the use of a work breakdown structure (WBS) is recommended.

[0003] The WBS (Work Breakdown Structure) is a system capable of fully clarifying all operations from the beginning to the end of a project. However, in the latest project, there is a tendency for the number of WBS work packages to increase, along with the complexity of the relationships between them. Furthermore, there is a case where products with differing quality characteristics are being manufactured concurrently by the same company. Differences in compliance with a standard, which places particular emphasis on performance, scalability, technical operation, maintenance, system environment, and environmental aspects of the quality characteristics, complicate the creation of the WBS for the project.

[0004] To solve such a problem, various technologies are proposed. For example, patent document 1 proposes a technique in which a management ID is assigned to the work to be performed, the management ID being accompanied by tracking information, and the tracking information being recorded in a traceability ID recording terminal. Patent document 2 proposes supervised learning in which a state variable and destination data are used as training data. Prior art documents (patent document(s)) Patent document 1: Japanese patent application disclosure number 2018-5802 Patent document 2: Japanese patent application, disclosure number 2017-188030 Brief description Problem to be solved by the invention

[0005] Traditionally, experts analyze a requirement received from a consumer and generate a basis for the PSP by appropriately reusing characteristics extracted from a company's previous project results. However, this generation process has many elements that depend on the experts' skills, and it is difficult to achieve consistent compliance with a standard within a limited timeframe while incorporating the consumer's requirements into the PSP basis.

[0006] Accordingly, conducting learning based on information of a standard is being considered; however, there is a problem that an appropriate PSP cannot be generated solely from the information of the standard.

[0007] The present disclosure was therefore made to solve the problems described above, and it is a purpose of the present disclosure to provide a technology capable of generating a suitable multi-purpose PSP. methods for problem solving

[0008] A machine learning device according to the present disclosure comprises: a receiving part that receives first information regarding a PSP work package of a standard, second information regarding a PSP work package of a project conforming to the standard, and third information regarding a process diagnostic result of a project conforming to the standard; and a (self-)learning control part that, based on the first information, the second information, and the third information, performs learning regarding a PSP work package, thereby generating association information that associates the first information, the second information, and the third information. Effects of the invention

[0009] According to the present disclosure, learning regarding the WBS work package is performed based on the first piece of information regarding the WBS work package of the standard, the second piece of information regarding the WBS work package of the project corresponding to the standard, and the third piece of information regarding the process diagnostic result of the project corresponding to the standard, in order to generate the association information. With such a structure, a suitable multi-purpose WBS can be generated.

[0010] These and other tasks, features, aspects and advantages of the present disclosure will become apparent from the following detailed description of the specification when considered together with the accompanying drawings. Brief description of the drawings Fig. 1 A block diagram illustrating an example of a configuration of a machine learning device according to embodiment 1. Fig. 2 A flowchart describing an example of a machine learning process according to embodiment 1. Fig. 3 A diagram illustrating an example of an ID management table according to embodiment 1. Fig. 4 A flowchart describing an example of a machine learning process according to embodiment 1. Fig. 5 A diagram illustrating an example of a terminology management table according to embodiment 1. Fig. 6 A flowchart to describe an example of a machine learning process according to embodiment 1. Fig. 7 A diagram illustrating an example of a multi-purpose PSP task table according to embodiment 1. Fig. 8 A diagram illustrating an example of a multi-purpose PSP linking table according to embodiment 1. Fig. 9 A diagram illustrating an example of a link type according to embodiment 1. Fig. 10 A diagram illustrating a configuration example of a neural network. Fig. 11 A block diagram illustrating an example of a configuration of a PSP generating device according to embodiment 2. Fig. 12 A diagram showing an example of an input screen for an information input part according to embodiment 2. Fig. 13 A diagram showing an example of an output screen of an information output part according to embodiment 2. Fig. 14 A diagram showing an example of an output screen of the information output section according to embodiment 2. Fig. 15 A block diagram showing a hardware configuration of the machine learning device according to another modification example. Fig. 16 A block diagram showing a hardware configuration of the machine learning device according to another modification example. Description of embodiments<Ausführungsform 1 >

[0011] Fig. Figure 1 is a block diagram illustrating an example configuration of a machine learning device according to the present embodiment 1. The machine learning device in Fig. 1 comprises a standard management part 101, a multi-purpose PSP control part 105, a property management part 111, and a diagnostic management part 121. At least one of the standard management part 101, the multi-purpose PSP control part 105, the property management part 111, or the diagnostic management part 121 can be deployed on a single terminal, or they can be distributed across multiple terminals. The machine learning device in Fig. 1 can be deployed on a cloud server.

[0012] Standard Management Part 101 manages initial information regarding a WBS work package of a standard. Property Management Part 111 manages additional information regarding a WBS work package of a project that conforms to a standard. Diagnostic Management Part 121 manages additional information regarding an evaluation result of a process diagnostic result of a project that conforms to a standard. The project that conforms to the standard includes, for example, a project that reuses a standard. In the following description, the project that conforms to the standard is sometimes simply referred to as "the project." The initial information may contain the WBS work package itself, and the second information may contain the WBS work package of the project itself.

[0013] The multi-purpose WBS control part 105, based on the first, second, and third pieces of information managed in the standard management part 101 and the property management part 111, performs learning with respect to a WBS work package, thereby generating association information that links the first, second, and third pieces of information. This learning with respect to the WBS work package can also include learning of the WBS work package itself.

[0014] The association information generated by this learning process is used to create a multi-purpose PSP. According to the present embodiment 1, the association information comprises at least one of the following: an ID management table 106, a terminology management table 107, a multi-purpose PSP task table 108, or a multi-purpose PSP linking table 109. However, the association information is not limited to these tables and may, for example, include a table different from these.

[0015] Standard Administration Part 101 manages, as its first piece of information, for example, a Standard Information Table 102, a Standard Terminology Table 103, and a Standard Association Table 104. Standard Information Table 102 registers a WBS work package, which, based on information defined by the standard, comprises a technical process measure, a management process measure, and an auxiliary process measure. The meaning of a terminology defined by the standard is registered in Standard Terminology Table 103. The name of the standard and the name of the other standard associated with it are registered in Standard Association Table 104.

[0016] Property Management Unit 111 manages, as its second piece of information, for example, a Property Information Table 112, a Property Terminology Table 113, and a Property Association Table 114. Property Information Table 112 contains a WBS work package, including a technical process action, a management process action, and an auxiliary process action, all of which were executed in a previous project. The meaning of the terminology defined by the project is recorded in Property Terminology Table 113. A project name, generated by reusing the standard to which the project conforms, is recorded in Property Association Table 114.

[0017] The diagnosis management section 121 manages, as its third piece of information, a diagnosis information table 122, a diagnosis terminology table 123, and a diagnosis association table 124. Diagnosis information table 122 contains an assessment result, which includes a technical process measure, a management process measure, and an auxiliary process measure diagnosed in a previous project. The meaning of the terminology defined by the project is recorded in diagnosis terminology table 123. A project name, generated by reusing the standards that describe the project, is recorded in diagnosis association table 124.

[0018] The multi-purpose PSP control unit 105 comprises a receive unit 105a and a (self-)learning control unit 105b. The receive unit 105a receives the first piece of information from the standard management unit 101, the second piece of information from the property management unit 111, and the third piece of information from the diagnostic management unit 121. The receive unit 105a can obtain the first, second, and third pieces of information from a network.

[0019] The learning control element 105b has a machine learning function and performs learning based on the first, second, and third information regarding the PSP work package, thereby generating association information that associates the first, second, and third information. According to the present embodiment 1, the association information comprises at least one of the ID management table 106, the terminology management table 107, the multi-purpose PSP task table 108, or the multi-purpose PSP linking table 109.

[0020] The ID management table 106 is a table that records: an identifier that identifies a standard or a project, an identifier of a standard or a project with respect to the standard or the project, a name of a standard or a project, and version control. Specifically, the ID management table 106 includes a name of at least one of the standard or the project.

[0021] Terminology Management Table 107 is a table that records: an identifier that identifies one standard or project, an identifier that identifies the other standard or project that uses the same terminology as the standard or project, and a defined terminology. Specifically, Terminology Management Table 107 includes a terminology from at least one of the standard or project.

[0022] The multipurpose WBS task table 108 is a table that records: an identifier that identifies a WBS work package used for a multipurpose WBS, a standard identifier in which the WBS work package used for the multipurpose WBS is defined, and the contents of the WBS work package used for the multipurpose WBS. Specifically, the multipurpose WBS task table 108 contains the contents of the WBS work package.

[0023] The multipurpose WBS link table 109 is a table that records: a link type between WBS work packages used for a multipurpose WBS, an identifier of a WBS work package used for the multipurpose WBS of a link source, and an identifier of the WBS work package used for the multipurpose WBS of a link destination. The multipurpose WBS link table 109 contains tracking information that associates the WBS work packages.

[0024] The learning control unit 105b performs supervised learning using the standard information table 102 and the standard association table 104, thereby generating the ID management table 106. The learning control unit 105b performs reinforcement learning using the property information table 112 and the property association table 114, thereby reinforcing the ID management table 106. The learning control unit 105b performs reinforcement learning using the diagnostic information table 122 and the diagnostic association table 124, thereby reinforcing the ID management table 106.

[0025] The learning control unit 105b performs supervised learning using the standard terminology table 103 and the standard association table 104, thereby generating the terminology management table 107. The learning control unit 105b performs reinforcement learning using the property terminology table 113 and the property association table 114, thereby reinforcing the terminology management table 107. The learning control unit 105b performs reinforcement learning using the diagnostic terminology table 123 and the diagnostic association table 124, thereby reinforcing the terminology management table 107.

[0026] The learning control part 105b performs supervised learning using the ID management table 106, the terminology management table 107, and the content of the WBS work package of a standard recorded in the standard information table 102, thereby generating the general-purpose WBS task table 108. The learning control part 105b also performs reinforcement learning using the ID management table 106, the terminology management table 107, and the content of the WBS work package of a project recorded in the property information table 112, thereby reinforcing the general-purpose WBS task table 108. The learning control part 105b performs reinforcement learning using the ID management table 106, the terminology management table 107 and contents of a process diagnostic result of a project recorded in the diagnostic information table 122, thereby reinforcing the multi-purpose PSP task table 108.

[0027] The learning control part 105b performs supervised learning using the ID management table 106, the terminology management table 107, and correlated information from the WBS work package of a standard recorded in the standard information table 102, thereby generating the multipurpose WBS linking table 109. The learning control part 105b also performs reinforcement learning using the ID management table 106, the terminology management table 107, and correlated information from the WBS work package of a project recorded in the property information table 112, thereby reinforcing the multipurpose WBS linking table 109.The learning control part 105b performs reinforcement learning using the ID management table 106, the terminology management table 107, and the mapping information of a process diagnostic result of a project recorded in the diagnostic information table 122, thereby reinforcing the multipurpose PSP linking table 109.

[0028] If there is a shortage of multipurpose PSP information in the multipurpose PSP task table 108 and the multipurpose PSP linking table 109, missing information can be recorded manually. Supervised learning, for example, can be used instead of the reinforcement learning described above.

[0029] Fig. Figure 2 is a flowchart to describe an example of the machine learning process in the learning control section 105b with respect to the ID management table 106. Fig. Figure 3 shows an example of ID management table 106.

[0030] Although not shown in the diagrams, when the machine learning process is initiated with respect to the ID Management Table 106, the referent part 105a retrieves a name of a standard or project from the Standard Association Table 104, the Property Association Table 114, or the Diagnostic Association Table 124. The name retrieved by the referent part 105a is referred to as the "referenced name" and is described below.

[0031] In step S1, the learning control unit 105b determines whether a name matching the referenced name is registered in data 303 and data 304 in the ID management table 106. If it is determined that they match, the process of Fig. 2 is completed, and if it is determined that they do not match, the process proceeds to step S2.

[0032] In step S2, the learning control unit 105b registers the referenced name in data 303 and data 304. For example, a common name of the referenced name is registered in a master of data 303, and a version of the referenced name is registered in a slave of data 304.

[0033] In step S3, the learning control unit 105b generates an identifier to identify the referenced name and registers the identifier in data 301.

[0034] In step S4, the learning control unit 105b determines whether a previous standard or project associated with the standard or project specified by the referenced name is registered in the ID management table 106. For example, if the learning control unit 105b determines, based on the standard information table 102, that the previous standard branched into the standard specified by the referenced name, it determines that the standard specified by the referenced name is associated with the previous standard. Similarly, if the learning control unit 105b determines, based on the standard information table 102, the property information table 112, and the diagnostic information table 122, that the previous project branched into the project specified by the referenced name, it determines that the project specified by the referenced name is associated with the previous project.

[0035] If it is determined that the standard or project specified by the referenced name is associated with the previous standard or project, the process proceeds to step S5. If it is determined that the standard or project specified by the referenced name is not associated with the previous standard or project, the process proceeds to step S5. Fig. 2 to its end.

[0036] In step S5, the learning control unit 105b registers the identifier of the previous standard or project in the data 302 of the referenced name, which is associated with the previous standard or project. The process then continues in Fig. 2 completed. The learning control unit 105b performs a learning process, which, for example, makes the determination in step S4 suitable, and the ID management table 106 in Fig. 3 is made suitable.

[0037] Fig. 4 is in flowchart to describe an example of the process for machine learning in the learning control section 105b in relation to the terminology management table 107. Fig. Figure 5 is an example of the terminology management table 107.

[0038] The machine learning process in the learning control part 105b with respect to the terminology management table 107 is applied to the standard or project in which the identifier, for example, in step S3 of Fig. 2 was generated. Although not shown in the diagrams, when the machine learning process begins, the referent part 105a retrieves terminology used, for example, for the standard or project in which the identifier was generated in step S3, from the standard terminology table 103, the property terminology table 113, or the diagnostic terminology table 123. This terminology, used, for example, for the standard or project in which the identifier was generated in step S3, is referred to as "referenced terminology" and is described below.

[0039] In step S11, the learning control unit 105b determines whether a terminology whose meaning matches that of the referenced terminology is registered in data 503 in the terminology management table 107. If it is determined that they match, the process continues to step S12, and if it is determined that they do not match, the process continues to step S13.

[0040] In step S12, the learning control part 105b retrieves from data 501 in the terminology management table 107 an identifier of a previous standard or project which uses a terminology whose meaning is determined to match the terminology obtained in step S11.

[0041] In step S13, the learning control unit 105b registers the input from step S3. Fig. 2 generated identifiers in data 501. When the learning control unit 105b executes the process in step S12, the learning control unit 105b registers the identifier obtained in step S12 in data 502.

[0042] In step S14, the learning control unit 105b registers the related terminology in data 503. The process then continues in Fig. 4 completed. The learning control unit 105b performs a learning process, thus, for example, the determination in step S11 is adjusted accordingly, and the terminology management table 107 in Fig. 5 will be adjusted to fit.

[0043] Fig. Figure 6 is a flowchart to describe an example of a machine learning process in the learning control part 105b in relation to the multi-purpose PSP task table 108 and the multi-purpose PSP linking table 109. Fig. Figure 7 presents an example of the multi-purpose PSP task table 108, and Fig. Figure 8 shows an example of the multi-purpose PSP linking table 109.

[0044] The machine learning process in the learning control part 105b with respect to the multi-purpose WBS task table 108 and the multi-purpose WBS linking table 109 is applied to the standard or project in which the identifier, for example, in step S3 of Fig. 2 is generated. Although not shown in the diagrams, when the machine learning process is started, the reference part 105a references the WBS work package used for the standard or project in which the identifier is generated in step S3 from, for example, the standard information table 102, the property information table 112, or the diagnostic information table 122. The WBS work package referenced in the reference part 105a is referred to as the "referenced work package" and is described below.

[0045] In step S21, the learning control unit 105b determines whether the WBS work package is registered as a link target of the related package in the multipurpose WBS task table 108 and the multipurpose WBS link table 109. If it is determined that the WBS work package is registered as the link target of the related package, the process continues with step S23; if it is determined that the WBS work package is not registered as the link target of the related package, the process continues with step S22.

[0046] In step S22, the learning control unit 105b determines whether the WBS work package that has the same action as the related package is registered in the multipurpose WBS task table 108 and the multipurpose WBS linking table 109. If it is determined that the WBS work package that has the same action as the related package is registered, the process continues to step S23; if it is determined that the WBS work package that has the same action as the related package is not registered, the process is in Fig. 6 completed.

[0047] In step S23, the learning control unit 105b generates a task ID and registers the task ID in data 701, and registers the one from step S3 in Fig. 2 generated identifiers in data 702.

[0048] In step S24, the learning control unit 105b registers contents of the received package in data 703.

[0049] In step S25, the learning control unit 105b registers a link type in data 801 based on the task information of the WBS work package. The learning control unit 105b registers the task ID of the WBS work package as the link target in data 701 in data 802, and registers the task ID of the WBS work package as the link source in data 701 in data 803. The process then continues in Fig. 6 completed. The learning control unit 105b performs a learning process, thus, for example, making the registration in step S24 suitable, and the multi-purpose PSP task table 108 in Fig. 7 and the multi-purpose PSP linking table 109 in Fig. 8 will be made suitable.

[0050] Fig. Figure 9 is a diagram to describe an example of the link type in Fig. 8. A state where the link type is 0 indicates that the link target is a next process for the link source. A state where the link type is 1 indicates that the link target and the link source are equivalent processes. A state where the link type is 2 indicates that the link target is part of a process of the link source.

[0051] The process in steps S1 to S5 in Fig. 2, the process in step S11 to step S14 in Fig. 4, and the process in step S21 to step S25 in Fig. The six steps are executed repeatedly until the learning control unit 105b has sufficiently learned a multi-purpose WBS work package. The standard or project WBS work package is preferably assembled (mounted) in patterns of several link types.

[0052] Fig. Figure 10 is a diagram illustrating a configuration example of a neural network. A design review success prediction can be learned in accordance with a neural network model, for example, during machine learning of the learning control section (Figure 105b). As shown in Fig. As shown in Figure 10, the neural network comprises an input layer, which includes one neuron x1, x2, x3, ... or x1; an intermediate layer (hidden layer), which includes m numbers of neurons y1, y2, y3, ... and ym; and an output layer, which includes n numbers of neurons z1, z2, z3, ... and zn. The intermediate layer, which comprises only one layer, is shown in the configuration example of the neural network in Figure 10. Fig. 10 layers are provided, but an intermediate layer comprising two or more layers may also be provided. A general-purpose computer or processor can be used for the machine learning (neural network) of the learning control part 105b, and the process can be carried out at a higher speed by, for example, using a large-scale PC cluster for machine learning.

[0053] The neural network can learn the ID management table 106, the terminology management table 107, the multi-purpose WBS task table 108, or the multi-purpose WBS linking table 109, which are associated with the multi-purpose WBS work package. This is essentially learning the association information. The neural network can learn the association information that connects the WBS work package of the standard and the project, for example, through supervised learning, in accordance with a test result based on a combination of the association information and the determining data, which indicate whether the association information is correct or incorrect.In this context, "supervised learning" refers to a learning process in which a learning device is fed many groups of a specific input and output (label), thereby learning characteristics contained in the test result, and which is capable of inductively obtaining a model for estimating the result from the input, in other words, a relationship between the input and the result.

[0054] The machine learning of the learning control unit 105b is not limited to "supervised learning." For example, it is possible to use only correct association information—namely, only the ID management table 106, the terminology management table 107, the multipurpose PSP task table 108, or the multipurpose PSP linking table 109, which typically meet a testing standard—to accumulate it using the neural network, and then learn the multipurpose work package through so-called "unsupervised learning." For example, if a measure of the success of the test result of the ID management table 106, the terminology management table 107, the multipurpose PSP task table 108, or the multipurpose PSP linking table 109 is extremely high, an "unsupervised learning" method is considered effective.

[0055] Unsupervised learning refers to a learning process that learns to compress, classify, and train input data by simply providing many pieces of input data without corresponding monitoring data. Based on this learning, clustering for classification can be performed, resulting in a similar test outcome. If a distribution of output is performed to optimize the data by providing a specific standard that uses this output, a prediction of the outcome can be achieved.

[0056] The multi-purpose PSP work package can be learned through a process known as "semi-supervised learning," which lies between "unsupervised learning" and "supervised learning." Semi-supervised learning is efficient when the data set includes both input and output data in some cases, and only input data in others. <Schlussfolgerung der Ausführungsform 1 >

[0057] According to the present embodiment 1, learning with respect to the PSP work package is performed based not only on the first piece of information relating to the PSP work package of the standard, but also on the second piece of information relating to the PSP work package of the project and the third piece of information relating to the process diagnostic result of the project, in order to generate the association information. With such a configuration, missing content of the first piece of information can be compensated for by the second piece of information or the third piece of information, so that the learning with respect to the PSP work package can be carried out appropriately, and as a result, the association information can be generated that is trained to produce the appropriate multi-purpose PSP.Even if one factor contributing to the success of the test result is complex, and it is difficult to define the project's PSP work package in advance, a predictive model can be generated that is capable of producing a highly accurate PSP.

[0058] The learning control unit 105b can learn the ID management table 106, the terminology management table 107, the multipurpose WBS task table 108, or the multipurpose WBS linking table 109, as learning with respect to the WBS work package, as described above. The learning control unit 105b can perform learning with respect to the WBS work package based on the first, second, and third pieces of information obtained from the multiple standard management units 101, the multiple property management units 111, and the multiple diagnostic management units 121 that are activated in the same development field.The learning control unit 105b can perform learning regarding the WSP work package based on the first, second, and third pieces of information obtained from the multiple standard management units 101, the multiple property management units 111, and the multiple diagnostic management units 121, which are activated independently in different development areas. Any of the standard management units 101, property management units 111, and diagnostic management units 121 can be added to or removed from a target from which the first, second, and third pieces of information are obtained, either during the process.

[0059] The test result can be shared (used jointly) across multiple multi-purpose PSP control parts 105. As a first example, a model of the same neural network can be shared across multiple multi-purpose PSP control parts 105. In particular, each network weight coefficient can be transmitted to reflect a difference between multiple multi-purpose PSP control parts 105 using a communication means. As a second example, a machine learning weight can be shared by sharing the test result of the neural network's input and output. As a third example, a state can be shared so that a similar model is used by providing access to a database that is pre-prepared to load an even more accurate model of the neural network.The sharing of the test result of the multiple multi-purpose PSP control parts 105 is not limited to the first to third examples. <Ausführungsform 2 >

[0060] Fig. Figure 1 is a block diagram illustrating an example configuration of a PSP generation device according to the present embodiment 2. In the following description, the PSP generation device according to the present embodiment 2 is a specific PSP generation device that generates a specific PSP work package of the project. Reference numerals identical or similar to those described above are assigned to identical or similar forming elements according to the present embodiment 2, and the different forming elements are described in particular below.

[0061] As in Fig. As shown in Figure 11, the PSP generation device according to the present embodiment 2 has one of the machine learning devices in Fig. 1, which has been described in embodiment 1, similar configuration except that an information input part 1101, a specific PSP control part 1102, and an information output part 1103 are added.

[0062] The information input part 1101, for example, receives a current project or a standard corresponding to the current project as input information. The specific WBS control part 1102 retrieves association information from the general-purpose WBS work package, which was generated, for example, by machine learning from the general-purpose WBS control part 105, based on the input information. The specific WBS control part 1102 generates a specific WBS work package or a specific WBS for the current project, which corresponds to a level of development or a quantity of change in the project, based on the input information and the association information. The information output part 1103 outputs a specific WBS work package or a specific WBS generated in the specific WBS control part 1102.

[0063] In the present embodiment 2, the project in which the second piece of information used in machine learning is obtained can be an earlier project, as in embodiment 1, or a current new project, which is specified in the information input part 1101.

[0064] In a configuration for generating the specific PSP work package, as in the present embodiment 2, the neural network model can be used. For example, the output layer can calculate a compatibility between a test result input to the input layer of the neural network and information indicating the success or failure of progress in the general-purpose PSP work package or the PSP work package that is to be valid. The output layer can then output the information indicating the success or failure of progress in the general-purpose PSP work package or the PSP work package that is to be valid.

[0065] Fig. Figure 12 is a diagram illustrating an example of an input screen for the information input section 1101. A main WBS selection screen 1201 displays a project name retrieved from the ID management table 106 of the multi-purpose WBS control section 105 to allow selection of the project name. The project name can be displayed as a list of all project names, or it can be shown in a drop-down menu.

[0066] An integration target WBS selection screen 1202 displays a standard name retrieved from the ID management table 106 of the multi-purpose WBS control part 105 to allow selection of its name. As with the standard name display, a list of names from all projects can be shown, or the project name can be displayed in a drop-down menu. When the project is displayed on the main WBS selection screen 1201, the integration target WBS selection screen 1202 can display the standard associated with the project (for example, the standard to which the project conforms), pre-selected with a selection marker. The integration target WBS selection screen 1202 can add or remove the standard associated with the project as needed.

[0067] Fig. Figure 13 is a diagram illustrating an example output screen of the information output section 1103. The specific WBS control section 1102 retrieves the standard associated with the project selected in the information input section 1101 from the multipurpose WBS work package, which is registered in the multipurpose WBS task table 108 and the multipurpose WBS link table 109 in the multipurpose WBS control section 105. The specific WBS control section 1102 then generates the specific WBS based on the input information and the retrieved standard, and the specific WBS output screen 1301 displays the generated specific WBS.

[0068] If the project includes a work package that was newly added to the project, the specific work package output screen 1301 can highlight and display the added work package. If multiple work packages were newly added to the project, the information output part 1103 can be configured to select one work package from among the multiple work packages.

[0069] The specific PSP output screen 1302 is a button for starting an output of the project of the selected PSP work package. The output format can be text or a file format corresponding to another application tool.

[0070] In the present embodiment 2, the information input part 1101 and the information output part 1103 are selectively displayed on one screen, and the specific PSP output screen 1303 is a button to access the display in Fig. 12 to return. The specific PSP output screen 1304 is a button to complete an operation of the PSP generating device.

[0071] Fig. Figure 14 is a diagram that illustrates an example of the output screen of the information output section 1103, and in particular, it is an example of the selection screen in a case involving multiple WBS work packages that have been newly added to the project. A main WBS definition screen 1401 displays the name of a project that was selected, for example, in the main WBS selection screen 1201 of the information input section 1101.

[0072] For example, a WBS work package selection screen 1402 displays a list of the names of the multiple WBS work packages that fall under the selected project, allowing their names to be selected. If the project is displayed on the main WBS determination screen 1401, the WBS work package selection screen 1402 can display the multiple WBS work packages that have been newly added to the project and were previously marked with a selection indicator.

[0073] There is a case where the WBS work package newly added to the project is not included in an actual operation. In such a case, the WBS work package selection screen 1402 can display the WBS work package from another standard and project that is closely related to the standard associated with the project. A work package used for the project can be selected from the multiple WBS work packages in the WBS work package selection screen 1402. <Schlussfolgerung zu Ausführungsform 2>

[0074] In the present embodiment 2, the PSP of the current project is generated based on the information of the current project and the association information. With such a configuration, a suitable specific PSP can be generated. <Weiteres Modifikationsbeispiel >

[0075] The receiver part 105a and the learning control part 105b as described above Fig. 1 are referred to below as "the recipient part 105a etc." The recipient part 105a etc. is represented by a Fig. The processing circuit 81 shown in Figure 15 achieves this. The processing circuit 81 comprises the receiver part 105a, which receives the first piece of information regarding the standard's PSP work package, the second piece of information regarding the project's PSP work package that conforms to the standard, and the third piece of information regarding the project's process diagnostic result that conforms to the standard. It also comprises the learning control part 105b, which performs learning regarding the PSP work package based on the first, second, and third pieces of information, thereby generating the association information that links the first, second, and third pieces of information. Dedicated hardware can be used for the processing circuit 81, or a processor executing a program stored in memory can also be used.Examples of processors include a central processing unit, a processing device, an arithmetic device, a microprocessor, a microcomputer, or a digital signal processor (DSP).

[0076] If the processing circuit 81 is dedicated hardware, then, for example, a single circuit, a complex circuit, a programmed processor, a parallel programmed processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination thereof fall under the processing circuit 81. Each function of the receiver part 105a, etc., can be achieved by circuits into which the processing circuit is subdivided, or each function can also be achieved jointly by a processing circuit.

[0077] If the processing circuit 81 is the processor, the functions of the receiver part 105a, etc., are achieved through a combination with software, etc. Software, firmware, or software and firmware, for example, fall under the category of software, etc. The software, etc., is described as a program and stored in memory. As in Fig.As shown in Figure 16, a processor 82 applied to the processing circuit 81 reads and executes a program stored in the memory 83, thereby achieving the function of each unit. The machine learning device includes the memory 83 for storing a program that consists of a step of obtaining the first piece of information regarding the PSP work package of the standard, the second piece of information regarding the PSP work package of the project corresponding to the standard, and the third piece of information regarding the process diagnostic result of the project corresponding to the standard; and a step of performing a learning process with respect to the PSP work package based on the first piece of information, the second piece of information, and the third piece of information, thereby generating the association information that associates the first piece of information, the second piece of information, and the third piece of information.In other words, this program is designed to cause a computer to execute a procedure or process of the recipient part 105a, etc. The memory 83 can be a non-volatile or volatile semiconductor memory, such as random access memory (RAM), read-only memory (ROM), flash memory, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), a hard disk drive (HDD), a magnetic disk, a floppy disk, an optical disk, a compact disk, a minidisk, a digital versatile disk (DVD), or a drive device for these, or any storage medium that may be used in the future.

[0078] The configuration described above is such that each function of the receiver unit 105a, etc., is achieved, for example, by hardware and software. However, the configuration is not limited to this; it is applicable, for example, to a configuration in which part of the receiver unit 105a, etc., is achieved by dedicated hardware and another part by software. For example, the function of the receiver unit 105a can be achieved by the processing circuit 81 as dedicated hardware, an interface, and a receiver, and the function of the other units can be achieved, for example, by the processing circuit 81 as the processor 82, which reads and executes the program stored in the memory 83.

[0079] As described above, the processing circuit 81 can, for example, achieve any of the functions described above by hardware, software, or a combination thereof.

[0080] Each embodiment and each modification example can be combined in any way desired, or each embodiment and each modification example can be suitably varied or omitted.

[0081] The foregoing description is in every respect explanatory and does not limit the invention. It should therefore be understood that numerous modification examples can be devised. Explanation of reference symbols 105a Recipient part, 105b (Self-)learning control unit. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] JP 2018-5802

[0004] JP 2017-188030

[0004]

Claims

[1] Apparatus for machine learning, comprising: a receiving part which receives first information regarding a WBS work package of a standard, second information regarding a WBS work package of a project conforming to the standard, and third information regarding a process diagnosis result of a project conforming to the standard; and a learning control part that performs learning on a PSP work package based on the first information, the second information and the third information, thereby generating association information that associates the first information, the second information and the third information. [2] The machine learning apparatus according to claim 1, wherein the association information includes contents of the PSP work package. [3] The machine learning apparatus according to claim 1 or 2, wherein the association information comprises tracking information associating the PSP work packages. [4] The machine learning apparatus according to any one of claims 1 to 3, wherein the association information comprises at least one terminology and at least one name of one of the standard or the project. [5] The machine learning apparatus according to any one of claims 1 to 4, wherein the association information includes a terminology and / or a name of the diagnostic information. [6] The machine learning apparatus according to any one of claims 1 to 5, wherein the obtaining part obtains the first information and the second information from a network. [7] The machine learning apparatus according to any one of claims 1 to 6, wherein the project in which the second information is obtained comprises a current project. [8] The machine learning apparatus according to any one of claims 1 to 7, wherein the machine learning apparatus is provided on a cloud server. [9] PSP generating device, comprising the machine learning device according to one of claims 1 to 6, where a WBS of a current project is created based on information of the current project and the association information. [10] A method for machine learning, comprising: Obtaining first information regarding a WBS work package of a standard, second information regarding a WBS work package of a project conforming to the standard, and third information regarding a process diagnosis result of a project conforming to the standard; and Performing learning on a PSP work package based on the first information, the second information, and the third information, thereby generating association information associating the first information, the second information, and the third information.

Images