JP2019105887A - Information processor, information processing method and program - Google Patents

Abstract

To automatically obtain an accurate unit work time as it is desirable to be able to set the unit work time as accurate as possible for each process when creating a production plan.SOLUTION: Included are estimation means which estimates a process in each time of a set time interval, and correction means which corrects an estimated value of a unit work time for every process based on an estimation accuracy. The estimation means uses an estimation model in which work performance data on a process acquired in the past are associated with a process actually performed. The correction means improves the estimation accuracy of the estimation model by using measures of recall and precision.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program.

At the production site, create a production plan. The production plan is a plan representing when to use the resources (manufacturing equipment and workers) to produce the product linked to the order. Planning information is required to create a production plan. The plan information refers to information including order information, product manufacturing information, and the like. Here, at a print site where printed matter is printed, the number of copies of an order, the number of pages per copy, the required time per page (unit work time) is multiplied by a process (printing, processing, inspection, etc.) Calculate the working time of Then, a production plan is created by determining the work start time of the process. In order to create an accurate production plan, it is necessary that accurate unit work time be obtained. The correct unit work time refers to the time calculated from the work time actually taken in the past.
As a technique for obtaining an accurate unit work time, there is a technique disclosed in Patent Document 1. In the technology disclosed in Patent Document 1, when determining the required time, the required time of the work calculated from the past record information is corrected based on the information specified by the user.

JP, 2015-5191, A

  However, in the technique disclosed in Patent Document 1, the user has to manually input information necessary to obtain an accurate unit work time.

  An information processing apparatus according to the present invention comprises estimation means for estimating a process at each time point of a set time interval, and correction means for correcting an estimated value of a unit work time for each process based on the accuracy of the estimation. .

  According to the present invention, accurate unit work time can be obtained automatically.

It is a figure which shows the example of the data regarding plan information and machine learning. It is a figure showing an example of the hardware constitutions of an information processor. FIG. 2 is a diagram showing an example of an information correction application of the first embodiment. It is a flowchart which shows an example of the information processing of an information correction application. It is a flowchart which shows an example of the information processing of a correction | amendment part. It is a figure which shows an example of the information correction application of Embodiment 2. FIG. It is a figure which shows an example of plan information. It is a flowchart which shows an example of the information processing of an information correction application. It is a flowchart which shows an example of the information processing of a representative value determination part. It is a flowchart which shows an example of the information processing which creates a plan. It is a figure which shows an example of a planning screen. It is a flowchart which shows an example of the information processing in an adjustment part. It is a Gantt chart which shows an example when it applies to equalization. It is a flowchart which shows an example of the information processing at the time of applying to optimization.

  Hereinafter, embodiments of the present invention will be described based on the drawings.

First Embodiment
A process of automatically determining an accurate unit work time and correcting it without human intervention will be described. The correction method of unit work time will be described below.
(Planning information)
First, an example of a data structure of plan information storing information necessary for automatically determining and correcting an accurate unit work time will be described using FIGS. 1 (a) to (f).
FIG. 1A shows an example of the data structure of the product table 101.
The field is configured of a product ID, a product name, and a process ID.
The item ID is an ID for identifying each item. The product ID is registered so as to be unique in this field. In other words, the same ID is registered so as not to appear in this field.
As a method of guaranteeing the uniqueness of the process ID, for example, a unique constraint may be set in this field using a database management system, and re-input may be prompted when duplicate contents are registered. Hereinafter, the first field of each table is a primary key, and uniqueness is guaranteed.
The product name is the name of the product.
Process ID is process ID of the first process for producing goods. The process ID refers to the process ID field of the process table 102 described later.

FIG. 1B shows an example of the data structure of the process table 102.
The fields are composed of process ID, process name, resource ID, next process ID, unit work time, and estimation accuracy ID.
The process ID is an ID for identifying a process.
The process name is the name of the process.
The resource ID is a resource ID of a resource used in the corresponding process. The resource ID refers to the resource ID field of the resource table 103 described later.
The next process ID is a process ID of the process next to the corresponding process. The next process ID refers to the process ID field of the process table 102.
The unit work time is a unit work time of the corresponding process.
The estimation accuracy ID is an estimation accuracy ID of an estimation accuracy table 105 described later. Represents the estimation accuracy of the corresponding process. The estimation accuracy ID refers to an estimation accuracy ID field of an estimation accuracy table 105 described later.

FIG. 1C shows an example of the data structure of the resource table 103.
The field consists of resource ID and resource name.
The resource ID is an ID for identifying each resource.
The resource name is the name of a resource.

FIG. 1D shows an example of the data structure of the unit work time estimated value table 104. This represents the estimated value of the unit work time of the process corresponding to the process ID field of this table.
The field is composed of a unit work time ID, a process ID, and a unit work time estimated value. The details of the unit work time estimated value will be described later.
The unit work time ID is an ID for identifying each unit work time.
The process ID is a process ID of a process in which a unit work time is linked. The process ID refers to the process ID field of the process table 102.
The unit work time estimated value is an estimated value of the unit work time of the process corresponding to the process ID.

FIG. 1E is an example of the data structure of the estimation accuracy table 105.
The field is composed of an estimated accuracy ID, a recall rate, and a precision rate. Details of the estimation accuracy, the recall rate, and the accuracy rate will be described later in the section on estimation accuracy.
The estimation accuracy ID is an ID for identifying each estimation accuracy.
The recall is a recall relating to the process associated with the estimation accuracy ID of the process table.
The accuracy rate is a accuracy rate related to the process linked by the estimation accuracy ID of the process table.

FIG. 1F shows an example of the data structure of the order table 106.
The fields include an order item ID, a product ID, an order quantity, a page number, a manufacturing deadline, and an order date and time.
The order item ID is an ID for identifying each order item.
The item ID is the item ID of the item ordered by the order item. The product can be identified by referring to the product ID field of the product table described above with reference to FIG. 1 (a).
The order quantity is the quantity of the ordered item. The unit is the number of copies.
The number of pages is the number of pages constituting a product.
The production deadline is the deadline for completing the production of the ordered product.
The order receipt date is the date when the order was accepted.
That is, the unit work time estimated value is managed for each product-specific process.

(Machine learning)
In this embodiment, the process for each order at each time point is estimated using an estimated model created by machine learning. Here, machine learning is a method of empirically learning data acquired in the past and interpreting newly acquired data based on a learning result. Further, the estimation model is information of a determination criterion generated by empirically learning the relationship between data acquired in the past and a process for each order at each time point.
First, the procedure for creating an estimation model will be described.
In order to create an estimation model, data (feature amount data) obtained by digitizing features included in the data and correct answer data are input as learning data for machine learning. In the present embodiment, the position, orientation, posture of the worker at a predetermined time interval, the order ID of the product under work at each work place, the operation log of the manufacturing apparatus, etc. are input as learning data of machine learning as feature quantity data. . The order ID of the item under work at each work place is acquired, for example, as follows. When the work process is a printing process, a camera is installed above the output tray of the printing press, and the printed matter output to the output tray is photographed. The photographed image is compared with the image of the print data on the printer server, and the matching print data is specified. The order ID is specified by specifying the product linked to the specified print data from the list of orders. If the work process is a processing process, the camera captures the printed matter output to the output tray to which the cut printed matter is output if it is a cutting machine, or to the output tray to which the bound printed matter is output if it is a bookbinding machine Do. Thereafter, as in the process of the printing process, the photographed image is compared with the image of the print data on the printer server, and the matching print data is specified. The order ID is specified by specifying the product linked to the specified print data from the list of orders. By doing this, the order ID of the product under work at each work place is acquired.
As another method, it acquires as follows. Print a barcode in which the order ID is encoded on the work instruction in which the work content is written. Place work instruction sheet in the work place of each process. When working at the work place, the workers set operation rules to place work instructions in the work instruction storage area. Then, a camera is installed at a position where the barcode of the work instruction can be read at each work instruction storage place. The camera is ready to read the barcode. At the time of operation, when a worker performs the work of the process, a work instruction is placed in the work instruction storage area. The camera reads the barcode of the work instruction sheet, and associates information on the process of work currently being performed with the order ID obtained by decoding the barcode. By providing such a mechanism, the order ID of the product under work at each work place may be acquired.

As another method, an order ID in operation at each work place is acquired using an RFID (Radio Frequency Identification) reader (an apparatus for reading an RF tag described later) and an RF tag (an information medium in which ID information is embedded). You may do so. More specifically, an RFID reader is installed at the work place of each process. Also, write the order ID on the RFID tag and attach it to the work instruction. A work instruction storage room will be provided at each work site, and an RFID reader will be provided at the work instruction storage area. As an operation rule, workers shall place work instructions at the work instruction storage area at the start of work in each process. At the end of the work, the work instruction sheet is taken out from the work instruction sheet storage and taken to the work place of the next process or the collection place of the work instruction sheet. By providing such a mechanism, it is possible to specify the order ID under work at each work place. In this way, the order ID of the item under work at each work place may be acquired.
At the same time, the process of the correct answer for each order at each time interval is input as learning data of machine learning as correct data. FIG. 1 (g) shows an example of learning data. FIG. 1 (g) is a learning data table showing an example of learning data. The learning data table is composed of the following fields. The record ID, time, x coordinate of start position, y coordinate, x coordinate of end position, y coordinate, direction at start, direction at end, total amount of change in direction at time interval, order ID, correct answer, etc. The result field is an ID for identifying each result. The time field is the time when each record was acquired. In the example of FIG. 1 (g), the results are acquired at one second intervals. The x-coordinate field and the y-coordinate field of the start position represent the position of the worker at the printing site at the start of each time. For example, if it is record TR000001, it will indicate the position of the worker at time 2017/6/1 09:00:00. Here, with respect to the coordinate values, the distance in the lateral direction from there is a x-coordinate, and the distance in the vertical direction is y-coordinate, with the position at the printing site as the origin. The coordinate value may be a value obtained by scaling the distance from the origin, or may be any expression as long as the position of the worker can be grasped. The x-coordinate field and y-coordinate field of the end position represent the position of the worker at the printing site at the end of each time. For example, in the case of the record TR000001, the position of the worker at time 2017/6/1 09:00: 01 is represented. The start orientation field and the end orientation field are information on the direction in which the worker is pointing. The start direction field indicates the direction in which the worker is facing at the start of each time. The end direction field indicates the direction in which the worker is facing at the end of each time. The total change in orientation field in the time interval represents the total change in worker orientation during each time interval, i.e. 1 second. The order ID field represents an order ID associated with the work instruction recognized at the work instruction storage area closest to the worker at each time point. The correct answer field is the process ID of the process that the worker is working at each time point, that is, the information on the correct answer that the machine learning is desired to learn. These learning data are one example, and feature amount data that may increase the estimation accuracy of the estimation model may be used.
By inputting these pieces of information as learning data for machine learning, it is possible to obtain an estimation model that estimates steps in a predetermined time interval for each order.

  Next, a method of obtaining an estimation result using the created estimation model will be described. In order to obtain an estimation result using an estimation model, feature amount data is input to the estimation model. The feature amount data is data excluding the correct answer field of the learning data table 111. By inputting the feature amount data, the processes at predetermined time intervals can be estimated for each order. The example of the table of estimation result data is shown in FIG. 1 (h). In FIG. 1 (h), the estimation result data table 112 is an estimation result data table showing estimation result data. A field consists of an achievement ID and an estimation result. The result ID field corresponds to the result ID field of the learning data table 111. The estimation result field represents the process ID of the process at each time point estimated by the estimation model. In the example of FIG. 1 (h), in the actual results IDTR000001 and TR000002, the estimation result is OP001, which matches the correct answer in the learning data table. Therefore, the estimation model can be estimated correctly. On the other hand, in the actual results IDTR000003 and TR000004, the estimation result is OP101, which does not match the correct answer in the learning data table. Therefore, the estimation model can not be estimated correctly.

(Estimated accuracy)
By inputting feature quantity data (test data) with a correct answer to the estimated model, it is possible to obtain the estimation accuracy of the estimated model from the estimation result estimated by the estimated model and the input correct answer. Among the correct answers, there is a recall rate as an index for evaluating how much the solution is estimated as a correct solution, and a precision rate as an index for evaluating how much the correct solution is included in the estimation. For example, when the number of inspection processes is six in a predetermined time interval, it is assumed that the estimation model estimates that the number of inspection processes is five. And suppose that four of these were correct. At this time, the recall rate is about 67% for "the number 4 of the correct answers that the estimation result was also the correct answer" ÷ "the number of correct answers 6". The accuracy rate is 80% as “the number of correct answers in the estimated correct solution 4” ÷ “the estimated number of correct solutions 5”. In this embodiment, the unit work time is corrected using this estimation accuracy to calculate an accurate unit work time. The calculation method will be described in the image of the process of the present embodiment described later.

(Image of processing of this embodiment)
The process of this embodiment consists of two phases. The first phase is a phase in which daily work results are accumulated. The second phase is a phase that corrects and updates the unit work time for each predetermined period from the accumulated work results.
In the first phase, work results obtained from production activities at a production site are accumulated for a predetermined period, for example, one day. The accumulated work results are input to an estimation model, and processes are estimated for each order at predetermined time intervals, such as every one second. Taking (g) in FIG. 1 and (h) in FIG. 1 as an example, the accumulated work record, that is, information excluding the correct answer field in the learning data table 111 is input to the estimation model. Then, as shown in the estimation result data table 112, information of a pair of the result ID and the estimation result is obtained.
Next, the number of processes is counted by order and by process, and the work time is calculated by order and process by multiplying the predetermined time interval. In the example of the estimation result data table 112 of FIG. 1 (h), the number is counted for each process type such as OP001, OP002, OP101, OP102. If it is assumed that there are 700 OP001's, then by multiplying 700 by a predetermined time interval, 1 second, the working time of the process OP001 can be calculated to be 700 seconds. The unit work time is calculated by dividing the calculated work time by the order quantity. For example, if the process OP001 is a printing process for printing 700 copies of a leaflet product, the operation time for printing one sheet by calculating 700 seconds per 700 seconds of operation time, that is, a unit operation time It can be calculated as one second. Finally, the calculated unit work time is stored in the database for each order and for each process. These processes are repeated until sufficient data are accumulated.
Next, in the second phase, a representative value of unit work time is calculated by statistical processing from the accumulated frequency distribution of unit work time for each order and process. The statistical processing may use desired statistical processing for calculating statistical values such as the mode value, the average value, and the median value. By applying the estimation accuracy for each product process to the calculated unit work time, it is possible to correct the correct unit work time. More specifically, the correct unit work time can be calculated by multiplying the calculated unit work time by the precision rate and dividing the recall rate.
This is from the definition of recall and precision
・ The lower the recall rate is, the more correct answers are missed. As the lower the precision rate, more correct answers are estimated. Therefore, the correct rate is calculated by applying the recall rate and the precision rate to correct the estimated value. be able to.
More specifically, since the calculated unit work time is an estimated value and the number of correct answers estimated by the denominator of the precision rate, that is, the estimated value, the estimated unit work time is estimated by multiplying the precision rate. The value and the number of correct answers estimated by the denominator cancel each other, and the number of correct answers in the estimated correct solution remains. By dividing the recall rate by this, the number of correct answers in the correct answer and the number of correct answers among the correct answers cancel each other out, and the number of correct answers in the denominator of the recall rate remains finally become. As a result, the correct unit work time can be calculated back by the above calculation.

(Hardware configuration of information processing apparatus)
FIG. 2 is a diagram illustrating an example of a hardware configuration of the information processing apparatus.
A CPU (Central Processing Unit) 201 reads, interprets, and executes a program stored on the storage device 206, and controls the RAM 203, the display 204, and the like connected to the bus via the bus 209.
A ROM (Read Only Memory) 202 is a read only memory that performs input / output control with hardware connected to a bus.
A random access memory (RAM) 203 is a random access memory that temporarily stores information when the CPU 201 performs an operation.
The display 204 is a display for displaying the calculation result by the CPU 201.
The keyboard 205 receives an input from the user. However, when the information processing apparatus is a smart device, since the touch panel is provided, the keyboard 205 may not be attached.
The storage device 206 is a storage device for storing programs and data used for calculation.
The network I / F 207 is a network I / F (I / F: interface) connected to a network such as a local network, an intranet, or the Internet.
The pointing device 208 is a pointing device such as a mouse or a trackball that receives coordinates designated by the user on the screen.
A bus 209 is a signal line to which hardware from the CPU 201 to the pointing device 208 is connected.
When the CPU 201 executes a process based on a program stored in the storage device 206, the function of an information correction application (hereinafter, information correction application) 301 shown in FIGS. 3 and 6 described later is realized. Also, the processing of the flowcharts of FIG. 4, FIG. 5, FIG. 8 to FIG. 10, FIG. 12 and FIG. 14 to be described later is realized by the CPU 201 executing the processing based on the program stored in the storage device 206.

(Information correction application)
FIG. 3 is a view showing an example of the information correction application 301. As shown in FIG.
The work result holding unit 302 holds the collected work results as a result of the production activities at the production site. Hold the estimated model. Note that the estimation model may be held in a storage device inside the information processing apparatus, or the information correction application may acquire one held in a storage device external to the information processing apparatus.
The process estimation unit 303 estimates processes in predetermined time intervals for each order using an estimation model.
The unit work time creation unit 304 creates a unit work time for each product-based process from the estimated process.
The plan information holding unit 305 holds the plan information.
The recording amount determination unit 306 determines the recording amount of the unit work time held in the plan information holding unit 305.
The representative value determination unit 307 determines a representative value from the unit work time stored in the plan information storage unit 305.
The correction unit 308 calculates an original unit work time (correction value) from the representative value of the unit work time and the estimation accuracy of the estimated model.
The updating unit 309 updates the unit work time per product-specific process held in the plan information holding unit 305 with the correction value.

(Flow of processing in the information correction application 301)
FIG. 4 is a flowchart showing an example of the information processing of the information correction application 301.
In step S401, the process estimation unit 303 inputs work results into the estimation model, and estimates the processes at predetermined time intervals for each order. More specifically, it is as follows. The process estimation unit 303 acquires the estimated model from the plan information holding unit 305, and acquires the work result from the work result holding unit 302. The process estimation unit 303 inputs the work result to the estimation model, and acquires an order-specific process at a predetermined time interval.
In S402, the unit work time creation unit 304 calculates an estimated value of the unit work time of the process for each order. More specifically, it is as follows. First, the unit work time creation unit 304 counts the number of processes estimated for each order. The unit work time creation unit 304 multiplies the number by a predetermined time interval, and calculates the work time of the process for each order. The unit work time creation unit 304 converts the calculated work time into a unit work time. For example, in the case of a printing process, the unit work time creation unit 304 divides the number of copies and the number of pages to calculate an estimated value of the unit work time. The unit work time creation unit 304 also divides the number of copies and the number of pages in the cutting process to calculate an estimated value of the unit work time.
In S403, the recording amount determination unit 306 determines whether or not the estimated value of the unit work time has been sufficiently stored for all steps. As a result of the determination, if the information is sufficiently stored, the recording amount determination unit 306 stores the process in the memory, and proceeds to S404. On the other hand, if there is no process accumulated enough, the recording amount determination unit 306 returns to S401. The recording amount determination unit 306 determines whether storage has been sufficiently performed by comparing a predetermined number of thresholds with the number of recording amounts, and determines that storage is sufficient if the recording amount is equal to or more than the threshold.
In S404, the representative value determining unit 307 determines a representative value from estimated values of a plurality of unit work times for each process stored in the memory, and stores the representative value in the memory. The representative value determination unit 307 sets the representative value as the average value of the estimated values.
In S405, the correction unit 308 calculates a correction value from the above-described representative value using the estimation accuracy of the estimation model. Details of the process of this step will be described later with reference to FIG.
In S406, the updating unit 309 stores the calculated correction value in the plan information holding unit 305 for each product-specific process.

(Flow of processing in correction unit 308)
FIG. 5 is a flowchart showing an example of the information processing of the correction unit 308.
In S501, the correction unit 308 acquires, from the memory, a representative value of the unit work time of the process to be corrected. Here, the process to be corrected refers to a process determined that the above-mentioned recording amount has been sufficiently stored.
In S502, the correction unit 308 acquires, from the plan information holding unit 305, the recall rate of the estimated model of the corresponding process.
In S503, the correction unit 308 acquires, from the plan information holding unit 305, the relevance ratio of the estimation model of the corresponding process.
In S504, the correction unit 308 corrects the representative value of the unit work time of the corresponding process from the acquired recall rate and the relevance rate, and calculates a correction value. More specifically, the correction unit 308 corrects by “the representative value × the precision ratio × the reproduction rate”. The precision and recall are real numbers from 0 to 1.

  As described above, according to the technology described above, it is possible to correct the unit work time calculated from the estimation result of the estimation model from the relevance ratio and the recall ratio of the estimation model for each product and process and calculate an accurate unit work time. it can.

Second Embodiment
In the first embodiment, a unique value (correction value) obtained by correcting a representative value in a frequency distribution is used as a unit work time at the time of creating a plan. In this embodiment, a value obtained by adjusting the above-mentioned correction value according to the degree of variation of the frequency distribution is used as a unit work time at the time of creating a plan. That is, it is possible to control the probability that the work will be completed as planned by creating a plan in consideration of how much the unit work time used in the first embodiment may exceed with how much probability. The second embodiment shows an example.

(Information correction application)
FIG. 6 is a diagram showing an example of the information correction application 601. As shown in FIG. About the processing part of the same processing content as the information correction application 301 of Embodiment 1 among the information correction application 601, the same code | symbol is attached | subjected and description is abbreviate | omitted.
In addition to the information held by the plan information holding unit 305, the plan information holding unit 605 holds the variance of the frequency distribution regarding the estimated value of the unit work time for each product-specific process. FIG. 7 is a diagram showing an example of the plan information. The process table 701 is an example of a process table held by the plan information holding unit 605. The process table has a process ID, a process name, a resource ID, a next process ID, a unit work time, a variance, and an estimation accuracy ID. The process ID is an ID for identifying a process. The process name is the name of the process. The resource ID is a resource ID of a resource used in the corresponding process. The next process ID is a process ID for identifying the next process of the corresponding process. The unit work time is a unit work time in the corresponding process. The variance is the variance of the frequency distribution with respect to the estimated value of the unit work time of the corresponding process. The estimation accuracy ID is an estimation accuracy ID of estimation accuracy for estimating a process corresponding to the estimation model. The resource table 702 is an example of a resource table held by the plan information holding unit 605. The resource table has a resource ID and a resource name. The resource ID is an ID for identifying each resource. The resource name is the name of each resource.

In addition to the processing of the representative value determination unit 307, the representative value determination unit 607 calculates the variance from the frequency distribution of the unit work time for each product process held in the plan information holding unit 605.
The updating unit 609 updates the unit work time and the variance held in the plan information holding unit 605 with the corrected unit work time and the variance.
The receiving unit 610 receives an instruction from the user regarding the planned safety level. The planned safety is an index indicating the degree of possibility that each process can be completed within the planned time.
The adjustment information acquisition unit 611 acquires adjustment information for adjusting a unit work time from a user interface (hereinafter, UI).
The adjustment unit 612 adjusts the unit work time based on the adjustment information.
The planning unit 613 creates a production plan using the adjusted unit work time. The planning unit 613 may be included in the information correction application 601, or may not be included as shown in FIG. When the planning unit 613 is not included in the information correction application 601, the function of the planning unit 613 is realized by the CPU 201 executing processing based on a program stored in the ROM 202 or the storage device 206.

(Flow of processing in the information correction application 601)
FIG. 8 is a flowchart showing an example of information processing in the information correction application 601.
In S701, the representative value determination unit 607 calculates the variance from the frequency distribution of the unit work time associated with the process for all the processes stored in the memory. The details of the process of this step will be described later with reference to FIG.
In S702, the updating unit 609 updates the value of the distribution field held by the plan information holding unit 605 with the value of the dispersion for all steps stored in the memory. The details of the process of this step will be described later with reference to FIG.

によって算出する。Ｎは分布の全要素数である。
Ｓ９０３において、代表値決定部６０７は、分布から分散を算出し、メモリに記憶する。代表値決定部６０７は、分散を、

によって算出する。 (Flow of processing in representative value determination unit 607)
FIG. 9A is a flowchart illustrating an example of information processing in the representative value determination unit 607.
In step S901, the representative value determination unit 607 obtains, from the plan information holding unit 605, the distribution of estimated values of unit work time by product type for all the processes stored in the memory. The unit work time for each product process is ti. Here, i is an element number of a set of unit work time distributions.
In S902, the representative value determination unit 607 calculates an average value from the distribution and stores the average value in the memory. The representative value determination unit 607 calculates the average value μ as

Calculated by N is the total number of elements of the distribution.
In S903, the representative value determination unit 607 calculates the variance from the distribution, and stores the variance in the memory. The representative value determination unit 607 determines the variance

Calculated by

(Flow of processing in update unit 609)
FIG. 9B is a flowchart illustrating an example of information processing in the updating unit 609.
In step S 911, the updating unit 609 acquires, from the plan information holding unit 605, the unit work time and the process ID of the process for which the distribution is to be updated.
In S912, the updating unit 609 updates the unit work time associated with the acquired process ID with the average value stored in the memory.
In S913, the updating unit 609 updates the distribution associated with the acquired process ID with the distribution stored in the memory.

(Flow of processing in the information correction application 601)
In the processing up to S702 in FIG. 8, the update of the plan information is completed. Next, the flow of processing for creating a plan using the updated plan information will be described below with reference to FIG. FIG. 10 is a flowchart illustrating an example of information processing for creating a plan.
In S1001, the reception unit 610 displays on the display 204 a planning screen for receiving an instruction of the planned safety degree and an instruction of the planning from the user. FIG. 11 is a diagram showing an example of the planning screen 1101.
The slider bar 1102 is a slider bar for causing the user to specify the degree of planned safety.
The indicator 1103 is an indicator for causing the user to designate the degree of planned safety. The indicator 1103 can be moved within the range of the slider bar, and the planned safety is set lower as it is located on the left, and higher as it is located on the right.
The plan creation button 1104 is a button for receiving a plan creation instruction from the user.
In S1002, the reception unit 610 stands by until the plan button is selected.
In S1003, the adjustment information acquisition unit 611 acquires the relative position corresponding to the position of the indicator of the slider bar as a coefficient for adjusting the unit work time.
In S1004, the adjustment unit 612 acquires the variance between the relative position and the product-by-product process, and uses them to adjust the unit work time. The details of the process of this step will be described later with reference to FIG.
In S1005, the planning unit 613 creates a plan.

(Flow of processing in adjustment unit 612)
FIG. 12 is a flowchart illustrating an example of information processing in the adjustment unit 612.
In S1201, the adjustment unit 612 obtains a relative position k corresponding to the position of the indicator. The relative position is a real number from 0 to 1. K = 0 when the indicator is at the far left, k = 1 when it is at the far right, and k = 0.5 when it is at the middle.
In S1202, the adjustment unit 612 acquires the product corresponding to the order information from the plan information holding unit 605.
In S <b> 1203, the adjustment unit 612 acquires, from the plan information holding unit 605, a process in which the product is linked.
In S1204, the adjustment unit 612 obtains a unit work time μ of the process.
In S1205, the adjustment unit 612 obtains the process variance σ.
In S1206, the adjustment unit 612 adjusts the unit work time of the process using the relative position k, the unit work time μ, and the variance σ. The adjustment unit 612 calculates the unit work time by t = μ + k√σ, and updates the unit work time of the plan information holding unit 605 with this value.

Other application examples of the technology of the second embodiment will be described below.
Depending on the production site, processing is performed to equalize the load for each operation day. An example of leveling will be described using FIG. FIG. 13A is a Gantt chart in which processes are allocated to one resource. In FIG. 13A, start dates and times 1301, 1302, and 1303 indicate start dates and times of the first week, the second week, and the third week, respectively. In addition, rectangles of steps 1304, 1305, 1306, etc. represent respective steps. In FIG. 13A, it can be seen that the total time of the steps allocated to the first and second weeks is almost the same, and the load is leveled. However, there is a possibility that each process may be delayed, and as a result of starting work according to this plan, it may be delayed as shown in FIG. 13 (b) (the second week process is omitted) . The operation results 1314, 1315, 1316 are operation results corresponding to the steps 1304, 1305, 1306, respectively. If it is delayed and replanned, the plans for the following week will also change. Then, adjustments will occur as the plan changes. Since adjustment takes time and effort, it is preferable to avoid changing the plan after the next week as much as possible. In such a case, it is conceivable to make a plan with room for the working time of the process.
However, it is difficult to determine how much margin should be given.
Therefore, by using the dispersion, the information correction application 601 can set the margin long for a process with large variance, and set the margin short for a process with small variance. The plan when leveling is performed in consideration of dispersion is shown in FIG. 13 (c). Step 1326 is a step corresponding to step 1306 in FIG. 13 (a), and is a step assigned to the second week in consideration of the possibility of delay.

In addition, another application example will be described below using FIG.
FIG. 14 is a flowchart showing an example of information processing applied to optimization.
In S1401, the adjustment unit 612 acquires the result of planning by the planning unit.
In S1402, the adjustment information acquisition unit 611 acquires a value corresponding to the position of the indicator on the slider bar.
In S1403, the adjustment unit 612 calculates a unit work time corrected from the unit work time, the variance, and the position of the indicator.
In S1404, the adjustment unit 612 corrects the plan result using the corrected unit work time. At the time of correction, the adjustment unit 612 maintains and corrects the process context. For example, the adjustment unit 612 maintains and corrects the context of the process in which the cutting process can be started only after the printing process is completed, and the bookbinding process can be started only after the cover printing and the text printing are completed.
In S1405, the adjustment unit 612 calculates the degree to which the corrected plan fits in one day. The calculated degree is used as one evaluation item of the objective function in scheduling using a combination optimization technique. An objective function is a function that is maximized or minimized under certain constraints. In the scheduling, evaluation items include that the completion date of the product does not exceed the delivery date of the order and that the lead time is short. The adjustment unit 612 simultaneously evaluates the above degree.

Introduction of planned safety is also effective in the following cases.
・ I would like to reduce the impact on the downstream process If there is a department in charge of the upstream process and the downstream process, and the work of the upstream process is delayed, the impact on the downstream process occurs. By setting and creating the planned safety degree of the upstream process plan high, even if the upstream process is delayed, the influence on the downstream process can be reduced. For example, when the printing company carries out the printing process at its own company and the post-processing such as die cutting is outsourced to the processing company, the processing company prepares the workers, the work place and the manufacturing apparatus that can only receive the outsourcing. Etc. However, if the work at the printing company is delayed and the provision of the printed matter to the processing company is delayed, the preparation at the processing company will be wasted until the arrival of the printed matter. When creating a plan at a printing company, it is possible to reduce waste at the processing company by setting the planned safety degree high.
・ I would like to start work immediately after the completion of work in a machine At a production site using a 3D printer, after modeling a 3D model with a 3D printer, if the model is left in the 3D printer for a long time, the modeling accuracy Can go down. Therefore, as for the timing of completion of modeling, the time zone in which the worker works is preferable. Therefore, at a production site using a 3D printer, a planned safety level is set high to plan so as to ensure time for which post-processing such as extraction can be performed within the working hours of workers. By doing this, it is possible to prevent a decrease in modeling accuracy.
・ I want to complete the work on time. In a production site where it is strongly required to complete the work on time, it is not desirable that the work planned on the day is not completed on time. If a plan is created without ample work time at such a site, once the work is delayed, it will be difficult to complete the work on time. Even in such a site, by setting the planned safety degree high and creating a plan, it is possible to create a plan with high possibility of completing the work within a fixed time.

  As described above, according to the technology described above, it is possible to automatically obtain an accurate unit work time by correcting the unit work time using the recall rate and the relevance rate of the estimated model.

Embodiment 3
In the embodiment described above, the estimation accuracy of the estimation model is acquired from the plan information holding unit. However, the method of obtaining the estimation accuracy is not limited to this. The information correction application causes the display 204 to display a UI (input receiving unit) that accepts specification of the recall ratio or the relevance ratio, or both from the user, and acquires the recall ratio or the relevance ratio, or both. It is also good.
Further, in the embodiment described above, the representative value for correcting the unit work time (in other words, the unit work time before the correction) is obtained by calculation. However, the present invention is not limited to this, and the information correction application may display a UI (input reception unit) that receives specification of an estimated value from the user on the display 204 and may also acquire the estimated value.
In the embodiment described above, although the example of specifying the planned safety degree of the planning screen with the slider bar is shown, the specified method is not limited to this. The information correction application may cause the user to select from candidate values in the pull-down list, or may directly input numerical values. In addition, the value of the planned safety degree may be described in a file, and the information correction application may read it. Furthermore, the information correction application may have any form as long as it can obtain the value of the planned safety.
In the embodiment described above, the adjustment unit 612 calculates the unit work time t by t = μ + k√σ, but the calculation method is not limited to this, and for example, using 2√σ or 3√σ, A factor may be multiplied by √σ, such as t = μ + 2k√σ, t = μ + 3k√σ. By doing this, it is possible to create a safer plan.
In the embodiment described above, the adjustment unit 612 does not particularly refuse the scope of the correction of the planning result of S1404. However, for example, the adjustment may be performed by limiting the process to the critical path. Alternatively, the adjustment unit 612 may correct only the large dispersion process. Of course, the adjustment unit 612 may combine both.
Further, in the embodiment described above, the variance of the distribution of the estimated value of the unit work time is used for the adjustment of the unit work time. However, the information correction application is not limited to the dispersion, and the information correction application may be the width between the maximum value and the average value of the unit work time distribution, the width between the average value and the minimum value, or the width between the maximum value and the median value Any value may be used as long as it is an index value representing the dispersion of the distribution.

<Other Embodiments>
The present invention supplies a program that implements one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium. And, it is also possible to realize the processing in which one or more processors in the computer of the system or apparatus read and execute the program. It can also be implemented by a circuit (eg, an ASIC) that implements one or more functions.

As mentioned above, although an example of an embodiment of the present invention was explained in full detail, the present invention is not limited to such a specific embodiment.
For example, as a hardware configuration of the information processing apparatus, a plurality of CPUs may exist, and the plurality of CPUs may execute the processing based on a program stored in the ROM, the storage device, or the like. Further, as a hardware configuration of the information processing apparatus, a GPU (Graphics Processing Unit) may be used instead of the CPU.
Also, the embodiments described above may be implemented in any combination.

  As mentioned above, according to each embodiment mentioned above, exact unit work time can be obtained automatically.

201 CPU
202 ROM
206 storage device

Claims (11)

Estimation means for estimating the process at each time point of the set time interval;
A correction unit that corrects an estimated value of a unit work time for each process based on the accuracy of the estimation;
An information processing apparatus having The image processing apparatus further comprises creation means for creating an estimated value of unit work time for each process based on the result of the estimation,
The information processing apparatus according to claim 1, wherein the correction unit corrects the estimated value of the unit work time based on the accuracy of the estimation.   The information processing apparatus according to claim 1, further comprising a first update unit configured to update a unit work time of plan information with the corrected estimated value of the unit work time.   The information processing apparatus according to any one of claims 1 to 3, wherein the estimation unit estimates a process at each time point of the set time interval in the estimation model.   The information processing apparatus according to claim 4, wherein the estimation unit estimates a process at each time point of the set time interval by inputting feature amount data collected at a production site to the estimation model.   The information processing apparatus according to claim 4, wherein the accuracy of the estimation is a recall ratio and a precision ratio of the estimated model. A determination means for determining the variance from the frequency distribution of unit work time for each process based on the result of the estimation;
Second updating means for updating the distribution of the plan information with the determined distribution;
The information processing apparatus according to claim 3, further comprising:   8. The information processing apparatus according to claim 7, further comprising: adjustment means for adjusting the unit work time of the plan information based on the information on the plan safety degree inputted via the plan screen.   9. The information processing apparatus according to claim 8, further comprising display means for displaying the plan screen. An information processing method executed by the information processing apparatus;
An estimation step of estimating a step at each time point of the set time interval;
A correction step of correcting an estimated value of a unit work time for each step based on the accuracy of the estimation;
Information processing method having.   A program for causing a computer to function as each means of the information processing apparatus according to any one of claims 1 to 9.

Images