JP2018067147A - Managing system, and method of controlling managing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a precise prediction of a replacement timing of a consumable part without a notification which indicates a consumption level of the consumable part and which is redundantly generated in accordance with, for example, a way of use by a user.SOLUTION: A replacement timing estimating unit 344 of a consumable part managing system 101 classifies near-full and full notifications received from an image forming device 103 and stored into multiple clusters on the basis of a notification cycle, and estimates a replacement timing of a consumable part on the basis of a generation timing of the notification classified into the specific cluster (FIG. 6 and FIG. 7). A model creating unit 346 calculates a feature quantity in a time period with respect to the replacement timing of the consumable part using consumable part information received from the image forming device 103 at a constant cycle and stored, and creates a model to predict the next replacement timing of the consumable part on the basis of the feature quantity (FIG. 8). A predicting unit 364 predicts the next replacement timing of the consumable part by applying the consumable part information received from the image forming device 103 to this model. A delivery instructing unit 365 gives a delivery instruction on the basis of this prediction result (FIG. 9).SELECTED DRAWING: Figure 3

Description

  The present invention relates to a management system for managing devices including consumables, and a control method for the management system.

  2. Description of the Related Art Conventionally, in a system for maintaining an image processing apparatus, the number of toner collection boxes that are one of replaceable parts is remotely managed and delivered according to the use of the customer who owns the image processing apparatus. Has been proposed (Patent Document 1).

  Here, the image processing apparatus has a mechanism for notifying a replaceable component according to the degree of wear or the like and prompting the replacement. With respect to such replaceable parts, there is a known problem that notification regarding the same wear level is performed a plurality of times for the same parts depending on the behavior of the user. For example, in the case of a toner collection box, a mechanism may be considered in which when the height of toner in the box exceeds a reference value, a near-full condition or a consumption level such as fullness is notified.

  Note that, after the full notification is given, the user shakes the toner collection box, so that the height of the toner in the toner collection box becomes uniform and temporarily lower than the height detected by the sensor. However, since a certain amount of toner is actually present in the collection box, the toner is detected again by a slight increase in the toner in the toner collection box. As a result, there are cases where notifications relating to the same degree of wear are performed a plurality of times.

  Therefore, there is a technique in which a maintenance system refers to the time between a plurality of notifications, the number of printed sheets, etc. in order to eliminate unnecessary notifications for the same component (Patent Document 2).

JP 2013-030153 A JP 2003-159854 A

  In the technique disclosed in Patent Literature 1, inventory information is updated when a signal indicating that the image forming apparatus is almost full is notified. Although it is possible to issue a delivery instruction for consumables at this timing, if a delivery instruction is given after a notification indicating that it is almost full, delivery may not be in time until it is full There are challenges. In such a case, it is necessary to predict the full timing and deliver consumables in advance in view of the time required for delivery.

  Here, in order to predict the timing of fullness, that is, the timing of replacement, it is necessary to calculate the feature amount related to the timing of fullness. To that end, it is necessary to accurately determine the timing of fullness. Necessary. However, as described above, depending on the user's behavior, there is a problem that a notification related to the same degree of wear is performed a plurality of times.

  In the technique shown in Patent Literature 2, by comparing the time between a plurality of notifications, the number of sheets printed in the meantime, and a predetermined threshold, among a plurality of notifications related to the same consumption level, Exclude unwanted notifications. However, some users put a large amount of toner on paper, and some users use only a small amount of toner. In such a case, for example, even if the time between a plurality of notifications is the same, there is a great difference in the amount of toner actually consumed, and hence the amount of toner discarded in the toner collection box. In such a case, it is difficult to set an appropriate threshold value in advance.

  The object of the present invention is to take into account the above-mentioned problems, and exclude duplicate notifications that occur according to the user's usage, etc. from notifications indicating the degree of consumption of consumables with which the device is provided, so that the consumable replacement timing is accurate. It is to provide a mechanism that makes it possible to predict well.

  The present invention is a management system for managing devices including consumables, and acquires monitoring information of the devices and notifications generated when the degree of wear of consumables included in the devices is in a specific state. An acquisition unit, a storage unit for storing the monitoring information and the notification acquired by the acquisition unit, and a notification stored in the storage unit are classified into a plurality of clusters based on a first feature amount, Identifying means for identifying the replacement timing of the consumables based on the generation timing of the classified notification, and the replacement timing of the consumables identified by the identifying means using the monitoring information stored in the storage means And a creation means for obtaining a second feature amount in a period related to and for creating a model for predicting a next replacement timing of the consumable based on the second feature amount. And features.

  According to the present invention, it is possible to accurately predict the replacement timing of consumables by excluding duplicate notifications that occur according to the user's usage and the like from the notification indicating the degree of consumption of the consumables provided in the device. As a result, it becomes possible to ship the consumables at an appropriate timing.

The figure which illustrates the structure of the system which shows a present Example Hardware configuration diagram of each device constituting the system of this embodiment Functional configuration diagram of the system of this embodiment The figure which illustrates mixed Gaussian distribution in a present Example The figure which illustrates the feature-value distribution in the feature-value space in a present Example Flowchart illustrating the estimated exchange timing calculation process of the present embodiment The flowchart which illustrates the exchange timing estimation process in a present Example Flowchart illustrating model creation processing in the present embodiment The flowchart which illustrates the exchange timing prediction process in a present Example

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

FIG. 1 is a diagram illustrating the configuration of a system showing an embodiment of the present invention.
In the system of this embodiment, the consumables management system 101 and the image forming apparatus 103 are connected via a network such as the Internet 102.

  The consumable management system 101 includes a data collection server 104, a prediction model creation server 105, and a delivery instruction server 106. Communication between each server and the image forming apparatus follows protocols such as HTTP, HTTPS, and SOAP.

Each server included in the consumables management system 101 is not limited to the three servers of the data collection server 104, the prediction model creation server 105, and the delivery instruction server 106, and includes four or more servers. Also good.
Further, each server constituting the consumables management system 101 may be configured on an individual device, or at least two servers may be configured on the same device. Moreover, each said server may be comprised by the cloud server etc.

  In the present embodiment, a toner collection box (a container for collecting used recording materials) will be described as an example of a consumable item to be managed by the consumable item management system 101. However, the consumables to be managed by the consumable management system 101 are not limited to the toner collection box. Depending on the behavior of the user, notifications according to the degree of consumption of the consumables are performed multiple times for the same part. Any consumable that may be possible. For example, the consumable item to be managed by the consumable item management system 101 may be a toner cartridge or the like.

  In the present embodiment, as an example of a network device to be managed by the consumable management system 101, an image forming apparatus such as a digital multifunction peripheral, a facsimile machine, or a printer will be described as an example. However, this is an example and does not limit the scope of application of the present invention. The network device managed by the consumables management system 101 is a network device that has consumables. Depending on the user's behavior, etc., there is a possibility that notifications depending on the degree of consumption of consumables may be sent multiple times for the same part. Any network device with no problem may be used. In this case, a counter value such as the number of printed sheets to be described later is a counter value obtained by counting a predetermined operation in the network device.

Hereinafter, the hardware configuration of each device constituting the system of this embodiment will be described with reference to FIG. 2. FIG. 2A is applicable to the data collection server 104, the prediction model creation server 105, and the delivery instruction server 106. It is a figure which illustrates the hardware constitutions of a simple information processing apparatus. These can be constituted by hardware of a general information processing apparatus (so-called PC (personal computer)).

  The CPU 201 executes various programs such as a program stored in the program ROM in the ROM 203 and an OS (operating system) and application programs loaded from the hard disk 210 to the RAM 202. That is, the CPU 201 functions as each processing unit that executes processing of each flowchart described later by executing a program stored in a readable storage medium.

A RAM 202 is a main memory of the CPU 201 and functions as a work area or the like.
The ROM 203 stores various programs and various data.

A keyboard controller 204 controls operation inputs from a keyboard 208 and a pointing device (not shown) (mouse, touch pad, touch panel, trackball, etc.).
A display controller 205 controls display on the display 209.

The disk controller 206 controls data access to the external memory 210 such as a hard disk (HD), solid state drive (SSD), and flexible disk (FD) that stores various data. The external memory 210 stores various programs such as an OS (Operating System) and application programs and various data.
A network I / F 207 is connected to the network and executes communication control processing with other devices connected to the network.

FIG. 2B is a diagram illustrating a hardware configuration of the image forming apparatus 103.
The image forming apparatus 103 includes a CPU 221, a RAM 222, a ROM 223, a storage device 224, a network I / F 225, an internal bus 226, a device control 227, a printing unit 228 including a fixing device, an input / output I / F 229, an input / output device 230, a toner. A recovery box remaining amount detection unit 231 and a sensor unit 232 are included.

  The CPU 221 includes a program stored in the ROM 223 (including a program for realizing each process described later), and comprehensively controls each device via the internal bus 226. The RAM 222 functions as a memory or work area for the CPU 221. The CPU 221 performs a program execution process together with the RAM 222 and the ROM 223 and performs a process of recording image data on a recording medium such as the storage device 224.

  The network I / F 225 exchanges data with an external network device or a personal computer (PC) in one or both directions. The device control 227 controls the printing unit 228.

The storage device 224 functions as an external storage device. The storage device 224 is, for example, an HDD or an SSD.
The input / output device 230 shows a plurality of configurations that perform input / output in the image forming apparatus 103. Specifically, an input (button input or the like) from the user is received, and a signal corresponding to the input is transmitted to each processing unit described above by the input / output I / F 229. In addition, the input / output device 230 includes a display device (such as a touch panel) for providing necessary information to the user or accepting a user operation. Further, the input / output device 230 may include a scanning device for reading a document and receiving electronic data as input.

  The toner collection box remaining amount detection unit 231 detects the amount of toner in the toner collection box via the sensor unit 232 and generates it as a signal such as an alarm. The signal generated here is transmitted to the consumables management system 101 via the network I / F 225. In this embodiment, the toner collection box remaining amount detection unit 231 generates a signal indicating the timing when the toner collection box is full (a state to be replaced) as a full signal, and is approaching full (a state before full). The timing signal is generated as a near full signal.

  The sensor unit 232 includes an image sensor that detects the amount of toner in the toner collection box, a sensor that detects temperature and humidity in the image forming apparatus 103, a sensor that detects the sheet conveyance status in the image forming apparatus 103, and the like. A sensor for detecting various states in the forming apparatus 103 is included.

FIG. 3 is a functional block diagram illustrating each function in the system of the present embodiment.
First, the functional configuration of the image forming apparatus 103 will be described. Each functional unit of the image forming apparatus 103 is realized and functions when the CPU 221 of the image forming apparatus 103 reads and executes a program recorded in the ROM 223 or the like. Details will be described below.

  The consumable information collection unit 301 acquires various pieces of consumable information acquired from the image forming apparatus 103 such as various counter information, the number of jobs, and the amount of applied toner from the image forming apparatus 103. The various consumable information is not limited to the consumable information, and may include any information as long as it is monitoring information of the image forming apparatus 103. On the other hand, the sensor information collection unit 302 collects various consumable information required in hardware, like the toner collection box remaining amount detection unit 231 described above. In the present embodiment, the sensor information collection unit 302 performs the above-described “near full” timing (timing when the toner collection box is almost full) and “full” timing (timing when the toner collection box is full). Sensor information is collected, but other information may be collected. When the image forming apparatus 103 transmits a signal such as “near full” or “full” to the server, the image forming apparatus 103 adds information such as counter information at that time and transmits the signal. That is, the notification such as “near full” or “full” includes monitoring information of the image forming apparatus 103 such as counter information. The monitoring information included in the notification is not limited to the counter information, and may be any information as long as it is monitoring information of the image forming apparatus 103, such as the number of jobs, the amount of applied toner, temperature, and humidity.

  The information transmission unit 303 transmits various information of the image forming apparatus 103 to the consumables management system 101. Note that the notification of the signal such as “near full” and “full” to the server is transmitted at the timing when the event occurs, but the monitoring information transmitted by the consumables information collection unit 301 such as various counters is periodically (And when the user etc. instructs). Information obtained by the above-described two collection units is transmitted to the information reception unit 321 of the data collection server 104 via the information transmission unit 303. In this embodiment, the collection unit is described in two from the viewpoint of software and hardware. However, either one of the collection units may collect all information, The above collection unit may be provided, and these collection units may share and collect each information.

Next, functions of each server constituting the consumable management system 101 will be described.
First, the functional configuration of the data collection server 104 will be described. Each function unit of the data collection server 104 is realized and functions when the CPU 201 of the data collection server 104 reads and executes a program recorded in the external memory 210 or the like. Details will be described below.

  The information receiving unit 321 receives information indicating various types of information of the image forming apparatus 103 such as counter information and sensor information transmitted from the information transmitting unit 303, thereby acquiring information related to the signal. The information receiving unit 321 stores the information related to the acquired signal in the consumable information storage unit 322 realized by a permanent medium such as a database or a file system. An example of data stored in the consumable information storage unit 322 is shown in Table 1.

  In Table 1, the device ID (a1) is an ID for uniquely identifying the image forming apparatus 103 that has transmitted the signal received by the information receiving unit 321. Time (a2) represents the date and time when the event corresponding to the signal occurred. The signal code (a3) represents the type of the signal. In this embodiment, the signal code “001” represents the near full signal, and the signal code “002” represents the full signal. The counter (a4) represents a counter at that time. Actually, there are various counters such as a counter indicating the number of A4 size printed sheets, a counter indicating the number of A3 size printed sheets, a counter indicating the number of black and white sheets, but in this embodiment, for the sake of simplification. Are collectively referred to as counters.

  Next, the functional configuration of the prediction model creation server 105 will be described. Each functional unit of the prediction model creation server 105 is realized and functions when the CPU 201 of the prediction model creation server 105 reads and executes a program recorded in the external memory 210 or the like. Details will be described below.

  The consumable information acquisition unit 341 acquires information stored in the consumable information storage unit 322. Based on the information acquired by the consumable information acquisition unit 341, the consumable information calculation unit 342 calculates the time difference between successive signals for each device ID (here, the time between signals from the full signal to the near full signal). The difference and the time difference from the near full signal to the near full signal) are calculated and stored in the consumable information storage unit 343 after the calculation. In addition, since it is a normal phenomenon that a full signal occurs after a near full signal, a time difference between signals from a near full signal to a full signal is not calculated. Taking the data shown in Table 1 as an example, an example of data stored in the post-calculation consumable information storage unit 343 is as shown in Table 2.

  The device ID (b1) is an ID for uniquely identifying the image forming apparatus 103. The start point time (b2) and the end point time (b3) are timings (date and time) at which signals for which a difference is to be obtained are generated, and the difference between these two timings is set as the time difference (b5). The type (b4) represents a set of signals for which a difference is obtained, “002001” represents a time difference between the full signal and the near full signal, and “001001” is between the near full signal and the near full signal. Represents the time difference of.

The replacement timing estimation unit 344 excludes unnecessary notifications generated by the user's behavior based on the difference data set in the time difference (b5), and estimates the timing at which the toner collection box is replaced. Hereinafter, the outline of the estimation process will be described with reference to Table 2 and FIG.
FIG. 4 is a diagram illustrating a mixed Gaussian distribution in this embodiment.

  The first and second rows in Table 2 both represent the time difference between the full signal and the near full signal, but the time difference in the first row is 14 hours, while the time in the second row. The difference is 50 days or more. As described above, when a plurality of time differences between the full signal and the near full signal are obtained and expressed in a histogram, the result is as shown in FIG.

  This histogram is assumed to be a mixed Gaussian distribution composed of two Gaussian distributions. One Gaussian distribution is a distribution in which a user's behavior such as shaking a toner collection box is generated, an unnecessary notification is generated, and a time difference between signals is shortened. The other Gaussian distribution is a distribution of time differences when there is no user behavior. This is because it is assumed that the time difference in the case of the user's behavior is shorter than that in the case of no user's behavior, and therefore the distribution of the time difference in each case is assumed to exist.

This mixed distribution is estimated using an EM algorithm or the like, and distributions N1 (M1, σ1) and N2 (M2, σ2) are estimated. M1 and M2 are averages of the respective distributions, and M1 <M2. Also, σ1 and σ2 are N1 and N2 variances. Here, the data including the time difference within the range of M2 ± 3σ2 is defined as the time difference when there is no user behavior, and the “start time” when the difference is obtained is estimated as the replacement timing of the toner collection box. To do. Hereinafter, this exchange timing is referred to as “estimated exchange timing”.
For example, if the time difference 50 days in the data in the second row in Table 2 is within the range of M2 ± 3σ2, the replacement timing is “2016/06/10 12: 30: 00.000”.

  In the above description, the time difference between the full signal and the near full signal has been described. However, the exchange timing estimation unit 344 performs the same process on the time difference between the near full signal and the near full signal to obtain the estimated exchange timing. For example, if the time difference 46 days in the data in the third row in Table 2 is within the range of M2 ± 3σ2, the replacement timing is “2016/07/31 10: 00: 00.000”. That is, the exchange timing estimation unit 344 classifies notifications such as near-full signals and full signals into a plurality of clusters based on feature quantities (in this embodiment, time differences between signals, that is, signal generation intervals), and a specific cluster. The replacement timing of the consumables is estimated (specified) based on the generation timing of the notification classified as follows.

  The exchange timing estimation unit 344 stores the estimated exchange timing estimated as described above in the estimated exchange timing storage unit 345. Specifically, data as shown in Table 3 is stored.

  The device ID (c1), time (c2), and signal code (c3) are the same as a1, a2, and a3 described in Table 1, and therefore will not be described. The exchange timing (c4) is a flag indicating whether or not the signal represents exchange, and “0” represents non-exchange and “1” represents exchange.

Next, the model creation unit 346 creates a model for the exchange timing estimation unit 344 to predict the estimated exchange timing. There are many model creation methods, but an example is described below.
First, when a consumable is delivered to a user, the number of days required for delivery is defined, and this is set to “α”. In order to eliminate the downtime of the device, a model for predicting whether or not replacement will occur after α days from a certain timing is created.

  First, the model creation unit 346 calculates the difference between the consumable information at the estimated replacement timing obtained by the replacement timing estimation unit 344 and the consumable information obtained periodically. A data example of the calculated difference information will be specifically described with reference to Table 4. Here, it is assumed that α = 7, and the estimated exchange timing is two points of “2016/06/10 12: 30: 00.000” and “2016/07/31 10: 00: 00.000”.

The device ID (d1) is an ID for uniquely identifying a device. The estimated replacement timing (d2) represents the date and time of the estimated replacement timing representing the start point. The counter time (d3) represents the date and time when the counter information is acquired. The counter difference (d4) represents the difference between the counter value at the estimated exchange timing (d2) and the counter value at the counter time (d3). That is, the counter difference (d4) stores the difference between the counter information at the estimated exchange timing (d2) and the counter information after the estimated exchange timing (D2). Finally, in the presence / absence of exchange (d5) after α day, “1” is set if there is a next estimated exchange timing within α (= 7) days after the counter time (d3). “0” is set. In the example of Table 4, since “2016/07/31 10: 00: 00.000” is set as the next estimated replacement timing, whether the counter is replaced after α days after “2016/07/24 16: 00: 00.000” or later (D5) becomes “1”.
When the counter difference is plotted on the horizontal axis and the exchange status after α days is plotted on the vertical axis, a graph as shown in FIG. 5 is obtained.
FIG. 5 is a diagram illustrating a feature quantity distribution in the feature quantity space in the present embodiment.

  Based on these data, the model creation unit 346 obtains a prediction model using an algorithm such as logistic regression, and stores the prediction model in the model storage unit 347. In the present embodiment, only one counter is selected as a feature quantity for the sake of simplicity of explanation and ease of illustration. However, in practice, a plurality of counter values and a plurality of values such as temperature and humidity are used. Is adopted as a feature quantity.

For example, there are n feature quantities as described above, and those values are x ₁ , x ₂ ,..., X _n, and the probability that the presence or absence (d5) after α days will be “1” is p. In this case, a prediction model (logistic regression model) using a logistic regression algorithm is expressed by the following equation (1).

In Equation 1, a is a constant, and b _i (i = 1 to n) is a regression coefficient of x _i , which is obtained from the logistic regression analysis result. When the feature amounts x ₁ , x ₂ ,..., X _n of an image forming apparatus 103 are input into Equation 1, the image forming apparatus 103 stores the amount of toner in the toner collection box after α days from the date and time when the feature amount was acquired. The probability that an exchange is required can be calculated. Further, based on the calculated probability that the toner collection box needs to be replaced after α days (for example, by comparing the probability with a threshold value), it is determined whether the toner collection box needs to be replaced after α days. It is also possible to predict. Note that the logistic regression algorithm is well known, and thus detailed description thereof is omitted.

  In addition to the logistic regression, other algorithms such as a neural network and a support vector machine (SVM) may be used for creating a prediction model.

  Finally, the functional configuration of the delivery instruction server 106 will be described. Each functional unit of the delivery instruction server 106 is realized and functions when the CPU 201 of the delivery instruction server 106 reads out and executes a program recorded in the external memory 210 or the like. Details will be described below.

  The model acquisition unit 361 acquires a prediction model from the model storage unit 347. Next, the consumable information acquisition unit 362 acquires information from the consumable information storage unit 322. Further, the estimated exchange timing acquisition unit 363 acquires information on the estimated exchange timing from the estimated exchange timing storage unit 345.

  The prediction unit 364 calculates the feature amount from the estimated replacement timing and the consumable information acquired by the consumable information acquisition unit 362, and applies the feature amount to the prediction model to determine when the consumable replacement timing occurs. Make predictions. In the above example, it can be obtained as a prediction result whether or not an exchange will occur after α days.

  Based on the prediction result of the prediction unit 364, the delivery instruction unit 365 notifies the related system and related persons of a consumables delivery instruction. These notification destinations are set in advance in the delivery instruction unit 365.

Next, the estimated exchange timing calculation process will be described with reference to FIG.
FIG. 6 is a flowchart illustrating the estimated replacement timing calculation process according to this embodiment. Note that the processing in this flowchart is executed periodically (for example, every day) under the control of the consumable information acquisition unit 341 and the like. The processing of the flowcharts shown in FIG. 6 and FIGS. 7, 8, and 9 described later is realized by the CPU 201 of the prediction model creation server 105 reading and executing the program recorded in the external memory 210 or the like.

In S601, the consumable information acquisition unit 341 acquires the consumable information from the consumable information storage unit 322.
Next, in S602, the consumable information calculation unit 342 sorts information corresponding to full and near full among the information acquired in S601, in order of occurrence time for each device ID, and sets the first device ID as the current device ID. The processing is shifted to S603 as the device ID.
In S603, the consumable information calculation unit 342, among the information sorted in S602, information on the head of the current device ID (hereinafter referred to as “i1”) and the next information (hereinafter referred to as “i2”). To get. Although not shown, in this case, when two pieces of information on the current device ID (“i1” and “i2”) cannot be acquired, the consumable information calculation unit 342 shifts the process to S612 as it is.

  Next, in S604, the consumable information calculation unit 342 verifies whether or not “i1 is information corresponding to near full and i2 is information corresponding to full”. As a result of the verification, if it is determined that “i1 corresponds to near full and i2 corresponds to full” (Yes in S604), the consumables information calculation unit 342 performs the process in S609 as it is. Transition. The process of S609 will be described later.

  On the other hand, as a result of the verification, if it is determined that the information is not “i1 corresponds to near full and i2 corresponds to full” (No in S604), the consumable information calculation unit 342 performs the process in S605. Transition.

  In S605, the consumable information calculation unit 342 verifies whether or not “information corresponding to i1 near near full and information corresponding to i2 near near full”. As a result of the verification, if it is determined that “i1 is near-full information and i2 is near-full information” (Yes in S605), the consumable information calculation unit 342 shifts the process to S606. Let

  In S606, the consumable information calculation unit 342 acquires the difference between the occurrence times of i1 and i2, and stores the difference result in the post-calculation consumables information storage unit 343. The type (B4) at this time is “001001”. After S606, the consumable information calculation unit 342 shifts the process to S609.

On the other hand, in S605, when it is determined that “i1 is information corresponding to near-full and i2 is information corresponding to near-full” (No in S605), the consumable information calculation unit 342 performs processing in S607. Transition.
In S607, the consumable information calculation unit 342 verifies whether or not “information corresponding to i1 being full and information corresponding to i2 being near full”.

As a result of the verification, if it is determined that “i1 corresponds to full information and i2 corresponds to near full” (Yes in S607), the consumable information calculation unit 342 shifts the process to S608. Let
In S608, the consumable information calculation unit 342 acquires the difference between the occurrence times of i1 and i2, and stores the difference result in the post-calculation consumables information storage unit 343. The type (b4) at this time is “002001”. After the S608, the consumable information calculation unit 342 shifts the process to S609.

  On the other hand, if it is determined in S607 above that “i1 is information corresponding to full and i2 is information corresponding to near full” (No in S607), the consumable information calculation unit 342 The process is directly shifted to S609.

  In S609, the consumable information calculation unit 342 verifies whether or not the processing indicated in S604 to S610 has been performed on all information of the current device ID among the information sorted in S602. If it is determined that there is information that has not yet been processed in S604 to S610 among the information sorted in S602 (No in S609), the consumable information calculation unit 342 performs the process in S610. Transition.

  In S610, the consumable information calculation unit 342 sets i1 as i2 and i2 as information next to i2. That is, control is performed to shift one piece of information. After S610, the consumable information calculation unit 342 shifts the process to S604.

On the other hand, if it is determined in S609 that the processing shown in S604 to S610 has been performed on all information of the current device ID among the information sorted in S602 (Yes in S609), consumables The information calculation unit 342 shifts the process to S611.
Through the processing of S601 to S609, data as shown in Table 2 is stored in the post-calculation consumable information storage unit 343.

  Next, in S611, the exchange timing estimation unit 344 performs an exchange timing estimation process. Details of the replacement timing estimation processing are shown in FIG. After S611, the replacement timing estimation unit 344 advances the process to S612.

  In S612, the consumable information calculation unit 342 verifies whether or not the processing shown in S604 to S610 has been performed on the information of all device IDs among the information sorted in S602. If it is determined that there is device ID information that has not yet been subjected to the processing shown in S604 to S610 among the information sorted in S602 (No in S612), the consumable information calculation unit 342 The next device ID of the information sorted in S602 is set as the current device ID, and the process proceeds to S603.

  On the other hand, if it is determined in S612 that the processing shown in S604 to S610 has been performed on the information of all device IDs among the information sorted in S602 (Yes in S612), the consumable information calculation is performed. The unit 342 ends the process of this flowchart.

FIG. 7 is a flowchart illustrating the replacement timing estimation process in the present embodiment.
In S701, the replacement timing estimation unit 344 determines that the type (b4) is “001001” (difference information between near full and near full) and “002001” (difference information between full and near full) from the post-calculation consumable information storage unit 343. After the calculation, the consumable information is acquired, and control is performed so that S702 to S705 are applied to each.

  In S702, the exchange timing estimation unit 344 forms a histogram of the difference information acquired in S701, and uses two gaussian distributions constituting the difference information distribution using an EM algorithm (Expectation-Maximization Algorithm) or the like. N1 (M1, σ1) and N2 (M2, σ2) are obtained. Here, M1 <M2. Since the EM algorithm is publicly known, detailed description is omitted. Further, the algorithm used in the process of S702 is not limited to the EM algorithm, and other algorithms may be used.

  Next, in S703, the replacement timing estimation unit 344 selects one of the post-calculation consumable information acquired in S701.

  Next, in S704, the replacement timing estimation unit 344 verifies whether or not the difference information of the post-calculation consumable information selected in S703 is within M2 ± 3σ2. As a result of the verification, when it is determined that it is within M2 ± 3σ2 (Yes in S704), the exchange timing estimation unit 344 shifts the process to S705.

  In S705, the replacement timing estimation unit 344 estimates the consumable information that is the starting point out of the consumable information that is the basis of the post-calculation consumable information selected in S703, as the replacement timing, and the estimation result is estimated replacement timing. Save in the storage unit 345. After S705, the exchange timing estimation unit 344 shifts the process to S706.

  On the other hand, if it is determined in S704 that it is not within M2 ± 3σ2 (in the case of No in S704), the exchange timing estimation unit 344 directly proceeds to S706.

  In S706, the replacement timing estimation unit 344 verifies whether or not all post-calculation consumable information acquired in S701 has been processed. If it is determined that there is information that has not yet been processed among the calculated consumable information acquired in S701 (No in S706), the replacement timing estimation unit 344 determines that there is an unprocessed calculated consumable. Information is selected, and the process proceeds to S704.

On the other hand, if it is determined in S706 that the processing has been performed on all post-calculation consumable information acquired in S701 (Yes in S706), the replacement timing estimation unit 344 ends the processing of this flowchart. .
Through the processing of FIG. 7, the data as shown in Table 3 is stored in the estimated exchange timing storage unit 345.

Next, a prediction model construction process using the estimated replacement timing will be described with reference to FIG.
FIG. 8 is a flowchart illustrating the model creation process in the present embodiment. Note that the processing of this flowchart is executed periodically (for example, every day) under the control of the model creation unit 346 and the like.

In S801, the model creation unit 346 acquires consumable information from the consumable information acquisition unit 341.
Next, in S802, the model creation unit 346 sets a variable α representing the number of days required for delivery to “7”. In this embodiment, α = 7 will be described, but α is not limited to “7”.

Next, in S803, the model creation unit 346 selects an arbitrary one from information corresponding to the exchange timing estimated by the exchange timing estimation unit 344.
Next, in S804, the model creation unit 346 sets the variable X used internally to “1”.

  Next, in S805, the model creation unit 346 calculates the difference (counter difference (d4)) between the consumable information at the estimated replacement timing selected in S803 and the consumable information between the estimated replacement timing and X days later. Each difference is calculated, and this difference is treated as a feature amount.

  Next, in S806, the model creation unit 346 determines whether or not the next estimated exchange timing exists within the estimated exchange timing + X + α days, and if so, corresponds to the feature amount obtained in S805. Set the objective variable (replacement after α days (d5)) to “1”. On the other hand, if it does not exist, the model creation unit 346 sets the objective variable to “0”. At this point, a set of feature quantities and their objective variables as described in Table 4 is created.

Next, in S807, the model creation unit 346 increments the internally used variable X by “1”.
Next, in S808, the model creation unit 346 compares the estimated exchange timing + X day with the next estimated exchange timing, and verifies whether the estimated exchange timing + X day is in the future from the next estimated exchange timing. . If it is determined that the next estimated exchange timing is in the future (after X days ≦ next estimated exchange timing) (No in S808), the model creation unit 346 shifts the process to S805.

  On the other hand, in S808, when it is determined that the estimated exchange timing + X day is in the future (the next estimated exchange timing after X days> the next estimated exchange timing) (Yes in S808), the model creation unit 346 The processing is shifted to S809.

  In S809, the model creation unit 346 verifies whether or not the processing of S804 to S808 has been performed for all the replacement timings estimated by the replacement timing estimation unit 344. If it is determined as a result of the verification that there is an estimated replacement timing that has not yet been processed (No in S809), the model creation unit 346 selects an estimated replacement timing that has not yet been selected, and the process proceeds to S804. Transition.

On the other hand, when it is determined in S809 that the process has been performed for all estimated replacement timings (Yes in S809), the model creation unit 346 shifts the process to S810.
Data as shown in Table 4 is generated by the processing of S801 to S809.

In S810, the model creation unit 346 creates a two-class classification model using an algorithm such as logistic regression based on a set of all the obtained feature values and the objective variable.
Next, in S811, the model creation unit 346 stores the model created in S810 in the model storage unit 347, and ends the process of this flowchart.

Finally, prediction processing using a model will be described with reference to FIG.
FIG. 9 is a flowchart illustrating the replacement timing prediction process in the present embodiment. Note that the processing of this flowchart is executed periodically (for example, every day) for each image forming apparatus 103 under the control of the consumable information acquisition unit 362 and the like.

  In S900, the consumable information acquisition unit 362 receives a variable α representing the number of days required for delivery from a recent consumable delivery instruction to the image forming apparatus 103 to be processed (α = 7 in this embodiment) days or more. Judge whether or not it has passed. If it is determined that α days or more have not yet elapsed (No in S900), the consumable information acquisition unit 362 ends the process of this flowchart.

On the other hand, when it is determined that α days or more have elapsed (Yes in S900), the consumable information acquisition unit 362 advances the process to S901.
In S901, the consumable information acquisition unit 362 acquires consumable information from the consumable information storage unit 322 of the image forming apparatus 103 to be processed.

  In S902, the prediction unit 364 verifies whether the information acquired in S901 includes a signal indicating near full. If it is determined that a signal indicating near full is included as a result of the verification (Yes in S902), the prediction unit 364 causes the process to transition to S908. In step S908, the delivery instruction unit 365 issues a consumable delivery instruction, and ends the processing of this flowchart. In this case, since α days have passed since the last delivery instruction (Yes in S900), the consumables should have been delivered to the customer. It is thought that it is in a state to be done. For this reason, it is considered that the possibility that the near full (notification concerning the same consumption level) will occur again due to the user's behavior is reduced. Therefore, if more than α days have passed since the last delivery instruction, even if a consumable delivery instruction is given in response to the near full signal, it is considered that there is a low possibility of multiple delivery.

On the other hand, when it is determined in S902 that a signal indicating near full is not included (Yes in S902), the prediction unit 364 causes the process to transition to S903.
In S903, the prediction unit 364 uses the estimated replacement timing acquisition unit 363 to acquire the estimated replacement timing information from the estimated replacement timing storage unit 345. The estimated exchange timing information acquired here is assumed to be the latest timing.

Next, in S904, the prediction unit 364 acquires (calculates) a difference between the consumable information acquired in S901 and the counter value included in the estimated replacement timing information acquired in S903, and uses this as a feature amount. To do.
Next, in S905, the model acquisition unit 361 acquires a model from the model storage unit 347, and shifts the processing to S906.

Next, in S906, the prediction unit 364 applies the feature amount acquired in S904 to the model acquired in S905, and obtains a prediction result.
Next, in S907, the prediction unit 364 determines whether or not the prediction result obtained in S906 is that a consumable item has been replaced within α days. If the prediction result is determined to be that replacement of consumables occurs within α days (Yes in S907), the prediction unit 364 shifts the process to S908.
In step S908, the delivery instruction unit 365 issues a consumable delivery instruction, and ends the processing of this flowchart.

  On the other hand, when it is determined in S907 that the prediction result does not indicate that the replacement of the consumable item occurs within α days (No in S907), the prediction unit 364 ends the processing of this flowchart as it is.

  As described above, by realizing the consumable item management system described in the present embodiment, it is possible to estimate an appropriate replacement timing from the notification of the degree of consumption that occurs redundantly depending on the user's behavior. As a result, a highly accurate prediction model can be created, and consumables can be delivered at an appropriate timing. That is, a threshold value for determining an unnecessary notification can be dynamically determined, and an appropriate notification can be estimated for each device. Moreover, the timing of exchange is specified based on the estimated notification, and a model for predicting delivery timing based on the exchange timing is constructed, and prediction using the model is performed. This makes it possible to notify delivery at an appropriate timing before the consumable is actually unavailable.

Hereinafter, although Example 2 is demonstrated, description is abbreviate | omitted about the same point as Example 1, and only a different point from Example 1 is demonstrated.
In the first embodiment, when estimating the replacement timing, the distribution is estimated by treating only the period (interval) between near-full and near-full as the feature amount. However, even if the period is long, if the number of jobs during that period is small or the amount of toner on the paper is small, there is a possibility that the replacement timing is not appropriate.

  Therefore, in the second embodiment, the consumable information calculation unit 342 calculates not only the period but also the number of jobs (accumulation executed by the image forming apparatus 103 so far) as information to be used when calculating the difference between the consumption levels. Number of jobs) and the amount of applied toner (total amount of recording agent used in the image forming apparatus 103 so far), and other factors, and multi-dimensional distribution estimation using each information as an axis. Configure to do. In the multi-dimensional distribution, the data included in the predetermined range is set as a time difference when there is no user behavior, and the “starting time” when the difference is obtained is estimated as the replacement timing of the toner collection box. To be configured.

  As described above, by realizing the consumables management system described in the second embodiment, the amount of difference between a plurality of types of information such as near-full and near-full intervals, full-near full intervals, the number of jobs, the amount of applied toner, and the like can be obtained. The multi-dimensional distribution is estimated based on the plurality of types of feature amounts, notifications such as near-full and full are classified into a plurality of clusters, and the exchange timing is estimated. Therefore, it is possible to estimate the replacement timing using information other than between near-full and near-full, and thereby it is possible to create a more accurate prediction model.

  In each of the above-described embodiments, the toner collection box has been described as an example of the consumable item. However, the replacement timing of the type of consumable item that reduces the amount of consumable item when used, such as a toner cartridge, is predicted. This embodiment can be realized by replacing the “near full” notification with the “consumable low” notification and the “full” notification with the “consumable empty” notification in the above embodiment.

  As described above, in this embodiment, the interval between notifications for notifying the degree of wear is obtained, and the distribution is divided into two distributions (the distribution due to the notification generated by the user's behavior and the distribution due to the notification without the user's behavior). ), The distribution is estimated, a notification in the absence of user behavior is identified using the estimation result, and a prediction model is created based on the notification. As a result, it is possible to exclude a notification concerning the same consumption level of the consumables, obtain a correct replacement timing, and predict the next replacement timing using the replacement timing. Therefore, by predicting the next replacement timing of the consumables included in the image forming apparatus and giving a delivery instruction or the like in advance, it is earlier than the notification regarding the degree of consumption from the image forming apparatus 103 (notification corresponding to replacement), In addition, the consumable item can be delivered to the customer at an appropriate timing.

In addition, the structure of the various data mentioned above and its content are not limited to this, You may be comprised with various structures and content according to a use and the objective.
Although one embodiment has been described above, the present invention can take an embodiment as, for example, a system, apparatus, method, program, or storage medium. Specifically, the present invention may be applied to a system composed of a plurality of devices, or may be applied to an apparatus composed of a single device.
Moreover, all the structures which combined said each Example are also contained in this invention.

(Other examples)
The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in the computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.
Further, the present invention may be applied to a system composed of a plurality of devices or an apparatus composed of one device.
The present invention is not limited to the above embodiments, and various modifications (including organic combinations of the embodiments) are possible based on the spirit of the present invention, and these are excluded from the scope of the present invention. is not. That is, the present invention includes all the combinations of the above-described embodiments and modifications thereof.

Claims (10)

A management system for managing devices with consumables,
Monitoring means for the device, and acquisition means for acquiring a notification generated in the device when the consumption level of the consumables provided in the device is in a specific state;
Storage means for storing the monitoring information and the notification acquired by the acquisition means;
Specifying means for classifying notifications stored in the storage means into a plurality of clusters based on a first feature quantity, and specifying replacement timing of the consumables based on the generation timing of the notification classified into a specific cluster; ,
Using the monitoring information stored in the storage unit, a second feature amount in a period related to the replacement timing of the consumable item specified by the specifying unit is acquired, and the consumable item is acquired based on the second feature amount. A creation means for creating a model for predicting the next replacement timing,
A management system comprising:   The management system according to claim 1, further comprising a predicting unit that predicts a next replacement timing of the consumable by applying the monitoring information acquired by the acquiring unit to the model.   The management system according to claim 2, further comprising a processing unit that performs processing related to delivery of the consumables based on a prediction result of the prediction unit.   3. The management system according to claim 1, wherein the specifying unit classifies the notification using an amount indicating the generation interval of the notification as a first feature amount. 4.   The specifying means uses an amount indicating the occurrence interval of the notification and a difference amount of one or a plurality of types of monitoring information respectively included in successive notifications as the first feature amount, respectively. The management system according to claim 1, wherein the notification is classified based on a feature amount. The notification is a first notification informing that the degree of wear of the consumables provided in the device has reached a first state to be replaced, or a second state in which the consumables provided in the device are in front of the first state. Is a second notification to let you know that
The specifying means includes the first feature amount obtained from the first notification and the second notification generated after the first notification, and the second feature generated after the second notification and the second notification. 6. The management system according to claim 5, wherein the classification is performed based on the first feature amount obtained from two notifications.   The creation unit includes a difference amount between the predetermined monitoring information included in the notification corresponding to the replacement timing of the consumables specified by the specifying unit and the predetermined monitoring information in a period related to the replacement timing of the consumables The management system according to any one of claims 1 to 6, wherein the model is created using the second feature amount.   8. The management system according to claim 7, wherein the predetermined monitoring information is a counter value indicated by at least one of one or a plurality of counter values obtained by counting a predetermined operation in the device. The device is an image forming apparatus,
The management system according to claim 1, wherein the consumable is a container for collecting a used recording agent or a recording agent. A control method of a management system for managing devices including consumables,
An acquisition step of acquiring the monitoring information of the device and a notification that occurs when the degree of wear of the consumables included in the device is in a specific state;
A storage step of storing the monitoring information and the notification acquired in the acquisition step in a storage unit;
A specifying step of classifying the notification stored in the storage unit into a plurality of clusters based on a first feature amount, and specifying a replacement timing of the consumable based on the generation timing of the notification classified into the specific cluster; ,
Using the monitoring information stored in the storage unit, a second feature amount in a period related to the replacement timing of the consumable item specified in the specifying step is acquired, and the consumable item is based on the second feature amount. A creation step to create a model for predicting the next replacement timing of
A control method for a management system, comprising:

Images