Evaluation device and evaluation method

Abstract

Appropriately evaluate the work capabilities of workers. [Solution] As an evaluation method, a work time prediction model that has learned the work times for multiple tasks performed by multiple workers during a first period is used to estimate the work time for standard tasks performed by each worker, and a first proficiency level of a specific worker during the first period is estimated based on the ranking of the work time for standard tasks performed by a specific worker in the distribution of work times for standard tasks performed by multiple workers.

Description

[Technical Field]

[0001] The present disclosure relates to an evaluation device and an evaluation method. [Background technology]

[0002] For example, when multiple workers need to work together to complete a task by a target time, such as picking and delivering packages from a warehouse, a plan is made to assign which tasks to which workers. In theory, assigning more tasks to faster workers will shorten the overall work time, but this alone will not improve the work capabilities of other workers.

[0003] Patent Document 1 discloses a worker allocation system and a worker allocation method that can allocate beginners so that beginners with low skills have more opportunities to improve their skills while ensuring work results. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-118359 Summary of the Invention [Problem to be solved by the invention]

[0005] In Patent Document 1, workers who work for a short time are considered to have a high skill level, and workers who work for a long time are considered to have a low skill level. However, because multiple workers do not necessarily perform the same tasks, simply comparing work times is not enough to properly evaluate the level of a worker's work ability.

[0006] Therefore, an object of the present disclosure is to provide a technique that can appropriately evaluate the working ability of a worker. [Means for solving the problem]

[0007] The present disclosure provides an evaluation device including a processor and a memory, wherein the processor works in cooperation with the memory to estimate a standard work time for each worker using a work time prediction model that has learned the work times for a plurality of tasks performed by each of a plurality of workers during a first period, and to estimate a first proficiency level of the specific worker during the first period based on the ranking of the work time for the standard work by the specific worker within the distribution of the work times for the standard work by each of the multiple workers.

[0008] The present disclosure provides an evaluation method that estimates the work time of a standard task by each worker using a work time prediction model that has learned the work times of multiple tasks performed by each of multiple workers during a first period, and estimates a first proficiency level of a specific worker during the first period based on the ranking of the work time of the standard task by a specific worker within the distribution of the work times of the standard task by each of the multiple workers.

[0009] These comprehensive or specific aspects may be realized as a system, an apparatus, a method, an integrated circuit, a computer program, or a recording medium, or may be realized as any combination of a system, an apparatus, a method, an integrated circuit, a computer program, and a recording medium. [Effects of the Invention]

[0010] According to the present disclosure, the work ability of a worker can be appropriately evaluated. [Brief explanation of the drawings]

[0011] [Figure 1] A block diagram showing an example of the configuration of a work planning system according to an embodiment of the present invention. [Figure 2] FIG. 1 is a sequence diagram illustrating an example of an operation of the work planning system according to the present embodiment. [Figure 3] FIG. 1 is a diagram for explaining a first example of a method for estimating a proficiency level according to an embodiment of the present invention. [Figure 4] FIG. 10 is a diagram for explaining a second example of a method for estimating a proficiency level according to the present embodiment. [Figure 5] FIG. 10 is a diagram for explaining a modified example of the second example of the proficiency level estimation method according to the present embodiment. [Figure 6] FIG. 1 is a diagram for explaining a learning method and a re-learning method for a proficiency improvement value prediction model according to an embodiment of the present invention. [Figure 7] FIG. 10 is a diagram showing an example of a calculation formula for optimally allocating tasks to workers so as to minimize the overall task time according to the present embodiment. [Figure 8] A diagram for explaining the variables in the equations shown in Figure 7. [Figure 9] FIG. 10 is a diagram for explaining a modified example of the proficiency level estimation method according to the present embodiment. [Figure 10] FIG. 10 is a diagram showing an example of a UI screen for selecting an allocation optimization policy on a management terminal according to the present embodiment; [Figure 11] FIG. 10 is a diagram showing an example of a proficiency level confirmation screen on the management terminal according to the embodiment; [Figure 12] FIG. 10 is a diagram showing an example of a work time evaluation screen on a worker terminal according to the present embodiment; DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present disclosure will be described in detail with appropriate reference to the drawings. However, more detailed description than necessary may be omitted. For example, detailed descriptions of well-known matters and redundant descriptions of substantially identical configurations may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art. Note that the accompanying drawings and the following description are provided to enable those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter described in the claims.

[0013] The functions of one configuration shown in this embodiment may be realized by two or more physical configurations, or the functions of two or more configurations may be realized by, for example, one physical configuration.

[0014] (Embodiment 1) <System configuration> FIG. 1 is a block diagram showing an example of the configuration of a work planning system 10 according to this embodiment.

[0015] The work planning system 10 is a system that plans which work should be assigned to workers who will be working on a work day.

[0016] The work planning system 10 includes an ordering system 11, a database 12, a management terminal 13, a worker terminal 14, and an evaluation device 20. These can send and receive information via a communication network 3. Examples of the communication network 3 include a wired LAN, a wireless LAN, the Internet, a mobile communication network, etc.

[0017] The ordering system 11 transmits product ordering information to the database 12 .

[0018] The database 12 manages various types of information as a database. The database 12 manages order information, work information, work performance information, proficiency information, workload information, work plan information, etc. The order information may include the order date for each product, the scheduled shipping date, the product name, etc. The work information may include a worker ID that identifies the worker, the worker's name, the worker's work days, etc. The work performance information may include information indicating the past work performance of each worker. The work performance information may include information indicating the work performed by the worker in the past and the work time required for that work, etc. The proficiency information may include information indicating the proficiency and proficiency improvement value of each worker for a specified period. Details of the proficiency and proficiency improvement value will be described later. The workload information may include information indicating the magnitude of the workload for each task. The work plan information may include information indicating which task to assign to each worker.

[0019] The evaluation device 20 generates work plan information by planning which tasks to assign to each worker working on a work day based on various information registered in the database 12. The evaluation device 20 then registers the work plan information in the database 12. The evaluation device 20 includes a processor 21, a memory 22, a storage 23, a communication unit 24, an input unit 25, and an output unit 26.

[0020] The processor 21 reads out and executes programs and data from the memory 22, etc., thereby realizing the functions of the evaluation device 20. The functions of the evaluation device 20 will be described later. The processor 21 may also be interpreted as a central processing unit (CPU), a microprocessor unit (MPU), a controller, etc. The processor 21 may also include a graphics processing unit (GPU) and / or a neural processing unit (NPU).

[0021] The memory 22 is configured with a volatile storage medium and / or a non-volatile storage medium, and stores programs and data for realizing the functions of the evaluation device 20.

[0022] The storage 23 is configured by a nonvolatile storage medium and stores programs and data for realizing the functions of the evaluation device 20. Examples of the storage 23 include a solid state drive (SSD), a flash memory, and a hard disk drive (HDD).

[0023] The communication unit 24 controls the transmission and reception of information to and from other devices via the communication network 3 .

[0024] The input unit 25 is, for example, a keyboard, a mouse, a touchpad, a microphone, etc., and receives input from the user.

[0025] The output unit 26 is, for example, a display that displays information, images, and a user interface (UI) screen, and is, for example, a speaker that outputs sound and the like.

[0026] The evaluation device 20 has, as its functions, a proficiency estimation unit 31, a proficiency improvement value estimation unit 32, a model learning unit 33, a task allocation unit 34, an operation time prediction model 41, and a proficiency improvement value prediction model 42. In the following description, processing mainly performed by the proficiency estimation unit 31, the proficiency improvement value estimation unit 32, the model learning unit 33, or the task allocation unit 34 may be interpreted as processing mainly performed by the processor 21. Furthermore, in this embodiment, the evaluation device 20 is described as having the operation time prediction model 41 and the proficiency improvement value prediction model 42, but the operation time prediction model 41 and the proficiency improvement value prediction model 42 may be provided by an external server device (not shown) or a service, etc. Details of these functions will be described as appropriate.

[0027] <Operation of the work planning system> 2 is a sequence diagram showing an example of the operation of the work planning system 10 according to the present embodiment. Next, an overview of the operation of the work planning system 10 will be described with reference to FIG.

[0028] <<Work Days Ahead>> The work planning system 10 performs the following operations before the work day. (S11) The user creates work information for the next work day through the management terminal 13 and registers it in the database 12. The work information includes information about the workers who will be working on that work day. (S12) The user instructs the evaluation device 20 via the management terminal 13 to create work plan information for the next work day. (S13) The evaluation device 20 calculates the demand forecast for the work day, the work information for the work day, and the number of people working on the work day. Based on the proficiency information of each worker, work plan information for the next work day is generated and registered in the database 12.

[0029] <<Working days>> The work planning system 10 performs the following operations on a work day. (S14) The ordering system 11 registers the ordering information for the work day in the database 12. (S15) The evaluation device 20 corrects the work allocation information based on the order information for the work day. (S16) The worker performs the work assigned to them based on the work assignment information. The worker terminal 14 records the work content and work time performed by the worker, and registers this in the database 12 as work performance information. (S17) The evaluation device 20 refers to the work performance information, and if the work assignment needs to be revised, for example, because the work is behind schedule, the evaluation device 20 revise the work assignment information.

[0030] <<After work>> The work planning system 10 performs the following operations after the end of the work day. (S18) The evaluation device 20 registers the work performance of all workers on the work day in the database 12 as work performance information. (S19) The evaluation device 20 estimates the proficiency level and proficiency improvement value of each worker based on the work performance information, and registers them in the database 12 as proficiency information.

[0031] <First example of a method for estimating proficiency> FIG. 3 is a diagram for explaining a first example of a method for estimating a proficiency level according to the present embodiment.

[0032] The evaluation device 20 estimates the worker's proficiency in a certain task by the following process. Proficiency is an index that indicates how proficient a worker is in a certain task. In the first example, a method for estimating proficiency that focuses on the distribution of the task time of each individual worker will be described.

[0033] (S101) The proficiency estimation unit 31 calculates a distribution of the work time required by the workers to perform tasks in a first period (hereinafter referred to as a first work time distribution) and a distribution of the work time required by the workers to perform tasks in a second period that is later than the first period (hereinafter referred to as a second work time distribution). The first period and the second period are, for example, one week or one month, and are typically the same length. The proficiency estimation unit 31 may generate the distribution of the work time required by the workers to perform tasks in each of the first period and the second period (i.e., the first and second work time distributions) based on the work performance information registered in the database 12.

[0034] (S102) The proficiency estimation unit 31 estimates the worker's task proficiency based on the difference between the first task time distribution and the second task time distribution. The difference may be the difference between the median of the first task time distribution and the median of the second task time distribution. Alternatively, the difference may be the area of ​​the portion where the first task time distribution (graph) and the second task time distribution (graph) do not overlap when they are superimposed. The proficiency estimation unit 31 estimates a higher proficiency as the difference becomes smaller, and a lower proficiency as the difference becomes larger. In other words, the more stable the task time distribution for different periods is, the higher the proficiency is estimated, and the more unstable the task time distribution for different periods is, the lower the proficiency is estimated.

[0035] By estimating proficiency using this standard, the influence of differences in the physical abilities of individual workers can be suppressed in the evaluation of proficiency. For example, when handling items in high places, the shortest time for a short worker is likely to be longer than the shortest time for a tall worker. However, a longer shortest time does not necessarily mean that the short worker is less skilled at the task than the tall worker. Generally, highly skilled workers accurately understand the task procedures and can perform the task at a stable speed, and their task times are likely to be stable and close to their shortest. On the other hand, less skilled workers may slow down due to unsureness about the task procedures, resulting in unstable task times. In other words, task proficiency is likely reflected in the degree of stability of the task time distribution rather than the length of the task time. Therefore, to estimate this proficiency, the first example uses the degree of stability of the task time distribution.

[0036] As such, the proficiency level in the first example (second proficiency level) is a different index from task speed (i.e., task time), and a worker with a fast task speed (i.e., a short task time) does not necessarily have a high level of proficiency. Proficiency levels based on this standard highly evaluate the proficiency of a worker with a slow but stable task speed. Therefore, if tasks are assigned to workers based on this proficiency level, the task speed may be slow under ideal conditions. However, from the perspective of managing the time required for a task, a stable task speed may be more important than a fast task speed. For example, if a task is assigned to a worker with a fast task speed but large variations in task time, there is a high risk that the task will not be completed within the specified time. In contrast, by assigning tasks to workers based on their proficiency level, as in this embodiment, the risk of completing the entire task on time can be reduced. Note that proficiency level may be interpreted as other terms such as proficiency level, task time stability level, task assignment risk, etc.

[0037] That is, according to the first example, the stability of the work can be evaluated as the level of proficiency, so that the degree to which the worker is familiar with the work can be evaluated excluding the influence of the individual's physical ability, etc.

[0038] <Second example of proficiency estimation method> 4 is a diagram for explaining a second example of a method for estimating a proficiency level according to the present embodiment. In the second example, a method for estimating a proficiency level focusing on the distribution of standard work times of workers is explained. Note that the proficiency level may be estimated by the method of the first example described above, or by the method of the second example described here.

[0039] The evaluation device 20 estimates the worker's proficiency in a certain task through the following process.

[0040] (SX1) The model learning unit 33 learns the work times of all workers for each work type for a predetermined period of time, and generates a work time prediction model for all workers.

[0041] (SX2) The proficiency level estimation unit 31 estimates the work time required for each worker to perform a task using the work time prediction model for all workers. Note that the work time estimated here is not the work time for any of the learned tasks, but the work time for a standard task set based on all tasks learned by the work time prediction model. Hereinafter, this standard task will also be referred to as an average task, but the standard task does not have to be a task for which some average has been calculated for the tasks to be learned. Furthermore, the work time for an average task can be estimated by inputting the worker ID of each worker into the work time prediction model and setting other parameters as neutral values ​​(missing values).

[0042] (SX3) The proficiency estimation unit 31 calculates the percentage ranking of the average work time of the specified worker in the average work time of each of all workers for the average work. In other words, the percentage ranking indicates the ranking of the average work time of the specified worker in the average work time of each of all workers for the average work. The proficiency estimation unit 31 determines the proficiency of the specified worker based on this percentage ranking. The proficiency may be the percentage ranking value itself, or a value obtained by processing the percentage ranking value. In this way, the proficiency estimation unit 31 estimates the proficiency (first proficiency) of the specified worker (specific worker) for a predetermined period.

[0043] Instead of percentage ranking, it is also possible to use the ranking value itself, which indicates the rank of a worker's work time among all the work times of all workers. However, ranking values ​​do not allow comparison between environments with different numbers of workers. For example, the same "third place" may be almost the top in a case where there are 100 workers, but it may be the bottom in a case where there are only three workers. Therefore, in an environment where the number of workers may change, using percentage ranking as in this embodiment allows for a more accurate evaluation of proficiency.

[0044] In addition, the proficiency estimation unit 31 can estimate the extent to which the proficiency of the specified worker has changed from the first period to the second period (i.e., the improvement value of the proficiency) by comparing the proficiency estimated by setting the predetermined period to the first period and the proficiency estimated by setting the predetermined period to the second period in the processes (SX1) to (SX3).

[0045] According to the second example, a standard task is estimated, and proficiency levels are evaluated based on the estimated work time of each worker for that standard task. Therefore, even if the tasks actually performed by each worker are different, the proficiency levels of each worker can be appropriately evaluated.

[0046] <Modification of the second example of the method for estimating proficiency> 5 is a diagram for explaining a modified example of the second example of the proficiency level estimation method according to the present embodiment. In the modified example of the second example, a proficiency level estimation method that focuses on the distribution of working times of a plurality of workers will be explained.

[0047] The evaluation device 20 estimates the worker's proficiency in a certain task through the following process.

[0048] (S201) The model learning unit 33 generates a task time prediction model for all workers using the task time distribution of all workers for each task type.

[0049] (S202) The proficiency level estimation unit 31 uses the task time prediction model for all workers to estimate the task time distribution for each specified task unit. The task unit for which this task time distribution is estimated may be a unit that combines a slip number and a task type, or may be a unit of task type (however, it may include information other than the task type). For example, if a slip number includes the task type "picking" and the product type "pen," and another slip number includes the task type "picking" and the product type "television," when the task unit is a unit that combines a slip number and a task type, these are treated as different task units, and when the task unit is a unit that combines a task type, these are treated as the same task unit. The task unit is the task that is the focus of proficiency estimation.

[0050] (S203) The proficiency estimation unit 31 obtains the actual work time of the specified worker (hereinafter referred to as the target worker) in the specified task unit during a first period and the actual work time during a second period different from the first period from the work performance information registered in the database 12.

[0051] (S204) The proficiency estimation unit 31 calculates a percentage ranking distribution of the actual work time in the first period with respect to the work time distribution in the specified task unit of all workers (hereinafter referred to as the percentage ranking distribution in the first period). In addition, the proficiency estimation unit 31 calculates a percentage ranking distribution of the actual work time in the second period with respect to the work time distribution in the specified task unit of all workers (hereinafter referred to as the percentage ranking distribution in the second period).

[0052] (S205) The proficiency estimation unit 31 estimates the proficiency of the target worker based on the difference between the percentage rank distribution for the first period and the percentage rank distribution for the second period calculated in step S204. As in the first example of the proficiency estimation method, the difference may be the difference between the median of the percentage rank distribution for the first period and the median of the percentage rank distribution for the second period. Alternatively, the difference may be the area of ​​the portion where the percentage rank distribution (graph) for the first period and the percentage rank distribution (graph) for the second period do not overlap when the distributions are superimposed. The proficiency estimation unit 31 estimates a higher proficiency level as the difference becomes smaller, and estimates a lower proficiency level as the difference becomes larger. In other words, the more stable the percentage rank distributions over different periods are, the higher the proficiency level is estimated, and the more unstable the percentage rank distributions over different periods are, the lower the proficiency level is estimated. In other words, the less the fluctuation in the rank of a worker's task speed for a specified task unit among all workers over different periods, the higher the proficiency level is estimated to be, and the more the fluctuation in the rank of a worker over different periods, the lower the proficiency level is estimated to be. In other words, even if a worker has a low rank among all workers, if the fluctuation in the rank is small, the proficiency level of the worker may be estimated to be high.

[0053] According to this modification, the worker's proficiency (second proficiency) is evaluated based on the deviation in the distribution of percent ranks, so the degree of stability of the work time is reflected in the proficiency, as in the first example. Therefore, according to this modification, the stability of the work can be evaluated as the proficiency, so that the degree to which the worker is accustomed to the work can be evaluated excluding the influence of the individual's physical ability, etc.

[0054] <Proficiency improvement value> Next, a method for estimating the proficiency improvement value will be described. The proficiency improvement value is an index that indicates the pace at which a worker's proficiency is improving or is predicted to improve. A higher proficiency improvement value means that the pace of proficiency improvement is faster, and a lower proficiency improvement value means that the pace of proficiency improvement is slower.

[0055] If we only consider shortening the overall work time of the most recent tasks, we can simply assign more work to highly skilled workers, but if we consider shortening the overall work time of future tasks, we need to allow workers with low proficiency to gain work experience and improve their proficiency. In this case, for example, by actively assigning work to workers with high proficiency improvement values ​​and improving their proficiency, it may be possible to shorten the overall work time of future tasks in a shorter period of time.

[0056] Furthermore, each task includes multiple task characteristics. Task characteristics are various characteristics related to the task, such as task time, number of tasks, task type, distance traveled during the task, destination during the task, route traveled during the task, and packaging method during the task. It should be noted that multiple different tasks may include common task characteristics. By analyzing how these task characteristics contribute to changes in the proficiency improvement value and actively assigning tasks to workers that include task characteristics that significantly contribute to improving the proficiency improvement value, the proficiency of the worker can be improved in a short period of time.

[0057] <Learning the proficiency improvement value prediction model> FIG. 6 is a diagram for explaining a learning method and a re-learning method for the skill improvement value prediction model 42 according to this embodiment.

[0058] The evaluation device 20 performs learning of the skill improvement value prediction model 42 through the following process.

[0059] (S301) The model learning unit 33 identifies task features for each of the multiple tasks. For example, the model learning unit 33 identifies task features based on task performance information registered in the database 12. At this stage, it is unknown which task feature among the multiple task features will contribute to improving proficiency. Therefore, the identified task features can be of any type and number as long as they are information obtained from the task performance information and have the potential to contribute to improving proficiency.

[0060] (S302) The model learning unit 33 obtains a distribution of work times during a first period (first work time distribution) and a distribution of work times during a second period (second work time distribution) required for each of a plurality of tasks by each of a plurality of workers. For example, the model learning unit 33 obtains the first work time distribution and the second work time distribution based on the work performance information registered in the database 12.

[0061] (S303) The model learning unit 33 calculates the difference between the first task time distribution and the second task time distribution. This difference is as described in the first or second example of the proficiency level estimation method above.

[0062] (S304) The model learning unit 33 causes the proficiency improvement value prediction model 42 to learn the correlation between the task features identified in step S301 and the difference calculated in step S303. The proficiency improvement value prediction model 42 is, for example, a linear model in which the improvement in proficiency is used as the objective function and each task feature constituting the task is used as an explanatory variable. Here, the linear model is expressed, for example, as the sum of each task feature multiplied by a coefficient. When an actual value of the task feature based on the task performance information is given, the model learning unit 33 learns a coefficient for each task feature such that the improvement value of proficiency approximates the actual value. As a result, a value indicating the degree to which each task feature contributed to the improvement of proficiency is learned as a coefficient.

[0063] As a result, the correlation between the task features and the difference is learned in the proficiency improvement value prediction model 42. In other words, the proficiency improvement value prediction model 42 learns how much each task feature contributes to reducing the difference (i.e., improving the proficiency rate).

[0064] <Re-training the proficiency improvement value prediction model> The evaluation device 20 may perform re-learning on the proficiency improvement value prediction model 42 trained as described above so as to be able to more clearly analyze how much each task feature contributes to reducing the difference (i.e., improving the proficiency rate). However, this re-learning does not have to be performed.

[0065] The evaluation device 20 re-learns the skill improvement value prediction model 42 through the following process.

[0066] (S401) The model learning unit 33 selects multiple task features from among the multiple task features based on the influence they have on changes in the proficiency improvement value in the proficiency improvement value prediction model 42. For example, the model learning unit 33 selects a predetermined number of task features in descending order of their influence on changes in the proficiency improvement value, or task features whose influence values ​​are equal to or greater than a predetermined threshold. The magnitude of the influence of a task feature on changes in the proficiency improvement value may be based on the absolute value of the coefficient of the task feature. For example, in the proficiency improvement value prediction model 42, which is a linear model with proficiency improvement as the objective function and each task feature constituting a task as an explanatory variable, the larger the absolute value of the coefficient of the explanatory variable, the greater the influence the task feature corresponding to that explanatory variable has on changes in the proficiency improvement value.

[0067] (S402) The model learning unit 33 causes the proficiency improvement value prediction model 42 to re-learn the correlation between the multiple task features selected in step S401 and the difference calculated in step S303.

[0068] As a result, the correlation between the difference and task features that have a large impact on changes in the proficiency improvement value is re-learned in the proficiency improvement value prediction model 42. In other words, the proficiency improvement value prediction model 42 is re-learned so that it can more clearly analyze how much each task feature contributes to reducing the difference (i.e., improving the proficiency rate).

[0069] <Estimation of proficiency improvement value> The evaluation device 20 estimates the proficiency improvement value of the worker by the following process using the proficiency improvement value prediction model 42 learned as described above.

[0070] (S501) The input unit 25 receives the designation of a worker whose proficiency improvement value is to be estimated.

[0071] (S502) The proficiency improvement value estimation unit 32 uses the proficiency improvement value prediction model 42 after relearning (or the proficiency improvement value prediction model 42 before relearning) to estimate the proficiency improvement value for each of the multiple tasks when the worker specified in step S501 is assigned to perform each of the multiple tasks.

[0072] This makes it possible to estimate the improvement value of the worker's proficiency for each task.

[0073] <Assigning work to workers> If the expected total work time for a work day is earlier than the target work time (for example, the time until the shipping deadline or the end of work day) by a predetermined amount of time or more (i.e., if there is sufficient work time), the work allocation unit 34 may make the next work allocation in order to improve the worker's proficiency.

[0074] For example, the work allocation unit 34 uses the work time estimated for multiple tasks and multiple workers using the work time prediction model 41 and the proficiency improvement value estimated using the proficiency improvement value prediction model 42 after relearning (or the proficiency improvement value prediction model 42 before relearning) to optimize the allocation of each of the multiple tasks to the multiple workers so as to maximize the proficiency improvement value for the multiple tasks.

[0075] For example, the task allocation unit 34 determines the optimal allocation of a plurality of tasks to a plurality of workers so that the sum (objective variable) of the following item values ​​A to C is minimized. (Item value A) The total working time of the worker estimated using the working time prediction model 41 when the work assignment is performed. (Item Value B) Total work time for standard work estimated based on the improvement in worker proficiency when the work assignment is implemented. (Standard work, for example, represents the average daily shipping content. This item represents the future reduction in work time based on the improvement in proficiency when the worker performs the assigned work being evaluated.) (Item Value C) A penalty value that is added when a constraint condition is violated. The constraint condition may be at least one of Equations 2 to 7, which will be described later.

[0076] Fig. 7 is a diagram showing an example of a calculation formula for optimally allocating tasks to workers so as to minimize the overall task time according to this embodiment. Fig. 8 is a diagram for explaining variables in the formula shown in Fig. 7.

[0077] The work allocation unit 34 allocates work to workers so that formula 1 shown in FIG. 7 satisfies and minimizes the constraints of formulas 2 to 7. Formula 2 represents the constraint that each work must be performed by one worker at a time. Formula 3 represents the constraint that one worker must not perform another work in parallel. Formula 4 represents the time constraint (for example, transmission / shipping time, or the completion / start time of the previous or next process). Formula 5 is 1 if worker k performs work i→j, and 0 if he does not perform the work. Formula 6 represents the constraint that the completion time of all work must be equal to or less than z. Formula 7 represents the constraint that the workload imbalance must be less than a threshold Wth.

[0078] For example, the task allocation unit 34 may first allocate tasks to each worker so as to shorten the overall task time as much as possible, and then modify the task allocation taking into account the improvement of the worker's proficiency by the method of the following steps (A101) to (A102).

[0079] (A101) The work allocation unit 34 sets training priorities for each worker on a work day based on their proficiency. For example, the work allocation unit 34 sets a higher priority for workers with lower proficiency.

[0080] (A102) The work allocation unit 34 reassigns tasks assigned to workers with low priority (for example, priority below a predetermined first threshold) to workers with high priority (for example, priority equal to or greater than a second threshold that is higher than the first threshold), and corrects the work plan information. In other words, the work allocation unit 34 reassigns tasks assigned to workers with high proficiency (for example, proficiency equal to or greater than a predetermined third threshold) to workers with low proficiency (for example, proficiency less than a fourth threshold that is lower than the third threshold), and corrects the work plan information.

[0081] In step (A102), the task allocation unit 34 adjusts the task allocation so that the total task time is equal to or less than the target task time and the workload imbalance constraint is satisfied. This increases the task experience of less skilled workers, and can improve the overall task performance of the workers.

[0082] For example, the work allocation unit 34 may first allocate tasks to each worker so as to shorten the overall work time as much as possible, and then modify the work allocation taking into account the improvement of the workers' proficiency by the method of the following steps (B101) to (B102).

[0083] (B101) The work allocation unit 34 sets a training priority for each worker on a work day based on the proficiency improvement value. For example, the work allocation unit 34 sets a priority so that a worker with a high proficiency improvement value has a high priority.

[0084] (B102) The work allocation unit 34 reassigns tasks assigned to workers with low priority (for example, priority less than a predetermined first threshold) to workers with high priority (for example, priority equal to or greater than a second threshold higher than the first threshold), and corrects the work plan information. In other words, the work allocation unit 34 reassigns tasks assigned to workers with low proficiency improvement values ​​(for example, proficiency improvement values ​​less than a predetermined third threshold) to workers with high proficiency improvement values ​​(for example, proficiency improvement values ​​equal to or greater than a fourth threshold higher than the third threshold), and corrects the work plan information.

[0085] In step (B102), the task allocation unit 34 adjusts the task allocation so that the total task time is equal to or less than the target task time and the workload imbalance constraint is satisfied. This increases the task experience of workers with high proficiency improvement values, and can improve the overall task ability of workers in a short period of time.

[0086] <Modification of the proficiency estimation method> FIG. 9 is a diagram illustrating a modified example of the proficiency estimation method according to the present embodiment. The proficiency may be estimated by the following method. Like the second example, this estimation method focuses on the distribution of work times of multiple workers. However, while the second example focuses on the distribution at different times, this modification estimates the proficiency by focusing on the distribution within each task.

[0087] (S601) The model learning unit 33 generates a task time prediction model 41 for all workers using the task time distribution of all workers for each task type.

[0088] (S602) The proficiency estimation unit 31 uses the task time prediction model 41 for all workers to estimate the task time distribution for all workers for each specified task unit. The task unit for which this task time distribution is estimated may be a unit that combines a slip number and a task type, or may be a task type unit (however, it may include information other than the task type). For example, if one slip number includes the task type "picking" and the product type "pen," and another slip number includes the task type "picking" and the product type "television," when the task unit is a unit that combines a slip number and a task type, these are treated as different task units, and when the task unit is a unit that combines a task type, these are treated as the same task unit. The task unit is the task that is the focus of proficiency estimation.

[0089] (S603) The skill level estimation unit 31 obtains the actual work time of the specified worker (hereinafter referred to as the target worker) for each task unit from the work performance information registered in the database 12.

[0090] (S604) The skill level estimation unit 31 calculates a percentage ranking of the actual work time of the target worker for the task unit relative to the predicted distribution of work times for all workers for the specified task unit.

[0091] (S605) The proficiency estimation unit 31 calculates the target worker's proficiency with the task unit based on the target worker's multiple percentage rankings for the task unit. For example, the proficiency may be the average or median of the multiple percentage rankings, and in this case, the proficiency corresponds to the ranking of the task speed when performing the task unit described above. For example, the proficiency may be the standard deviation of the multiple percentage rankings, and in this case, the proficiency corresponds to the variation in the ranking of the task time when performing the task unit described above. Therefore, in this modified example, it is possible to estimate a proficiency that indicates how proficient the worker is with the task unit compared to all other workers.

[0092] In the above example, the proficiency level is estimated by focusing on one task unit, but the proficiency level may be estimated by focusing on multiple task units simultaneously.

[0093] <UI screen of the management terminal> FIG. 10 is a diagram showing an example of a UI screen for selecting an allocation optimization policy on the management terminal 13 according to this embodiment.

[0094] The evaluation device 20 displays an allocation optimization policy selection UI screen 100, as shown in Fig. 10, on the management terminal 13. The allocation optimization policy selection UI screen 100 displays options of minimizing work time 101, considering worker skill improvement 102, and hybrid 103.

[0095] When the user selects task time minimization 101, the evaluation device 20 assigns tasks to each worker so that the task time for the entire task is minimized. In other words, improvement in the worker's proficiency is not taken into consideration.

[0096] When the user selects "Consider worker skill improvement" 102, the evaluation device 20 assigns tasks to each worker so that the proficiency of each worker is improved as much as possible, within the scope of the total task time not exceeding the target task time. In other words, the improvement of the worker's proficiency is taken into consideration.

[0097] When the user selects Hybrid 103 and inputs an allocation rate (%) in the skill improvement allocation field 104, the evaluation device 20 allocates tasks to each worker so that the proficiency of each worker is improved within a range in which the total task time does not exceed the target task time and the input allocation rate. In other words, the improvement of the worker's proficiency is taken into consideration.

[0098] This allows the user to instruct the evaluation device 20 how to allocate tasks in consideration of improving the proficiency of the workers.

[0099] FIG. 11 is a diagram showing an example of a skill level confirmation screen 110 on the management terminal 13 according to this embodiment.

[0100] The evaluation device 20 displays a proficiency confirmation screen 110 as shown in Fig. 11 on the management terminal 13. The proficiency confirmation screen 110 displays a task selection field 111, a period selection field 112, a graph 113 showing changes in proficiency for each worker, a proficiency improvement value 114, and a priority 115 of the training target.

[0101] When a user selects a task from the task selection field 111 and a period from the period selection field 112, the proficiency confirmation screen 110 displays a graph 113 showing the change in proficiency over the period for several (e.g., three) workers with high proficiency improvement values ​​for the selected task, proficiency improvement values ​​114, and priorities 115.

[0102] This allows the user to easily check workers with high proficiency improvement values.

[0103] The evaluation device 20 may cause the management terminal 13 to display the following screens D1 to D4.

[0104] (D1) Workload confirmation / adjustment screen -You can check the workload of each worker after work assignment. ·Reassign work by adjusting the workload of designated workers to reduce it.

[0105] (D2) Workload imbalance threshold confirmation / adjustment screen Based on the work assignment, the magnitude of the workload (deviation from the average) for each worker is displayed along with the upper limit of the workload imbalance. -If the assigned workload exceeds the upper limit of the skew, a warning is displayed. The upper limit of the skew can be changed to adjust the parameter. - An upper limit for bias can be set for each worker.

[0106] (D3) Plan deviation alert display screen - View progress as work is being performed, comparing it with the estimated time to complete the work when assigned. - The display of planned time for completed work is replaced with actual time, and the deviation from the plan is displayed. An alert is displayed depending on the size of the deviation.

[0107] (D4) Correction plan confirmation / adjustment screen · You can check the estimated completion time and workload imbalance for the correction plan proposed by the evaluation device.

[0108] <UI screen of worker's terminal> FIG. 12 is a diagram showing an example of a task time evaluation screen 200 on the worker terminal 14 according to this embodiment.

[0109] When the worker completes a task and presses the task completion button displayed on the worker terminal 14, the evaluation device displays a task time evaluation screen 200 as shown in Fig. 12 on the worker terminal 14. The task time evaluation screen 200 displays a task time distribution 201 for similar tasks and a task time change 202 for similar tasks.

[0110] The task time distribution for similar tasks 201 displays a graph 203 showing the task time distribution for tasks similar to the task performed by the worker, and a mark 204 showing where the worker's current task time is located in the distribution. This allows the worker to recognize whether their current task time is faster or slower than the task times of other workers.

[0111] The Work Time Change for Similar Tasks 202 displays a graph 205 that shows how the distribution of the worker's work time has changed compared to the past. This allows the worker to recognize how their work ability is improving.

[0112] The evaluation device 20 may cause the worker terminal 14 to display the following screens E1 to E3.

[0113] (E1) Work instruction screen -Display instructions for the next task to the worker.

[0114] (E2) Display screen of work plan and deviation of current work from plan - Displays how far behind (or ahead of) the plan is based on the difference between the work progress and the work time based on the work time forecast. -If work is behind schedule, you can alert your manager.

[0115] (E3) Proficiency display screen -Displays the proficiency level of the worker among all workers. -Displays the level of proficiency for each type of work. -Displays the improvement value of the worker's proficiency. In this way, displaying the proficiency level and the proficiency improvement value motivates the worker to improve their work ability. Furthermore, the proficiency level and the proficiency improvement value may be used to evaluate the worker and reflect them in compensation.

[0116] <Other variations> The proficiency level may be calculated by integrating the proficiency levels estimated by each of the above-described proficiency estimation methods. For example, integrating the proficiency levels may involve calculating a weighted sum of the proficiency levels estimated by each estimation method. In this way, a value that comprehensively evaluates the proficiency levels calculated from multiple perspectives can be used as the proficiency level. For example, the proficiency level based on the first example focuses on the stability of each worker's work time, making it suitable for evaluations where the worker wants to complete the work within the planned time. On the other hand, the proficiency level based on the second example focuses on the worker's work speed, making it suitable for evaluations where the worker wants to complete the work as quickly as possible. However, in a typical work environment, there is a mixture of situations where the worker wants to complete the work within the planned time and situations where the worker wants to complete the work as quickly as possible. Therefore, it is difficult to accurately evaluate the proficiency level estimated using only one of the criteria. In such cases, by setting weights according to the points that the worker wants to prioritize in the work environment, the worker's proficiency can be evaluated using criteria appropriate for the work environment. In the above, an example was described in which a weighted sum of proficiency based on the first example and proficiency based on the second example was taken, but proficiency based on other estimation methods may also be integrated, or three or more proficiencies may be integrated.

[0117] (Summary of this embodiment) The above description of the present embodiment discloses the following techniques.

[0118] <Technology 1> In one embodiment, in an evaluation device (20) having a processor (21) and a memory (22), the processor works in cooperation with the memory to estimate a work time for a standard work by each worker using a work time prediction model that has learned the work times for a plurality of tasks by each of a plurality of workers during a first period, and estimates a first proficiency level of the specific worker during the first period based on the ranking of the work time for the standard work by the specific worker in the distribution of the work times for the standard work by each of the multiple workers. This allows the evaluation device to estimate the first proficiency level of the specific worker based on the ranking of the specific worker's standard work time.

[0119] <Technology 2> In the evaluation device described in Technology 1, the processor further estimates a working time for a standard task by each worker using a working time prediction model that has learned the working times for a plurality of tasks by each of a plurality of workers in a second period that is after the first period, and estimates an improvement value of the first proficiency of the specific worker based on a change between the first proficiency of the specific worker in the first period and the first proficiency of the specific worker in the second period. This makes it possible to estimate the improvement value of the first proficiency level of the specific worker.

[0120] <Technology 3> In the evaluation device described in Technology 1 or 2, the processor further calculates a first distribution which is a distribution of work times required for the specific worker to perform a task in a third period, and a second distribution which is a distribution of work times required for the specific worker to perform a task in a fourth period after the third period, estimates a second proficiency level of the specific worker for the task based on the difference between the first distribution and the second distribution, and integrates the first proficiency level and the second proficiency level to estimate the proficiency level of the specific worker for the task. This allows the evaluation device to estimate the second proficiency level of the specific worker based on the difference in the distribution of the working time of the specific worker. Furthermore, the evaluation device can estimate a proficiency level that is an integration of the first proficiency level and the second proficiency level.

[0121] <Technology 4> In the evaluation device described in Technology 3, the processor estimates the distribution of work times for the specific worker in the third period and the fourth period using a work time prediction model that learns the correlation between workers, tasks, and work time distributions based on the work times required for each of a plurality of tasks by each of a plurality of workers in the third period and the fourth period, and estimates a second proficiency level of the specific worker in the task based on the difference between the estimated distribution of work times in the third period and the estimated distribution of work times in the fourth period. This allows the evaluation device to estimate the second proficiency level of the specific worker based on the difference in the distribution of the working time of the specific worker.

[0122] <Technology 5> In the evaluation device described in Technology 4, the processor estimates an improvement value of the second proficiency level for each of the plurality of tasks when the specific worker performs the tasks, using an improvement value prediction model that has learned a correlation between the characteristics of the tasks and the difference in distribution of work time between the third period and the fourth period, based on the characteristics of each of the plurality of tasks and the distribution of work time required for each of the plurality of tasks by each of the plurality of workers in the third period and the fourth period. This allows the evaluation device to estimate an improvement value of the second proficiency level for each of the plurality of tasks when the tasks are performed by a specific worker.

[0123] <Technology 6> In the evaluation device according to Technology 5, at least one of the task features is a common feature among a plurality of tasks. This allows an estimate of the improvement in a particular worker's proficiency with respect to common task characteristics.

[0124] <Technology 7> In the evaluation device described in Technology 6, the processor generates an improvement value prediction model by re-learning the correlation between a plurality of features selected from the characteristics of the work based on the influence they have on changes in the improvement value in the improvement value prediction model and the difference in distribution of work time between the third period and the fourth period, and uses the re-learned improvement value prediction model to estimate an improvement value of the second proficiency level for each of the plurality of work tasks when performed by the specific worker. This allows the evaluation device to estimate the improvement value of the second proficiency level of the worker by focusing on the characteristics of the task that have a large influence on the change in the improvement value.

[0125] <Technology 8> In the evaluation device described in Technology 7, the improvement value prediction model is a model that re-learns the correlation between a predetermined number of features that have a large influence on changes in the improvement value, or features whose values ​​indicating the influence are equal to or greater than a predetermined threshold, and the difference in distribution of the work time. This allows the evaluation device to estimate the improvement value of the second proficiency level of the worker by focusing on the characteristics of the task that have a large influence on the change in the improvement value.

[0126] <Technology 9> In the evaluation device described in Technology 7 or 8, the processor optimizes the allocation of each of the multiple tasks to the multiple workers by using the task times estimated using the task time prediction model and the improvement values ​​estimated using the improvement value prediction model for the multiple tasks and the multiple workers so as to maximize the improvement values ​​for the multiple tasks. This allows the evaluation device to optimize the allocation of a plurality of tasks to a plurality of workers that maximizes the improvement values ​​for the tasks.

[0127] <Technology 10> In the evaluation device according to any one of Techniques 3 to 9, the difference between the first distribution and the second distribution is the difference between the median of the first distribution and the median of the second distribution. This allows the evaluation device to calculate the difference between the first distribution and the second distribution.

[0128] <Technology 11> In the evaluation device described in any one of Techniques 3 to 9, the difference between the first distribution and the second distribution is the area of ​​the portion where the first distribution and the second distribution do not overlap when the distributions are superimposed. This allows the evaluation device to calculate the difference between the first distribution and the second distribution.

[0129] <Technology 12> In the evaluation device described in any one of Techniques 1 to 11, the processor reallocates a task assigned to a worker whose first proficiency level is equal to or greater than a predetermined first threshold to a worker whose first proficiency level is less than a second threshold that is smaller than the first threshold. In this way, by reassigning the work of a worker with a first high level of proficiency to a worker with a first low level of proficiency, the work experience of the worker with the first low level of proficiency can be increased, and the first level of proficiency can be improved.

[0130] <Technology 13> In the evaluation device described in any one of Techniques 2 to 12, the processor reallocates a task assigned to a worker whose improvement value of the first proficiency level is less than a predetermined first threshold to a worker whose improvement value of the first proficiency level is equal to or greater than a second threshold that is greater than the first threshold. In this way, by reassigning the work of a worker with a low improvement value of the first proficiency to a worker with a high improvement value of the first proficiency, the work experience of a worker whose first proficiency is easy to improve can be increased, and the first proficiency can be improved efficiently.

[0131] <Technology 14> An evaluation method according to one embodiment estimates the work time of a standard task by each worker using a work time prediction model that has learned the work times of multiple tasks performed by each of multiple workers during a first period, and estimates a first proficiency level of a specific worker during the first period based on the ranking of the work time of the standard task by a specific worker within the distribution of the work times of the standard tasks performed by each of the multiple workers. In this way, the evaluation method can estimate the first proficiency level of the specific worker based on the ranking of the specific worker's standard work time.

[0132] Although the embodiments have been described above with reference to the accompanying drawings, the present disclosure is not limited to such examples. It is clear that a person skilled in the art can conceive of various modifications, alterations, substitutions, additions, deletions, and equivalents within the scope of the claims, and it is understood that these also fall within the technical scope of the present disclosure. Furthermore, the components in the above-described embodiments may be combined in any manner without departing from the spirit of the invention. [Industrial Applicability]

[0133] The technology disclosed herein is useful for an apparatus and method for appropriately allocating tasks to each worker. [Explanation of symbols]

[0134] 3. Communication Network 10 Work Planning System 11 Ordering System 12 Databases 13 Management terminal 14 Worker terminal 20 Evaluation equipment 21 processors 22 Memory 23. Storage 24 Communications Department 25 Input section 26 Output section 31 Proficiency Estimation Unit 32 Proficiency improvement value estimation unit 33 Model Learning Department 34 Work Allocation Department 41 Work time prediction model 42 Proficiency improvement value prediction model 100 Selection UI screen 110 Proficiency Check Screen 200 Work time evaluation screen

Claims

1. An evaluation device comprising a processor and a memory, The processor, in cooperation with the memory, estimating a work time for a standard work by each worker using a work time prediction model that has learned work times for a plurality of works by each of a plurality of workers during a first time period; estimating a first proficiency level of the specific worker for the first time period based on a ranking of the work time of the specific worker for the standard task in a distribution of work times of the standard task by each of the plurality of workers; Evaluation equipment.

2. The processor further comprises: estimating a work time for a standard work by each worker using a work time prediction model that has learned work times for a plurality of works by each of a plurality of workers during a second period that is after the first period; estimating an improvement value of the first proficiency level of the specific worker based on a change in the first proficiency level of the specific worker during the first period and a change in the first proficiency level of the specific worker during the second period; The evaluation device according to claim 1 .

3. The processor further comprises: a first distribution that is a distribution of work times required for work by the specific worker during a third period; and a second distribution that is a distribution of the work time required for the specific worker to perform work in a fourth period that is after the third period; estimating a second proficiency level of the specific worker for the task based on a difference between the first distribution and the second distribution; estimating the proficiency of the specific worker for the task by integrating the first proficiency and the second proficiency; The evaluation device according to claim 1 .

4. The processor: using a task time prediction model that learns a correlation between a worker, a task, and a task time distribution based on task times required for each of a plurality of tasks by a plurality of workers in the third time period and the fourth time period, to estimate a task time distribution for the specific worker in the third time period and a task time distribution for the specific worker in the fourth time period; estimating a second proficiency level of the specific worker for the task based on a difference between the estimated distribution of task times in the third period and the estimated distribution of task times in the fourth period; The evaluation device according to claim 3 .

5. The processor: an improvement value prediction model that learns a correlation between the characteristics of each of the plurality of tasks and a difference in distribution of task times between the third period and the fourth period, based on the characteristics of each of the plurality of tasks and a distribution of task times required for each of the plurality of tasks by each of the plurality of workers in the third period and the fourth period, and estimates an improvement value of the second proficiency level for each of the plurality of tasks when the specific worker performs the tasks; The evaluation device according to claim 4 .

6. At least one of the characteristics of the task is a common characteristic among a plurality of tasks. The evaluation device according to claim 5 .

7. The processor: generating an improvement value prediction model by re-learning the correlation between a plurality of features selected from the characteristics of the work based on the influence on the change in the improvement value in the improvement value prediction model and the difference in distribution of the work time between the third period and the fourth period; using the retrained improvement value prediction model to estimate an improvement value of the second proficiency level for each of the plurality of tasks when the tasks are performed by the specific worker; The evaluation device according to claim 6 .

8. The improvement value prediction model is a model that re-learns the correlation between a predetermined number of features that have a large influence on a change in the improvement value, or a feature whose value indicating the influence is equal to or greater than a predetermined threshold, and the difference in distribution of the work time. The evaluation device according to claim 7 .

9. The processor: optimizing allocation of each of the plurality of tasks to the plurality of workers so as to maximize the improvement values ​​for the plurality of tasks, using the task times estimated using the task time prediction model and the improvement values ​​estimated using the improvement value prediction model for the plurality of tasks and the plurality of workers; The evaluation device according to claim 7 .

10. the difference between the first distribution and the second distribution is the difference between the median of the first distribution and the median of the second distribution. The evaluation device according to claim 3 .

11. The difference between the first distribution and the second distribution is an area of ​​a portion where the first distribution and the second distribution do not overlap when the distributions are superimposed. The evaluation device according to claim 3 .

12. The processor: reassigning a task assigned to a worker whose first proficiency level is equal to or greater than a predetermined first threshold to a worker whose first proficiency level is less than a second threshold that is smaller than the first threshold; The evaluation device according to claim 1 .

13. The processor: reassigning a task assigned to a worker whose improvement value in the first proficiency level is less than a predetermined first threshold to a worker whose improvement value in the first proficiency level is equal to or greater than a second threshold that is greater than the first threshold; The evaluation device according to claim 2 .

14. estimating a work time for a standard work by each worker using a work time prediction model that has learned work times for a plurality of works by each of a plurality of workers during a first time period; estimating a first proficiency level of the specific worker for the first time period based on a ranking of the work time of the specific worker for the standard task in a distribution of work times of the standard task by each of the plurality of workers; Evaluation method.

Images