JP2011170898A - User evaluation apparatus depending on hardware usage status - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus which can create a user evaluation model which presumes a repairing modulus from a use situation of hardware by a user by removing an influence of a failure rate due to a product. <P>SOLUTION: A user evaluation apparatus includes: a sample-dependent weight value decision section which decides a weight value of each sample to cancel a difference of a repairing modulus estimated value, between machines when two or more machines which are measuring objects are defined as two or more samples; a weight value storage section for each sample, which memorizes each weight value of two or more samples; a parameter decision section which decides a parameter of the user evaluation model for computing a using influence degree, a user evaluation model storage section which memorizes the parameter of the user evaluation model; and a user evaluation section which computes the using influence degree using the measurement data and the parameter of the user evaluation model. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an apparatus for performing user evaluation according to hardware usage conditions.

  2. Description of the Related Art Conventionally, research on technology for measuring vibration and temperature of hardware such as a personal computer (hereinafter referred to as “PC” or “machine”) and calculating the degree of performance degradation and fatigue has been underway. There has also been proposed an apparatus that obtains information from a sensor or the like provided inside a PC housing and presents the risk of failure occurrence to the user with respect to the PC main body or individual components (see, for example, Patent Document 1). ). In addition, a technology for evaluating the product repair rate from the repair data is being developed for the hardware quality control department.

JP 2006-127070 A

  However, in the conventional technology, when evaluating the usage method of the user, it is not possible to accurately evaluate whether or not the PC has failed in the past, and it is necessary to create a model that goes into the usage of the user. . On the other hand, unlike car accidents, PC failures are more likely to occur due to problems with the product itself. However, it has been difficult to distinguish between user usage issues and product issues.

  It is an object of the present invention to provide an apparatus capable of creating a user evaluation model for estimating a repair rate from a hardware usage state by a user by removing the influence of a failure rate caused by a product.

  In order to solve the above problems, a failure information collection result storage unit that stores failure information collection results for a plurality of machines having the same characteristics as the plurality of machines to be measured, and each machine based on the failure information collection results A repair rate calculation unit that calculates a repair rate estimate value, a repair rate estimate value storage unit that stores the repair rate estimate value, and measurement data reception that receives measurement data relating to the usage status of the plurality of machines that are the measurement targets A measurement data storage unit that stores the measurement data, a plurality of machines that are the measurement targets, and a load value that cancels a difference between the machines of the estimated repair rate, A load value determination unit for each sample determined for each of the samples, a load value storage unit for each sample for storing the load values of each of the plurality of samples, and the load value for each sample Generating an objective function using the measurement data, the failure information collection result, and a parameter of the user evaluation model for calculating the use influence degree, and maximizing the objective function with the parameter of the user evaluation model A parameter determination unit that determines the user evaluation model, a user evaluation model storage unit that stores parameters of the user evaluation model, a user evaluation unit that calculates the usage influence using the measurement data and the parameters of the user evaluation model And a processing unit for displaying the use influence degree.

  According to the present invention, usage can be fed back to the user, and an appropriate extended guarantee fee can be calculated. In addition, it is possible to calculate an appropriate guarantee fee according to the responsibility of the user, and it is possible to provide a PC extended warranty service at a lower price to a user who uses it safely.

The block diagram of the user evaluation apparatus according to the hardware usage condition which concerns on 1st Embodiment The flowchart which shows the process sequence of the model creation which concerns on 1st Embodiment. The flowchart which shows the process sequence of the user evaluation which concerns on 1st Embodiment. The figure which shows the example of data of the measurement data memory | storage part which concerns on 1st Embodiment. The figure which shows the data example of the shipping information storage part which concerns on 1st Embodiment. The figure which shows the example of data of the manufacturing information storage part which concerns on 1st Embodiment. The figure which shows the example of data of the failure information total result storage part which concerns on 1st Embodiment The figure which shows the example of the data for model creation which concerns on 1st Embodiment The figure which shows the example of the decision tree which concerns on 1st Embodiment The figure which shows the example of data of the repair rate estimated value memory | storage part which concerns on 1st Embodiment. The figure which shows the example of data of the load value memory | storage part classified by sample which concerns on 1st Embodiment. The figure which shows the example of data of the user evaluation model memory | storage part which concerns on 1st Embodiment. The block diagram of the user evaluation apparatus according to the hardware use condition which concerns on 2nd Embodiment The flowchart which shows the process sequence of the model creation which concerns on 2nd Embodiment The figure which shows the example of data of the daily measurement data storage part which concerns on 2nd Embodiment. The figure which shows the example of data of the measurement data memory | storage part which concerns on 2nd Embodiment. The block diagram of the user evaluation apparatus according to the hardware usage condition which concerns on 3rd Embodiment The flowchart which shows the process sequence of the model creation which concerns on 3rd Embodiment The figure which shows the example of data of the category information specific part which concerns on 3rd Embodiment The figure which shows the example of data for model creation which concerns on 3rd Embodiment The figure which shows the example of data of the repair rate estimated value memory | storage part classified by category which concerns on 3rd Embodiment The block diagram of the user evaluation apparatus according to the hardware use condition which concerns on 4th Embodiment The block diagram of the user evaluation apparatus according to the hardware use condition which concerns on 5th Embodiment The figure which shows the example of data of the model category memory | storage part which concerns on 5th Embodiment The figure which shows the data example of the user evaluation model memory | storage part which concerns on 5th Embodiment The block diagram of the user evaluation apparatus according to the hardware usage condition which concerns on 6th Embodiment The figure which shows the data example of the tolerance | permissible_range memory | storage part which concerns on 6th Embodiment. The block diagram of the user evaluation apparatus according to the hardware use condition which concerns on 7th Embodiment The figure which shows the example of data of the user ranking model memory | storage part which concerns on 7th Embodiment The block diagram of the user evaluation apparatus according to the hardware use condition which concerns on 8th Embodiment The figure which shows the example of data of the standard repair cost memory | storage part which concerns on 8th Embodiment The figure which shows the example of data of the guarantee category memory | storage part which concerns on 8th Embodiment The block diagram of the user evaluation apparatus according to the hardware use condition which concerns on 9th Embodiment The figure which shows the screen image transmitted from the process part which concerns on 9th Embodiment.

In the present invention, information specifying the products quality is constant, previously calculated based on the failure rate p _m in the range m quality is constant in the failure information. The estimated value of the repair rate described later is an estimated value based on this failure rate. On the other hand, the PC is equipped with a function for monitoring temperature and vibration, and there is a device for storing these, and if the user permits it, the information is transmitted. This monitoring information is transmitted to the server together with the serial number, and this information is recorded on the server. Here, the recorded information, that the relevant product is to be a certain period if the information as to whether the failed failed based on the load value 1-p _m, the load value p _m unless failed Thus, the influence of the failure rate caused by the product is removed. The presence or absence of a failure may be the presence or absence of a repair history. Then, by learning a sample with a load value, a user evaluation model for estimating the repair rate from the user's usage is created, and the user's usage is evaluated using this user evaluation model.

  Thereby, in the present invention, it is also possible to feed back usage to the user and to calculate an appropriate extended guarantee fee. It is possible to calculate an appropriate guarantee fee according to the user's responsibility, and it is possible to provide a PC extended warranty service at a lower price for users who use it safely.

(First embodiment)
FIG. 1 is a block diagram of a user evaluation apparatus according to the hardware usage status according to the first embodiment. This apparatus includes a measurement data receiving unit 101, a measurement data storage unit 102, a failure information collection result storage unit 105, a repair rate calculation unit 106, a repair rate estimated value storage unit 107, and a sample-specific load value determination unit 108. A sample-specific load value storage unit 109, a parameter determination unit 110, a user evaluation model storage unit 111, and a processing unit 113.

The failure information collection result storage unit 105 stores failure information collection results for a plurality of machines having the same characteristics as the plurality of machines to be measured. The repair rate calculation unit 106 calculates a repair rate estimation value for each machine based on the failure information collection result stored in the failure information collection result storage unit 105. The repair rate estimated value storage unit 107 stores the repair rate estimated value calculated by the repair rate calculation unit 106. The measurement data receiving unit 101 receives measurement data related to the usage status of a plurality of machines that are measurement targets. The measurement data storage unit 102 stores the measurement data received by the measurement data receiving unit 101. The load value determination unit by sample 108 sets a plurality of machines to be measured as a plurality of samples, and sets a load value that cancels the difference between the machines in the repair rate estimation value stored in the failure information collection result storage unit 105. Determine for each of the samples. The load value storage unit 109 stores the load values of the plurality of samples determined by the sample-specific load value determination unit 108. The parameter determination unit 110 includes the sample-specific load value determined by the sample-specific load value determination unit 108, the measurement data received by the measurement data reception unit 101, and the failure stored in the failure information collection result storage unit 105. Using the information collection result as an objective function, the parameters of the user evaluation model for calculating the usage influence degree are determined so as to maximize the objective function described above. The user evaluation model storage unit 111 stores the parameters of the user evaluation model determined by the parameter determination unit 110. The user evaluation unit 112 calculates the use influence degree using the measurement data received by the measurement data receiving unit 101 and the parameters of the user evaluation model determined by the parameter determination unit 110. The processing unit 113 displays the usage influence degree calculated by the user evaluation unit 112.

  In the following, after describing the model creation processing procedure according to the flowchart of FIG. 2, the user evaluation processing procedure will be described according to the flowchart of FIG. 3.

[Model creation]
In step F101 of FIG. 2, the measurement data receiving unit 101 receives measurement data. Data is sent via the Internet and received. Further, the received content is stored in the measurement data storage unit 102. An example of data stored in the measurement data storage unit 102 is shown in FIG. As shown in FIG. 4, information such as “machine start time”, “machine start time”, and “high temperature time” is stored in the measurement data storage unit 102 for each serial number. These pieces of information are caused by how the machine is used, but are likely to affect the repair rate of the machine. Such information is referred to as “index” in this specification.

  In the present embodiment, a laptop computer is assumed as target hardware. The measurement data is collected by using the number of impacts, CPU accumulated operating time, memory accumulated usage rate, and the like as indexes. When the target hardware is a product other than a computer, the index is also different. For example, in television, the number of channel switching may be used as one of the indexes. In the case of a refrigerator, power consumption, refrigerator temperature, and the like are important indicators.

  Next, in step F102, the repair rate calculation unit 106 calculates the repair rate based on information in the shipping information storage unit 103, the manufacturing information storage unit 104, and the failure information storage unit 105. The shipping information storage unit 103 stores the data shown in FIG. 5, and the manufacturing information storage unit 104 stores the data shown in FIG. 6, and stores data of all manufactured products. On the other hand, the data shown in FIG. 7 is stored in the failure information tabulation result storage unit 105, and information regarding failures is stored for all machines that have been repaired.

  According to the shipping information stored in the shipping information storage unit 103, the total number of operating days to date of the number of machines that have been shipped is, for example, (total operating days = current date−sales base shipping date−30). Can be estimated. In this example, 30 is subtracted on the assumption that the use of the machine starts about 30 days after the sales base is shipped. The current date may be set according to an actual calendar, or may be set to a date before the current date in consideration of a time lag until data is collected. In particular, if the total number of working days becomes negative, it must be replaced with 0.

  In this embodiment, considering that the manufacturer is requested to repair 100% within the warranty period (for example, one year), if the total number of working days is longer than the warranty period, the total number of working days is replaced with the warranty period. To calculate the number of working days within the warranty period. This is expressed as follows.

Operating days within warranty period = min [max [total operating days, 0], warranty period]
Further, the failure information storage unit 105 stores information on machines that have been repaired so far.

  As a result, the model creation data as shown in FIG. 8 is stored in the model creation data storage unit 114. The model creation data includes a manufacturing number, the number of machine operating days within the warranty period, the number of failures within the warranty period, the model name, the manufacturing date, and the component configuration information.

  Next, by classifying machines according to the information in the fourth column in FIG. 8, that is, the column after “model name”, a plurality of categories having different failure rates are created. For example, in FIG. 8, the manufacturing plant separates the machines of CHINA2, that is, serial numbers “23057V125J” and “23057V127J” from the whole. These two products are set as the first category, and the remaining products are set as the second category. In this case, it has been operated for a total of 131 (= 65 + 66) days and failed twice, but the other machines have been operated for a total of 262 (= 65 + 65 + 66 + 66) days and have never failed, and the failure rates are different from each other. One category was created. A category can be detected by classifying information in the fourth and subsequent columns of model creation data as shown in FIG. 8 using a method called a decision tree.

  As for how to create a decision tree, the failure rate of products manufactured during a specific period is high (initial failure, mixing of defective parts, elimination of problems due to improvement measures), and the quality of the product depends on the manufacturing plant. You may carry out from a viewpoint of being different and having a problem in a specific component. The classification using category specifying information that can specify a product group that provides each constant quality will be described later in another embodiment.

  Based on the created decision tree, a decision rule is created based on the items in the fourth and subsequent columns of the model creation data as shown in FIG. An example of the decision rule is shown in FIG.

  Once the decision tree is created, an estimate of the repair rate can be calculated for each machine. For example, if the model name is “PIPLUP” and the date of manufacture is June 20, 2008, it can be seen from the decision rule in FIG. 9 that this machine belongs to CATEGORY B, and the repair rate is estimated based on the failure rate of the category. The value p = 0.015267 is obtained. However, for example, the probability is expressed in units of PPM (parts per million).

  As described above, the repair rate calculation unit 106 receives data other than the serial number and the number of failures within the warranty period as input, determines parameters to be given to the model that outputs probability values, and determines from the model characterized by the parameters. The output probability value is used as an estimated repair rate. The repair rate estimated value obtained in this way is stored in the repair rate estimated value storage unit 107. An example of data in the repair rate estimated value storage unit 107 is as shown in FIG. 10, and the serial number and the estimated value of the repair rate are stored in pairs.

  Next, in step F103, the load value for each sample is determined. Samples accumulated in the measurement data storage unit 102 are stored in a format as shown in FIG. The “sample” here refers to a machine (sample) from which measurement data is obtained. The estimated value p of the repair rate stored in the repair rate estimated value storage unit 107 is obtained from the serial number in the measurement data of FIG. Also, based on the data stored in the failure information tabulation result storage unit 105, is the machine with the serial number repaired within, for example, one year from the machine use start date stored in the measurement data storage unit 102? Whether or not there is a repair history is determined. For example, the value of c is 1 when there is a repair history, and the value of c is 0 when there is no repair history.

  Furthermore, the load value w is determined by the following equation (1).

w = c (1-p) + (1-c) p (1)
This load value is the probability (1-p) that no repair is made if there is an actual repair history, and the probability (p) that repair is made if there is no repair history. This value is stored in the sample-specific load value storage unit 109 in association with the manufacturing number. A specific example of the load value for each sample is shown in FIG.

Next, in the parameter determination of step F104, using the data stored in the sample-by-sample load value storage unit 109, the measurement data storage unit 102, and the failure information tabulation result storage unit 105, for example, an objective function represented by the following equation (2) is used. Set.

However,

Is called the use influence degree and is given by the following equation (3).

  The subscript i corresponds to the serial number and is used to identify the machine. U is a set of machines stored in the measurement data storage unit 102 used for the model creation this time. The U may not necessarily be the entire data, but may be a subset such as the latest one year.

Also the indicator, ie

The j-th component of is u _ji . The index u _ji is the jth explanatory variable based on the data in the measurement data storage unit 102. In this embodiment, for example, (u _1i , u _2i , u _3i , u _4i , u _5i ) = (machine start time, machine continuous start time (average), number of shocks, vibration total, CPU cumulative operation rate / machine start time ).

Here, in order to explain the significance of multiplying the load value by Equation (2), a set V of machines having a repair rate of p is considered, and a discussion about this is shown.

However, n ₁ and n ₀ are the number of machines with repair history and the number of machines without repair history respectively included in V.

Here, a value obtained by multiplying a machine having a repair history belonging to the set V by a load value, that is, a value c _i (= 1) indicating presence / absence of a repair history, and a value obtained by multiplying a load value w _i by i∈V. The sum is

It is. Further, a sum of i∈V of a value obtained by multiplying a machine having no repair history belonging to the set V by a load value, that is, a value c _i (= 0) indicating presence / absence of a repair history and a load value w _i. Is

It is. Here, if the model created by the repair rate calculation unit 106 is correct,

You may think. Using these equations,

It becomes. As a result, the repair rate p ′ when the number of machines is counted by multiplying the load value is

It becomes. This does not depend on the original repair rate p. That is, the repair in the case where the load value is taken into consideration for both the set V of machines having a large output value p and the set V ′ of machines having a small repair rate p created by the repair rate calculation unit 106. The rate will be the same. This means that the difference in the repair rate p has been canceled by multiplying by the load value. By multiplying the load value by the formula (2), the difference in the repair rate estimated from the conventional quality information (103 to 105) is eliminated, and the difference in the repair rate caused by the difference depending on the use situation of the user is evaluated. You can build a model.

In the parameter determination in step F104, the parameter is set so that the objective function determined by the equations (2) and (3) is maximized.

To decide.

  The parameters determined here are stored in the user evaluation model storage unit 111. An example of data stored in the user evaluation model storage unit 111 is shown in FIG.

[User Evaluation]
A user evaluation processing procedure will be described with reference to the flowchart of FIG. First, in step F111, the user evaluation unit 112 receives measurement data (items in FIG. 4) from the measurement data reception unit 101, processes them, and performs the same processing as in the parameter determination in step F104.

Is calculated. As described above, the index (u _1i , u _2i , u _3i , u _4i , u _5i ) = (machine start time, machine continuous start time (average), number of shocks, vibration total, CPU cumulative operation rate / machine start time ).

In addition, parameters stored in the user evaluation model storage unit 111

Is read, and the degree of use influence is calculated by equation (3).

Is calculated. In step F112, the processing unit 113 displays the usage influence degree calculated by the user evaluation unit 112 on a screen such as a display.

(Second Embodiment)
FIG. 13 is a block diagram of the user evaluation apparatus according to the hardware usage status according to the second embodiment, and FIG. 14 is a flowchart showing a processing procedure of model creation according to the second embodiment.

  Instead of the measurement data storage unit 102 in the configuration according to the first embodiment shown in FIG. 1, as shown in FIG. 13, in the second embodiment, the daily measurement data storage unit 215, the measurement data aggregation processing unit 216, A measurement data storage unit 217 is provided. In the second embodiment, the model creation processing procedure is changed as shown in FIG. The user evaluation is the same as that in the first embodiment, and is the same as that shown in FIG. Here, step F205 (measurement data aggregation) in the flowchart (model creation) in FIG. 14 will be described mainly.

  First, in step F201, measurement data from the user is received every day, and the measurement date and the measurement data are recorded in the daily measurement data storage unit 215 in the format shown in FIG. The measurement data is received every day, but there are days when the user does not connect the machine to the net or cannot be received due to circumstances such as the machine being turned off at the transmission / reception set time. In this case, the data that could not be received will be received together at a later date. Further, when data is recorded every time the machine is turned on / off, data of the same date may be stored in a plurality of rows as on September 1 in FIG.

  In the measurement data tabulation in step F205, when one year has passed since the machine use start date stored in the daily measurement data storage unit 215, the data for the year is tabulated. For example, using the machine continuous startup time of the daily measurement data storage unit 215, the average or variance is calculated and stored in the measurement data storage unit 217. In parallel with the tabulation work, it is checked whether there is any contradiction in the data. For example, a check is performed such that the maximum number of impacts is not 0 even when the number of impacts is 0, or the CPU cumulative operation rate is too large compared to the machine continuous startup time. The presence or absence of contradiction is recorded in the measurement data storage unit 217 together with the total result. An example of data in the measurement data storage unit 217 is shown in FIG.

  As a result, inconsistent data is not used in determining the load value for each sample in step F203 and determining the parameter in step F204. Also, regarding user evaluation, the user who provided the contradictory data is not evaluated.

(Third embodiment)
FIG. 17 is a block diagram of the user evaluation apparatus according to the hardware usage status according to the third embodiment, and FIG. 18 is a flowchart showing a processing procedure of model creation according to the third embodiment.

  Compared with the configuration of the first embodiment shown in FIG. 1, a category specifying information storage unit 315 is added in the third embodiment as shown in FIG. Data in the category specifying information storage unit 315 is shown in FIG. In FIG. 19, if the model, date of manufacture, and factory are determined, the corresponding category can be specified. If they are in the same category, they can be considered to have the same quality and the same repair rate. For example, the model PIPLUP has been manufactured from September 14, 2008, but the quality has improved after the manufacturing change instruction is issued on October 15. For this reason, since the quality from September 14 to October 15 is the same, it is set as the same category. PIPLUP and EMPOLEON are almost the same product, but have different names for business reasons. Therefore, EMPOLEON manufactured at the same time is in the same category 1 as PIPLUP. As described above, the category specifying information may include a category number, a model name, and a manufacturing period.

  The category-specific repair rate calculation in step F302 is executed by the category-specific repair rate calculation unit 316. From the shipping information storage unit 303, the manufacturing information storage unit 304, and the failure information tabulation result storage unit 305, data as shown in FIG. 8 can be created by the same processing procedure as in the first embodiment. Further, based on the category specifying information stored in the category specifying information storage unit 315, the category corresponding to each serial number is determined, and the model generation data shown in FIG. 20 is generated and stored in the model generation data storage unit 314. To do. Furthermore, using this data, the repair rate by category is calculated by calculating the ratio between the number of working days within the warranty period and the number of failures. The category-specific repair rate calculated here is stored in the category-specific repair rate estimated value storage unit 307. An example of data in the category-specific repair rate estimated value storage unit 307 is shown in FIG. The user evaluation processing procedure is the same as in the first embodiment.

(Fourth embodiment)
In the fourth embodiment, another method of realizing the sample-specific load value determination units (108, 208, 308) in the first to third embodiments will be described.

FIG. 22 is a block diagram of a user evaluation device according to the hardware usage status according to the fourth embodiment. In addition to the configuration shown in FIG. 1, this apparatus further includes a load value parameter storage unit 417 that stores a load value parameter s. The load value determination unit 408 by sample sets the load value to (1-p) ^s / p if there is a repair history, depending on the presence or absence of the repair history represented by the failure information tabulation result when the repair rate estimation value is p. ^1-s decided to, to determine the load value if no repair history to ^{p s / (1-p)} 1-s.

Specifically, the load value is obtained as in the following equation (4).

Here, s is an arbitrary real number.

  Thus, as a load value for canceling the difference in the repair rate p, there can be the above formula (4) in addition to the above formula (1).

When the same calculation as in the first embodiment is performed,

And using [Equation 8],

Therefore, the difference in the repair rate p can be canceled by multiplying by the load value.

Therefore, in the present embodiment, the load value is calculated according to the preset load value parameter s. Assuming that the load value parameter s = 0 is set in the load value parameter storage unit 417, the load value determination unit 408 for each sample

The load value is calculated according to Other processes are the same as those in the third embodiment described above.

(Fifth embodiment)
FIG. 23 is a block diagram of a user evaluation device according to the hardware usage status according to the fifth embodiment. The model category storage unit 517 further stores a model category that provides a corresponding relationship with at least one model, and the parameter determination unit 510 includes a user evaluation model for calculating the use influence degree for each model category. The user evaluation model storage unit 511 stores parameters of the user evaluation model for calculating the use influence degree for each model category, and the user evaluation unit 512 uses the use influence degree for each model category. The use influence degree is calculated using the parameter of the user evaluation model for calculating.

  As described above, in this embodiment, the parameter determination unit 510 creates separate user evaluation models for each category stored in the model category storage unit 517 in advance. There are several ways to categorize user evaluation models in this way.

(Significance 1) Since general users and corporate users are expected to use differently, it is better to create user evaluation models separately. Therefore, the categories are divided and registered in the model category storage unit 517 from the viewpoint of whether the machine is for a general user or a machine for a corporate user.

(Significance 2) It is better to create separate user evaluation models because the expected usage is different between machines that sell robust designs and machines that do not. For this reason, categories are divided and registered in the model category storage unit 517 from the viewpoint of robust design or general design.

An example of data in the model category storage unit is as shown in FIG. 24, and the parameter determination unit 510 creates a user evaluation model for each set of samples belonging to each category. For example, the model belonging to the company robust model category is the "PP001", "PP002", a set of samples of these two models and U _1. Similarly, a set of samples for the enterprise general model is U ₂ , a set of samples for the consumer model is U _3, and the following objective function is set for each set.

However, there are three categories k = 1, 2, and 3. A parameter is collected for each k,

Get. These are stored in the user evaluation model storage unit 511. Moreover, the example of the data of the user evaluation model memory | storage part 511 which concerns on this embodiment is shown in FIG.

  The general procedure for user evaluation is the same as that shown in FIG. 3, but the specific processing for user evaluation in step F111 is different from that in the first embodiment.

The user evaluation unit 512 specifies the model from the manufacturing number received from the measurement data receiving unit 501 with reference to the manufacturing information storage unit 504, and further specifies the category with reference to the model category storage unit 517. Next, by reading out the parameter of the corresponding category stored in the user evaluation model storage unit 511 and using this parameter, the use influence degree is obtained by Expression (3).

Is calculated.

(Sixth embodiment)
FIG. 26 is a block diagram of a user evaluation device according to the hardware usage status according to the sixth embodiment. This embodiment is a modification of the fifth embodiment. The difference in configuration from the fifth embodiment is that an allowable range storage unit 518 is provided. The permissible range storage unit 518 stores a permissible range composed of a set of index name, upper limit value, and lower limit value. The parameter determination unit 510 calculates an index whose value does not change between the upper limit value and the lower limit value corresponding to the index name from the measurement data and the allowable range, and calculates the index, the load value for each sample, and the failure information aggregation When the result is the objective function, the parameter of the user evaluation model for calculating the usage influence degree is determined so as to maximize the objective function. In the present embodiment, a model category storage unit 515 may be provided that stores a model category that gives a corresponding relationship with at least one model. In this case, the allowable range storage unit 518 includes, for each model category, An allowable range composed of a set of index name, upper limit value, and lower limit value is stored. The parameter determination unit 510 calculates an index whose value does not change between the upper limit value and the lower limit value corresponding to the model category and the index name from the measurement data and the allowable range, and the calculated index When the objective function is the load value by sample, measurement data, and failure information aggregation result, the objective function parameter for calculating the usage impact for each model category is maximized. decide.

  The allowable range is an upper limit value and a lower limit value for each index given for each category, and FIG. 27 shows an example of the allowable range.

If the category of machine i is k and the upper limit of index j is

And set the lower limit to

The index shown in the first embodiment

Is converted as follows.

Further, a new index Z _ji is obtained by the following equation.

The parameter determination unit 510 sets the objective function using Z _ji as follows.

However,

Is called the degree of influence of use and is given by the following equation.

In the user evaluation, the user evaluation unit 512 calculates the index Z _ji according to the equations (6) and (7).

(Seventh embodiment)
FIG. 28 is a block diagram of a user evaluation device according to the hardware usage status according to the seventh embodiment.

  The structural difference from the first embodiment shown in FIG. 1 is that a user rank model storage unit 715 for storing values corresponding to ranks for each index is provided.

  The parameter determination unit 710 calculates the index value rank from the user rank model for each of a plurality of samples, and uses the index value, the load value by sample, and the failure information aggregation result as the objective function, The parameter of the user evaluation model for calculation is determined so as to maximize the objective function. The user evaluation unit 712 calculates the use influence degree by using the measurement data, the order of the index values, and the parameters of the user evaluation model. The value corresponding to the rank for each index may be a percentile point in the sample. Or it is good also considering the value corresponding to the order | rank for every parameter | index as the deviation value in a sample.

Specifically, when the parameter determination unit 710 defines the index, the data in the measurement data storage unit 702 are arranged in order from the smallest index value j, and the order v _ji of the sample of the production number i is obtained.

further,

As

Ask for. This is a percentage of the total from the smallest. In addition, the value of u _{ji in which} the value of s _ji is 1%, 2%, 3% _,. An example of data to be recorded is shown in FIG.

The parameter determination unit 710 sets the objective function as follows using s _ji .

However,

Is called the degree of influence of use and is given by the following equation.

In addition, parameters are set so that this objective function is maximized.

Is stored in the user evaluation model storage unit 711.

(Eighth embodiment)
FIG. 30 is a block diagram of the user evaluation device according to the hardware usage status according to the eighth embodiment. The difference in configuration from the first embodiment shown in FIG. 1 is that a standard price storage unit 815, a guarantee category storage unit 816, and a guarantee fee calculation unit 817 are added. Data of the standard price storage unit 815 is shown in FIG. As shown in FIG. 32, the data in the warranty category storage unit 816 is a group of models corresponding to the categories, and is also given a standard repair rate for each category. The reference repair rate is a numerical value determined in advance when providing insurance services.

The guarantee fee calculation unit 817 first refers to the shipping information storage unit 803 and obtains the model name from the production number. Further referring to guarantee category storage unit 816, and the assurance category g corresponding to the model name, obtains the reference repair rate p _g in that category. Further, the user evaluation unit 812 uses the degree of use influence.

Get. Using these, the assumed repair rate P _i is calculated by the following equation.

In this equation, the portion of the repair rate by the product-specific quality differences regarded as reference repair rate p _g, because it calculates the repair rate by considering only use influence, referred to as a repair factor considers the calculation result.

The standard repair cost C _i is obtained from the standard price storage unit 815. Using this, an appropriate guarantee fee is calculated as D _i = C _i (P _i + r). Here, r is a ratio for adding profit, and is given in advance. The calculated D _i is displayed on the screen as an insurance premium.

  As described above, the present embodiment includes the standard price storage unit 815 that stores the standard price for each machine, and the guarantee fee calculation unit 817 that calculates the guarantee fee from the degree of influence of use and the standard price. The standard price may be a selling price of the machine. Alternatively, the standard price may be a standard repair price for the machine.

  Moreover, you may further comprise the security category memory | storage part 816 which memorize | stores the reference | standard repair rate provided for every security category. In this case, the guarantee fee calculation unit 817 calculates a guarantee fee based on the usage impact, the standard price, and the reference repair rate.

  Further, the guarantee fee calculation unit 817 has A / (A + B) where A is the product of the standard repair rate and the usage impact level, and B is the product of (1-standard repair rate) and (1-usage impact level). It is also possible to obtain the assumed repair rate expressed by the following, and calculate the guarantee fee from the assumed repair rate and the standard price.

(Ninth embodiment)
FIG. 33 is a block diagram of the user evaluation device according to the hardware usage status according to the ninth embodiment. The configuration shown in FIG. 33 is a combination of the sixth embodiment and the seventh embodiment. As a result, the user evaluation unit 912 can obtain information on the allowable range and rank for each index. By collecting these pieces of information, it is possible to present a screen as shown in FIG. 34 to the user. That is, the index name is specified on the left side, and S _ji is displayed next to it. Further, the upper limit value and the lower limit value of the index corresponding to the machine can be acquired from the allowable range storage unit 918, the upper limit value and the lower limit value can be converted into percentile points by the user rank model storage unit 919, and the result can be displayed. it can.

  As described above, in the present embodiment, the user rank model storage unit 919 that stores a value corresponding to the rank for each index, and the allowable range storage unit 918 that stores a set of an index name, an upper limit value, and a lower limit value for each category of the model. The user evaluation unit 912 calculates a machine percentile point and an upper limit value and a lower limit value percentile point for each index. The processing unit 913 can display a graph representing these values on the screen.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

101: Measurement data receiving unit;
102: Measurement data storage unit;
103. Shipping information storage unit;
104 ... manufacturing information storage unit;
105 ... failure information tabulation result storage unit;
106: Repair rate calculation unit;
107 ... repair rate estimated value storage unit;
108: Load value determination unit for each sample;
109 ... Sample-specific load value storage unit;
110 ... parameter determination unit;
111 ... user evaluation model storage unit;
112 ... User evaluation section;
113 ... Processing unit

Claims (18)

A failure information collection result storage unit for storing failure information collection results for a plurality of machines having the same characteristics as the plurality of machines to be measured;
A repair rate calculator for calculating a repair rate estimate for each machine based on the failure information collection result;
A repair rate estimated value storage unit for storing the repair rate estimated value;
A measurement data receiving unit that receives measurement data related to the usage status of a plurality of machines to be measured;
A measurement data storage unit for storing the measurement data;
A load value determination unit for each sample that determines a load value that cancels a difference between machines of the repair rate estimation value as a plurality of samples as a plurality of machines to be measured, and for each of the plurality of samples,
A load value storage unit for each sample that stores the load value of each of the plurality of samples;
An objective function is generated using the load value for each sample, the measurement data, the failure information collection result, and a parameter of a user evaluation model for calculating a use influence degree, and the parameter of the user evaluation model is A parameter determination unit that determines to maximize the objective function;
A user evaluation model storage unit for storing parameters of the user evaluation model;
A user evaluation unit that calculates the usage impact using the measurement data and parameters of the user evaluation model;
A user evaluation device according to a hardware usage state, comprising: a processing unit that displays the usage influence degree.   The load value determination unit for each sample determines the load value to be 1-p if there is a repair history depending on the presence or absence of the repair history represented by the failure information aggregation result when the repair rate estimated value is p. The user evaluation apparatus according to claim 1, wherein the load value is determined as p if there is no repair history. A load value parameter storage unit for storing the load value parameter s;
The load value determination unit for each sample sets the load value to (1-p) if there is a repair history, depending on the presence or absence of a repair history represented by the failure information tabulation result when the repair rate estimation value is p. determined in ^{s /} p ^1-s, according to the load value if no repair history in hardware usage of claim 1, wherein the determining the ^{p s / (1-p)} 1-s User evaluation device. A category identification information storage unit for storing category identification information for identifying a set with a fixed repair rate;
A category-specific repair rate calculation unit that calculates the repair rate estimate for each category specified by the category specifying information;
The user evaluation according to the hardware usage state according to claim 2, further comprising: a category-specific repair rate estimated value storage unit that stores the repair rate estimated value for each category. apparatus. A model creation data storage unit for storing model creation data including a serial number, machine operation days within the warranty period, number of failures within the warranty period, model name, production date, and part configuration information;
The repair rate calculation unit receives data other than the serial number and the number of failures within the warranty period, determines a parameter to be given to a model having a probability value as an output, and outputs from the model characterized by the parameter 4. The user evaluation apparatus according to the hardware usage status according to claim 2, wherein a probability value is used as the repair rate estimation value.   The user evaluation apparatus according to the hardware usage status according to claim 4, wherein the category specifying information includes a category number, a model name, and a manufacturing period. A daily measurement data storage unit for storing daily measurement data based on the measurement data received multiple times from each machine by the measurement data receiving unit;
A measurement data totaling processing unit for counting the daily measurement data and detecting the presence or absence of data inconsistency;
A total data storage unit that stores the results totaled by the measurement data totalization processing unit and the presence or absence of data inconsistency;
The load value determination unit for each sample obtains the load value for the machine having no contradiction,
The user evaluation apparatus according to claim 2 or 3, wherein the user evaluation unit calculates the use influence degree with respect to the machine having no contradiction. A model category storage unit for storing a model category each of which gives a correspondence relationship with at least one model;
The parameter determination unit determines a parameter of a user evaluation model for calculating the use influence degree for each model category,
The user evaluation model storage unit stores, for each model category, parameters of a user evaluation model for calculating the usage influence degree,
7. The hardware usage status according to claim 6, wherein the user evaluation unit calculates a usage influence degree using a parameter of a user evaluation model for calculating the usage influence degree for each model category. Responding user evaluation device. It further comprises an allowable range storage unit that stores an allowable range consisting of a set of an index name, an upper limit value, and a lower limit value,
The parameter determination unit
From the measurement data and the allowable range, calculate an index whose value does not change between the upper limit value and the lower limit value corresponding to the index name,
An objective function is generated using the index, the load value for each sample, the failure information tabulation result, and a parameter of a user evaluation model for calculating a use influence degree, and the parameter of the user evaluation model is set as an objective function 7. The user evaluation device according to the hardware usage status according to claim 6, wherein the user evaluation device is determined so as to maximize the value. It further comprises a user rank model storage unit that stores values corresponding to ranks for each index,
The parameter determination unit calculates an index value rank from the user rank model for each of the plurality of samples, the index value, the load value by sample, the failure information aggregation result, and the user evaluation model. And generating an objective function using the parameters of and determining the parameters of the user evaluation model to maximize the objective function,
4. The hardware according to claim 2, wherein the user evaluation unit calculates a use influence degree by using the measurement data, the ranking of the index value, and a parameter of the user evaluation model. 5. User evaluation device according to usage.   11. The user evaluation apparatus according to hardware usage status according to claim 10, wherein the value corresponding to the rank for each index is a percentile point in a sample.   11. The user evaluation device according to the hardware usage status according to claim 10, wherein the value corresponding to the rank for each index is a deviation value in a sample. A standard price storage unit that stores the standard price for each machine;
The user evaluation apparatus according to the hardware usage status according to claim 1, further comprising: a guarantee fee calculation unit that calculates a guarantee fee from the use influence level and the standard price. A security category storage unit for storing a standard repair rate given for each security category is further provided.
14. The user evaluation apparatus according to the hardware usage status according to claim 13, wherein the guarantee fee calculation unit calculates a guarantee fee based on the use influence level, the standard price, and the reference repair rate.   14. The user evaluation apparatus according to the hardware usage status according to claim 13, wherein the standard price is a selling price of a machine.   14. The user evaluation device according to the hardware usage status according to claim 13, wherein the standard price is a standard repair price of the machine.   When the product of the standard repair rate and the usage impact is A, and the product of (1-standard repair rate) and (1-usage impact) is B, the guarantee fee calculation unit is A / (A + B) 15. The user evaluation apparatus according to the hardware usage state according to claim 14, wherein a deemed repair rate represented by the following is obtained, and a guarantee fee is calculated from the deemed repair rate and the standard price. A user rank model storage unit for storing values corresponding to the ranks for each index;
And an allowable range storage unit that stores a set of index name, upper limit value, and lower limit value for each category of the model,
The user evaluation unit calculates the percentile point of the machine for each of the indicators, and the upper and lower percentile points of the indicator,
4. The user evaluation apparatus according to claim 2 or 3, wherein the processing unit displays a graph representing these values on a screen.

Images