WO2018150539A1

WO2018150539A1 - Data acquisition device, data acquisition method, and data acquisition program

Info

Publication number: WO2018150539A1
Application number: PCT/JP2017/005900
Authority: WO
Inventors: 祐司小関; 直樹脇阪
Original assignee: 株式会社日立製作所
Priority date: 2017-02-17
Filing date: 2017-02-17
Publication date: 2018-08-23
Also published as: WO2018150639A1; JP6591080B2; JPWO2018150639A1

Abstract

A data acquisition device for acquiring data from a database according to a data acquisition request from a requester requesting data from the database, which stores a data group, and returning said data to the requester, the data acquisition device being configured so as to: learn the data acquisition request type and the acquisition timing defining the data acquisition request chronological order for each of the data acquisition requests among a series of data acquisition requests received from the requester; identify, in advance of a first data acquisition request following the series of data acquisition requests, the acquisition timing and type of a first subsequent estimated data acquisition request which is a data acquisition request which could be received from the requester, on the basis of the learning results obtained from the learning processing; acquire, from the database, data corresponding to the first subsequent estimated data acquisition request for which the type was identified; determine, when the first data acquisition request is received, whether or not there is a match between the type included in the first data acquisition request and the type of the first subsequent estimated data acquisition request identified by the identification processing; and when determined there is a match, transmit the advance acquisition results to the requester at the acquisition timing of the identified first subsequent estimated data acquisition request.

Description

Data acquisition device, data acquisition method, and data acquisition program

The present invention relates to a data acquisition device, a data acquisition method, and a data acquisition program for acquiring data.

Common Information Model (CIM) is a standard information model and in-system / application data in a system that operates power facilities (power stations, substations, power transmission lines, etc. that constitute the power system) in the backbone energy management system (EMS). This model specifies exchange. By sharing power system information with CIM, data exchange between power system monitoring, control, and analysis applications is facilitated. Thereby, the connectivity between applications and the expandability of the system can be improved.

For example, there is a power network model database (hereinafter referred to as a power NW model DB) managed in a relational database format (Relational Database (RDB) format). The power NW model DB stores information on power facilities in the backbone system EMS. For example, in a user operation or business process, information satisfying a set of attributes whose position information is “point A” and whose other attribute is “high voltage equipment” is acquired from the power NW model DB. In business using the power NW model DB, it is required to respond quickly to data acquisition requests, but data conversion processing from RDB format to Extensible Markup Language (XML) format is performed each time a data acquisition request is made. It becomes difficult to respond quickly.

As a technique for quickly responding to a data acquisition request, for example, there is Patent Document 1 below. In the cache method of Patent Document 1, when the search request data from the client does not exist in the cache database, the master database returns the search result record to the distributed cache database, and turns on the corresponding bit on the reference bitmap table. Thereafter, when there is a search request for the same record, the cache method of Patent Document 1 returns data on the distributed cache database. When the master database is updated, the cache method of Patent Document 1 updates data only to the distributed cache database whose reference bit is ON.

JP 2000-20385 A

However, the data cached by the cache method of Patent Document 1 is not necessarily the data required at that time for user operation or business processing in a system that operates power equipment in the power system. Therefore, the cache method disclosed in Patent Document 1 does not necessarily improve the data response.

The object of the present invention is to suppress response delay of data acquired in response to a request.

A data acquisition device, a data acquisition method, and a data acquisition program that are one aspect of the invention disclosed in this application are accessible to a database that stores a data group, and respond to a data acquisition request from a request source for the database A data acquisition apparatus, a data acquisition method, and a data acquisition program for acquiring data from the database and returning the data to the request source, the processor executing the program, and a storage device for storing the program, Learning to learn a type of the data acquisition request and an acquisition timing that defines a time series order of the data acquisition request for each data acquisition request of the series of data acquisition requests received from the request source Processing and the first data acquisition request after the series of data acquisition requests Standing on the basis of the learning result of the learning process, a specifying process for specifying the type and acquisition timing of the first next estimated data acquisition request that is a data acquisition request that can be acquired from the request source, and a type by the specifying process The data included in the first next data acquisition request when the first data acquisition request is received, and the pre-acquisition process for acquiring the data corresponding to the first next estimated data acquisition request identified from the database And the type of the first next estimated data acquisition request specified by the specifying process, and the pre-acquisition when it is determined by the determining process that the type of the first next estimated data acquisition request is the same The pre-acquisition result by the process is sent to the request source at the acquisition timing of the first next estimated data acquisition request specified by the specifying process. Characterized by a transmission processing for the execution.

According to a typical embodiment of the present invention, it is possible to suppress response delay of data acquired in response to a request. Problems, configurations, and effects other than those described above will become apparent from the description of the following embodiments.

FIG. 1 is an explanatory diagram of an example of data pre-acquisition according to the first embodiment. FIG. 2 is an explanatory diagram illustrating a system configuration example of the data acquisition system. FIG. 3 is a block diagram illustrating a hardware configuration example of a computer. FIG. 4 is an explanatory diagram showing an example of the contents stored in the power NW model DB. FIG. 5 is an explanatory diagram illustrating an example of an acquisition pattern. FIG. 6 is an explanatory diagram showing an example of the stored contents of the data acquisition history accumulation table. FIG. 7 is an explanatory diagram of an example of the contents stored in the data acquisition estimation table. FIG. 8 is an explanatory diagram of an example of stored contents of the step information management table. FIG. 9 is an explanatory diagram of an example of stored contents of the pre-acquired information management table. FIG. 10 is an explanatory diagram of an example of stored contents of the activation timing management table. FIG. 11 is a block diagram illustrating a functional configuration example of the data acquisition apparatus. FIG. 12 is an explanatory diagram illustrating an example of a learning result table. FIG. 13 is a flowchart illustrating an example of a learning process procedure by the learning unit. FIG. 14 is a flowchart showing a detailed processing procedure example of the learning processing (step S1303) shown in FIG. FIG. 15 is a flowchart illustrating a detailed processing procedure example of the aggregation processing (step S1409) illustrated in FIG. FIG. 16 is a flowchart showing a detailed processing procedure example of the selection probability calculation process (step S1510) shown in FIG. FIG. 17 is a flowchart showing a detailed processing procedure example of the representative value calculation processing (step S1410) shown in FIG. FIG. 18 is a flowchart showing a detailed processing procedure example of the data acquisition estimation table update processing (step S1304) shown in FIG. FIG. 19 is a flowchart illustrating a detailed processing procedure example 1 of the data acquisition processing by the data acquisition device. FIG. 20 is a flowchart illustrating a detailed processing procedure example 2 of the data acquisition processing by the data acquisition device. FIG. 21 is an explanatory diagram of an example of transfer of the acquired pattern sequence. FIG. 22 is a flowchart illustrating a detailed processing procedure example of the data acquisition / history accumulation processing (steps S1906 and S2006) illustrated in FIGS. 19 and 20. FIG. 23 is a flowchart illustrating a detailed processing procedure example of the data pre-acquisition processing. FIG. 24 is an explanatory diagram of an example of stored contents of the data acquisition history accumulation table according to the second embodiment. FIG. 25 is an explanatory diagram of an example of stored contents of the second data acquisition estimation table. FIG. 26 is a flowchart of a detailed process procedure example 1 of the data acquisition process performed by the data acquisition apparatus according to the second embodiment. FIG. 27 is an explanatory diagram of an example of preliminary data acquisition according to the third embodiment. FIG. 28 is a flowchart of a detailed process procedure example of the data pre-acquisition process according to the third embodiment. FIG. 29 is an explanatory diagram of an example of starting timing adjustment according to the fourth embodiment. FIG. 30 is a flowchart illustrating an example of a startup timing adjustment processing procedure.

<Example of prior data acquisition>
FIG. 1 is an explanatory diagram of an example of data pre-acquisition according to the first embodiment. The data acquisition system 100 includes a terminal 101, an application 102, a data acquisition function 103, and a power NW model DB 104. The terminal 101 transmits a data acquisition request to the data acquisition function 103 via the application 102, and acquires data corresponding to the data acquisition request from the power NW model DB 104 via the application 102 and the data acquisition function 103. For example, the terminal 101 is provided in a power facility (a power plant or a substation constituting a power system) and is communicably connected to a server (data acquisition device) having a data acquisition function 103 via a communication network.

Application 102 is an interface between terminal 101 and data acquisition function 103. The application 102 is mounted on the terminal 101 or the data acquisition device. The application 102 is software that acquires data from the power NW model DB 104, although the application 102 differs depending on the corresponding power facility.

The data acquisition function 103 has a data access function 131, a pattern learning function 132, and a pre-acquisition function 133. The data access function 131 acquires data corresponding to the data acquisition request from the terminal 101 from the power NW model DB 104 and returns it to the request source application 102. Specifically, for example, the data access function 131 acquires, from the power NW model DB 104, entries of a plurality of attribute tables corresponding to the data acquisition request, converts them into one XML data, and converts the converted XML data to Return to the requesting application 102.

The pattern learning function 132 learns a series of data acquisition request patterns input from the terminal 101. For example, the worker A transmits a series of data acquisition requests including a data acquisition range of “circle (3 km)”, “cone”, and “cylinder” from the terminal 101. Therefore, in the case of the worker A, the pattern learning function 132 learns to select the data acquisition range in the order of “circle (3 km)” → “cone” → “cylinder”.

When the data acquisition request is received from the terminal 101, the pre-acquisition function 133 specifies the data acquisition range that the user of the terminal 101 will select next from the learning result of the pattern learning function 132, and the data access function 131 Is acquired before the next data acquisition request for requesting data within the specified data acquisition range.

The power NW model DB 104 is a database that stores a CIM that is a power NW model. The power NW model is a network model in which nodes of power equipment are used and electric wires are links. The power NW model is, for example, a three-dimensional model expressed by two-dimensional position information and a voltage of a node. The power NW model DB 104 has a table for each attribute such as system (high voltage or low voltage), facility type (power plant, substation, etc.), area, and position. As a result, the data access function 131 acquires a corresponding entry from the table for each attribute included in the data acquisition request and converts the entry into one XML data. Note that this conversion processing to XML data takes time and causes a reduction in response performance.

(1) Assume that worker A inputs a pattern. Specifically, for example, the worker A transmits a first data acquisition request from the terminal 101 in which the data acquisition range is a circle having a diameter of 3 km, and then the second data acquisition in which the data acquisition range is a cone. Assume that a request is transmitted from the terminal 101 and then a third data acquisition request is transmitted from the terminal 101 whose data acquisition range is a cylinder.

(2) The data access function 131 (1) accesses the power NW model DB 104 each time a data acquisition request is received, acquires entries in a plurality of attribute tables, and merges them into one XML data. The data access function 131 returns the XML data to the terminal 101.

(3) The pattern learning function 132 learns the data acquisition range in the order of “circle (3 km)” → “cone” → “column” selected by the worker A.

(4) After this, it is assumed that the worker A once logs out and logs in to the terminal 101 again at another opportunity. Then, it is assumed that the worker A transmits a first data acquisition request from the terminal 101 whose data acquisition range is a circle having a diameter of 3 km. In this case, as shown in (2) and (3) above, data access and pattern learning are executed for the first data acquisition request, but the prior acquisition function 133 is used for the worker A so far. From the learning result, the data acquisition range selected after the “circle with a diameter of 3 km” is specified as “cone”, and an acquisition formula including the specified data acquisition range “cone” is created, A prefetch instruction is given to the data access function 131. Thereby, the data access function 131 can acquire the data included in the specified data acquisition range “cone” from the power NW model DB 104 and convert it into XML data before the arrival of the second data acquisition request. . Thereby, the worker A can obtain an immediate response without waiting for the time required for the data acquisition process from when the second data acquisition request is made until the response is received.

Thus, since the data acquisition system 100 predicts what data acquisition request will arrive and prefetches the corresponding data, the response performance can be improved. In addition, by predicting data to be prefetched, it is possible to reduce the processing load due to prefetching.

<System configuration example>
FIG. 2 is an explanatory diagram showing a system configuration example of the data acquisition system 100. The data acquisition system 100 has a configuration in which a data acquisition device 200 that is a server and a terminal 101 are communicably connected via a communication network 201. The data acquisition device 200 has a power NW model DB 104. However, the power NW model DB 104 may be included in another server accessible from the data acquisition apparatus 200 via the communication network 201.

<Hardware configuration example>
FIG. 3 is a block diagram illustrating a hardware configuration example of a computer. The computer is, for example, the data acquisition device 200 and the terminal 101 shown in FIG. The computer 300 includes a processor 301, a storage device 302, an input device 303, an output device 304, and a communication interface (communication IF 305). The processor 301, the storage device 302, the input device 303, the output device 304, and the communication IF 305 are connected by a bus. The processor 301 controls the computer 300. The storage device 302 serves as a work area for the processor 301. The storage device 302 is a non-temporary or temporary recording medium that stores various programs and data. Examples of the storage device 302 include a ROM (Read Only Memory), a RAM (Random Access Memory), an HDD (Hard Disk Drive), and a flash memory. The input device 303 inputs data. Examples of the input device 303 include a keyboard, a mouse, a touch panel, a numeric keypad, and a scanner. The output device 304 outputs data. Examples of the output device 304 include a display and a printer. The communication IF 305 is connected to the communication network 201 and transmits / receives data.

<Power NW model DB 104>
FIG. 4 is an explanatory diagram showing an example of the contents stored in the power NW model DB 104. As described above, the power NW model DB 104 is a database that stores a CIM that is a power NW model, and the power NW model is a network model that uses nodes of power equipment and links electric wires. The power NW model DB 104 is realized by the storage device 302 described above. Specifically, the power NW model DB 104 is, for example, a set of tables in which node IDs are associated with each attribute (system, facility type, area, position information,...). Description will be made by paying attention to information held by the node N and the link L constituting the model.

The node N is identified by the node ID 401 and has a system 402, an equipment type 403, an area 404, position information 405, electrical characteristic information 406, and a connection link ID 407. The node ID 401 is identification information that uniquely identifies the node N indicating the power equipment. The system 402 is identification information indicating the system of the node N. The system | strain 402 is the information which shows a high voltage | pressure system or low voltage | pressure systemization, for example. The equipment type 403 is information indicating the type of equipment of the node N. The equipment type 403 includes, for example, a transformer and a steel tower.

Area 404 is information indicating an area to which node N belongs. The area 404 is defined, for example, by a municipality range. The position information 405 is information indicating the position of the node N stretched on the XY plane. The position information 405 is defined by, for example, latitude (value in the Y-axis direction) and longitude (value in the X-axis direction). The electrical characteristic information 406 is used as a value of the node N in the Z-axis direction. The electrical characteristic information 406 is, for example, a voltage value at the node N. The connection link ID 407 is identification information that uniquely identifies the connection link L. The connection link L is a link L connected to the node N. The link L is an electric wire connected to the power facility that is the node N.

The link L is specified by the link ID 411 and the link length 412. The link ID 411 is identification information that uniquely identifies the link L. The link length 412 is information indicating the length of the electric wire that is the link L.

<Example of acquisition pattern>
FIG. 5 is an explanatory diagram illustrating an example of an acquisition pattern. The acquisition pattern indicates the acquisition range of the power NW model. The operator operates the terminal 101 to select an acquisition pattern, and transmits a data acquisition request including the selected acquisition pattern to the data acquisition device 200. In FIG. 5, five types of acquisition patterns will be described as an example.

The acquisition pattern P1 is a circular acquisition pattern. The acquisition pattern P1 is, for example, an acquisition pattern for acquiring data related to a node group included in a system within a circle having a diameter D [km] with the point A as the center. For example, when the user gives the terminal 101 position information of the point A that is the node N, designation of a circle, diameter D, electrical characteristic information (voltage value), and equipment type, the terminal 101 generates the acquisition pattern P1. Then, a data acquisition request is transmitted.

The acquisition pattern P2 is a conical acquisition pattern. The acquisition pattern P2 relates to, for example, a node group included in a system in a cone whose height is a voltage (B1-B2) and whose bottom circle is a circle having a diameter D [km] with the point A as the center. This is an acquisition pattern for acquiring data. For example, when the user gives the terminal 101 position information of the point A, designation of the cone, diameter D, electrical characteristic information (voltage value range), and equipment type, the terminal 101 generates the acquisition pattern P2. A data acquisition request is transmitted.

The acquisition pattern P3 is a cylindrical acquisition pattern. For example, the acquisition pattern P3 acquires data related to a node group included in a system in a cylinder having a height of voltage B and a bottom circle having a diameter D [km] range centered on the point A. It is an acquisition pattern for. For example, when the user gives the terminal 101 position information of the point A, designation of a cylinder, diameter D, electrical characteristic information (voltage value range), and equipment type, the terminal 101 generates an acquisition pattern P3. A data acquisition request is transmitted.

The acquisition pattern P4 is a frustoconical acquisition pattern. The acquisition pattern P4 is, for example, an acquisition pattern for acquiring data related to a node group included in the system in the truncated cone that is in the range of n hops from the point A. For example, when the user gives the terminal 101 position information of the point A, designation of the truncated cone, the number of hops, electrical characteristic information (voltage value), and the equipment type, the terminal 101 generates the acquisition pattern P2, A data acquisition request will be sent. The number of hops is the number of nodes that can be traced from the link L with the point A as the departure point.

The acquisition pattern P5 is a spherical acquisition pattern. The acquisition pattern P5 is, for example, an acquisition pattern for acquiring data related to a node group included in a system in a sphere that is n hops from the point A where the voltage value is B. For example, when the user gives the terminal 101 position information of the point A, designation of the truncated cone, the number of hops, electrical characteristic information (voltage value), and the equipment type, the terminal 101 generates the acquisition pattern P2, A data acquisition request will be sent. The number of hops is the number of nodes that can be traced from the link starting from the point A.

Note that the acquisition patterns P1 to P5 are examples. The range designation from the point A may be either a distance or the number of hops. Further, the shape of the acquisition pattern is not limited to the shapes of the acquisition patterns P1 to P5. The acquisition pattern may be generated by the terminal 101 or may be generated by the application 102.

<Data acquisition history accumulation table>
FIG. 6 is an explanatory diagram showing an example of the contents stored in the data acquisition history accumulation table 600. In the following description of the database or table, the value of the AA field bbb (AA is a field name and bbb is a code) may be expressed as AAbbb. For example, the value of the log ID field 601 is expressed as a log ID 601.

The data acquisition history accumulation table 600 is a table updated by the data acquisition device 200. The data acquisition history storage table 600 is a table in which the data acquisition device 200 stores a history regarding data acquisition from the power NW model DB 104 in response to a data acquisition request from an operator.

The data acquisition history accumulation table 600 includes a log ID field 601, a user ID field 602, a session ID field 603, a data acquisition request time field 604, a data processing time field 605, a pattern shape field 606, and an attribute value field. 607. An entry indicating data acquisition history information is configured by the values of the fields 601 to 607 in the same row.

The log ID field 601 is a storage area for storing a log ID. The log ID 601 is identification information that uniquely identifies an entry indicating data acquisition history information. The data acquisition history information corresponds to one data acquisition of a certain worker. The user ID field 602 is an area for storing a user ID. The user ID 602 is identification information that uniquely identifies a worker who is a user. A user ID (which may include a password) 602 is given to the worker by pre-registration. When the user ID 602 is authenticated by the data acquisition device 200 (the data acquisition function 103 or the application 102), the user can log in to the data acquisition device 200 and can use the data acquisition device 200.

The session ID field 603 is a storage area for storing a session ID. The session ID 603 is identification information that uniquely identifies a session. Here, the session refers to a state from when an operator logs in to the data acquisition apparatus 200 until logout. That is, during the session, the worker can transmit a data acquisition request from the terminal 101 many times.

The data acquisition request time field 604 is a storage area for storing the data acquisition request time. The data acquisition request time 604 is the time when the data acquisition request is made. Specifically, for example, it may be a transmission time of a data acquisition request from the terminal 101 or a reception time of a data acquisition request in the data acquisition device 200. The time t1 in the time chart of FIG. 6 is the data acquisition request time 604 in the first data acquisition request R1 after a certain user logs in, and the time t2 is the data acquisition request in the second data acquisition request R2 of a certain user. It is time 604.

The data processing time field 605 is a storage area for storing the data processing time. The data processing time 605 is the time from when the data access function 131 accesses the power NW model DB 104 by a data acquisition request until the corresponding data is read out, converted into XML data, and output (data conversion Time). The time TP1 in the time chart of FIG. 6 is the data processing time 605 from the data acquisition request time 604 (t1) in the first data acquisition request R1 after login of a certain user to the output of XML data, and the time TP2 is This is the data processing time 605 from the data acquisition request time 604 (t2) in the second data acquisition request R2 of a user to the output of XML data.

The pattern shape field 606 is a storage area for storing the pattern shape. The pattern shape 606 is identification information indicating, for example, the graphic shown in FIG. The attribute value field 607 is a storage area for storing attribute values. The attribute value 607 includes a voltage that is the electrical characteristic information 406, a point serving as a reference for the pattern shape 606 (point A in FIG. 5), and an acquisition range. The acquisition range is the range of the XY plane encompassed by the pattern shape 606. For example, if the pattern shape 606 is a circle, it is the diameter D of the circle. That is, the pattern shape 606 determines the shape of the acquired pattern, and the attribute value 607 determines the range of the acquired pattern.

<Data acquisition estimation table>
FIG. 7 is an explanatory diagram of an example of the contents stored in the data acquisition estimation table. The data acquisition estimation table 700 is a table updated by the data acquisition device 200. The data acquisition history accumulation table 600 is used to estimate data acquisition by the next data acquisition request before the next data acquisition request when the data acquisition device 200 receives a data acquisition request from an operator. It is a table.

The data acquisition estimation table 700 includes a step number field 701, a user ID field 602, a pattern shape field 606, a representative attribute value field 702, and a next acquisition timing array field 703. An entry indicating data acquisition estimation information for estimating the next acquisition pattern is configured by the value of each field in the same row.

The step number field 701 is a storage area for storing a step number. Step number 701 is identification information for uniquely identifying a data acquisition request. The representative attribute value field 702 is a storage area for storing the representative attribute value 702. The representative attribute value 702 is a value that represents the attribute value 607, and is, for example, a statistical attribute value. The statistical attribute value is an average value, a minimum value, a maximum value, or a median value of the attribute values 607. The representative attribute value 702 includes a representative voltage and a representative acquisition range. The representative voltage is a statistical voltage value, for example, an average value, a minimum value, a maximum value, or a median value of the voltage values. The representative acquisition range is a statistical acquisition range, for example, an average value, minimum value, maximum value, or median value of the acquisition range (for example, if the pattern shape 606 is a circle, the diameter D of the circle). .

Note that the attribute value 607 in the data acquisition history accumulation table 600 of FIG. 6 includes a voltage, a point, and an acquisition range, but the representative attribute value 702 does not include a point. Therefore, in the case of acquiring in advance, the position information specified in the immediately preceding data acquisition request is used.

The next acquisition timing array field 703 is a storage area for storing zero or more next acquisition timing arrays. The next acquisition timing array 703 includes a step number 701 indicating a next estimated data acquisition request that is a data acquisition request estimated to arrive next, and a representative acquisition time (unit: minutes, for example) by the next estimated data acquisition request. , Selection probability, and array data. For example, when the next acquisition timing array 703 is [U003] [1] [70], the top [U003] is the step number 701, the central [1] is the representative acquisition time, and the last [70]. Is the selection probability. In the next acquisition timing array field 703, one or more next acquisition timing arrays 703 are stored, but the last data acquisition request is “NULL” because there is no next estimated data acquisition request.

The representative acquisition time is a representative value of acquisition time. The acquisition time is from the data acquisition request time 604 (for example, t1) in a certain acquisition request until the next estimated data acquisition request is received, that is, until the data acquisition request time 604 (for example, t2) in the next estimated data acquisition request. (E.g., Ta2). However, at the time of login in each session, the time from the login time t0 to the first data acquisition request time 604 (t1) is defined as an acquisition time Ta1. Further, since there is no next data acquisition request for the last data acquisition request, there is no acquisition time. For example, when the last data acquisition request is the data acquisition request R2, the next estimation data acquisition request R3 does not exist, and therefore the acquisition time Ta3 does not exist. The representative value of the acquisition time is a statistical acquisition time, for example, an average value, a minimum value, a maximum value, or a median value of the acquisition times.

The selection probability is a probability that an acquisition pattern (a combination of the pattern shape 606 and the representative attribute value 702) that will be specified in the next estimation data acquisition request is selected next time by the user. For example, there are two next acquisition timing arrays 703 in the entry whose step number 701 is “U001”: {[U003] [1] [70], [U004] [1] [30]}. In the first next acquisition timing array 703 {[U003] [1] [70]}, the data acquisition request of [U003] having the step number 701 is logged in (the pattern shape 606 and the representative attribute value 702 are “NULL”). 1] means that it is selected with a probability of [70]% when the time elapses. Also, in the next acquisition timing array 703 {[U004] [1] [30]} at the end, the data acquisition request of [U004] having the step number 701 is logged in (the pattern shape 606 and the representative attribute value 702 are “NULL”). Means that it is selected with a probability of [30]% when the time of [1] elapses.

Similarly, there are two next acquisition timing arrays 703 in the entry whose step number 701 is “U003”, {[U008] [5] [90], [U010] [7] [10]}. In the next next acquisition timing array 703 {[U008] [5] [90]}, the data acquisition request of [U008] with step number 701 is [5] from the data acquisition request of “U003” with step number 701. Means that it is selected with a probability of [90]%. Further, the next acquisition timing array 703 {[U010] [7] [10]} at the end indicates that the data acquisition request of [U010] with the step number 701 is changed from the data acquisition request of “U003” with the step number 701 [ 7] means that it is selected with a probability of [10]% when the time elapses.

Note that an entry whose pattern shape 606 and representative attribute value 607 is “NULL” is an entry that the user refers to for the first time when starting a session with the data acquisition apparatus 200. The entry whose next acquisition timing array 703 is “NULL” is an entry indicating that there is no next data acquisition request (step number 701).

<Step information management table>
FIG. 8 is an explanatory diagram of an example of stored contents of the step information management table. The step information management table 800 is a table that is updated by the data acquisition device 200. The step information management table 800 is a table for managing a step number 701 indicating a data acquisition request for each user.

The step information management table 800 has a user ID field 602, a step number field 701, and a previous step number field 801. An entry indicating step management information is configured by the value of each field in the same row.

The previous step number field 801 is a storage area for storing the previous step number 801. The previous step number 801 is a step number 701 indicating a data acquisition request immediately before the data acquisition request corresponding to the step number 701.

<Pre-acquired information management table>
FIG. 9 is an explanatory diagram of an example of stored contents of the pre-acquired information management table. The pre-acquisition information management table 900 is a table that is updated by the data acquisition apparatus 200 through a pre-acquisition process (described later in FIG. 22). The prior acquisition information management table 900 is a table for managing data acquired in advance for each user. The prior acquisition information management table 900 includes a user ID field 602, a pattern shape field 606, an attribute value field 607, and a preliminary acquisition data field 901. An entry indicating pre-acquired information is configured by the value of each field in the same row.

The advance acquisition data field 901 is a storage area for storing advance acquisition data. The pre-acquisition data 901 is data acquired in advance by the data access function 131 accessing the power NW model DB 104, and an acquisition pattern (pattern shape 606 and the same entry of the same entry) specified by the user ID 602 is specified. Data included in the attribute value 607 combination).

<Startup timing management table>
FIG. 10 is an explanatory diagram of an example of stored contents of the activation timing management table. The activation timing management table 1000 is a table that is updated by the data acquisition device 200. The activation timing management table 1000 is a table for managing activation timing for each user. The activation timing management table 1000 has a user ID field 602 and an activation timing field 1001. An entry indicating activation timing information is configured by the value of each field in the same row.

The activation timing field 1001 is a storage area for storing activation timing. When the session ends, the value of the activation timing field 1001 of the user ID 602 is deleted. The activation timing 1001 is a timing at which the pre-acquisition function 133 is activated to cause the data access function 131 to acquire data from the power NW model DB 104. Specifically, for example, in the data acquisition estimation table 700, the representative acquisition time of the next acquisition timing array 703 having the highest selection probability among the next acquisition timing arrays 703 of the entry of the corresponding step number 701 is the start timing 1001. Applied.

For example, as shown in the time chart of FIG. 10, an entry whose step number 701 is “U001” includes two next acquisition timing arrays 703 {[U003] [1] [70], [U004] [1] [1] [ 30]}. Since the highest selection probability is [70] in the next next acquisition timing array 703 {[U003] [1] [70]}, the highest next selection probability array 703 {[U003] [1] [70]} The representative acquisition time [1] is applied to the activation timing 1001. Specifically, for example, the point of time 701 is the time point that is the same user ID 602 and the average value of the data processing time 605 in the same pattern shape 606 from the time when the representative acquisition time [1] has elapsed since the login. This is the start timing 1001 of the advance acquisition process by the data acquisition request corresponding to U003.

Also, the entry whose step number 701 is “U003” has two next acquisition timing arrays 703 {[U008] [5] [90], [U010] [7] [10]}. Since the highest selection probability is [90] of the next next acquisition timing array 703 {[U008] [5] [90]}, the highest next selection probability array 703 {[U008] [5] [90]} The representative acquisition time [5] is applied to the activation timing 1001. Specifically, for example, the same user ID 602 and the same pattern shape from the elapse of the representative acquisition time [5] from the time when the XML data is output by the data acquisition request corresponding to the step number 701 “U003” The point in time of the average value of the data processing time 605 in 606 is the start timing 1001 of the pre-acquisition processing by the data acquisition request corresponding to U008 which is step number 701.

Specifically, the tables 600, 700, 800, and 900 shown in FIGS. 6 to 9 are realized by the storage device 302 shown in FIG. 3, for example.

<Example of Functional Configuration of Data Acquisition Device 200>
FIG. 11 is a block diagram illustrating a functional configuration example of the data acquisition device 200. The data acquisition device 200 includes a reception unit 1101, a conversion unit 1102, a transmission unit 1103, a learning unit 1104, a specifying unit 1105, a pre-acquisition unit 1106, and a setting unit 1107. Specifically, the reception unit 1101 to the setting unit 1107 are functions realized by causing the processor 301 to execute the program stored in the storage device 302 illustrated in FIG. 3 or by the communication IF 305, for example. .

The receiving unit 1101 receives a data acquisition request from the terminal 101. The conversion unit 1102 accesses the power NW model DB 104, acquires data corresponding to the data acquisition request received by the reception unit 1101 (a plurality of data having different attributes) from the power NW model DB 104, and stores data in different formats. Convert to (XML data). The conversion unit 1102 corresponds to the data access function 131 described above. The transmission unit 1103 transmits the XML data obtained by the conversion unit 1102 to the request source, that is, the terminal 101 that is the transmission source of the data acquisition request.

The learning unit 1104 learns the type of data acquisition request and the acquisition timing that defines the time series order of the data acquisition request for each data acquisition request of the series of data acquisition requests received from the request source. The learning unit 1104 corresponds to the pattern learning function 132 described above. The request source is the terminal 101 used by the operator to log in to the data acquisition device 200. The type of data acquisition request is, for example, an acquisition pattern (a combination of pattern shape 606 and attribute value 607). The acquisition timing is, for example, the data acquisition request time 604. In addition, the learning unit 1104 identifies the user ID 602 and the session ID 603 included in the data acquisition request. Further, the learning unit 1104 measures the data processing time 605 in response to the notification of the completion of the output of the XML data from the conversion unit 1102. Thereby, the learning unit 1104 registers an entry in the data acquisition history accumulation table 600 for each data acquisition request.

Further, the learning unit 1104 uses the entry registered in the data acquisition history accumulation table 600 to generate and register an entry in the data acquisition estimation table 700. Specifically, the learning unit 1104 generates a learning result using a learning result table.

FIG. 12 is an explanatory diagram showing an example of a learning result table. A learning result table 1200 is generated for each user ID 602. The learning result table 1200 includes an acquisition pattern ID column field 1201, a selection count field 1202, a selection probability field 1203, a representative acquisition time field 1204, and a representative attribute value field 702.

The acquisition pattern ID column field 1201 is a storage area for storing the acquisition pattern ID column. The acquisition pattern ID string 1201 is an ID string that identifies a time-series acquisition pattern. For example, “P1” is an acquisition pattern ID column 1201 indicating that the acquisition pattern P1 is selected in one data acquisition request and the session is ended. “P1-P1-P2” is an acquisition pattern P1 for the first data acquisition request, an acquisition pattern P1 for the second data acquisition request, and an acquisition pattern P2 for the third data acquisition request. It is an acquisition pattern ID column 1201 indicating that the processing has ended.

The selection number field 1202 is a storage area for storing the number of selections. The number of selections 1202 is the number of times that the session has ended in the acquisition pattern ID column 1201, that is, the number of times the acquisition pattern ID column 1201 has been selected.

The selection probability field 1203 is a storage area for storing the selection probability. The selection probability 1203 is a probability that the acquired pattern ID string 1201 is selected. The learning unit 1104 calculates a selection probability 1203 using an acquisition pattern ID string 1201 in which a series of acquisition pattern IDs from the head to the previous one are common. However, when the acquisition pattern ID string 1201 includes only one acquisition pattern ID, there is no series of acquisition pattern IDs from the head to the previous one, so the denominator of the selection probability 1203 is the acquisition pattern ID string 1201 “P1 ”To“ P5 ”, and the numerator is the number of selections 1202 of the target acquisition pattern ID column 1201 among the acquisition pattern ID columns 1201“ P1 ”to“ P5 ”.

For example, the selection probability 1203 of the acquisition pattern ID string 1201 “P1” is 7/10 (70%), the selection probability 1203 of the acquisition pattern ID string 1201 “P2” is 3/10 (30%), and the acquisition pattern ID string The selection probabilities 1203 of 1201 “P3” to “P5” are 0/10 (0%), respectively.

For the acquisition pattern ID column 1201 having two or more IDs, for example, when the number of IDs is three, there are 125 acquisition pattern ID columns 1201. Of these, the learning unit 1104 groups the acquired pattern IDs 1201 having the same acquired pattern ID from the first (first) to the previous (second), and calculates a selection probability 1203 within each group. For example, “P1-P1-P1”, “P1-P1-P2”, “P1-P1-P3”, “P1-P1-P4”, “P1-P1-P5” are 1 from the top (first time). The acquisition pattern ID up to the previous (second) is common to “P1-P1”. In this case, the denominator of the selection probability 1203 is the sum of the selection times 1202 of the acquisition pattern ID column 1201 “P1-P1-P1” to “P1-P1-P5”, and the numerator is the acquisition pattern ID column 1201 “P1-P1”. This is the number of selections 1202 of the acquisition pattern ID column 1201 that is the target among “-P1” to “P1-P1-P5”.

For example, the selection probability 1203 of the acquisition pattern ID column 1201 “P1-P1-P1” is 4/10 (40%), and the selection probability 1203 of the acquisition pattern ID column 1201 “P1-P1-P2” is 6/10 ( 60%), the selection probabilities 1203 of the acquired pattern ID columns 1201 “P1-P1-P3” to “P1-P1-P5” are 0/10 (0%), respectively.

The representative acquisition time field 1204 is a storage area for storing the representative acquisition time. For the representative acquisition time 1204, the learning unit 1104 refers to the data acquisition history accumulation table 600, and for each acquisition pattern ID column 1201, from the data acquisition request time 604 of the data acquisition request corresponding to the last acquisition pattern ID, The time obtained by subtracting the data acquisition request time 604 of the data acquisition request corresponding to the previous acquisition pattern ID is calculated as the acquisition time. Then, the learning unit 1104 calculates, for example, the average value of the calculated acquisition times as the representative acquisition time 1204.

For example, when the acquired pattern ID column 1201 is “P1-P2-P3”, the learning unit 1104 sets the pattern shape 606 of the acquired pattern to “P1” → “P2” → “P3” for each session ID 603 of the user. The entry ending in is identified from the data acquisition history accumulation table 600. Then, the learning unit 1104 registers the data acquisition request time 604 of the entry in which the last acquisition pattern ID “P3” is registered as the pattern shape 606 and the previous acquisition pattern ID “P2” as the pattern shape 606. The data acquisition request time 604 of the registered entry is acquired, and the acquisition time is calculated. The learning unit 1104 executes this acquisition time calculation process for each session whose acquisition pattern ID column 1201 is “P1-P2-P3”, obtains the average value of the calculated acquisition times, and acquires the acquisition pattern ID column 1201. The representative acquisition time 1204 of “P1-P2-P3” is assumed.

For the acquisition pattern ID column 1201 with one ID, there is no previous acquisition pattern ID, so instead of the data acquisition request time 604 of the data acquisition request corresponding to the previous acquisition pattern ID. The login time is applied.

The representative attribute value 702 is a combination of a representative voltage and a representative acquisition range. The representative attribute value 702 is, for example, the average value of the voltage value and the acquisition range when the acquisition pattern indicated by the acquisition pattern ID at the end is selected in each acquisition pattern ID column 1201.

For example, when the acquired pattern ID column 1201 is “P1-P2-P3”, the learning unit 1104 sets the pattern shape 606 of the acquired pattern to “P1” → “P2” → “3” for each session ID 603 of the user. The entry ending in is identified from the data acquisition history accumulation table 600. Then, the learning unit 1104 acquires the attribute value 607 (excluding the point) of the entry in which the acquisition pattern ID “P3” at the end is registered as the pattern shape 606. The learning unit 1104 executes the acquisition processing of the attribute value 607 (excluding the point) for each session whose acquisition pattern ID string 1201 is “P1-P2-P3”, and acquires the acquired attribute value 607 (excluding the point). Is obtained as the representative attribute value 702 of the acquired pattern ID column 1201 “P1-P2-P3”. The selection probability 1203, the representative acquisition time 1204, and the representative attribute value 702 are used for entry registration of the data acquisition estimation table 700.

Returning to FIG. 11, when the first data acquisition request is received from the request source for a series of other data acquisition requests different from the series of data acquisition requests, the specifying unit 1105 displays the learning result by the learning process of the learning unit 1104. Based on the first data acquisition request, the type of the next estimated data acquisition request that can be acquired from the request source and the acquisition timing are specified. The specifying unit 1105 corresponds to the advance acquisition function 133 described above.

The identifying unit 1105 identifies the step number 701 of the next estimated data acquisition request with the selection probability 1203 from the learning result, specifically, the next acquisition timing array 703 of the data acquisition estimation table 700. Then, the specifying unit 1105 specifies the pattern shape 606 and the representative attribute value 702 corresponding to the entry of the step number 701 of the next estimated data acquisition request from the data acquisition estimation table 700 as the next estimated data acquisition request type. Further, the specifying unit 1105 specifies the next estimated data acquisition request time 604 that is the time when the representative acquisition time 1204 has elapsed from the data acquisition request time 604 of the first data acquisition request as the acquisition timing of the next estimated data acquisition request.

The prior acquisition unit 1106 acquires data corresponding to the next estimated data acquisition request whose type is specified by the specifying process of the specifying unit 1105 from the power NW model DB 104 before the acquisition timing of the next estimated data acquisition request. The advance acquisition unit 1106 corresponds to the advance acquisition function 133 described above. Specifically, for example, the advance acquisition unit 1106 controls the conversion unit 1102 to acquire data from the power NW model DB 104 before the acquisition timing of the second data acquisition request.

The setting unit 1107 starts the advance acquisition process so that the next estimated data acquisition request acquisition timing is after the elapse of the data processing time 605 required to convert the data into a predetermined format (for example, XML format). Timing 1001 is set. Specifically, for example, as described with reference to the time chart of FIG. 10, the setting unit 1107 includes the average value of the data processing time 605 for the same user ID 602 and the same pattern shape 606 from the time when the representative acquisition time 1204 has elapsed. The retroactive time is set as the activation timing 1001. Thereby, the advance acquisition unit 1106 can execute the advance acquisition process at the activation timing 1001.

When the second data acquisition request is actually received after the first data acquisition request, the determination unit 1108 includes the type included in the second data acquisition request and the type of the next estimated data acquisition request specified by the specifying process. And whether they match. If they match, the pre-acquisition unit 1106 sends the pre-acquisition result to the transmission unit 1103. In the case of mismatch, the conversion unit 1102 acquires data corresponding to the second data acquisition request received by the reception unit 1101 from the power NW model DB 104, converts the data into XML data, and sends the XML data to the transmission unit 1103. In the case of mismatch, the identifying unit 1105 identifies the step number 701 of the next estimated data acquisition request that was not selected last time from the next acquisition timing array 703 of the data acquisition estimation table 700 with the selection probability 1203.

<Learning process>
FIG. 13 is a flowchart illustrating an example of a learning process procedure performed by the learning unit 1104. The learning unit 1104 waits for learning timing (step S1301: No). The learning timing is, for example, irregular timing designated by the user, or regular timing such as daily, weekly, monthly or the like. If it is the learning timing (step S1301: Yes), the learning unit 1104 acquires the target entry group from the data acquisition history accumulation table 600 (step S1302). The target entry group is an entry group to be learned and may be all entries in the data acquisition history accumulation table 600, and is limited in advance within a range of values of a certain field in the data acquisition history accumulation table 600. It may be a group of entries.

Then, the learning unit 1104 executes a learning process (step S1303). Details of the learning process (step S1303) will be described later with reference to FIG. Thereafter, the learning unit 1104 executes a data acquisition estimation table update process (step S1304). Details of the data acquisition estimation table update process (step S1304) will be described later with reference to FIG. Thereby, the learning unit 1104 ends the series of processes.

<Learning process (step S1303)>
FIG. 14 is a flowchart showing a detailed processing procedure example of the learning processing (step S1303) shown in FIG. The learning unit 1104 determines whether there is an unselected user ID 602 for the learning target entry group acquired from the data acquisition history accumulation table 600 (step S1401). When there is an unselected user ID 602 (step S1401: Yes), the learning unit 1104 selects an unselected user ID 602 and selects an entry group of the selected user ID 602 from the learning target entry group (step S1402). The learning unit 1104 determines whether there is an unselected session ID 603 for the entry group of the selected user ID 602 (step S1403). If there is an unselected session ID 603 (step S1403: Yes), the learning unit 1104 selects an unselected session ID 603 from the entry group of the selected user ID 602, and selects the entry group of the selected session ID 603 (step S1403). S1404).

The learning unit 1104 determines whether there is an unselected entry among the entry group of the selected session ID 603 selected in step S1404 (step S1405). When there is an unselected entry (step S1405: Yes), the learning unit 1104 selects an entry that is not selected and has the oldest data acquisition request time 604 from the entry group of the selected session ID 603 (step S1406). Then, the learning unit 1104 identifies the acquisition pattern ID (pattern shape 606), the data acquisition request time 604, and the attribute value 607 from the selection entry in step S1406, and adds them to the end of the acquisition pattern array (step S1407). . Then, the process returns to step S1405. An acquisition pattern array is generated for each entry of the selected session ID 603 through the loop of steps S1405 to S1407.

If there is no unselected entry in step S1405 (step S1405: No), the learning unit 1104 stores the acquired pattern array of the selected user ID 602 and the selected session ID 603, which is finally generated, in the storage device 302. (Step S1408), the process returns to Step S1403.

In step S1403, when there is no unselected session ID 603 for the selected user ID 602 (step S1403: No), the learning unit 1104 executes the aggregation process (step S1409) and the representative value calculation process (step S1410). Details of the aggregation process (step S1409) and the representative value calculation process (step S1410) will be described later with reference to FIGS. 16 and 17, respectively. After the representative value calculation process (step S1410), the process returns to step S1401.

In step S1401, when there is no unselected user ID 602 (step S1401: No), the learning unit 1104 ends the learning process (step S1303) and proceeds to the data acquisition estimation table update process (step S1304).

<Aggregating process (step S1409)>
FIG. 15 is a flowchart illustrating a detailed processing procedure example of the aggregation processing (step S1409) illustrated in FIG. The tabulation process (step S1409) is a process of counting the number of selections 1202 in the learning result table 1200 and calculating an acquisition time that is a calculation source of the representative acquisition time 1204.

The learning unit 1104 determines whether or not there is an unselected session ID 603 for the acquired pattern array group generated in step S1407 (step S1501). When there is an unselected session ID 603 (step S1501: Yes), the learning unit 1104 selects an unselected session ID 603 and selects an acquisition pattern array (step S1502). Then, the learning unit 1104 identifies a series of acquisition pattern IDs (hereinafter referred to as an acquisition pattern ID string 1201) included in the selection acquisition pattern array (step S1503).

The learning unit 1104 adds 1 to the selection count 1202 of the specified acquisition pattern ID string 1201 in the learning result table 1200 (step S1504).

Next, the learning unit 1104 sets the variable i to i = 1 (step S1505). The learning unit 1104 determines whether or not the first to i-th acquisition pattern ID strings 1201 exist in the selection acquisition pattern array (step S1506). If it exists (step S1506: Yes), the learning unit 1104 subtracts the (i-1) -th data acquisition request time 604 from the i-th data acquisition request time 604 of the selected acquisition pattern array to obtain (i-1 ) The acquisition time from the first data acquisition to the i-th data acquisition is calculated (step S1507).

The learning unit 1104 stores the acquisition time calculated in step S1507 at the i-th data acquisition request time 604 of the selected acquisition pattern array (step S1508). Then, the learning unit 1104 increments i (step S1509) and returns to step S1506.

In step S1506, when there is no i-th acquisition pattern ID string 1201 from the top in the selected acquisition pattern array (step S1506: No), the process returns to step S1501. In step S1501, when there is no unselected session ID 603 for the acquired pattern array group (step S1501: No), the learning unit 1104 executes a selection probability calculation process (step S1510) and a representative value calculation process (step S1410). Migrate to

<Selection probability calculation process>
FIG. 16 is a flowchart showing a detailed processing procedure example of the selection probability calculation process (step S1510) shown in FIG. The learning unit 1104 sets the initial value of i to i = 1 (step S1601). i is a variable indicating the number of IDs included in the acquisition pattern ID string 1201 defined in the learning result table 1200.

The learning unit 1104 determines whether there is an acquisition pattern ID string 1201 with i IDs (step S1602). The case where there are i acquisition pattern ID columns 1201 means a case where there are i acquisition pattern ID columns 1201 having a selection count 1202 of 1 or more.

When there are i acquisition pattern ID strings 1201 (step S1602: Yes), the acquisition pattern ID string 1201 group having the number of IDs is the same as the acquisition pattern ID from the top to the (i-1) th acquisition pattern ID. Grouping is performed on the column 1201 (step S1603). For example, when the number of IDs is i = 3, the acquired pattern ID strings 1201 “P1-P1-P1” and “P1-P1-P2” are “P1-P1” from the top to the second (i−1). "Are common, they are grouped into the same group. On the other hand, the acquisition pattern ID columns 1201 “P1-P1-P1” and “P1-P3-P2” have the acquisition pattern IDs “P1-P1” and “P1-P3” from the top to the second (i−1) th. Are grouped into different groups. When the number of IDs is i = 1, since there are no (i−1) th acquired pattern IDs from the top, “P1” to “P5” are grouped into the same group.

The learning unit 1104 determines whether there is an unselected group (step S1604). When there is an unselected group (step S1604: Yes), the learning unit 1104 selects one unselected group (step S1605). Then, the learning unit 1104 calculates the selection probability 1203 of each acquired pattern ID string 1201 with the number of IDs i in the selected group, and stores it in the learning result table 1200 (step S1606). Then, the process returns to step S1604.

In step S1604, when there is no unselected group (step S1604: No), the learning unit 1104 increments i (step S1607) and returns to step S1602. In step S1602, when there is no acquisition pattern ID string 1201 with i IDs (step S1602: No), the learning unit 1104 ends the selection probability calculation process (step S1510) and the representative value calculation process (step S1410). ).

<Representative value calculation processing (step S1610)>
FIG. 17 is a flowchart showing a detailed processing procedure example of the representative value calculation processing (step S1410) shown in FIG. In the representative value calculation process (step S1410), the learning unit 1104 calculates a representative acquisition time 1204 and a representative attribute value 702.

Learning unit 1104 sets variable i to i = 1 (step S1701). The learning unit 1104 determines whether there is an i-th acquired pattern ID string 1201 from the top for the acquired pattern array of all session IDs 603 (step S1702).

If there is an i-th acquisition pattern ID string 1201 from the top (step S1702: Yes), the learning unit 1104 selects the corresponding acquisition pattern ID string 1201 (step S1703). For example, if there is an acquisition pattern array including the acquisition pattern ID string 1201 “P1-P2-P3”, if i = 1, “P1” is selected as the acquisition pattern ID string 1201 from “P1-P2-P3”. If i = 2, “P1-P2” is selected as the acquisition pattern ID column 1201 from “P1-P2-P3”. If i = 3, the acquisition pattern ID from “P1-P2-P3” is selected. “P1-P2-P3” is selected as the column 1201.

The learning unit 1104 performs grouping using the same selection / acquisition pattern ID string 1201 (step S1704). Then, the learning unit 1104 determines whether there is an unselected group (step S1705). When there is an unselected group (step S: Yes), the learning unit 1104 selects one unselected group (step S1706).

Then, the learning unit 1104 specifies each acquisition time from the (i-1) th data acquisition to the i-th data acquisition from the selection acquisition pattern ID string 1201 in the selection group, and from the specified acquisition time, the representative An acquisition time 1204 is calculated (step S1707). Similarly, the learning unit 1104 identifies each i-th attribute value 607 from the selection acquisition pattern ID string 1201 in the selected group, and calculates a representative attribute value 607 from each identified attribute value 607 (step S1708). . Then, the process returns to step S1705.

In step S1705, when there is no unselected group (step S1705: No), it returns to step S1702. If there is no i-th acquired pattern ID string 1201 from the beginning in step S1702 (step S1702: No), the learning unit 1104 ends the representative value calculation process (step S1610), returns to step S1401, and sets the user ID 602. change.

<Data acquisition estimation table update process (step S1304)>
FIG. 18 is a flowchart showing a detailed processing procedure example of the data acquisition estimation table update processing (step S1304) shown in FIG. The data acquisition estimation table update process (step S1304) is a process for updating the data acquisition estimation table 700 shown in FIG.

The learning unit 1104 determines whether there is an unselected user ID 602 (step S1801). When there is an unselected user ID 602 (step S1801: Yes), the learning unit 1104 selects the learning result table 1200 of the unselected user ID 602 (step S1802). Next, the learning unit 1104 sets a variable i = 0 (step S1803). The variable i indicates the time series order of the acquisition pattern ID and also corresponds to the time series order of the step number 701. The learning unit 1104 numbers the initial step number 701 and sets it as the registration target step number 701 (step S1804). For example, as shown in FIG. 7, when the selected user ID 602 is “A0001”, “U001” is assigned as the initial step number 701.

The learning unit 1104 determines whether or not there is an acquired pattern ID string 1201 in which the number of selections 1202 is 1 or more and the number of IDs is (i + 1) in the learning result table 1200 selected in step S1802 (step S1802). S1805). If it exists (step S1805: No), the learning unit 1104 takes the next step number 701 for each of the acquired pattern ID strings 1201 having the selection count 1202 of 1 or more and the number of IDs (i + 1). The next acquisition timing array 703 is generated (step S1806).

Specifically, for example, if i = 0, the acquired pattern ID string 1201 having one ID corresponding to step S1805 is “P1” (the selection count 1202 is seven, the selection probability 1203 is 70%, representative The acquisition time 1204 is “1”) and “P2” (the number of selections 1202 is 3, the selection probability 1203 is 30%, and the representative acquisition time 1204 is “1”). The learning unit 1104 assigns the next step number 701 {U003, U004} to each of “P1” and “P2”, and uses {[U003] [1] [70], [[ U004] [1] [30]}.

Thereafter, the learning unit 1104 determines whether i = 0 (step S1807). When i = 0 (step S1807: Yes), the learning unit 1104, the initial step number 701 (registration target step number 701) numbered in step S1804, the selected user ID 602, the acquisition pattern ID (NULL), and the representative attribute An entry including the value 702 (NULL) and the next acquisition timing array 703 generated in step S1806 is registered in the data acquisition estimation table 700 (step S1808). Then, control goes to a step S1810.

On the other hand, if i is not 0 in step S1807 (step S1807: No), the registration target step number 701, the selected user ID 602 set in step S1810 of the previous loop, and the acquisition pattern confirmed to exist in step S1805. An entry consisting of the i-th acquisition pattern ID in the ID column 1201, the i-th representative attribute value 702 in the acquisition pattern ID column 1201 whose existence has been confirmed in step S1805, and the next acquisition timing array 703 generated in step S1806 is stored as data. Register in the acquisition estimation table 700 (step S1809). Then, the process proceeds to step S1810.

Note that the i-th acquisition pattern ID in the acquisition pattern ID column 1201 is the acquisition pattern ID at the end of the acquisition pattern ID column 1201. The learning unit 1104 acquires the i-th representative attribute value 702 in the acquisition pattern ID string 1201 whose existence is confirmed in step S 1805 from the acquisition pattern array including the acquisition pattern ID string 1201.

In step S1810, the learning unit 1104 sets the next step number 701 numbered in step S1806 as the registration target step number 701 (step S1810). Then, the learning unit 1104 increments i (step S1811) and returns to step S1805. In step S1805, in the learning result table 1200 selected in step S1802, when the number of selections 1202 is 1 or more and the acquired pattern ID string 1201 with the number of IDs (i + 1) does not exist (step S1805: No), step The process moves to S1811.

In step S1811, the learning unit 1104 has the registration target step number 701, the selected user ID 602 set in step S1810 of the previous loop, the acquired pattern ID string 1201 whose existence has been confirmed in step S1805 of the previous loop. An entry consisting of the i-th acquisition pattern ID in i, the i-th representative attribute value 702 in the acquisition pattern ID string 1201 whose existence was confirmed in step S1805 of the previous loop, and the next acquisition timing array 703 (NULL), It registers in the data acquisition estimation table 700 (step S1809). Then, the process proceeds to step S1801.

In step S1801, when there is no unselected user ID 602, the learning unit 1104 ends the data acquisition estimation table update process (step S1304).

<Data acquisition process>
FIG. 19 is a flowchart illustrating a detailed processing procedure example 1 of the data acquisition processing by the data acquisition device 200. First, the data acquisition apparatus 200 waits for login detection of the terminal 101 (step S1901: No). When the login is detected (step S1901: Yes), the data acquisition device 200 acquires the user ID 602 from the terminal 101 and newly issues a session ID 603 (step S1902). Thereafter, the data acquisition device 200 determines whether or not the entry of the acquisition user ID 602 in step S1902 exists in the data acquisition estimation table 700 (step S1903).

In step S1903, when the entry of the acquisition user ID 602 in step S1902 exists in the data acquisition estimation table 700 (step S1903: Yes), data can be acquired in advance. In this case, the data acquisition apparatus 200 acquires an entry whose acquisition pattern ID and representative attribute value 607 are NULL and matches the acquisition user ID 602 from the data acquisition estimation table 700 (step S1904). Then, the data acquisition apparatus 200 sets the activation timing 1001 of the advance acquisition process (FIG. 23) (step S1905).

The data acquisition apparatus 200 specifies the step number 701 of the next acquisition timing array 703 including the highest selection probability from the next acquisition timing array field 703 in the entry acquired in step S1904, and corresponds to the entry of the specified step number 701. An acquisition pattern (pattern shape 606 and representative attribute value 702) is specified from the data acquisition estimation table 700. The data acquisition device 200 sets the activation timing 1001 for the specified acquisition pattern. The set activation timing 1001 is managed for each user ID 602 in the activation timing management table 1000 of FIG. Details of the setting of the activation timing 1001 of the pre-acquisition process (FIG. 23) have been described with reference to the timing chart of FIG.

The data acquisition apparatus 200 stores the step information in the step information management table 800 shown in FIG. 8 (step S1906). Here, the step information includes the acquisition user ID 602, the step number 701 (the step number 701 of the highest selection probability in the next acquisition timing array 703 of the acquisition entry in step S1904), and the previous step number 801 (in step S1904). Step number 701) of the acquisition entry. Then, the process proceeds to step S2001 in FIG.

Note that, as shown by the dotted line arrow, only after login detection (step S1901: Yes), that is, when no data acquisition request has been received from the login terminal 101 even after login detection (step S1907: No), Instead of step S2001 in step 20, the process proceeds to step S1907.

In step S1903, if it does not exist (step S1903: No), the data acquisition apparatus 200 determines whether a data acquisition request is received from the login terminal 101 (step S1907). If received (step S1907: YES), the data acquisition device 200 executes data acquisition / history accumulation processing (step S1908), and proceeds to step S1909. Details of the data acquisition / history accumulation processing (step S1908) will be described later with reference to FIG. If not received (step S1907: NO), the process proceeds to step S1909.

In step S1909, when logout is not detected (step S1909: No), the process returns to step S1903, and when logout is detected (step S1909: Yes), a series of processing ends.

FIG. 20 is a flowchart showing a detailed processing procedure example 2 of the data acquisition processing by the data acquisition device 200. After step S1906 in FIG. 19, the data acquisition apparatus 200 stores the user ID 602 and the acquisition pattern (pattern shape 606 and attribute value included in the data acquisition request received in step S1907 in the pre-acquisition information management table 900 illustrated in FIG. 9. It is determined whether there is an entry corresponding to (607) (step S2001). In other words, the data acquisition device 200 determines whether or not the acquisition pattern in which the activation timing 1001 is set in steps S1905 and S1906 matches the acquisition pattern included in the data acquisition request received in step S1907. become.

If there is an entry (step S2001: Yes), it means that the pre-acquisition process with the activation timing 1001 set in step S1905 has been executed. Therefore, the data acquisition device 200 acquires the corresponding pre-acquisition data 901 from the pre-acquisition information management table 900 and transmits it to the login terminal 101 (step S2002).

The data acquisition device 200 registers data acquisition history information (user ID 602, session ID 603, data acquisition request time 604, data processing time 605, pattern shape 606, attribute value 607) in the data acquisition history accumulation table 600 (step S2003). ). Among the data acquisition history information, the user ID 602 is the user ID 602 acquired in step S1902 and included in the data acquisition request in step S1907. Of the data acquisition history information, the session ID 603 is the session ID 603 issued in step S1902. Of the data acquisition history information, the data acquisition request time 604 is the time of the data acquisition request in step S1907. Of the data acquisition history information, the data processing time 605 is the time from the data acquisition request time 604 to the completion of data acquisition, that is, output of XML data. Of the data acquisition history information, the pattern shape 606 and the attribute value 607 are the pattern shape 606 and the attribute value 607 included in the data acquisition request in step S1907.

The data acquisition device 200 sets the next activation timing 1001 of the pre-acquisition process (FIG. 23) (step S2004) and stores the step information in the step information management table 800 (step S2005), as in steps S1905 and S1906. The set activation timing 1001 is managed for each user ID 602 in the activation timing management table 1000 of FIG. In step S2005, the data acquisition apparatus 200 updates the step number 701 in the entry of the acquisition user ID 602 to the step number 701 of the highest selection probability in the next acquisition timing array 703 of the acquisition entry of the step number 701. The previous step number 801 is updated to the previous step number 701. Then, the process returns to step S1907 in FIG.

On the other hand, when there is no entry in the pre-acquisition information management table 900 in step S2001 (step S2001: No), other acquisition patterns that are not acquisition patterns of the highest selection probability 1203 in the data acquisition request received in step S1907. Means that is selected. In this case (step S2001: No), the data acquisition device 200 executes data acquisition / history accumulation processing (step S2006). Data acquisition / history storage processing (step S2006) is the same processing as step S1908.

The data acquisition device 200 acquires the previous step number 801 in the entry of the acquisition user ID 602 from the step information management table 800 (step S2007). Then, the data acquisition apparatus 200 determines whether or not the next acquisition timing array 703 corresponding to the previous step number 801 acquired in step S2007 exists in the entry acquired in S1904 (step S2008).

If it does not exist (step S2008: No), the process returns to step S1907 in FIG. 19, and the data acquisition device 200 waits for a data acquisition request. On the other hand, when it exists (step S2008: Yes), the data acquisition device 200 acquires one next acquisition timing array 703 corresponding to the previous step number 801 acquired in step S2007 (step S2009). Then, the data acquisition device 200 determines whether there is an acquisition pattern requested for acquisition in the entry of the transfer destination step number 701 in the next acquisition timing array 703 acquired in step S2009 in the data acquisition estimation table 700. (Step S2010).

If there is no acquisition pattern (step S2010: No), the process returns to step S1907 in FIG. 19, and the data acquisition apparatus 200 waits for a data acquisition request. On the other hand, when there is an acquisition pattern (step S2010: Yes), the acquisition pattern corresponds to the entry registered in the data acquisition / history accumulation process (step S2006). Therefore, the data acquisition apparatus 200 updates the step number 701 in the step information management table 800 to the transfer destination step number 701 in step S2010 to acquire in advance the data acquired next to the pattern shape 606 (step S2011). ), The next activation timing 1001 of the advance acquisition process (FIG. 23) is set using the updated transfer destination step number 701 (step S2012). Then, control goes to a step S1907. As a result, the next activation timing 1001 of the pre-acquisition process is set for another acquisition pattern that is not the acquisition pattern of the highest selection probability.

In other words, it is switched to an acquisition pattern sequence of another acquisition pattern that is not the acquisition pattern of the highest selection probability. In this way, even if the acquisition pattern estimated for the pre-acquisition is different from the acquisition pattern actually requested for data acquisition, if it exists in the next acquisition timing array 703, the subsequent data acquisition request is estimated. Data pre-acquisition can be continued.

<Transfer example of acquisition pattern sequence>
FIG. 21 is an explanatory diagram of an example of transfer of the acquired pattern sequence. FIG. 21 is a tree structure of step number 701 in the next acquisition timing array 703 of the user ID: A001 shown in FIG. 7 (step number 701 is a node, and transition of step field 701 is a link). The number associated with the link is the selection probability.

The step number [U001] which is the highest node is the step number 701 when the user with the user ID: A001 logs in. The user selects the acquisition pattern of step [U003] with a probability of 70%, and selects the acquisition pattern of step [U004] with a probability of 30%.

The data acquisition apparatus 200 acquires the entry in the first row of FIG. 7 from the data acquisition estimation table 700 in step S1904 of FIG. 19, and acquires the step number [U003] that is the highest selection probability (70%) in step S1905. The start timing 1001 of the advance acquisition process is set for the pattern. In step S1906, the data acquisition apparatus 200 stores the selected step number [U003] as the step number 701 in the step information management table 800, and the previous step number 801 as the step number [U001] before the selection. Is stored.

Then, in step S2001, the data acquisition device 200 determines whether or not the acquisition pattern in the actually received data acquisition request matches the acquisition pattern in step [U003]. If they match (step S2001: Yes), the data acquisition apparatus 200 transmits data corresponding to the acquisition pattern of step [U003] acquired in advance at the start timing 1001 in step S1905 to the terminal 101 of the user with the user ID: A001. (Step S2002).

On the other hand, if it is determined that there is a mismatch in step S2001 (step S2001: No), the data requested by the user is not data corresponding to the acquisition pattern of step [U003] acquired in advance at activation timing 1001 in step S1905. Means. Therefore, the data acquisition device 200 acquires data corresponding to the actual data acquisition request from the power NW model DB 104 and accumulates the history (step S2006). As a result, even when the prediction is lost, the subsequent prediction accuracy (selection probability 1203) can be improved by accumulating and learning as past data.

Thereafter, the data acquisition device 200 acquires [U001] as the previous step number 801 of the acquisition user ID 602 from the step information management table 800 in step S2007. In step S2008, the data acquisition apparatus 200 determines whether the next acquisition timing array 703 corresponding to the acquired previous step number [U001] is in the data acquisition estimation table 700.

In this case, since the next acquisition timing array 703 of the step number [U003] with a selection probability of 30% that has not been selected last time exists (step S2008: Yes), the data acquisition device 200 has the next acquisition timing array 703 { Step number [U004] is selected from [U004] [1] [30]} (step S2009). The selected step number [U004] is the “migration destination step number 701”.

In step S1907, the data acquisition apparatus 200 sets the entry of the transfer destination step number [U004] in {[U004] [1] [30]} that is the next acquisition timing array 703 in the data acquisition estimation table 700 in step S1907. It is determined whether there is an acquisition pattern for which a data acquisition request has been made (step S2010). If it exists (step S2010: Yes), it means that the acquisition pattern requested for data acquisition in step S2005 matches the acquisition pattern of step number [U004]. Therefore, the data acquisition apparatus 200 updates the step number 701 from the step number [U003] to the transfer destination step number [U004] in the entry of the user ID: A001 in the step information management table 800.

As a result, the data acquisition apparatus 200 performs an advance acquisition process for the acquisition pattern for the entry corresponding to the step number 701 ([U011]) of the highest selection probability (65%) in the next acquisition timing array 703 of the step number [U004]. The activation timing 1001 is set (step S2012). As a result, the data acquisition apparatus 200 can switch from the route [U001] → [U003] to be acquired in advance to the route [U001] → [U004] in accordance with the data acquisition request actually received. In this way, in order to execute the pre-acquisition process following the user's actual data acquisition request, it is possible to continue the pre-acquisition by the transfer in the case where the acquisition pattern to be acquired in advance is not predicted. it can.

<Data acquisition / history storage processing>
FIG. 22 is a flowchart illustrating a detailed processing procedure example of the data acquisition / history accumulation processing (steps S1908 and S2006) illustrated in FIGS. 19 and 20. The data acquisition / history accumulation processing (steps S1908, S2006) acquires data corresponding to the acquisition pattern (a combination of the pattern shape 606 and the attribute value 607) selected in the data acquisition request, and uses the history as the data acquisition history. This is a process of accumulating in the accumulation table 600.

The data acquisition device 200 transmits a data acquisition request to the power NW model DB 104 and acquires the data acquisition request time 604 (step S2201). The data acquisition device 200 acquires data corresponding to the acquisition pattern (pattern shape 606 and attribute value 607) of the data acquisition request from the power NW model DB 104 (step S2202), and converts it into XML data (step S2203). The data acquisition device 200 calculates the time from the data acquisition request time 604 to the output of the XML data in step S2203 as the data processing time 605 (step S2204).

The data acquisition device 200 transmits the converted XML data to the login terminal 101 (step S2205). Then, the data acquisition device 200 registers data acquisition history information (user ID 602, session ID 603, data acquisition request time 604, data processing time 605, pattern shape 606, attribute value 607) in the data acquisition history accumulation table 600 (step S2206). As a result, a series of processing ends.

<Data pre-acquisition processing>
FIG. 23 is a flowchart illustrating a detailed processing procedure example of the data pre-acquisition processing. The data pre-acquisition process is started at the start timing 1001 set in steps S1905, S2004, and S2012, and starts the process.

The data acquisition device 200 waits until the start timing 1001 is reached (step S2301: No). The data acquisition apparatus 200 acquires the step number 701 in the entry of the acquisition user ID 602 in step S1902 from the step information management table 800 (step S2302). The data acquisition device 200 acquires the pattern shape 606 and the representative attribute value 702 in the entry of the acquisition step number 701 from the data acquisition estimation table 700 (step S2303).

The data acquisition device 200 acquires data satisfying the pattern shape 606 and the representative attribute value 702 from the power NW model DB 104 (step S2304). Then, the data acquisition device 200 converts the acquired data into XML data and sets it as pre-acquisition data 901 (step S2305). The data acquisition apparatus 200 registers the pre-acquisition information (user ID 602, pattern shape 606, attribute value 607, pre-acquisition data 901) in the prior acquisition information management table 900 (step S2306).

Thereafter, the data acquisition device 200 waits for the pre-acquisition data 901 to be processed, that is, read from the pre-acquisition information management table 900 in step S2002 and transmitted to the login terminal 101 (step S2307: No). ). When the pre-acquisition data 901 is processed (step S2307: Yes), the data acquisition device 200 deletes the entry from the pre-acquisition information management table 900 (step S2308). As a result, a series of processing ends. When the session ends, the data acquisition device 200 may delete the entry of the user ID 602 from the pre-acquisition information management table 900.

According to the first embodiment, since the data required at that time is prefetched, it is possible to immediately respond to the request source terminal 101 when a data acquisition request is made. In particular, it is useful when it takes time to acquire data such as conversion to XML data.

Embodiment 2 is an example in which data required by a new user who has not accumulated data acquisition history information is estimated, and data is acquired in advance at a necessary timing. Specifically, for example, when the new user transmits a data acquisition request by learning the acquisition pattern for each ID of the application 102 used by the new user in the business, the data acquisition device 200 learns for each application ID. Data required by a new user is estimated based on the result, and data is acquired in advance at the necessary timing. That is, since the tendency of a new user is not known, the data acquisition apparatus 200 estimates the tendency by the application 102 to be used and acquires in advance. This speeds up the data response to the new user. In addition, since it demonstrates based on the process of Example 1 here, it demonstrates paying attention to a different location.

<Data acquisition history accumulation table 600>
FIG. 24 is an explanatory diagram of an example of stored contents of the data acquisition history accumulation table 600 according to the second embodiment. The data acquisition history accumulation table 600 is a table in which an application ID field 2401 is added to the data acquisition history accumulation table 600 of the first embodiment. An application ID field 2401 is a storage area for storing an application ID. The application ID 2401 is identification information that uniquely identifies the application 102.

<Second data acquisition estimation table>
FIG. 25 is an explanatory diagram of an example of stored contents of the second data acquisition estimation table. In the second embodiment, the data acquisition estimation table 700 of the first embodiment is referred to as a “first data acquisition estimation table 700”. The second data acquisition estimation table 2500 is a table in which the user ID field 602 of the first data acquisition estimation table 700 is replaced with an application ID field 2401.

<Data acquisition process>
FIG. 26 is a flowchart of a detailed process procedure example 1 of the data acquisition process performed by the data acquisition apparatus 200 according to the second embodiment. First, the data acquisition device 200 waits for login detection of the terminal 101 (step S2601: No). When the login is detected (step S2601: Yes), the data acquisition apparatus 200 acquires the user ID 602 and the application ID 2401 being used from the terminal 101, and newly issues a session ID 603 (step S2602).

Thereafter, the data acquisition apparatus 200 determines whether the entry of the acquisition user ID 602 in step S2602 exists in the first data acquisition estimation table 700 (step S2603). If it does not exist (step S2603: NO), data cannot be acquired in advance for the acquired user ID 602, and the process proceeds to step S2604. When it exists (step S2603: Yes), the process proceeds to step S2605.

In step S2604, the data acquisition device 200 determines whether an entry of the acquisition application ID 2401 exists in the second data acquisition estimation table 2500 (step S2604).

In step S2603 or S2604, when the entry of the acquisition user ID 602 or application ID 2401 in step S2602 exists in the first data acquisition estimation table 700 or the second data acquisition estimation table 2500 (step S2603: Yes, S2604: Yes), Data can be acquired in advance. In this case, the data acquisition apparatus 200 includes an entry whose acquisition pattern ID and representative attribute value 702 are NULL and matches the acquisition ID (step S2603: user ID 602 if Yes, step S2604: application ID 2401 if Yes). Obtained from the data acquisition estimation table 700 (first data acquisition estimation table 700 for user ID 602, second data acquisition estimation table 2500 for application ID 2401) (step S2605).

And the data acquisition device 200 sets the start timing 1001 of the pre-acquisition process (FIG. 23) (step S2606). The set activation timing 1001 is managed for each user ID 602 in the activation timing management table 1000 of FIG. Details of the setting of the activation timing 1001 of the pre-acquisition process (FIG. 23) have been described with reference to the timing chart of FIG.

And the data acquisition device 200 stores the step information in the step information management table 800 shown in FIG. 8 (step S2607). Here, the step information includes the acquisition user ID 602, the step number 701 (the step number 701 of the highest selection probability in the next acquisition timing array 703 of the acquisition entry in step S2605), and the previous step number 801 (in step S2605). Step number 701) of the acquisition entry. Then, the process proceeds to step S2001 in FIG.

Note that, as shown by the dotted line arrow, only after login detection (step S2601: Yes), that is, when no data acquisition request has been received from the login terminal 101 even after login detection (step S2608: No), Instead of 20 step S2001, the process proceeds to step S2608.

In step S2604, when there is no entry of the acquisition application ID 2401 (step S2604: No), the process proceeds to step S2608. The data acquisition device 200 determines whether a data acquisition request has been received from the login terminal 101 (step S2608). If not received (step S2608: No), the process proceeds to step S2600. If received (step S2608: Yes), the data acquisition device 200 executes data acquisition / history accumulation processing (step S2609), and proceeds to step S2600. The data acquisition / history storage process (step S2609) is the same as the data acquisition / history storage process (step S1908) shown in FIG. 22 except that the application ID is also registered as data acquisition history information. .

Thereafter, it is determined whether or not logout of the login terminal 101 has been detected (step S2600). When logout is not detected (step S2600: No), the process returns to step S2603, and when logout is detected (step S2600: Yes), a series of processing ends.

As described above, in the second embodiment, even for a new user whose tendency is not known, the data acquisition apparatus 200 estimates the tendency using the application 102 to be used and acquires in advance, thereby speeding up the data response to the new user. be able to. In the second embodiment, the application ID 2401 and the application ID field 2401 have been described. However, the function ID and function ID field of the application 102, or the organization ID and organization ID field to which the user belongs may be replaced.

[Embodiment 3] Embodiment 3 is an example (difference acquisition) in which data excluding duplicate portions is acquired in advance when there is already acquired data in the data acquired in advance in Embodiment 1. If there is already acquired data in the data acquired in advance in the next step, the data conversion processing acquired from the power NW model DB 104 can be reduced if the data is acquired in advance excluding the overlapping part with the data in the previous step. The response can be speeded up. In addition, since it demonstrates based on the process of Example 1 here, it demonstrates paying attention to a different location.

FIG. 27 is an explanatory diagram of an example of preliminary data acquisition according to the third embodiment. For example, it is assumed that the user acquires data by selecting an acquisition pattern of a circle having a diameter of 3 km in the first data acquisition request. In the second data acquisition request, when an acquisition pattern of a circle with a diameter of 5 km is selected at the same position, the data acquisition device 200 changes the acquisition pattern of a circle with a diameter of 5 km from the acquisition pattern of a circle with a diameter of 5 km. Data is acquired from the power NW model DB 104 for the excluded range (in this case, a donut shape). Then, the data acquisition device 200 merges the first acquired data (XML data) and the current (second) XML data and returns the merged data to the terminal 101. In addition, in the third data acquisition request, when an acquisition pattern of a circle with a diameter of 7 km is selected at the same position, the data acquisition device 200 acquires a second circle with a diameter of 5 km from the acquisition pattern of a circle with a diameter of 7 km. Data is acquired from the power NW model DB 104 for a range excluding the pattern (in this case, a donut shape is also obtained). Then, the data acquisition device 200 merges the acquired data (XML data) after the second merge and the current (third) XML data and returns the merged data to the terminal 101.

<Data pre-acquisition processing>
FIG. 28 is a flowchart of a detailed process procedure example of the data pre-acquisition process according to the third embodiment. The data pre-acquisition process is started at the start timing 1001 set in steps S1905, S2004, and S2012, and starts the process.

The data acquisition device 200 waits until the start timing 1001 (step S2801: No). The data acquisition device 200 acquires the step number 701 and the previous step number 801 in the entry of the acquisition user ID 602 in step S1902 from the step information management table 800 (step S2802). The data acquisition apparatus 200 includes the pattern shape 606 and the representative attribute value 702 (acquisition pattern) in the entry of the acquisition step number 701, and the pattern shape 606 and the representative attribute value 702 (previous acquisition pattern) in the acquired entry of the previous step number 801. Is acquired from the data acquisition estimation table 700 (step S2803).

The data acquisition apparatus 200 determines whether the acquisition pattern and the previous acquisition pattern overlap (step S2804). When not overlapping (step S2804: No), the data acquisition device 200 acquires data satisfying the pattern shape 606 and the representative attribute value 702 from the power NW model DB 104 (step S2805). Then, the data acquisition device 200 converts the acquired data into XML data and sets it as pre-acquisition data 901 (step S2806). The data acquisition apparatus 200 registers the pre-acquisition information (user ID 602, pattern shape 606, attribute value 607, pre-acquisition data 901) in the pre-acquisition information management table 900 (step S2807). Thereafter, the process proceeds to step S2812.

On the other hand, if it is determined in step S2804 that they overlap (step S2804: Yes), the data acquisition device 200 generates a conditional expression for acquiring a difference using a difference acquisition pattern obtained by subtracting the previous acquisition pattern from the acquisition pattern (step S2804). S2808). The data acquisition device 200 acquires data satisfying the difference conditional expression from the power NW model DB 104 (step S2809). Then, the data acquisition device 200 converts the acquired data into XML data, and merges it with the previously acquired XML data to obtain the pre-acquired data 901 (step S2810). The data acquisition device 200 registers the pre-acquisition information (user ID 602, pattern shape 606, attribute value 607, pre-acquisition data 901) in the prior acquisition information management table 900 (step S2811). Thereafter, the process proceeds to step S2812.

This completes a series of processing. When the session ends, the data acquisition device 200 may delete the entry of the user ID 602 from the pre-acquisition information management table 900.

According to the third embodiment, it is not necessary to convert the data in the overlapping portion, and an immediate response can be made to the request source terminal 101.

Example 4 is an example of adjusting the start timing 1001 when the data processing times 605 of a plurality of users overlap. When a plurality of data processing times 605 overlap, the data acquisition apparatus 200 determines that the plurality of start timings 1001 (start of the data processing time 605) so that the duplication number of the data processing time 605 is less than the duplication number before adjustment. Time). Thereby, the load of the prior acquisition process can be reduced as compared with before adjustment.

FIG. 29 is an explanatory diagram of an example of adjusting the start timing 1001 according to the fourth embodiment. (A) shows the state before adjustment, and (B) shows the state after adjustment. In the example of FIG. 28, the multiplicity (an integer of 1 or more) is set to 2. The multiplicity is the maximum number of average data processing times that allow duplication, and is equal to or less than the user duplication number−1, in other words, equal to or less than the duplication number−1 of the average data processing time. Here, dpa to dpd are average data processing times of the four users (see the time chart in FIG. 10), and ta to td are the activation timings 1001. It should be noted that ta and tb that are the start timing 1001 are the same time (for example, 10:21) (may be different times). In (A), the average data processing times dpa to dpd overlap in the period alt.

Since the multiplicity is 2, the data acquisition apparatus 200 adjusts each average data processing time so that the average data processing time overlap number is 2 or less. As shown in (B), the data acquisition device 200 keeps the average data processing times dpa and dpb fixed, and shifts the average data processing times dpc and dpd so as not to overlap with the average data processing times dpa and dpb. .

Further, this shift amount may be adjusted so that the difference from the start timing 1001 before the shift becomes small. Specifically, for example, the time next to the end time of the average data processing time to be shifted may be the start timing 1001 of the average data processing time that is not to be shifted. The time next to the end time of the average data processing time may be the start timing 1001 of the average data processing time to be shifted. As a result, it is possible to minimize the deviation of the activation timing 1001 while reducing the processing load on the data acquisition device 200.

Note that since the multiplicity is 2, in this case, the data acquisition apparatus 200 determines that the multiplicity is 1 (that is, does not overlap) unless the average data processing time of other users other than the four users overlaps. You may adjust so that it may become. As described above, when the data acquisition device 200 is set so that the processing load of the data acquisition device 200 is more important than the deviation of the activation timing 1001, the multiplicity may be adjusted to be lower.

In the fourth embodiment, the example in which the activation timing 1001 is adjusted by detecting the overlap of the average data processing time has been described. However, the activation timing 1001 is adjusted by detecting the overlap of the activation timing 1001. Also good. In this case, since the calculation of the average data processing time is limited to the overlapping activation timing 1001, the adjustment process can be simplified.

<Adjustment processing procedure example>
FIG. 30 is a flowchart illustrating an example of the adjustment processing procedure of the activation timing 1001. The flowchart is executed, for example, at the timing when the latest activation timing 1001 is set. First, the data acquisition device 200 acquires the pattern shape 606 from the data acquisition estimation table 700 for each of the user IDs 602 corresponding to the plurality of activation timings 1001 that have been set and have not arrived (step S3001).

Next, the data acquisition apparatus 200 refers to the data acquisition history accumulation table 600, acquires the data processing time 605 in the pattern shape 606 acquired in step S3001, and average data processing in the acquired pattern shape 606 for the user. The time is calculated for each user ID 602 (step S3002).

Then, the data acquisition apparatus 200 determines whether or not the average data processing time from the activation timing 1001 of each user ID 602 overlaps (step S3003). Here, if the overlap of the average data processing time is less than or equal to the multiplicity, the data acquisition device 200 determines that there is no overlap (step S3003: No), and if the multiplicity is exceeded, the data acquisition device 200 It is determined that they overlap (step S3003: Yes).

If there is no overlap (step S3003: No), the data acquisition device 200 ends the adjustment process of the activation timing 1001. On the other hand, if they overlap (step S3003: Yes), the data acquisition device 200 adjusts each activation timing 1001 based on the multiplicity, and updates the activation timing management table 1000 for the adjusted activation timing 1001 (step S3004). . Thereby, the data acquisition device 200 ends the adjustment process of the activation timing 1001.

As described above, according to the data acquisition device 200 described above, a data acquisition request is made by estimating a data acquisition request that the user will request and acquiring data corresponding to the data acquisition request in advance. Data can be provided to the user immediately. Therefore, it is possible to suppress response delay of data acquired in response to a request.

Further, when there are a plurality of types of data acquisition requests that the user will request, by obtaining the selection probability 1203 for each data acquisition request that the user will select from the learning result of the past data acquisition requests. Therefore, it is possible to improve the estimation accuracy of data pre-acquisition.

Further, even when a data acquisition request different from the data acquisition request selected according to the selection probability 1203 is actually received, the data acquisition device 200 performs data acquisition in advance by changing the acquisition pattern sequence. Can be run continuously. Further, by learning the different data acquisition requests, it is possible to improve the estimation accuracy of data prior acquisition thereafter.

Also, since the request source is a terminal that identifies the user (worker), it is possible to learn an acquisition pattern unique to the user. The request source may be the application 102 used by a plurality of users. Thereby, an acquisition pattern unique to the application 102 can be learned. As a result, even for a new user, data can be acquired in advance by specifying the application 102 to be used.

Also, when the acquisition patterns of a series of data acquisition requests are duplicated, the advance acquisition process can be speeded up by acquiring in advance excluding the overlapping part.

When the time periods of data processing times of a plurality of users overlap, the load on the data acquisition device 200 can be reduced by adjusting the number of overlaps so as to be less than the number of users according to the multiplicity.

In the first to fourth embodiments described above, the power NW model using power equipment as a node has been described as an example, but the present invention can also be applied to other data. For example, if you want to acquire store-specific data for the purpose of analyzing sales of a retail company, the node is a store, and the attributes are sales per month, number of customers per month, age group of customers with the highest sales, and store location. That's fine. As a result, for example, for the classified data, data of stores having the largest number of customers in their 60s, sales of 10 million yen or more, and the number of customers of 1,000 or less are acquired and analyzed. Usage is possible. Further, when the analysis is continued by changing the attributes such as the 50s and 40s, the acquisition pattern step can be estimated.

The present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the appended claims. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to those having all the configurations described. A part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Moreover, you may add the structure of another Example to the structure of a certain Example. Moreover, you may add, delete, or replace another structure about a part of structure of each Example.

In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, with an integrated circuit, and the processor realizes each function. It may be realized by software by interpreting and executing the program to be executed.

Information such as programs, tables, and files for realizing each function is recorded on a memory, a hard disk, a storage device such as SSD (Solid State Drive), or an IC (Integrated Circuit) card, SD card, DVD (Digital Versatile Disc). It can be stored on a medium.

Also, the control lines and information lines indicate what is considered necessary for the explanation, and do not necessarily indicate all control lines and information lines necessary for mounting. In practice, it can be considered that almost all the components are connected to each other.

Claims

A data acquisition device that is accessible to a database storing a data group, acquires data according to a data acquisition request from a request source for the database, and returns the data to the request source.
A processor that executes the program; and a storage device that stores the program;
The processor is
For each data acquisition request of the series of data acquisition requests received from the request source, a learning process for learning the type of the data acquisition request and the acquisition timing that defines the time series order of the data acquisition request;
Prior to the first data acquisition request after the series of data acquisition requests, the type of the first next estimated data acquisition request that is a data acquisition request that can be acquired from the request source based on the learning result by the learning process and Specific processing to identify acquisition timing;
Pre-acquisition processing for acquiring data corresponding to the first next estimated data acquisition request whose type is specified by the specifying processing from the database;
If the first data acquisition request is received, whether or not the type included in the first data acquisition request matches the type of the first next estimated data acquisition request specified by the specifying process A judgment process for judging
A transmission process for transmitting to the request source the acquisition result of the first next estimated data acquisition request specified by the specifying process, when the determination process determines that they match. When,
The data acquisition apparatus characterized by performing.
The data acquisition device according to claim 1,
In the learning process, the processor selects, from the request source, each data acquisition request of the series of data acquisition requests from the request source in chronological order for a series of past data acquisition requests from the request source. The probability of
In the specifying process, the processor specifies the type of the first next estimated data acquisition request based on the probability calculated by the learning process.
A data acquisition apparatus characterized by that.
The data acquisition device according to claim 2,
In the specifying process, when there are a plurality of first next estimated data acquisition requests, the processor specifies the type and acquisition timing of the first next estimated data acquisition request having the highest probability. A data acquisition device.
The data acquisition device according to claim 3,
In the specifying process, the processor specifies the type and the acquisition timing of the first next estimated data acquisition request that has not been selected, when the determination process determines that there is a mismatch,
In the determination process, the processor determines whether or not the type included in the first data acquisition request matches the type of the first next estimated data acquisition request not selected.
In the specific process, when the processor determines that the type included in the first data acquisition request matches the type of the first next estimated data acquisition request that has not been selected, Based on this, the data acquisition device is characterized by identifying the type and acquisition timing of the second next estimated data acquisition request, which is a data acquisition request that can be acquired from the request source after the first data acquisition request.
The data acquisition device according to claim 1,
In the learning process, the processor learns the acquisition timing based on an elapsed time from the previous data acquisition request for each of a series of past data acquisition requests from the request source.
A data acquisition apparatus characterized by that.
The data acquisition device according to claim 1,
The processor is
The acquisition timing of the first next estimated data acquisition request is after the elapse of the data conversion time required for converting the data corresponding to the first next estimated data acquisition request into a predetermined format. Execute the setting process to set the start timing of the pre-acquisition process,
In the pre-acquisition process, the processor acquires data corresponding to the first next estimated data acquisition request from the database at the start timing set by the setting process.
A data acquisition apparatus characterized by that.
The data acquisition device according to claim 1,
The data acquisition apparatus according to claim 1, wherein the request source is specified by user identification information.
The data acquisition device according to claim 1,
The request source is specified by identification information of an application that issues the data acquisition request at the request source.
The data acquisition device according to claim 1,
In the pre-acquisition process, the processor corresponds to a second next estimated data acquisition request that is a data acquisition request that can be acquired from the request source prior to a second data acquisition request subsequent to the first data acquisition request. If the data to be overlapped with the data corresponding to the first data acquisition request, the data not overlapping with the data corresponding to the first data acquisition request among the data corresponding to the second next estimated data acquisition request Acquired from a database and merged with data that overlaps with data corresponding to the first data acquisition request;
A data acquisition apparatus characterized by that.
The data acquisition device according to claim 6,
In the setting process, when the time zone of the prior acquisition process for acquiring data corresponding to the first next estimated data acquisition request from each of the plurality of request sources from the database overlaps, Adjusting the activation timing of the prior acquisition processing of the plurality of request sources so that the number of overlapping time zones is less than the number of request sources;
A data acquisition apparatus characterized by that.
A data acquisition method by a data acquisition device that is accessible to a database storing a data group, acquires data according to a data acquisition request from a request source for the database, and returns the data to the request source,
The data acquisition device includes a processor that executes a program, and a storage device that stores the program,
The processor is
For each data acquisition request of the series of data acquisition requests received from the request source, a learning process for learning the type of the data acquisition request and the acquisition timing that defines the time series order of the data acquisition request;
Prior to the first data acquisition request after the series of data acquisition requests, the type of the first next estimated data acquisition request that is a data acquisition request that can be acquired from the request source based on the learning result by the learning process and Specific processing to identify acquisition timing;
Pre-acquisition processing for acquiring data corresponding to the first next estimated data acquisition request whose type is specified by the specifying processing from the database;
If the first data acquisition request is received, whether or not the type included in the first data acquisition request matches the type of the first next estimated data acquisition request specified by the specifying process A judgment process for judging
A transmission process for transmitting to the request source the acquisition result of the first next estimated data acquisition request specified by the specifying process, when the determination process determines that they match. When,
The data acquisition method characterized by performing.
A data acquisition program that is accessible to a database that stores a data group and that is executed by a data acquisition device that acquires data according to a data acquisition request from a request source for the database from the database and returns the data to the request source. And
In a processor that executes the data acquisition program,
For each data acquisition request of the series of data acquisition requests received from the request source, a learning process for learning the type of the data acquisition request and the acquisition timing that defines the time series order of the data acquisition request;
Prior to the first data acquisition request after the series of data acquisition requests, the type of the first next estimated data acquisition request that is a data acquisition request that can be acquired from the request source based on the learning result by the learning process and Specific processing to identify acquisition timing;
Pre-acquisition processing for acquiring data corresponding to the first next estimated data acquisition request whose type is specified by the specifying processing from the database;
If the first data acquisition request is received, whether or not the type included in the first data acquisition request matches the type of the first next estimated data acquisition request specified by the specifying process A judgment process for judging
A transmission process for transmitting to the request source the acquisition result of the first next estimated data acquisition request specified by the specifying process, when the determination process determines that they match. When,
A data acquisition program for executing