JP7645705B2 - Man-hour management device, man-hour management method, and program - Google Patents

Description

This disclosure relates to a labor-hour management device, a labor-hour management method, and a program.

At production sites, including factories, systems that record work results are used to manage work hours or improve work efficiency (for example, Patent Document 1). The work results recording system described in Patent Document 1 has a voice recognition unit that recognizes input voice, and records a work results log that includes the start time of the work, the completion time of the work, and keywords recognized from the start time of the work to the completion time, based on the voice recognized by the voice recognition unit.

The work performance recording system described in Patent Document 1 is explained as improving work efficiency by allowing workers to input limited keywords by voice rather than inputting data manually, eliminating the need for workers to interrupt their work to input work performance.

JP 2017-167391 A

The work performance recording system described in Patent Document 1 is intended for work at production sites, including factories, but there is a demand not only for production departments, but also for other departments, including administrative departments and indirect departments, to record work performance and improve work based on that.

To improve work, it is necessary to compare planned man-hours, which are the planned work hours, with actual man-hours, which are the time actually required by workers to perform the work, and to extract problems with the work or areas for improvement based on the difference between the planned man-hours and actual man-hours. However, there was a problem in that the planned man-hours did not match the actual work, making it difficult to evaluate correctly.

This disclosure has been made in consideration of the above-mentioned circumstances, and aims to provide a labor-hour management device, labor-hour management method, and program that can obtain planned labor hours that correspond to the actual work conditions based on the characteristics of the workers.

In order to achieve the above object, the man-hour management device of the present disclosure includes a man-hour information acquisition unit that acquires planned man-hours, which are the task contents of planned tasks and the time required for each task content, worker information including the characteristics of the worker, and actual man-hours, which are the task contents of actual tasks that are tasks performed by the worker and the time required for each task content. The man-hour management device further includes a storage unit that classifies and stores the planned man-hours acquired by the man-hour information acquisition unit by the characteristics of the worker and the task content , and classifies and stores the actual man-hours acquired by the man-hour information acquisition unit by the characteristics of the worker related to the actual man-hours and the task content, and an optimal planned man-hour calculation unit that generates a trained model by machine learning using learning data including the characteristics and the actual man-hours when the actual man-hours for the characteristics and task content stored in the storage unit become equal to or less than the planned man-hours for the characteristics and task content, and uses the trained model to calculate an optimal planned man-hour for the task content when the characteristics of the worker are input . The man-hour information acquisition unit updates the planned man-hours related to the characteristics and the work content to the optimal planned man-hours for the characteristics and the work content calculated by the optimal planned man-hour calculation unit.

According to the present disclosure, the planned man-hours are updated based on the actual man-hours classified by the characteristics of the workers, making it possible to obtain planned man-hours that correspond to the actual work conditions based on the characteristics of the workers.

FIG. 1 is a block diagram showing a configuration example of a man-hour management device according to an embodiment of the present disclosure. Flowchart of man-hour information acquisition process Flowchart of optimal planned man-hour calculation process FIG. 1A is a block diagram showing a learning device of an optimal planned man-hour calculation unit; FIG. 1B is a flowchart showing a learning process; FIG. 1A is a block diagram showing an inference device of an optimal planned man-hour calculation unit; FIG. 1B is a flowchart of an inference process; Table showing worker characteristics and actual man-hours (a) A table showing data on planned work; (b) A table showing data on actual work; (c) A time chart of planned work; (d) A time chart of actual work. Table showing points of focus and proposals for work improvement

(Embodiment)
Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an example of the configuration of a man-hour management device 1 according to an embodiment of this disclosure. The man-hour management device 1 is a device that records planned work, planned man-hours, actual work, and actual man-hours, and outputs improvement proposals for work. The planned work and actual work include a wide variety of work content in any department, including the production department, management department, or indirect department. In this embodiment, a case will be described in which the work content of the planned work and actual work is work content in an indirect department, including document creation and data analysis.

The labor-hour management device 1 includes a calculation processing unit 110 that acquires labor-hour information, optimizes planned labor hours, and executes a process to generate improvement proposals for actual work, a storage unit 120 that stores data acquired or generated by the calculation processing unit 110, an operation input unit 130 that accepts operation input from an operator, and a display unit 140 that displays information acquired or generated by the calculation processing unit 110.

The calculation processing unit 110 is any calculation processing device, for example a CPU (Central Processing Unit). By executing a program stored in the memory unit 120, the calculation processing unit 110 functions as a man-hour information acquisition unit 111, an optimal planned man-hour calculation unit 112, a planned/actual man-hour difference extraction unit 113, an improvement focus extraction unit 114, and an improvement proposal output unit 115.

The man-hour information acquisition unit 111 of the calculation processing unit 110 acquires worker information including the characteristics of the workers, and man-hour information including the planned man-hours and the actual man-hours, and stores the information in the memory unit 120. The optimal planned man-hour calculation unit 112 calculates the optimal planned man-hours using artificial intelligence (AI) based on the characteristics of the workers and the actual man-hours.

The planned/actual man-hour difference extraction unit 113 extracts the difference between the planned man-hours and the actual man-hours, and the improvement focus extraction unit 114 extracts improvement focus points based on the difference between the planned man-hours and the actual man-hours. The improvement proposal output unit 115 obtains and outputs improvement proposals for the business from the memory unit 120 based on the extracted improvement focus points.

The storage unit 120 is any storage device, such as a non-volatile semiconductor memory including a flash memory or an erasable programmable read only memory (EPROM), or a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, or a digital versatile disk (DVD).

As shown in FIG. 1, the memory unit 120 includes a planned/actual man-hours database 121 that stores data including planned man-hours and actual man-hours, a learned model memory unit 122 that stores learned models used by the optimal planned man-hours calculation unit 112, and an improvement plan database 123 that stores data on improvement plans for business operations. The memory unit 120 also stores programs executed by the calculation processing unit 110.

The functions of the calculation processing unit 110 and the contents of the storage unit 120 will be described in detail below. First, the operation of the man-hour information acquisition unit 111 of the calculation processing unit 110 will be described with reference to the flowchart in Figure 2. Figure 2 is a flowchart of the man-hour information acquisition process.

The man-hour information acquisition unit 111 uses a scheduler and a work performance recording tool to acquire worker information including worker characteristics, planned tasks, planned man-hours, actual tasks, and actual man-hours based on information input by an operator to the operation input unit 130. Here, the operator is a person permitted to operate the man-hour management device 1, and is, for example, a worker who performs tasks, or a manager who manages tasks and the man-hour management device 1. Furthermore, the characteristics of a worker are qualities that affect the work of a worker, such as age, position, years of experience, qualifications, and skills.

The scheduler is any software with a schedule management function, and may be general-purpose schedule management software. The scheduler is software that can write out data on planned tasks and planned man-hours that can be read by a work performance recording tool. For example, a worker uses the scheduler to input the tasks planned for the day and the scheduled start and completion times for each task at the start of each day or for each predetermined period.

Here, because the names of the work contents may vary depending on the operator or region, the work contents are listed in advance and the operator can select from them. This makes it possible to avoid tasks being tabulated as different work data when tallying up man-hour data, even though they are the same work, due to differences in the names of the work contents.

The work performance recording tool is a tool that has the function of importing data from the scheduler and collecting data related to the actual work, and is, for example, a tool created using the macro function of EXCEL (registered trademark) VBA or other general-purpose spreadsheet software.

The work performance recording tool retrieves the planned work content and the scheduled start and completion times for each work content from the scheduler, and also obtains the work content of the actual work and the start and completion times for each work content based on the operator's input to the operation input unit 130. For example, the work performance recording tool has an input form that allows the operator to input the work content and to press or click a button at the start and completion timing of each work content.

The work performance recording tool has the same list of work contents as the list used by the scheduler, and the operator selects the work contents of the performed work from this list of work contents. In addition, the work performance recording tool records the start time and completion time of the selected work content when the operator presses and clicks a button at the timing of starting and completing the work.

In the labor-hour information acquisition process shown in FIG. 2, first, the labor-hour information acquisition unit 111 acquires the work content of the planned work and the scheduled start time and scheduled completion time from the scheduler (step S101). Next, the labor-hour information acquisition unit 111 imports the work content of the planned work and the scheduled start time and scheduled completion time acquired from the scheduler into the work performance recording tool (step S102).

Next, the man-hour information acquisition unit 111 acquires the start time and completion time of each task content of the actual work based on the information input by the operator using the work performance recording tool (step S103). The man-hour information acquisition unit 111 calculates the planned man-hours, which are the time required for each task content, from the scheduled start time and scheduled completion time of each task content of the planned work acquired using the work performance recording tool. The man-hour information acquisition unit 111 also calculates the actual man-hours, which are the time required for each task content, from the start time and completion time of each task content of the actual work acquired using the work performance recording tool (step S104).

Here, the planned man-hours or actual man-hours are the time required for the relevant task, so if other tasks or breaks are performed in between, the time for the other tasks or breaks is subtracted from the time from the scheduled start time to the scheduled completion time or from the time from the start time to the completion time. In step S104, in order to facilitate prioritizing task improvements or converting the effects of task improvements into a monetary value, the man-hour information acquisition unit 111 may use a work performance recording tool to multiply the actual man-hours by the worker's hourly processing cost rate to calculate and visualize the actual input amount.

Then, the man-hour information acquisition unit 111 classifies the planned man-hours by the characteristics of the worker and the work content (step S105) and stores them in the planned and actual man-hours database 121 of the storage unit 120. The man-hour information acquisition unit 111 also classifies the actual man-hours by the characteristics of the worker and the work content (step S105) and stores them in the planned and actual man-hours database 121 of the storage unit 120 (storage step). If the actual investment amount was calculated in step S104 by multiplying the hourly processing cost rate of the worker, the actual investment amount is stored in the planned and actual man-hours database 121 together with the actual man-hours.

By the above-mentioned man-hour information acquisition process, the planned man-hours and actual man-hours are stored in the planned/actual man-hours database 121 of the storage unit 120, grouped by the task content and the characteristics of the worker. For example, the planned man-hours and actual man-hours for the task content of creating document A are stored in groups by the worker's age, job title, years of experience, qualifications, and skills.

The optimal planned man-hour calculation unit 112 of the calculation processing unit 110 uses artificial intelligence (AI) to calculate the optimal planned man-hours based on the data stored in the planned and actual man-hours database 121. The operation of the optimal planned man-hour calculation unit 112 of the calculation processing unit 110 will be described with reference to the flowchart in Figure 3. Figure 3 is a flowchart of the optimal planned man-hours calculation process.

First, the optimal planned man-hour calculation unit 112 acquires the actual man-hours stored in the planned and actual man-hours database 121 (step S201). Next, the acquired actual man-hours are compared with the planned man-hours classified into the same work content and the same worker characteristics (step S202).

If the actual man-hours acquired in step S201 are shorter than the planned man-hours classified into the same task content and the same worker characteristics (step S203: Yes), the data of the actual man-hours is added to the learning data together with the data of the task content and the worker characteristics (step S204). If the actual man-hours acquired in step S201 are equal to or greater than the planned man-hours classified into the same task content and the same worker characteristics (step S203: No), the process proceeds to step S205.

The optimal planned man-hour calculation unit 112 calculates the optimal planned man-hour by AI using learning data including the actual man-hours, the work content, and the characteristics of the workers (step S205: optimal planned man-hour calculation step). The calculation of the optimal planned man-hours in step S205 may be performed each time the actual man-hours are added to the learning data in step S204, or may be performed when a predetermined amount of learning data has been accumulated. Alternatively, it may be performed at any predetermined time interval.

The AI that calculates the optimal planned man-hours is realized in FIG. 1 by the optimal planned man-hours calculation unit 112 of the computation processing unit 110 and the learned model storage unit 122 of the storage unit 120. The optimal planned man-hours calculation unit 112 includes a learning device 217 and an inference device 218 (see FIG. 4(a) and FIG. 5(a)).

In this embodiment, the learning device 217 and the inference device 218 are included in the labor-hour management device 1. The learning device 217 and the inference device 218 are used to learn the optimal planned labor hours for the target work content. The target work content is any work content performed in any department, including the production department, the management department, and the indirect department, and is, for example, the work of creating productivity analysis materials or monthly report materials in the indirect department.

The learning device 217 and the inference device 218 may take any form, and may be, for example, part of the labor-hour management device 1, or may exist on a server externally connected to the labor-hour management device 1 via a communication network.

First, the learning process will be described. FIG. 4(a) is a block diagram showing the learning device 217 of the optimal planned labor-hour calculation unit 112, and FIG. 4(b) is a flowchart of the learning process. Here, learning when the work content is a document creation work will be described.

As shown in FIG. 4(a), the learning device 217 includes a data acquisition unit 2171 and a model generation unit 2172. The data acquisition unit 2171 acquires the actual man-hours for the document creation work and the characteristics of the workers related to the actual man-hours, which were added to the learning data in step S204 of the optimal planned man-hour calculation process shown in FIG. 3.

The model generation unit 2172 learns based on the learning data including information on the combination of new actual man-hours and worker characteristics acquired by the data acquisition unit 2171. In other words, a trained model is generated for inferring the optimal planned man-hours from the new actual man-hours required for the target document creation task and the characteristics of the worker. The learning data in this learning process is data that associates the new actual man-hours and the characteristics of the worker with each other.

The learning algorithm used by the model generation unit 2172 may be a conventional algorithm. For example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning may be used.

You can also perform deep learning, which learns to extract the features themselves, or you can perform machine learning according to genetic programming, functional logic programming, or support vector machines.

In this embodiment, the application of a neural network will be described as an example. In this case, the model generation unit 2172 learns the optimal planned man-hours for the characteristics of the workers by supervised learning according to the neural network model. Here, supervised learning refers to a method in which a learning device is provided with pairs of input and result (label) data, and the device learns the characteristics of the learning data and infers the result from the input.

A neural network is composed of an input layer with multiple neurons, an intermediate layer (hidden layer) with multiple neurons, and an output layer with multiple neurons. The intermediate layer may be one layer or two or more layers.

In the neural network of this embodiment, the model generation unit 2172 learns the optimal planned man-hours for the characteristics of the worker through supervised learning based on the learning data including information on the combination of new actual man-hours and the characteristics of the worker acquired by the data acquisition unit 2171.

In other words, the model generation unit 2172 adjusts the weight W1 between the input layer and the middle layer and the weight W2 between the middle layer and the output layer for each characteristic of the worker, and learns by determining the weights W1 and W2 that can bring the output from the output layer when the characteristics of a new worker are input to the input layer closer to the actual labor hours, which is the result (correct answer). For example, when the characteristics of a worker who performed a document creation task, including the age and years of experience, are input to the input layer, an index indicating the amount of waste in the action is obtained in the middle layer, and the optimal planned labor hours are output to the output layer, which is compared with the actual labor hours, which is the correct answer, to determine the weights W1 and W2.

The trained model storage unit 122 stores the trained model output by the model generation unit 2172.

The flow of the learning process is as shown in the flowchart in Figure 4 (b). First, the data acquisition unit 2171 of the learning device 217 acquires data on new actual man-hours and worker characteristics (step S301). Here, the new actual man-hours and worker characteristics are acquired simultaneously, but it is sufficient if the new actual man-hours and worker characteristics are acquired in association with each other, and the data acquisition unit 2171 may acquire the data on the new actual man-hours and worker characteristics at different times.

Then, the model generation unit 2172 performs a learning process based on the combination of the new actual man-hours and the worker characteristics acquired by the data acquisition unit 2171 (step S302) to generate a trained model. Then, the generated trained model is stored in the trained model storage unit 122 (step S303).

Next, the inference process will be described. FIG. 5(a) is a block diagram showing the inference device 218 of the optimal planned labor-hour calculation unit 112, and FIG. 5(b) is a flowchart of the inference process.

Here, we will explain the case of inferring the optimal planned man-hours for a document creation task. As shown in FIG. 5(a), the inference device 218 includes a data acquisition unit 2181 and an inference unit 2182. The data acquisition unit 2181 acquires, as input data, the characteristics of a new worker acquired by the man-hour information acquisition unit 111.

The inference unit 2182 uses the learned model stored in the learned model storage unit 122 to infer the optimal planned man-hours for the characteristics of the worker acquired by the data acquisition unit 2181. In other words, the inference unit 2182 inputs the characteristics of the worker into the learned model, and outputs the optimal planned man-hours inferred from the characteristics of a new worker.

In this embodiment, a configuration has been described in which the optimal planned man-hours are output using a trained model generated by the model generation unit 2172 of the learning device 217, but it is also possible to obtain a trained model from outside the man-hour management device 1 and infer the optimal planned man-hours based on this trained model.

The flow of the inference process is as shown in the flowchart in FIG. 5(b). First, the data acquisition unit 2181 of the inference device 218 acquires data on the characteristics of the new worker (step S401). Next, the inference unit 2182 inputs the characteristics of the new worker into the trained model stored in the trained model storage unit 122 (step S402), and infers and outputs the optimal planned man-hours (step S403).

The optimal planned man-hour calculation unit 112 updates the planned man-hours stored in the planned and actual man-hours database 121 to the optimal planned man-hours calculated in step S403 (step S404: update step). At this time, the characteristics of the workers related to the planned man-hours before and after the update are the same.

In this way, the planned man-hours that are optimal for the characteristics of the worker can be updated, and the currently optimal planned man-hours can be used to manage man-hours when a worker with the same characteristics works later. In other words, by using man-hour data categorized by the characteristics of the worker, including age, position, years of experience, qualifications, and skills, for man-hour management when a worker is performing work that he or she has no experience with, or when the worker is a new employee and it is difficult to estimate the planned man-hours, it becomes possible to set a schedule that matches the capabilities or abilities of the worker.

In addition, because the new planned man-hours are values calculated from the actual man-hours to date, when workers are performing work that they have no experience with or when they are new employees, they can improve the speed of work processing by keeping this value in mind. Furthermore, for repeat work, which is work that is performed repeatedly, applying the new planned man-hours makes it possible to determine that the previous planned man-hours are not optimal, and to bring the actual man-hours closer to the optimal planned man-hours. If the planned man-hours used until now are long, revising them to the optimal planned man-hours increases the amount of planned work that can be processed in a day, making it possible to improve worker productivity.

The planned/actual man-hour difference extraction unit 113 compares the actual man-hours acquired by the man-hour information acquisition unit 111 with the planned man-hours calculated by the optimal planned man-hour calculation unit 112, and extracts the difference in man-hours. For example, the work performance recording tool of the man-hour information acquisition unit 111 acquires the actual tasks and actual man-hours from the scheduler and graphs them for each worker, and graphs the planned man-hours and planned tasks calculated by the optimal planned man-hour calculation unit 112, to visualize the difference. Any method of visualization may be used, but for example, a time chart or Gantt chart may be automatically created using the work performance recording tool and presented to the worker.

The improvement focus extraction unit 114 selects a task content for which the difference between the planned and actual man-hours extracted by the planned/actual man-hour difference extraction unit 113 is equal to or greater than a predetermined threshold, and extracts an improvement focus for the task content. The improvement focus may be selected from a predetermined category. For example, improvement focus points related to the seven wastes lurking in indirect departments described in a non-patent document (Hirano Hiroyuki, Nikkan Kogyo Shimbun, "Visually Understandable Indirect Just-in-Time"), namely, waste of doing too much, waste of stagnation, waste of movement, waste of mistakes, waste of work itself, waste of movement, and waste of waiting, may be stored in association with the task content, and improvement focus points may be extracted based on the task content of the actual task. Alternatively, improvement focus points may be extracted from information input by workers based on past task performance.

In the improvement plan database 123 of the storage unit 120, improvement plans are stored, categorized by business content and improvement viewpoint. For example, for the business content of document creation, improvement plans associated with improvement viewpoints related to the seven types of waste that lurk in indirect departments, namely, waste of overdoing, waste of stagnation, waste of movement, waste of mistakes, waste in the work itself, waste of motion, and waste of waiting, are stored in the improvement plan database 123.

The improvement proposal output unit 115 extracts improvement proposals stored in the improvement proposal database 123 based on the improvement points extracted by the improvement point extraction unit 114 and the information on the actual work and actual labor hours, and displays the list on the display unit 140.

The operation of the man-hour management device 1 configured as above will be explained using the example shown in Figures 6-8.

The man-hour information acquisition unit 111 uses a scheduler and a work performance recording tool to acquire worker information including worker characteristics, the work content of the planned work and the actual work, and the planned man-hours and actual man-hours based on the information input by the operator to the operation input unit 130, and stores them in the planned/actual man-hours database 121 of the storage unit 120. In the planned/actual man-hours database 121, the planned man-hours and actual man-hours are stored, grouped according to the work content and the characteristics of the worker.

Figure 6 shows an example of actual man-hours stored in the planned and actual man-hours database 121. Here, it shows the actual man-hours for the task of creating document A, which is performed in an indirect department, and is categorized by the worker's characteristics, such as age, position, and years of experience.

The optimal planned man-hour calculation unit 112 uses AI to calculate the optimal planned man-hours based on the data stored in the planned/actual man-hour database 121. Specifically, the optimal planned man-hour calculation unit 112 generates a trained model based on the actual man-hours classified by the work content and worker characteristics stored in the planned/actual man-hour database 121, and uses the trained model to calculate the optimal planned man-hours for the work content and worker characteristics.

The planned man-hours in the planned/actual man-hours database 121 are updated with the optimal planned man-hours calculated by the optimal planned man-hours calculation unit 112, and are used as the planned man-hours for the tasks in the subsequent work. The planned/actual man-hours difference extraction unit 113 compares the planned man-hours to which the optimal planned man-hours calculated by the optimal planned man-hours calculation unit 112 are applied with the newly acquired actual man-hours, and outputs the comparison result.

Figure 7 shows an example of data on the work content and time of planned work and actual work, where (a) is data on the work content and time of planned work, (b) is data on the work content and time of actual work, (c) is a time chart of planned work, and (d) is a time chart of actual work.

The planned/actual labor-hour difference extraction unit 113 acquires the planned labor hours and actual labor hours calculated based on the work content and time data of the planned work and actual work, and extracts work with a large difference between the planned labor hours and the actual labor hours. In addition, the planned/actual labor-hour difference extraction unit 113 can visualize the difference between the planned labor hours and the actual labor hours by displaying it in any form, including a time chart or Gantt chart.

The improvement focus extraction unit 114 selects work content for which the difference between the planned man-hours and the actual man-hours extracted by the planned/actual man-hour difference extraction unit 113 is equal to or greater than a predetermined threshold, and extracts improvement focus points for the work content of the actual work. Figure 8 shows improvement focus points extracted for seven types of waste that lurk in indirect departments: waste of doing too much, waste of stagnation, waste of movement, waste of mistakes, waste of the work itself, waste of action, and waste of waiting. For example, improvement focus points may be extracted from information input by workers based on the results of past improvement measures, which has been accumulated.

In the improvement plan database 123 of the storage unit 120, improvement measures implemented in the past are classified according to the work content and the focus point and stored as improvement plans. For example, for the work of creating documents, the improvement plan database 123 stores improvement plans associated with the improvement focus points related to the seven wastes that lurk in indirect departments: waste of doing too much, waste of stagnation, waste of movement, waste of mistakes, waste in the work itself, waste of motion, and waste of waiting.

The improvement proposal output unit 115 extracts improvement proposals stored in the improvement proposal database 123 based on the improvement viewpoints and information on the business content extracted by the improvement viewpoint extraction unit 114, and displays the list on the display unit 140. Figure 8 shows an example of the display of the improvement viewpoints and improvement proposals.

By having workers recognize the improvement ideas obtained in this way and carry out subsequent work, the workers' productivity can be improved.

As described above, in the man-hour management device 1 according to the present embodiment, the man-hour information acquisition unit 111 acquires the task content and planned man-hours of the planned task, the worker information including the characteristics of the worker, and the task content and actual man-hours of the actual task performed by the worker, and classifies the acquired actual man-hours according to the characteristics and task content of the worker and stores them in the planned/actual man-hour database 121 of the storage unit 120. The optimal planned man-hour calculation unit 112 calculates the optimal planned man-hours based on the characteristics and actual man-hours when the actual man-hours for the worker characteristics and task content stored in the planned/actual man-hour database 121 are equal to or less than the planned man-hours for the characteristics and task content. The man-hour information acquisition unit 111 then updates the planned man-hours for the characteristics and task content to the optimal planned man-hours for the characteristics and task content calculated by the optimal planned man-hour calculation unit 112. As a result, the optimal planned man-hours are updated based on the actual man-hours that are equal to or less than the planned man-hours, so that it is possible to use planned man-hours that are in line with the actual work state due to the characteristics of the worker.

In the above embodiment, the planned and actual man-hours database 121, the learned model storage unit 122, and the improvement proposal database 123 are included in the storage unit 120 of the man-hours management device 1, but they may also be stored in an external storage device connected to and communicating with the man-hours management device 1.

In addition, in the optimal planned man-hour calculation process shown in FIG. 3, if the actual man-hour is shorter than the planned man-hour, the actual man-hour is used as learning data for learning. However, if there are multiple actual man-hours that are shorter than the planned man-hour within a predetermined period, only the shortest actual man-hour may be used as learning data.

In addition, in the above embodiment, the trained model generated in the optimal planned labor hours calculation process is a model generated for a task content selected from task contents including document creation, and the optimal planned labor hours are inferred based on the actual labor hours related to the task content, but the trained model generated for the task content may also be used to infer the optimal planned labor hours for related or similar task content.

The hardware configuration and the flowchart shown in the above embodiment are merely examples and can be modified or adapted. The functions realized by the calculation processing unit 110 and the storage unit 120 can be realized by using a normal computer system, not by a dedicated system.

For example, a computer capable of implementing each function may be configured by distributing a program for executing the operations of the above-described embodiments stored on a computer-readable recording medium such as a CD-ROM (Compact Disc Read-Only Memory), a DVD (Digital Versatile Disc), an MO (Magneto Optical Disc), or a memory card, and installing the program on a computer. In addition, if each function is implemented by sharing the work between an OS (Operating System) and an application, or by cooperation between an OS and an application, only the parts other than the OS may be stored on the recording medium.

This disclosure is not limited to the above-described embodiment, and various modifications are of course possible without departing from the spirit of this disclosure.

1 Man-hour management device, 110 Calculation processing unit, 111 Man-hour information acquisition unit, 112 Optimal planned man-hour calculation unit, 113 Planned/actual man-hour difference extraction unit, 114 Improvement focus extraction unit, 115 Improvement proposal output unit, 120 Storage unit, 121 Planned/actual man-hour database, 122 Learned model storage unit, 123 Improvement proposal database, 130 Operation input unit, 140 Display unit, 217 Learning device, 218 Inference device, 2171 Data acquisition unit, 2172 Model generation unit, 2181 Data acquisition unit, 2182 Inference unit.

Claims (8)

a man-hour information acquisition unit that acquires planned man-hours, which are the work contents of planned work and the time required for each work content, worker information including the characteristics of the worker, and actual man-hours, which are the work contents of actual work that are work performed by the worker and the time required for each work content;
a storage unit that classifies and stores the planned man-hours acquired by the man-hour information acquisition unit according to the characteristics of the workers and the work content, and that classifies and stores the actual man-hours acquired by the man-hour information acquisition unit according to the characteristics of the workers related to the actual man-hours and the work content;
an optimal planned man-hour calculation unit that generates a trained model by machine learning using learning data including the characteristics and the actual man-hours when the actual man-hours for the characteristics and the work content stored in the storage unit become equal to or less than the planned man-hours for the characteristics and the work content, and calculates an optimal planned man-hours for the work content when the characteristics of the worker are input using the trained model ;
The man-hour information acquisition unit updates the planned man-hours related to the characteristics and the work content to the optimal planned man-hours calculated by the optimal planned man- hour calculation unit for the characteristics and the work content.
Man-hours management device. The characteristics of the worker include at least one of the worker's age, job title, years of work experience, qualifications, and skills;
The man-hour management device according to claim 1 . The actual man-hours are calculated from the start time and completion time of the work content input by the worker using a work performance recording tool that imports the work content and the planned man-hours of the planned work from schedule management software.
The man-hour management device according to claim 1 or 2. The optimal planned man-hour calculation unit is
A learning device that generates the trained model by machine learning using the training data; and
an inference device that infers the optimal planned man-hours for the work content when the characteristics of the worker are input, using the trained model generated by the learning device;
having
The man-hour management device according to any one of claims 1 to 3. The storage unit further stores an improvement viewpoint associated with the business content and an improvement plan classified according to the business content and the improvement viewpoint;
an improvement focus extraction unit that compares the actual man-hours for the characteristics and the work content of the worker with the planned man-hours for the characteristics and the work content, and extracts an improvement focus corresponding to the work content of the actual work stored in the storage unit when there is a difference equal to or greater than a predetermined threshold;
and an improvement proposal output unit that outputs an improvement proposal associated with the business content and the improvement viewpoint of the performance business stored in the storage unit .
The man-hour management device according to any one of claims 1 to 4. The improvement focus point is a focus point related to at least one of the following: waste of doing too much, waste of stagnation, waste of movement, waste of mistakes, waste of the work itself, waste of movement, and waste of on-hand time.
The man-hour management device according to claim 5 . A computer-implemented labor-hour management method, comprising:
A first storage step of classifying and storing the content of the planned work and the planned man-hours, which are the time required for each content of the planned work, according to the characteristics of the workers and the content of the work;
a second storage step of classifying and storing the actual man-hours, which are the time required for each task content of the actual tasks performed by the worker, according to the characteristics of the worker and the task content;
an optimal planned man-hour calculation step of generating a trained model by machine learning using learning data including the characteristics and the actual man-hours when the actual man-hours for the characteristics and the work content of the worker stored in the second storage step become equal to or less than the planned man-hours for the characteristics and the work content stored in the first storage step, and calculating an optimal planned man-hours for the work content when the characteristics of the worker are input using the trained model;
an updating step of updating the planned man-hours related to the characteristics and the work content to the optimal planned man-hours for the characteristics and the work content calculated in the optimal planned man-hour calculation step. Computer,
a storage unit that classifies and stores planned man-hours, which are the work contents of planned work and the time required for each work content, by worker characteristics and the work content, and that classifies and stores actual man-hours, which are the time required for each work content of actual work performed by the worker, by worker characteristics and the work content;
an optimal planned man-hour calculation unit that generates a trained model by machine learning using learning data including the characteristics and the actual man-hours of the worker stored in the storage unit when the actual man-hours for the characteristics and the work content become equal to or less than the planned man-hours for the characteristics and the work content, and calculates an optimal planned man-hours for the work content when the characteristics of the worker are input using the trained model;
a man-hour information acquisition unit that acquires the optimal planned man-hours for the characteristic and the task content calculated by the optimal planned man-hour calculation unit as the planned man-hours related to the characteristic and the task content;
A program that functions as a

Images