JP2019159981A - Information processing device, work instruction evaluation method, and program - Google Patents

Abstract

To assist in creating appropriate work instructions that correspond to the proficiency of a worker without incurring excess quality.SOLUTION: The present invention comprises: a model construction unit for classifying incident data including an instruction sheet for repair and maintenance work and past work records in accordance with the proficiency of a worker in charge and causing it learned by machine learning, and thereby constructing a machine learning model capable of outputting an evaluation result per worker proficiency regarding one instruction sheet; an instruction evaluation unit for evaluating a new instruction sheet using the machine learning model and outputting an evaluation result per worker proficiency; and a quality determination unit for determining whether or not an evaluation result among evaluation results per worker proficiency that corresponds to the proficiency of an assigned worker satisfies a prescribed quality standard stipulated with respect to proficiency.SELECTED DRAWING: Figure 2

Description

  The present disclosure relates to a technique for supporting creation of a work instruction.

  Conventionally, a technique for searching for a fault report similar to fault information from past response history, a technique for automatically creating a farm work report by converting work data by a predetermined method, and the like have been proposed. (See Patent Documents 1 and 2).

JP-A-2006-051447 International Publication No. 2014/050674

  Conventionally, in repair and maintenance work, a person in charge of work instruction has determined the cause of a phenomenon based on a contact from a customer received at a call center, and has created a work instruction document that describes how to deal with the phenomenon. Then, a customer engineer (hereinafter referred to as “worker”) working on site performs the work in accordance with this work instruction. Here, since the work instruction is created manually, the specific expression differs depending on the creator even if it is a method for dealing with the same phenomenon. And if the expression in the work instruction is an expression that is difficult to understand for the on-site worker and the contents of the instruction cannot be understood, the work at the site may be delayed, and the operator will not be able to read without getting the guidelines. If it is an expression that causes important descriptions to be skipped, trouble may occur on site. For this reason, it is expected to reduce the number of work instructions that are difficult for the worker to understand by mechanically checking the quality of the work instructions before giving them to the worker.

  However, what type of work instruction is described as a high quality work instruction depends on the skill level of the worker. For example, an exhaustive and detailed work instruction sheet is preferable for a worker with a low level of skill, but a simple work instruction sheet focused on the main points is preferable for a worker with a high level of skill. For this reason, if the quality standard of work instructions is a single standard, it is recommended that the content of excessive quality be matched to the worker with the lowest level of skill. However, it may become a wasteful and long work instruction for a highly skilled worker, which may be redundant and reduce work efficiency.

  In view of the problems described above, an object of the present disclosure is to support creation of an appropriate work instruction according to the skill level of an operator without causing excessive quality.

  In an example of the present disclosure, machine learning is performed by classifying incident data including instructions for repair and maintenance work and work results associated with the instructions according to the skill level of the worker in charge of the work related to the instructions. To build a machine learning model that can output an evaluation result for each worker's skill level for one instruction, and input of a new instruction that is newly created in response to a request for repair and maintenance An instruction sheet receiving means for receiving the instruction sheet, an instruction sheet evaluating means for outputting an evaluation result for each worker skill level of the instruction sheet by evaluating the instruction sheet using the machine learning model, and the new instruction sheet Skill level acquisition means for acquiring the skill level of the worker assigned to the work related to the work, and among the evaluation results for each worker skill level by the instruction evaluation means, the skill level of the assigned worker Response to the evaluation result, an information processing apparatus and a determining quality judging means to judge whether to satisfy a predetermined quality standards set for the proficiency.

  The present disclosure can be understood as an information processing apparatus, a system, a method executed by a computer, or a program executed by a computer. The present disclosure can also be understood as a program recorded on a recording medium readable by a computer, other devices, machines, or the like. Here, a computer-readable recording medium refers to a recording medium that stores information such as data and programs by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer or the like. Say.

  According to the present disclosure, it is possible to support the creation of an appropriate work instruction according to the skill level of an operator without falling into excessive quality.

It is the schematic which shows the hardware constitutions of the system which concerns on embodiment. It is a figure which shows the outline of a function structure of the information processing apparatus which concerns on embodiment. It is a figure which shows the outline | summary of the incident data used in embodiment. It is a figure which shows the recommendation description item list | wrist in an instruction sheet set for every worker rank used in embodiment. It is a flowchart which shows the outline | summary of the flow of the data update process which concerns on embodiment. It is a flowchart which shows the outline | summary of the flow of the instruction creation preparation process which concerns on embodiment. It is a flowchart which shows the outline | summary of the flow of the data storage process which concerns on embodiment.

  Hereinafter, embodiments of an information processing apparatus, a work instruction evaluation method, and a program according to the present disclosure will be described with reference to the drawings. However, the embodiment described below exemplifies the embodiment, and the information processing apparatus, the work instruction evaluation method, and the program according to the present disclosure are not limited to the specific configurations described below. . In implementation, a specific configuration according to the embodiment is appropriately adopted, and various improvements and modifications may be performed.

  In the description of the embodiment, the information processing apparatus, the work instruction evaluation method, and the program according to the present disclosure are implemented in a system operated in a call center that dispatches workers to the site in response to contact from a customer. Will be described. Note that the information processing apparatus, the work instruction evaluation method, and the program according to the present disclosure can be widely used for techniques for supporting the creation of a work instruction, and the application target of the present disclosure is shown in the embodiment. It is not limited to the example.

<System configuration>
FIG. 1 is a schematic diagram illustrating a hardware configuration of a system 1 according to the present embodiment. The system 1 according to the present embodiment includes a plurality of information processing terminals 90 (hereinafter referred to as “nodes 90”) and a server 50 that is communicably connected to the nodes 90 via a network.

  In the system 1 according to the present embodiment, the server 50 connected from the node 90 is connected at a remote location via the Internet or a wide area network, and is provided by, for example, an ASP (Application Service Provider). 50 need not necessarily be connected to a remote location. For example, the server 50 may be connected to a local network where the node 90 exists.

  The server 50 includes a CPU (Central Processing Unit) 11, a RAM (Random Access Memory) 13, a ROM (Read Only Memory) 12, an EEPROM (Electrically Erasable Memory and Programmable Read Only Memory) 14, and the like. , A computer including a communication unit such as a NIC (Network Interface Card) 15.

  FIG. 2 is a diagram illustrating an outline of a functional configuration of the server 50 according to the present embodiment. The server 50 reads out the program recorded in the storage device 14 to the RAM 13 and executes it by the CPU 11 to control each hardware provided in the server 50, so that the skill level determination unit 21, the incident Data storage unit 22, model construction unit 23, instruction sheet reception unit 24, instruction sheet evaluation unit 25, presentation unit 26, skill level acquisition unit 27, quality determination unit 28, correction instruction unit 29, decomposition unit 30, results reception unit 31 And functions as an information processing apparatus including the creator evaluation unit 32. In this embodiment and other embodiments described later, each function provided in the server 50 is executed by the CPU 11 that is a general-purpose processor. However, some or all of these functions are executed by one or more dedicated processors. May be executed.

  The skill level determination unit 21 reads at least one of the years of experience for each worker, the qualification of ownership, the number of corresponding cases of the device, and the number of troubles from the storage device 14, and based on the read information, the worker The skill level (hereinafter, “rank”) is determined. The storage device 14 holds the rank assigned to each combination of the worker ID and the target device, which is determined by the skill level determination unit 21.

In this embodiment, the rank of the worker for a certain target device is determined by the skill level determination unit 21 by assigning points of experience years, ownership qualifications, the number of experience cases corresponding to the corresponding device, the number of trouble occurrences, etc. This is done by assigning any rank from A to C. For example, the skill level determination unit 21 calculates points for each worker based on the following items, and workers who are 5 points or more are ranked A, workers who are 3 points or more are rank B, workers who are less than 3 points Is C rank.
・ Experience years: 1 year experience is 1 point (rounded down)
・ Number of experiences with the device: +1 point for every 5 cases ・ Quality of ownership: +4 points if you have qualifications according to the work ・ Number of troubles: If the operator has caused trouble in the past , -1 point

  The incident data accumulating unit 22 classifies incident data including instructions for repair and maintenance work and work results associated with the instructions according to the target equipment for repair and maintenance contents, and a storage device. 14 accumulate.

  FIG. 3 is a diagram showing an outline of incident data used in the present embodiment. The work results include initial solution availability, trouble occurrence, inquiry occurrence, work time, worker evaluation, worker rank, and the like. Here, whether or not the initial solution is possible is information indicating whether or not the problem has been solved by the first response as a result of performing the repair and maintenance work according to the instruction. The presence / absence of trouble is information indicating whether or not some trouble has occurred during repair and maintenance work in accordance with the instruction. The presence / absence of an inquiry is information indicating whether or not an inquiry has occurred to the call center from an operator during repair and maintenance work according to the instruction. The work time is information indicating the time required for repair and maintenance work according to the instruction. The worker evaluation is information indicating the evaluation of the instruction by the worker who is in charge of the repair and maintenance work related to the instruction. The worker rank is information indicating the rank at the time of evaluation of the worker who has evaluated the instruction.

  The model construction unit 23 is responsible for the incident data stored in the incident data storage unit 22 for the equipment related to repair maintenance, the contents of repair maintenance (for example, parts to be replaced or cleaned), and the work related to the instructions. By classifying according to the rank of the worker and machine learning, a machine learning model capable of outputting an evaluation result for each worker rank with respect to one instruction sheet is constructed. At this time, the model construction unit 23 changes the description items to be machine learning target among the plurality of description items in the instruction sheet included in the incident data for each worker rank. Build a machine learning model that can output each evaluation result.

  Note that the number of machine learning models to be constructed is not limited when constructing a machine learning model that can output an evaluation result for each worker rank. An evaluation result for each worker rank may be output by a single machine learning model, or for example, a plurality of machine learning models different for each rank may be constructed.

  FIG. 4 is a diagram showing a list of recommended description items in the instruction sheet set for each worker rank used in the present embodiment. The list is a list for determining the items required for the instruction for each worker rank. For example, the list is unnecessary for the worker of rank A, but for the worker of rank C. Information that can discriminate necessary explanation items and the like is shown, and is used as a guideline for determining the necessity of description in the instruction sheet according to the rank of the assigned worker. By referring to this list, the model construction unit 23 changes the description items to be machine learning target among the plurality of description items in the instructions included in the incident data for each worker rank. Build a machine learning model that can output the evaluation results for each rank.

The evaluation criteria for learning data obtained from incident data during machine learning by the model construction unit 23 are as follows.
[Highly rated learning data]
・ Instructions that solved the problem at the site (instructions that are “Yes” or not)
・ Instructions for which the work time was shorter than the work of the same content ・ Instructions with a high overall evaluation from the workers of the corresponding rank ・ Regarding the rank of the description item set as the machine learning target for the worker rank Instructions with high individual evaluations by workers [learning data for low evaluations]
・ Instructions that could not solve the problem on site (Instructions for which “No” is the first resolution possibility)
・ Instructions in which work trouble occurred at the site (instructions that indicate whether or not trouble occurred)
・ Instructions for inquiries from callers to the call center (instructions for which the existence of inquiries is “Yes”)
・ Instructions that take longer to work than work of the same content ・ Instructions with a low overall evaluation from workers of the corresponding rank ・ Applicable ranks of the items set as machine learning targets for worker ranks Instructions with low individual evaluation by workers

  The instruction sheet reception unit 24 receives an input of a work instruction sheet (hereinafter referred to as “new instruction sheet”) newly created by a work instruction person (creator) in response to a request for repair and maintenance.

  The instruction sheet evaluation unit 25 evaluates the instruction sheet using the machine learning model, and outputs an evaluation result for each worker rank for the instruction sheet. That is, in the present embodiment, the instruction sheet evaluation unit 25 evaluates the instruction sheet by using the machine learning model constructed by the model construction unit 23 and capable of outputting the evaluation result for each worker rank, so that the rank A The evaluation results based on the worker criteria, the evaluation results based on the worker criteria of rank B, and the evaluation results based on the worker criteria of rank C are output. The instruction evaluation unit 25 not only evaluates a new instruction that is newly created and input, but also re-evaluates the new instruction that has been corrected in response to the correction instruction.

  The presentation unit 26 receives the evaluation result from the instruction evaluation unit 25 and presents to the user (assignee) information that can determine whether or not a predetermined quality standard defined for each rank of the worker is satisfied. To do. Specifically, the presentation unit 26 outputs the evaluation result based on the rank A worker criterion, the evaluation result based on the rank B worker criterion, and the rank C worker criterion output by the instruction sheet evaluation unit 25. These evaluation results may be presented to the user as they are, or the quality determination unit 28 determines whether or not these evaluation results satisfy a predetermined quality standard in each rank. A: determination result such as pass, rank B: fail, rank C: fail) may be presented to the user. By presenting such information, the user (assignee) assigns a worker with a rank that satisfies the quality standards when assigning the worker, and omits the trouble of correcting the instructions. I can do it.

  The skill level acquisition unit 27 acquires the rank of the worker related to the new instruction, which is assigned by the user who confirmed the content presented by the presentation unit 26.

  The quality determination unit 28 determines whether the evaluation result corresponding to the rank of the assigned worker among the evaluation results for each worker rank by the instruction evaluation unit 25 satisfies a predetermined quality standard defined for the rank. Determine whether or not. For example, the quality determination unit 28 determines whether or not the quality standard is satisfied by determining whether or not the evaluation result is a predetermined threshold value (for example, 70 points) or more. Specifically, when the evaluation result is an evaluation result such as rank A: 73 points, rank B: 50 points, rank C: 40 points, the result of the quality determination is rank A: pass, rank B: Fail, Rank C: Fail. The quality determination unit 28 satisfies the predetermined quality standard not only for the evaluation result of the newly created and entered new instruction but also for the evaluation result of the new instruction that has been modified in response to the modification instruction. Determine whether or not.

  The correction instruction unit 29 instructs the correction of the new instruction when the quality determination unit 28 determines that the evaluation result does not satisfy the quality standard. By doing in this way, until the quality standard required for the assigned worker is satisfied, the correction / re-evaluation of the instruction is repeated, and the quality of the instruction is guaranteed. In the case of the correction instruction, the correction instruction unit 29 may present the instruction sheet extracted from the incident data and highly evaluated by the operator together with the correction instruction for the new instruction sheet to the user. For example, in the case where the description items described in the instruction form are not excessive or insufficient with respect to the recommended description items shown in the recommended description item list, but the quality does not satisfy the standard, the correction instruction unit 29 Instructions for applicable equipment, parts, and worker rank that have met quality in the past are obtained from the incident data and presented to the user to encourage rewriting.

  The disassembling unit 30 disassembles the instructions text into a plurality of sections. In the present embodiment, the worker evaluation is performed for each predetermined section included in the instruction sheet (in this embodiment, for each sentence). In order to acquire such an operator evaluation, the disassembling unit 30 mechanically disassembles the contents written in the instruction sheet included in the incident data into sentences one by one using clues or periods as clues.

  The result receiving unit 31 receives an input of work results including evaluation of the new instruction by the worker after the assigned worker finishes the work in accordance with the new instruction. The evaluation of the instruction sheet accepted here includes comprehensive evaluation of work instructions and individual evaluation designated by the worker for each section disassembled by the disassembling unit 30. Specifically, for each sentence disassembled by the disassembling unit 30, an evaluation by an on-site worker is input. The evaluation values to be entered include, for example, “good”, “bad”, “none (no entry)”, “necessary information”, “unnecessary information”, “insufficient information”, etc. Good. The input individual evaluation is related to each section to be evaluated.

  The work results including the comprehensive evaluation and the individual evaluation received by the result receiving unit 31 are associated with the instruction corresponding to the work result and the rank of the assigned worker, and incident data as new incident data. Accumulated by the accumulating unit 22. Then, the model construction unit 23 periodically updates the machine learning model for evaluating the instruction based on the accumulated incident data including new incident data.

  The creator evaluation unit 32 associates the instruction sheet with the creator of the instruction sheet, and outputs at least one of the evaluation and the work performance related to the instruction sheet as the evaluation index of the creator.

<Process flow>
Next, the flow of processing executed by the information processing apparatus 1 according to this embodiment will be described. Note that the specific contents and processing order of the processing described below are examples for implementing the present disclosure. Specific processing contents and processing order may be appropriately selected according to the embodiment of the present disclosure.

  FIG. 5 is a flowchart showing an outline of the flow of data update processing according to the present embodiment. Whether the processing shown in this flowchart is periodically executed at a preset timing in order to set / update the worker rank and to build / update the machine learning model for instruction evaluation. Or at a timing instructed by the administrator.

  In step S101, the rank of the worker is determined. The skill level determination unit 21 reads the years of experience, the qualification of ownership, the number of cases corresponding to the device, and the number of trouble occurrences from the storage device 14, and based on these information, the rank of the worker (this embodiment) Then, any one of A to C) is determined. The specific processing content for determining the rank is as described above in the description of the skill level determination unit 21. Thereafter, the process proceeds to step S102.

  In step S102 and step S103, a machine learning model is constructed. The model construction unit 23 reads out the incident data accumulated by the incident data accumulation unit 22, and for each combination of the target device for repair and maintenance, the contents of repair and maintenance, and the rank of the worker in charge of the work related to the instruction. Grouping is performed (step S102). Then, the model construction unit 23 learns the instruction and work results (first resolution possible, trouble occurrence, inquiry occurrence, work time, worker evaluation, worker rank, etc.) included in the grouped incident data. Then, a machine learning model that can output an evaluation result for each worker rank for one instruction is constructed. Specific processing contents of the model construction are as described above in the description of the model construction unit 23. Thereafter, the processing shown in this flowchart ends.

  FIG. 6 is a flowchart showing an overview of the flow of the instruction book creation support process according to the present embodiment. The processing shown in this flowchart is executed when a work instruction is created by a work instruction person in charge.

  In step S201, an input of a new instruction is accepted. When a customer needs to contact the call center for repair and maintenance work, the work instruction person creates a work instruction using the node 90 and inputs it to the server 50. The instruction sheet receiving unit 24 receives an input of a work instruction sheet (new instruction sheet) newly created by a work instruction person (creator) in response to a request for repair and maintenance. Thereafter, the process proceeds to step S202.

  In step S202, the new instruction is evaluated. When the created new instruction is input, the instruction evaluation unit 25 evaluates the new instruction using the machine learning model constructed by the data update process of FIG. The evaluation result for each worker rank is output. In the present embodiment, since the worker rank is any one of A to C, the instruction sheet evaluation unit 25 performs the evaluation result based on the rank A worker criterion, the evaluation result based on the rank B worker criterion, and the rank. The evaluation result based on the C worker standard is output. Thereafter, the process proceeds to step S203.

  In steps S203 and S204, an instruction to correct a new instruction sheet whose quality does not satisfy a predetermined standard is given. The quality determination unit 28 determines whether or not the evaluation result for each worker rank by the instruction sheet evaluation unit 25 satisfies a predetermined quality standard defined for at least one of the ranks (step S203). For example, the quality determination unit 28 determines whether or not the quality standard is satisfied by determining whether or not the evaluation result is a predetermined threshold value (for example, 70 points) or more.

  Specifically, when the evaluation result is an evaluation result such as rank A: 73 points, rank B: 50 points, rank C: 40 points, the result of the quality determination is rank A: pass, rank B: Fail, Rank C: Fail. The quality standard to be satisfied may differ depending on the rank, or a standard other than the score may be adopted. For example, a criterion such as “within the top N% of past instructions corresponding to the rank” or a criterion such as “deviation value N or more among past instructions corresponding to the rank” may be employed.

  Further, in the above, an example is given in which the rank A criterion indicating a highly skilled worker is passed and the rank B and C criteria indicating a relatively low worker is rejected. In the machine learning model used in the form, since different evaluations are performed for each rank, for example, when the created instruction sheet is exhaustive and redundant, the rank criterion indicating a highly skilled worker is used. There may also be situations where the rank criteria indicating an unsatisfactory worker with low skill level is acceptable. When the quality standard is satisfied in any rank (in the above example, the quality standard of rank A is satisfied), the process proceeds to step S205.

  On the other hand, when the quality standard is not satisfied in any rank, the correction instruction unit 29 notifies the node 90 of the creator of the new instruction that the new instruction needs to be corrected (step S204). ). If a correction instruction has been issued, the process returns to step S201 to wait for input of a corrected new instruction sheet.

  In step S205 and step S206, the worker rank corresponding to the new instruction is presented, and the assigned worker rank is acquired. The presenting unit 26 presents information to the user (assignee) that can determine whether or not a predetermined quality standard defined for each rank is satisfied (step S205). For example, the presentation unit 26 notifies the assignment person's node 90 of information indicating that “rank A: pass, rank B: fail, rank C: fail” as a result of the quality determination in step S203. .

  The assignee who sees this notice will, if possible, assign workers with ranks that have a quality evaluation result of “Pass” from the assignable workers to repair and maintenance work related to the instructions. And the ID of the assigned worker is transmitted from the node 90 to the server 50. However, since it may occur that the worker with the rank of the quality judgment result “pass” cannot be assigned, the assignee should have the rank of the result of the quality judgment “fail”. Even an operator can be assigned. When the server 50 receives the worker ID, the skill level acquisition unit 27 acquires the worker rank associated with the worker ID from the storage device 14 (step S206). However, the skill level acquisition unit 27 may acquire the worker rank input by the assignee as it is. Thereafter, the process proceeds to step S207.

  In step S207, it is determined whether or not the new instruction sheet corresponds to the rank of the assigned worker. The quality determination unit 28 determines whether the evaluation result corresponding to the rank of the assigned worker among the evaluation results for each worker rank by the instruction evaluation unit 25 satisfies a predetermined quality standard defined for the rank. Determine whether or not. Since the specific determination method is the same as that described in step S203, the description thereof is omitted. If the assigned worker's rank satisfies the quality criterion (in the above example, if the assigned worker is rank A, the quality criterion is satisfied), the process proceeds to step S211. On the other hand, if the assigned worker's rank does not satisfy the quality standard, the process proceeds to step S208.

  From step S208 to step S210, correction and re-evaluation of a new instruction not corresponding to the assigned worker's rank is performed. The correction instruction unit 29 notifies the node 90 of the creator of the new instruction that the new instruction needs to be corrected (step S208). Here, the correction instruction unit 29 extracts an instruction sheet highly evaluated by the worker of the corresponding rank from the incident data, and presents it to the user together with the correction instruction for reference for correction.

  The work instruction person who has received the correction instruction corrects the work instruction using the node 90 and inputs it to the server 50. The instruction sheet reception unit 24 receives an input of a new instruction sheet corrected by a work instruction person in charge (creator) (step S209).

  When the corrected new instruction is input, the instruction evaluation unit 25 re-evaluates the new instruction using the machine learning model, thereby evaluating the corrected new instruction for each worker rank. The result is output (step S210). Thereafter, the process returns to step S207. By doing in this way, until the quality standard required for the assigned worker is satisfied, the correction and evaluation of the instruction are repeated, and the quality of the instruction is guaranteed. However, in actual operation, an upper limit is set for the number of correction / re-evaluation operations, and if the number of correction / re-evaluation operations reaches the upper limit even if the quality standard is not met, It is good also as ending evaluation work and progressing to Step S211.

  In step S211, a new instruction is confirmed. If the quality of the new instruction meets the quality standard required for the assigned worker or the number of correction / re-evaluation operations reaches the upper limit, the server 50 determines and confirms the content of the new instruction. Output as work instructions. The output destination is the creator's node 90, the assigner's node 90, the worker's portable terminal (not shown), the incident data storage unit 22, and the like. Thereafter, the processing shown in this flowchart ends.

  FIG. 7 is a flowchart showing an outline of the flow of data accumulation processing according to the present embodiment. The processing shown in this flowchart is executed when an input indicating that the assigned worker has finished the work in accordance with the new instruction is received.

  In step S301 and step S302, the operator evaluates the new instruction. The disassembling unit 30 disassembles the contents written in the new instruction into one sentence (hereinafter referred to as one section) by using a phrase or a period as a clue (step S301). Then, the result receiving unit 31 outputs an interface capable of inputting worker evaluation (including comprehensive evaluation of instructions and evaluation for each disassembled section) to the worker terminal, and prompts evaluation by the worker. . In addition, in order to encourage the evaluation of instructions by workers, the higher the ratio of instructions evaluated relative to the number of work, the higher the degree of contribution to improving the quality of instructions and A mechanism may be employed.

  The operator inputs an overall evaluation of the instruction sheet and an evaluation for each section through the output interface. About the evaluation value used for evaluation, it is as having mentioned above in description of the results reception part 31. FIG. When the input by the worker is completed, the worker terminal transmits to the server 50 the work results related to the work (whether or not the initial solution is possible, whether or not a trouble occurs, whether or not an inquiry occurs, work time, worker evaluation, worker rank, etc.). To do. The result receiving unit 31 receives the input of the work result including the evaluation of the new instruction by the worker received by the server 50 (step S302). Thereafter, the process proceeds to step S303.

  In step S303 and step S304, the creator evaluation is output and incident data is accumulated. The creator evaluation unit 32 outputs the work record accepted in step S302 to the node 90 used by the administrator or the like as an evaluation index of the creator of the instruction (step S303). In addition, the incident data storage unit 22 classifies incident data including the instruction and the work result associated with the instruction according to the target device and the content of repair maintenance, and stores the incident data in the storage device 14. (Step S304). The incident data accumulated here is used for updating the machine learning model in the data updating process described with reference to FIG. Thereafter, the processing shown in this flowchart ends.

<Effect>
According to the embodiment described above, a new instruction is evaluated using an evaluation model capable of outputting an evaluation for each worker rank (skill level), and the quality standard for the skill level of the assigned worker is satisfied. If there is not, it is possible to support the creation of an appropriate work instruction according to the skill level of the worker without causing excessive quality by instructing the correction of the instruction.

1 system 50 servers 90 nodes

Claims (20)

By classifying the incident data including the instructions for repair and maintenance work and the work results associated with the instructions according to the skill level of the worker in charge of the work related to the instructions, Model construction means for constructing a machine learning model capable of outputting an evaluation result for each worker skill level for the instruction sheet;
An instruction receiving means for receiving an input of a new instruction newly created in response to a request for repair and maintenance;
By evaluating the instruction using the machine learning model, an instruction evaluation unit that outputs an evaluation result for each worker skill level for the instruction;
Skill level acquisition means for acquiring the skill level of the worker assigned for the work related to the new instruction;
Whether the evaluation result corresponding to the skill level of the assigned operator among the evaluation results for each worker skill level by the instruction evaluation means satisfies a predetermined quality standard defined for the skill level. Quality judging means for judging;
An information processing apparatus comprising: In response to the evaluation result from the instruction evaluation means, the information processing apparatus further includes a presentation means for presenting information to the user that can determine whether or not a predetermined quality standard defined for each skill level of the worker is satisfied.
The skill level acquisition means acquires the skill level of the worker assigned by the user who confirmed the content presented by the presentation means.
The information processing apparatus according to claim 1. When the quality determination means determines that the evaluation result does not satisfy the quality standard, the quality determination means further comprises a correction instruction means for instructing correction of the new instruction sheet,
The instruction evaluation means evaluates the new instruction that has been corrected in response to the correction instruction,
The quality determination means determines whether the evaluation result corresponding to the skill level of the assigned worker for the corrected new instruction document satisfies a predetermined quality standard defined for the skill level. judge,
The information processing apparatus according to claim 1 or 2. The work results in the incident data include an evaluation by the worker in charge of work related to the corresponding instruction,
The correction instruction means presents an instruction extracted from the incident data, which is highly evaluated by an operator, together with an instruction to correct the new instruction, to the user.
The information processing apparatus according to claim 3. Incident data storage means for storing the incident data;
A record receiving means for receiving an input of a work record including an evaluation of the new instruction by the worker after the assigned worker has finished the work in accordance with the new instruction; and
The incident data accumulating means adds and accumulates new incident data including work results accepted by the result accepting means, instructions corresponding to the work results, and skill levels of the assigned workers,
The model construction means updates a machine learning model for evaluating the instructions based on accumulated incident data including the new incident data.
The information processing apparatus according to any one of claims 1 to 4. A disassembling means for disassembling the instructions text into a plurality of sections;
The achievement acceptance means accepts an input of evaluation designated by the worker for each disassembled section,
The information processing apparatus according to claim 5. The model construction means constructs the machine learning model by classifying the incident data according to a target object of repair maintenance, repair maintenance contents, and skill level of a repair maintenance worker to perform machine learning.
The information processing apparatus according to any one of claims 1 to 6. The model construction means changes the description items to be machine learning targets among the plurality of description items in the instructions included in the incident data for each worker skill level. Build a machine learning model that can output evaluation results,
The information processing apparatus according to any one of claims 1 to 7. A skill level determination means for determining the skill level of the worker based on at least one of the years of experience, the qualification of ownership, the number of cases of the corresponding device and the number of troubles,
The information processing apparatus according to any one of claims 1 to 8. Further comprising a creator evaluation means for associating the instruction document with the creator of the instruction document, and outputting at least one of the evaluation and the work performance according to the instruction document as an evaluation index of the creator;
The information processing apparatus according to any one of claims 1 to 9. Computer
By classifying the incident data including the instructions for repair and maintenance work and the work results associated with the instructions according to the skill level of the worker in charge of the work related to the instructions, A model building step for building a machine learning model that can output an evaluation result for each skill level of the instruction sheet,
An instruction reception step for receiving input of a newly created new instruction in response to a request for repair and maintenance;
By evaluating the instruction using the machine learning model, an instruction evaluation step for outputting an evaluation result for each worker skill level for the instruction;
A proficiency level acquisition step of acquiring the proficiency level of the worker assigned for the work related to the new instruction;
Whether the evaluation result corresponding to the skill level of the assigned worker among the evaluation results for each worker skill level in the instruction evaluation step satisfies a predetermined quality standard defined for the skill level. A quality judgment step for judging;
Work instruction evaluation method to execute. Computer
In response to the evaluation result in the instruction evaluation step, further executing a presenting step of presenting to the user information that can be determined whether or not a predetermined quality standard defined for each skill level of the worker is satisfied,
In the skill level acquisition step, acquire the skill level of the worker assigned by the user who confirmed the presentation content in the presentation step.
The work instruction evaluation method according to claim 11. Computer
In the quality determination step, when it is determined that the evaluation result does not satisfy the quality standard, a correction instruction step for instructing correction of the new instruction sheet is further executed,
In the instruction evaluation step, the new instruction that has been corrected in response to the correction instruction is evaluated,
In the quality determination step, it is determined whether or not the evaluation result corresponding to the skill level of the assigned worker for the modified new instruction document satisfies a predetermined quality standard defined for the skill level. judge,
The work instruction evaluation method according to claim 11 or 12. Computer
Incident data storage step for storing the incident data;
After the assigned worker finishes the work in accordance with the new instruction, a record receiving step for receiving an input of a work record including evaluation of the new instruction by the worker is further executed,
In the incident data accumulation step, the work results received in the result acceptance step, instructions corresponding to the work results, and new incident data including the skill level of the assigned worker are added and accumulated,
In the model building step, based on accumulated incident data including the new incident data, a machine learning model for evaluating the instructions is updated.
The work instruction evaluation method according to any one of claims 11 to 13. In the model construction step, among the plurality of description items in the instruction sheet included in the incident data, the description items to be machine-learned are changed for each worker skill level, so that each worker skill level is changed. Build a machine learning model that can output evaluation results,
The work instruction evaluation method according to any one of claims 11 to 14. Computer
By classifying the incident data including the instructions for repair and maintenance work and the work results associated with the instructions according to the skill level of the worker in charge of the work related to the instructions, Model construction means for constructing a machine learning model capable of outputting an evaluation result for each worker skill level for the instruction sheet;
An instruction receiving means for receiving an input of a new instruction newly created in response to a request for repair and maintenance;
By evaluating the instruction using the machine learning model, an instruction evaluation unit that outputs an evaluation result for each worker skill level for the instruction;
Skill level acquisition means for acquiring the skill level of the worker assigned for the work related to the new instruction;
Whether the evaluation result corresponding to the skill level of the assigned operator among the evaluation results for each worker skill level by the instruction evaluation means satisfies a predetermined quality standard defined for the skill level. Quality judging means for judging;
Program to function as. Computer
In response to the evaluation result by the instruction evaluation means, it further functions as a presentation means for presenting information to the user that can determine whether or not a predetermined quality standard determined for each skill level of the worker is satisfied,
The skill level acquisition means acquires the skill level of the worker assigned by the user who confirmed the content presented by the presentation means.
The program according to claim 16. Computer
When the quality determination means determines that the evaluation result does not satisfy the quality standard, it further functions as a correction instruction means for instructing correction of the new instruction sheet,
The instruction evaluation means evaluates the new instruction that has been corrected in response to the correction instruction,
The quality determination means determines whether the evaluation result corresponding to the skill level of the assigned worker for the corrected new instruction document satisfies a predetermined quality standard defined for the skill level. judge,
The program according to claim 16 or 17. Computer
Incident data storage means for storing the incident data;
After the assigned worker finishes the work in accordance with the new instruction, it further functions as a result receiving means for receiving an input of work results including evaluation of the new instruction by the worker,
The incident data accumulating means adds and accumulates new incident data including work results accepted by the result accepting means, instructions corresponding to the work results, and skill levels of the assigned workers,
The model construction means updates a machine learning model for evaluating the instructions based on accumulated incident data including the new incident data.
The program according to any one of claims 16 to 18. The model construction means changes the description items to be machine learning targets among the plurality of description items in the instructions included in the incident data for each worker skill level. Build a machine learning model that can output evaluation results,
The program according to any one of claims 16 to 19.

Images