JP7121923B2 - Data management device, data management method and program, and simulation system - Google Patents

Description

The present invention relates to a data management device, a data management method and program, and a simulation system.

There is a technique of acquiring a large amount of data via a network or the like and performing some processing using the acquired data. In relation to this technique, Patent Literature 1 discloses an information processing device that extracts compound words with high frequency of appearance from large-scale data. The information processing apparatus according to Patent Document 1 acquires large-scale data such as blog text and web page data from a data management apparatus. The information processing apparatus according to Patent Document 1 performs morphological analysis on large-scale data, calculates the appearance frequency value of the morpheme obtained as the analysis result and/or the morpheme string that is a set of morphemes, and calculates the appearance frequency value character groups that are candidates for compound words are extracted based on Then, the information processing apparatus according to Patent Document 1 holds the extracted character group in a predetermined storage area, and, from the held character group, a character group that satisfies a condition for determining a valid compound word appears. Extract as a group of characters with high frequency.

JP 2012-141783 A

It is assumed that data (simulation data) output from at least one simulator that performs various simulations is stored in a database and managed in an integrated manner. This simulation data is of various types and has various data types. Here, in the technique according to Patent Document 1, it is assumed that large-scale data to be acquired is text data (character string type). Therefore, in the technique disclosed in Patent Document 1, when simulation data of various data types are stored in a database and centrally managed, the data types must be specified by the user before being stored in the database.

An object of the present disclosure is to solve such problems, and includes a data management device, a data management method and a program capable of automatically and centrally managing data output from a simulator, and It is to provide a simulation system.

A data management device according to the present disclosure includes data acquisition means for acquiring first data related to simulation from at least one simulator, and based on a character string related to the first data acquired by the data acquisition means, the first data type determination means for determining a data type of one data; and data management means for managing the first data by a management method according to the data type determined by the data type determination means.

Further, a simulation system according to the present disclosure includes at least one simulator that executes a simulation, a data management device that manages first data related to the simulation by the simulator, and an arithmetic device that determines parameters related to the simulation. , the data management device comprises: data acquisition means for acquiring the first data from the simulator; and data of the first data based on a character string related to the first data acquired by the data acquisition means data type determination means for determining a type; and data management means for managing the first data according to a management method according to the data type determined by the data type determination means; The parameters are determined using the first data managed by the data management means, and the simulator executes the simulation again using the parameters determined by the arithmetic device.

Further, the data management method according to the present disclosure acquires first data related to simulation from at least one simulator, and based on a character string related to the acquired first data, the data type of the first data is determined, and the first data is managed by a management method according to the determined data type.

Further, the program according to the present disclosure includes a step of acquiring first data related to simulation from at least one simulator, and a data type of the first data based on a character string related to the acquired first data and managing the first data by a management method according to the determined data type.

According to the present disclosure, it is possible to provide a data management device, a data management method and program, and a simulation system that can automatically and centrally manage data output from simulators.

1 is a diagram showing an overview of a simulation system according to an embodiment of the present disclosure; FIG. 1 is a diagram showing a configuration of a simulation system according to Embodiment 1; FIG. 1 is a diagram showing a configuration of a data management device according to Embodiment 1; FIG. 1 is a diagram showing a configuration of a simulator according to Embodiment 1; FIG. 1 is a diagram showing a configuration of an arithmetic device according to a first embodiment; FIG. 4 is a sequence diagram showing the flow of processing in the simulation system according to the first embodiment; FIG. 4 is a flow chart showing a data management method executed by the data management device according to the first embodiment; 4 is a flowchart showing data type determination processing performed by the data management device according to the first embodiment; 4 is a diagram visually exemplifying the hierarchical structure of the database in the data management unit according to the first embodiment; FIG.

(Overview of Embodiments According to the Present Disclosure)
Prior to describing the embodiments of the present disclosure, an outline of the embodiments of the present disclosure will be described. FIG. 1 is a diagram showing an overview of a simulation system 1 according to an embodiment of the present disclosure. The simulation system 1 has a data management device 10 , at least one simulator 20 and at least one arithmetic device 40 . The data management device 10, the simulator 20, and the arithmetic device 40 have functions as computers, for example.

Simulator 20 executes a simulation. The data management device 10 manages simulation data (first data) regarding simulation by the simulator 20 . Arithmetic unit 40 determines parameters for the simulation. Arithmetic processing by the arithmetic unit 40 can be realized using, for example, artificial intelligence (AI), data assimilation, or a statistical method, but is not limited to this.

The data management device 10 has a data acquisition section 12 , a data type determination section 14 and a data management section 16 . The data acquisition unit 12 functions as data acquisition means. The data type determination unit 14 functions as data type determination means. The data management unit 16 functions as data management means.

The data acquisition unit 12 acquires simulation data from the simulator 20 . The data type determination unit 14 determines the data type of the simulation data based on the character string regarding the simulation data acquired by the data acquisition unit 12 . Here, the "character string related to the simulation data" includes, for example, the extension of the data file of the simulation data, and the character string included in the simulation data (that is, the character string in the contents of the simulation data), but is limited to this. not. The data management unit 16 manages simulation data by a management method according to the data type determined by the data type determination unit 14 . Here, data types include, but are not limited to, image (binary) types, character string types, numeric types, and the like.

The computing device 40 uses the simulation data managed by the data management unit 16 to determine parameters related to simulation. The simulator 20 uses the parameters determined by the computing device 40 to perform the simulation again.

As described above, the data management device 10 according to the present embodiment determines the data type of simulation data acquired from the simulator 20, and manages the simulation data by a management method according to the determined data type. As a result, the data management device 10 according to the present embodiment can automatically determine data types of various simulation data and manage these simulation data by a management method suitable for the data type. can. Therefore, the data management device 10 according to the present embodiment can automatically and centrally manage data in various formats output from the simulator. It should be noted that the data output from the simulator can be automatically and centrally managed by using the data management method and the program for executing the data management method executed by the data management device 10 .

Furthermore, in the simulation system 1 having the data management device 10, simulation data is managed by a management method suitable for its data type. As a result, the data management device 10 can manage the simulation data in such a manner that it can be efficiently used by other nodes such as the arithmetic device 40 . Therefore, the computing device 40 can appropriately determine the parameters related to the simulation using the simulation data of the data type suitable for the computation (algorithm). Therefore, the simulation system 1 according to this embodiment can improve the accuracy of simulation by the simulator 20 .

(Embodiment 1)
Hereinafter, embodiments will be described with reference to the drawings. For clarity of explanation, the following descriptions and drawings are omitted and simplified as appropriate. Moreover, in each drawing, the same elements are denoted by the same reference numerals, and redundant description is omitted as necessary.

FIG. 2 is a diagram showing the configuration of the simulation system 1 according to the first embodiment. The simulation system 1 has a user interface 4 (UI; User Interface), a simulator 20 , a data management device 100 , an arithmetic device 40 , a data removal device 60 , a network edge 70 and a data monitoring device 80 . The user interface 4 is communicably connected to the simulator 20 . The simulator 20, the data management device 100, the arithmetic device 40, the data removal device 60, the network edge 70, and the data monitoring device 80 may be communicably connected to each other via the network 2 such as the Internet. The simulator 20, the data management device 100, the arithmetic device 40, the data removal device 60, the network edge 70, and the data monitoring device 80 are computers having processors, memories, I/O (Input/Output), user interfaces, etc. can have the function of

The user interface 4 sends an instruction to the simulator 20 to automatically execute a simulation in response to a user or computer-controlled operation. The instructions sent by the user interface 4 may include input values (initial values) to be input to the model used in the simulation.

The simulator 20 automatically performs simulation according to instructions from the user interface 4 . The simulator 20 performs simulation according to the model. The simulator 20 automatically transmits simulation data, which is data relating to the executed simulation, to the data management device 100 . The simulator 20 will be described later with reference to FIG. 4 and the like. Here, "simulation data" is a data file, and may include model input values, data indicating simulation results, and data indicating the model. Note that the simulator 20 may be controlled by an automated program such as a bot, for example.

A data management device 100 corresponds to the data management device 10 in FIG. The data management device 100 has a database. As described above, the data management device 100 determines the data type of the simulation data acquired from the simulator 20, and manages the simulation data by a management method according to the determined data type. The data management device 100 will be described later with reference to FIG. 3 and the like. Note that the data management device 100 may be realized by a cluster system configured by connecting a plurality of computer devices via a network.

As described above, the arithmetic device 40 uses the simulation data managed by the data management device 100 to calculate parameters related to the simulation performed by the simulator 20 by artificial intelligence, data assimilation, statistical techniques, or the like. decide. The computing device 40 will be described later with reference to FIG. 5 and the like. The “parameters related to simulation” may include analytical expressions (functions) that describe the model, variables and constants that make up the analytical expressions that describe the model, input values that are input to the model, and output values of the model. The simulator 20 uses the parameters determined by the computing device 40 to perform the simulation again. Thereby, the accuracy of the simulation by the simulator 20 can be improved.

The data removal device 60 removes unnecessary data from the simulation data managed by the data management device 100 (data cleansing). For example, the data removal device 60 removes data that are unlikely to contribute to improving the accuracy of the simulation. Here, the data removal device 60 may, for example, analyze the simulation data by machine learning to determine unnecessary data. The computing device 40 can acquire simulation data that has undergone data cleansing.

The network edge 70 generates actual data (actual data) of the events targeted in the simulation. In other words, the network edge 70 functions as a real data generation device. Note that the network edge 70 may be, for example, an actual machine implementing a model, a sensor observing the actual machine, or a monitoring device monitoring the actual machine. The network edge 70 transmits the generated actual data to the computing device 40 .

The data monitoring device 80 may have functions as a block chain, for example. The data monitoring device 80 manages, for example, consistency of data on a network such as the Web, prevention of falsification of data, and the like.

FIG. 3 is a diagram showing the configuration of the data management device 100 according to the first embodiment. The data management device 100 has a control unit 102, a storage unit 104, a communication unit 106, and an interface unit 108 (IF: Interface) as main hardware components. The control unit 102, storage unit 104, communication unit 106 and interface unit 108 are interconnected via a data bus or the like.

The control unit 102 is a processor such as a CPU (Central Processing Unit), for example. The control unit 102 has a function of executing control processing, arithmetic processing, and the like. The storage unit 104 is, for example, a storage device such as memory or hard disk. The storage unit 104 is, for example, a ROM (Read Only Memory) or a RAM (Random Access Memory). The storage unit 104 has a function of storing a control program, an arithmetic program, and the like executed by the control unit 102 . The storage unit 104 also has a function of temporarily storing processing data and the like. Storage unit 104 may include a database.

The communication unit 106 performs processing necessary for communicating with other devices (such as the simulator 20 and the arithmetic device 40) via the network 2. FIG. Communication unit 106 may include communication ports, routers, firewalls, and the like. The interface unit 108 is, for example, a user interface (UI). The interface unit 108 has an input device such as a keyboard, touch panel, or mouse, and an output device such as a display or speaker. The interface unit 108 receives a data input operation by a user (operator) and outputs information to the user. The interface unit 108 may display a diagram or the like showing the structure (hierarchy) of the database in the data management device 100 .

The data management device 100 also has a data storage unit 110 , a data acquisition unit 112 , a data size determination unit 114 , and a metadata extraction unit 116 . The data management device 100 also has a data type determination unit 120 , a management method determination unit 130 , a data management unit 140 , a data request reception unit 150 and a data transmission unit 152 . Hereinafter, the data storage unit 110 and the like described above may be referred to as "components". The data storage unit 110, data acquisition unit 112, data size determination unit 114, and metadata extraction unit 116 function as data storage means, data acquisition means, data size determination means, and metadata extraction means, respectively. Also, the data type determination unit 120, the management method determination unit 130, and the data management unit 140 function as data type determination means, management method determination means, and data management means, respectively. Also, the data request receiving unit 150 and the data transmitting unit 152 function as data request receiving means and data transmitting means, respectively.

A data acquisition unit 112 corresponds to the data acquisition unit 12 shown in FIG. The data type determination unit 120 corresponds to the data type determination unit 14 shown in FIG. The data management section 140 corresponds to the data management section 16 shown in FIG.

Each component can be realized by executing a program under the control of the control unit 102, for example. More specifically, each component can be implemented by control unit 102 executing a program stored in storage unit 104 . Further, each component may be realized by recording necessary programs in an arbitrary non-volatile recording medium and installing them as necessary. Moreover, each component may be implemented by any combination of hardware, firmware, and software, without being limited to being implemented by program software. Also, each component may be implemented using a user-programmable integrated circuit such as an FPGA (field-programmable gate array) or a microcomputer. In this case, this integrated circuit may be used to implement a program composed of the above components. This is the same for components of other devices (simulator 20, arithmetic device 40, etc.) described later.

Specific functions of each component of the data management device 100 will be described later. The data management unit 140 functions as a database (DB). The data management unit 140 conforms to Hadoop, for example, and can store and manage data so that distributed processing is possible. In this case, the data management unit 140 may be realized by HDFS (Hadoop Distributed File System). On the other hand, the data storage unit 110 functions as a directory. The data storage unit 110 stores data (for example, image data, etc.) that is not stored in the database. Note that the data storage unit 110 may be implemented as a storage device physically separate from the data management device 100 .

FIG. 4 is a diagram showing the configuration of the simulator 20 according to the first embodiment. The simulator 20 has a control unit 21, a storage unit 22, a communication unit 23, and an interface unit 24 as main hardware components. The control unit 21, storage unit 22, communication unit 23, and interface unit 24 are connected to each other via a data bus or the like.

The control unit 21 is a processor such as a CPU, for example. The control unit 21 has a function of executing control processing, arithmetic processing, and the like. The storage unit 22 is, for example, a storage device such as memory or hard disk. The storage unit 22 is, for example, ROM or RAM. The storage unit 22 has a function of storing control programs, arithmetic programs, and the like executed by the control unit 21 . The storage unit 22 also has a function of temporarily storing processing data and the like.

The communication unit 23 performs processing necessary for communicating with other devices via the network 2 . The communication unit 23 may include communication ports, routers, firewalls, and the like. The interface unit 24 is, for example, a user interface. The interface unit 24 has an input device such as a keyboard, touch panel, or mouse, and an output device such as a display or speaker. The interface unit 24 receives a data input operation by a user (operator) and outputs information to the user. The interface unit 24 may display a diagram or the like showing the simulation results. Note that the user interface 4 shown in FIG. 2 may be implemented by the interface section 24 .

The simulator 20 also has a simulation execution unit 26 , a simulation data transmission unit 28 and an algorithm execution result reception unit 30 . The simulation execution unit 26 may have, for example, a simulation A execution unit 26A, a simulation B execution unit 26B, a simulation C execution unit 26C, a simulation D execution unit 26D, and a simulation E execution unit 26E. The simulation execution unit 26, the simulation data transmission unit 28, and the algorithm execution result reception unit 30 function as simulation execution means, simulation data transmission means, and algorithm execution result reception means, respectively. Operations of the simulation execution unit 26, the simulation data transmission unit 28, and the algorithm execution result reception unit 30 will be described later.

The simulator 20 may be composed of a plurality of physically separate devices each having a simulation A executing section 26A to a simulation E executing section 26E. Also, the simulation A executing section 26A to the simulation E executing section 26E can execute different simulations A to E, respectively. For example, the simulation A executing section 26A may execute a structural analysis (strength analysis) simulation as the simulation A. The simulation B executing section 26B may execute a fluid simulation as the simulation B. The simulation C executing section 26C may execute an electromagnetic field simulation as the simulation C. The simulation D executing unit 26D may execute a weather forecast simulation as the simulation D. The simulation E executing section 26E may execute a production control simulation as the simulation E.

FIG. 5 is a diagram showing the configuration of the arithmetic device 40 according to the first embodiment. The arithmetic unit 40 has a control unit 41, a storage unit 42, a communication unit 43, and an interface unit 44 as main hardware components. The control unit 41, the storage unit 42, the communication unit 43 and the interface unit 44 are interconnected via a data bus or the like. Functions of the control unit 41, the storage unit 42, the communication unit 43, and the interface unit 44 are substantially the same as the functions of the control unit 21, the storage unit 22, the communication unit 23, and the interface unit 24, respectively. do.

The computing device 40 has an algorithm execution unit 46 , a simulation data acquisition unit 48 and an algorithm execution result transmission unit 50 . The algorithm execution unit 46 may have, for example, an algorithm A execution unit 46A, an algorithm B execution unit 46B and an algorithm C execution unit 46C. The algorithm execution unit 46, the simulation data acquisition unit 48, and the algorithm execution result transmission unit 50 function as algorithm execution means, simulation data acquisition means, and algorithm execution result transmission means, respectively. Operations of the algorithm execution unit 46, the simulation data acquisition unit 48, and the algorithm execution result transmission unit 50 will be described later.

Arithmetic device 40 may be composed of a plurality of physically separate devices or virtually separate devices each having algorithm A execution unit 46A, algorithm B execution unit 46B, and algorithm C execution unit 46C. good. Algorithm A executing section 46A, algorithm B executing section 46B, and algorithm C executing section 46C can execute different algorithms A to C, respectively. Also, algorithms A to C may differ in the data types that can be handled.

For example, the algorithm A execution unit 46A may execute a search algorithm as algorithm A. Algorithm B executing section 46B may execute a learning algorithm as Algorithm B. FIG. Algorithm C executing section 46C may execute a data assimilation algorithm as Algorithm C. FIG. The search algorithm may be, for example, a meta-heuristic algorithm such as a genetic algorithm, simulated annealing or Bayesian optimization. The learning algorithm may be a deep neural network, a Bayesian neural network, or the like. The data assimilation algorithm may be an algorithm using a particle filter, a Kalman filter, or the like.

FIG. 6 is a sequence diagram showing the flow of processing in the simulation system 1 according to the first embodiment. First, the simulator 20 executes a simulation (step S100). Specifically, the simulation execution unit 26 of the simulator 20 executes the simulation according to instructions from the user interface 4 .

Then, the simulation data transmission unit 28 of the simulator 20 transmits the simulation data (Sim data) including the simulation result to the data management device 100 (step S112). Here, as described above, the simulation data transmitted from the simulator 20 to the data management device 100 can be data in a wide variety of formats. The data management device 100 manages the simulation data as described later (step S120).

FIG. 7 is a flow chart showing the data management method (S120) executed by the data management device 100 according to the first embodiment. The data acquisition unit 112 acquires simulation data from the simulator 20 (step S122). Here, what the data acquisition unit 112 receives from the simulator 20 may be raw data of the simulation data. The format (data type) of the simulation data may differ from simulation to simulation (simulations A to E). Furthermore, the simulation data may differ depending on the stage of the simulation, even for the same simulation. Therefore, the data acquisition unit 112 acquires simulation data (raw data) in various formats.

Next, the data size determination unit 114 determines whether the data size of the simulation data acquired in S122 is equal to or greater than a predetermined threshold (step S124). Here, the threshold value is determined according to the capacity of each resource that performs parallel processing when data (management data) managed by the data management device 100 is subjected to distributed processing by Hadoop or the like. For example, the threshold is several hundred megabytes, but is not limited to this.

If the data size of the simulation data (raw data) is not equal to or greater than the threshold value (that is, is less than the threshold value) (NO in S124), the management method determining unit 130 manages the simulation data by storing the metadata in the database. decide. Metadata is data associated with simulation data. For example, data such as the file name in which the simulation data is stored, the date and time when the file was created, or keywords included in the data. be. Therefore, the metadata extraction unit 116 extracts the metadata of this simulation data (step S126). Then, the data management unit 140 (database) stores the metadata extracted in the process of S126 (step S128). At this time, simulation data whose data size is less than the threshold is stored in the data storage unit 110, which is a directory. As a result, the data management unit 140 manages simulation data whose data size is less than the threshold value by storing the metadata.

Here, the simulation data whose data size is less than the threshold may be, for example, CSV (Comma Separated Value) data. Also, the metadata extracted in the process of S126 may include an identifier indicating which simulation the corresponding simulation data relates to. Also, this metadata may include a storage destination path of a directory in which corresponding simulation data is stored. The data type of simulation data whose data size is less than the threshold may be determined by the user opening the data file, or by using the data type determination process (S130) described later. Also, the simulation data whose data size is less than the threshold can be transmitted by the data transmission unit 152 to the constituent elements of the algorithm execution unit 46 that execute algorithms (algorithms A to C) according to the determined data type. .

The reason why simulation data whose data size is less than a threshold value is managed by storing metadata in this way is to improve processing efficiency. That is, the data management unit 140 (database) can be configured to be applicable to distributed processing as described above. Therefore, if this simulation data is stored in the data management unit 140, distributed processing will be performed on this simulation data even though the data size is smaller than the capacity of each resource that performs parallel processing. In the case of a small data size, the overhead due to distributed processing becomes obvious, so it is difficult to achieve high distributed efficiency, and resources may be wasted. Therefore, by storing and managing simulation data whose data size is less than a threshold value, wasteful use of resources can be suppressed.

On the other hand, if the data size of the simulation data (raw data) is equal to or greater than the threshold (YES in S124), the data management device 100 (data type determination unit 120) determines the data type of the simulation data (step S130). Then, the data management unit 140 manages the simulation data according to the data type determined in the process of S130 (step S150).

FIG. 8 is a flowchart showing data type determination processing (S130) performed by the data management device 100 according to the first embodiment. The data type determination unit 120 determines the extension of the data file of the simulation data (step S132). If the extension is related to image data (binary data) such as "jpg" or "png" ("image data" in S132), the management method determination unit 130 converts this simulation data into Decide to manage by management method. In this case, management method determination unit 130 determines not to store the simulation data in data management unit 140 but to store it in data storage unit 110 as an image. Then, the management method determination unit 130 determines to manage the simulation data by having the data management unit 140 store the path of the storage destination directory (storage destination path) in the data storage unit 110 of the simulation data, which is an image. . Therefore, the data storage unit 110 stores image-type simulation data. The data management unit 140 stores the storage destination path of this simulation data (step S134). Thereby, the data management section 140 manages this simulation data by a management method according to the image type (S150). In this way, the data type determination unit 120 can easily determine whether or not the data type is the image type by first determining the extension of the data file of the simulation data.

On the other hand, if the extension is related to ASCII data (text data) such as "csv" or "txt" ("ASCII data" in S132), the data type determination unit 120 determines that the data type of this simulation data is text. Determined to be column type or numeric type. Then, the management method determining unit 130 determines to manage the simulation data by a management method corresponding to the character string type or numerical value type. In this case, first, the data management unit 140 temporarily stores data (character strings or numerical values) separated by delimiters (commas, tabs, spaces, etc.) in the simulation data as character strings (step S136). This makes it easy to manage data delimited by delimiters. Also, this makes it possible to read this data at high speed.

The data type determination unit 120 determines whether or not a predetermined amount of data regarding the same simulation is stored by repeating the same simulation as the data stored in S136 a predetermined number of times (step S138). An identifier that identifies the simulation may be added to the simulation data acquired from the simulator 20, and the data type determination unit 120 may refer to the identifier to determine "same simulation". Here, the "predetermined amount" is, for example, the amount of data obtained by repeating the simulation 100 times, but is not limited to this. If the predetermined amount of data is not stored (NO in S138), the processes of S132 and S136 are repeated.

On the other hand, if a predetermined amount of data is stored (YES in S138), the data type determination unit 120 determines whether the data (character string) temporarily stored in the process of S136 can be handled as a numerical value ( step S140). Specifically, the data type determination unit 120 determines whether or not it is possible to sum all the stored predetermined amounts of data. The data type determination unit 120 determines that the stored data can be handled as a numerical value when all the stored predetermined amount of data can be summed. The determination as to whether or not it is possible to add all the predetermined amount of data stored may be performed by a method similar to general spreadsheet calculation. If it is not possible, error processing will occur. The reason why it is determined in the process of S138 whether or not the predetermined amount of data has been stored is that it is impossible to determine whether or not summation can be performed unless the predetermined amount of data is stored. In other words, if the amount of data is small and the stored data happens to be only numbers, it is impossible to judge whether the data can really be summed (whether it can be handled as a number). In other words, the "predetermined amount" in S138 can be set in advance by the user or the like to such an extent that it is possible to reliably determine whether or not the data can be summed in the process of S140.

If it is determined that the stored data can be handled as numerical values (YES in S140), the management method determination unit 130 determines to manage this simulation data using a management method according to the numerical type. Therefore, the data management unit 140 stores the data delimited by the delimiter as a numerical value (step S142). At this time, this numerical value may be converted into an integer type or a decimal type (floating point type, etc.) according to the number of digits.

On the other hand, if it is determined that the stored data cannot be handled as numerical values (NO in S140), the management method determination unit 130 determines to manage this simulation data using a management method according to the character string type. Therefore, the data management unit 140 stores the data delimited by the delimiter as a character string (step S144). In this way, the data type determination unit 120 determines whether the data (character strings) constituting the simulation data can be treated as numerical values, thereby determining whether the data type of the simulation data is character string type or numerical value type. can be easily and reliably determined.

FIG. 9 is a diagram visually exemplifying the hierarchical structure of the database in the data management unit 140 according to the first embodiment. In the example shown in FIG. 9, the database in the data management unit 140 is composed of three layers, layer #1 to layer #3. Hierarchy #1 is the highest hierarchy, and hierarchy #3 is the deepest hierarchy.

Hierarchy #1 stores simulation names, which are simulation identifiers. In the example of FIG. 9, simulation names such as "XXXX", "YYYY" and "ZZZZ" are stored. Hierarchy #2 stores the storage destination of the raw data of the simulation data corresponding to the simulation indicated in hierarchy #1 (“XXXX” in the example of FIG. 9). Specifically, in layer #2, the execution date and time of simulation "XXXX" is stored in association with storage destination path #1 of raw data of simulation data obtained by execution at that time. .

Hierarchy #3 stores data corresponding to the simulation shown in hierarchy #1 (“XXXX” in the example of FIG. 9). That is, the data shown in hierarchy #3 corresponds to the data in the database (data management unit 140). For example, in hierarchy #3, in the column of the item "data #1 (image)", "save destination path #X", "save destination path #Y", " Save destination path #Z” is stored. In the hierarchy #3, the character strings "ABCDEF", "DDEF", and "ADEF" are stored in the column of the item "data #2 (character string)". Further, in the hierarchy #3, numerical values "1234", "134", and "124" are stored in the column of the item "data #3 (numerical value)". Similarly, in hierarchy #3, numerical values are stored in columns of items “data #4 (numerical value)” and “data #5 (numerical value)”.

Here, in layer #3, each column of data in each row can be related to each other. For example, one row may correspond to data obtained in one simulation. For example, "data #2" ("ABCDEF" in the example of the first line) is stored in the directory specified by the save destination path in data #1 ("save destination path #X" in the example of the first line). may indicate the attributes of the image. Also, "data #3" ("1234" in the example of the first line) and "data #4" ("5678" in the example of the first line) are the save destination paths in data #1 ("1234" in the example of the first line). It may indicate the resolution of the image stored in the directory designated by "storage destination path #X"). Also, if the simulation data does not contain an image, the data #1 item may be blank.

Returning to FIG. 6, the simulation data acquisition unit 48 of the arithmetic device 40 requests the data management device 100 for simulation data of a data type suitable for the algorithm to be executed (step S160). For example, when the algorithm A execution unit 46A executes the algorithm A, the simulation data acquisition unit 48 instructs the data management device 100 to request simulation data of a data type conforming to the algorithm A (data request instruction). Send. The data request reception unit 150 of the data management device 100 receives the data request instruction transmitted by the simulation data acquisition unit 48 .

Here, the data request instruction may include information designating the data of hierarchy #3 (data #1 to #5 in the example of FIG. 9) shown in FIG. For example, if the data type compatible with Algorithm A is a numeric type, the data request indication sent when Algorithm A is executed specifies at least one of Data #3, #4, #5. If the data type compatible with Algorithm B is an image type, the data request indication sent when Algorithm B is executed specifies data #1. If the data type conforming to Algorithm C is a numeric type, the data request indication sent when Algorithm C is executed specifies at least one of Data #3, #4, #5. In this way, the computing device 40 can acquire simulation data at high speed by designating at least one of the data of hierarchy #3.

In response to the data request instruction, the data management device 100 transmits simulation data of a data type suitable for the algorithm to be executed to the arithmetic device 40 (step S162). Specifically, when data request receiving unit 150 receives a data request instruction, data transmitting unit 152 transmits the data of hierarchy #3 (data #1 to #5, etc.) specified in the received data request instruction to It transmits to the arithmetic unit 40 . Thereby, the simulation data acquisition unit 48 acquires simulation data of a data type suitable for the algorithm to be executed.

Arithmetic device 40 executes an algorithm using the obtained simulation data (step S170). Thereby, the algorithm execution unit 46 determines parameters for the simulation. A specific example of the processing of the algorithm execution unit 46 will be described later.

Then, the algorithm execution result transmission unit 50 of the arithmetic unit 40 transmits the execution result of the algorithm to the simulator 20 (step S180). Here, the algorithm execution result includes the determined parameters. The algorithm execution result receiving unit 30 of the simulator 20 receives the execution result of the algorithm. At this time, the algorithm execution result transmission unit 50 may transmit the execution result of the algorithm to the data management device 100 depending on the executed algorithm.

Then, the processing of S100 to S180 is repeated. At this time, the simulator 20 executes the simulation again using the parameters included in the received execution result of the algorithm (S100). A specific example will be described later.

As described above, the data management device 100 according to the first embodiment determines the data type of simulation data acquired from the simulator 20 and manages the simulation data. Thereby, the data management device 100 according to the first embodiment can automatically and centrally manage the data output from the simulator 20 . Therefore, other nodes (arithmetic device 40, etc.) can efficiently use the data managed by the data management device 100. FIG.

In particular, the computing device 40 according to the first embodiment is configured to efficiently determine parameters related to simulation by the simulator 20 using the simulation data managed as described above. Thereby, the computing device 40 can improve the accuracy of the simulation in the simulator 20 . In other words, in the simulation system 1 according to the first embodiment, the data management device 100 automatically centrally manages the data output from the simulator 20, thereby improving the accuracy of the simulation.

For example, assume that the data type compatible with Algorithm A is "numeric", and Algorithm A is configured to determine the parameters of Simulations A, B, and C. In this case, if the simulation data is not managed according to the data type, there is a possibility that the "image type" simulation data will be erroneously input to the algorithm A execution unit 46A. However, the algorithm A executor 46A cannot properly process "image type" simulation data. Therefore, the algorithm A executing section 46A cannot execute the algorithm A properly. Therefore, it is difficult to improve the accuracy of the simulations A, B, and C because the parameters of the simulations A, B, and C are not properly determined.

Also, for example, assume that the data type suitable for Algorithm B is "image type", and Algorithm B is configured to be able to determine the parameters of Simulations A, B, and C. FIG. In this case, if the simulation data is not managed according to the data type, there is a possibility that "character string type" simulation data will be erroneously input to the algorithm B execution unit 46B. However, the algorithm B execution unit 46B cannot properly process "character string type" simulation data. Therefore, the algorithm B executing section 46B cannot execute the algorithm B properly. Therefore, it is difficult to improve the accuracy of the simulations A, B, and C because the parameters of the simulations A, B, and C are not properly determined.

Further, for example, assume that the data type suitable for Algorithm C is "numerical type" and Algorithm C is configured to be able to determine the parameters of Simulations D and E. In this case, if the simulation data is not managed according to the data type, there is a possibility that the "image type" simulation data will be erroneously input to the algorithm C execution unit 46C. However, the algorithm C executor 46C cannot properly process "image type" simulation data. Therefore, the algorithm C execution unit 46C cannot execute algorithm C properly. Therefore, it is difficult to improve the accuracy of the simulations A, B, and C because the parameters of the simulations A, B, and C are not properly determined.

On the other hand, the data management device 100 according to the first embodiment determines the data type of the simulation data acquired from the simulator 20 and manages the simulation data, as described above. Therefore, arithmetic unit 40 can execute algorithms A, B, and C appropriately. In the above example, the algorithm A executing section 46A can execute the algorithm A by acquiring the “numerical type” simulation data. Algorithm B execution unit 46B can execute Algorithm B by acquiring “image type” simulation data. Algorithm C executing section 46C can execute Algorithm C by acquiring “numerical type” simulation data. Accordingly, computing device 40 is suitably configured to determine parameters for simulation by simulator 20 . Thereby, the computing device 40 can improve the accuracy of the simulation in the simulator 20 .

(Specific example of algorithm)
Next, a specific example of processing of the simulation system 1 when the arithmetic device 40 executes the search algorithm will be described.
Algorithm A executing section 46A executes a search algorithm as Algorithm A. Then let Algorithm A determine the parameters for Simulation A (structural analysis simulation), Simulation B (fluid simulation), or Simulation C (electromagnetic field simulation). In the following description, an example in which the algorithm A executing section 46A determines parameters relating to the simulation A (structural analysis simulation) will be described, but the simulations B and C are the same.

The simulator 20 executes a structural analysis simulation as simulation A (S100), and transmits simulation data to the data management device 100 (S112). The data management device 100 manages the simulation data according to the data type by the method described above (S120). Here, in layer #3 illustrated in FIG. 9, as parameters related to simulation A, an input value X1 that is an input (explanatory variable) of simulation A and an output value Y1 that is a simulation result (objective variable) are set according to the data type. can be managed by That is, simulation data includes input values and output values. Note that the input values for Simulation A may be design parameters. The output value of Simulation A can be intensity. Note that the input value and the output value may be in a vector format consisting of a plurality of components (dimensions) (the same applies to examples described later). Here, the data types of the input value and the output value are data types suitable for algorithm A (search algorithm).

When the algorithm A execution unit 46A executes the search algorithm, the simulation data acquisition unit 48 requests data corresponding to the simulation A from the data management device 100 (S160). The data management device 100 (data transmission unit 152) transmits the input value X1 and the output value Y1 to the calculation device 40 as parameters related to the simulation A (S162).

Then, the algorithm A execution unit 46A executes the search algorithm (S170). Specifically, the algorithm A execution unit 46A executes a search algorithm such as a meta-heuristics algorithm, and uses the input value X1 and the output value Y1 to obtain a more objective (for example, a higher strength) search algorithm. ) search for an input value X2 that yields an output. Then, the arithmetic unit 40 (algorithm execution result transmission unit 50) transmits the input value X2 as the algorithm execution result to the simulator 20 (S180). The simulator 20 performs structural analysis simulation again using the input value X2 as an input value (S100). In this way, the simulation A executing section 26A can efficiently search for target design parameters that satisfy the necessary strength conditions.

Next, a specific example of processing of the simulation system 1 when the computing device 40 executes the learning algorithm will be described.
Algorithm B executing section 46B executes a learning algorithm as Algorithm B. FIG. Algorithm B then determines the parameters for Simulation A (structural analysis simulation), Simulation B (fluid simulation), or Simulation C (electromagnetic field simulation). In the following description, an example in which the algorithm B executing section 46B determines parameters relating to simulation B (fluid simulation) will be described, but the same applies to simulations A and C as well.

The simulator 20 executes a fluid simulation as simulation B (S100). At this time, the function representing the model f1 used in the fluid simulation is assumed to be Y=f1(X). Here, the input value X is the input (explanatory variable) of the simulation B, and the output value Y is the simulation result (objective variable). Note that the input values for Simulation B may be design parameters. The output value of Simulation B can be flow velocity.

The simulation data is transmitted to the data management device 100 (S112). The data management device 100 manages the simulation data according to the data type by the method described above (S120). Here, in hierarchy #3 illustrated in FIG. 9, as parameters related to simulation B, the input value X and the output value Y can be managed according to the data type. That is, simulation data includes input values and output values. Here, the data types of the input value and the output value are data types suitable for algorithm B (learning algorithm). The simulator 20 then executes the simulation B multiple times and transmits the data required for the learning algorithm to the data management device 100 . In other words, the processes of S100 to S120 are repeated until enough data for executing the learning algorithm is obtained.

When the algorithm B execution unit 46B executes the learning algorithm, the simulation data acquisition unit 48 requests data corresponding to the simulation B from the data management device 100 (S160). The data management device 100 (data transmission unit 152) transmits a predetermined number of sets of input values X and output values Y to the calculation device 40 as parameters related to simulation B (S162).

Then, the algorithm B executing section 46B executes the learning algorithm (S170). Specifically, the algorithm B execution unit 46B executes a learning algorithm such as a neural network to determine a more suitable function (analytic formula) f2(X) from a large number of pairs of input values X and output values Y. do. Then, the arithmetic unit 40 (algorithm execution result transmission unit 50) transmits the function f2(X) as the algorithm execution result to the simulator 20 (S180). The simulator 20 performs fluid simulation again using the function f2(X) as a model (S100). In this way, the simulation B execution section 26B can efficiently search for the target design parameters that satisfy the necessary flow velocity conditions. Note that the learning algorithm may be combined with the search algorithm described above. This makes it possible to search for the target design parameters more efficiently.

Next, a specific example of processing of the simulation system 1 when the arithmetic unit 40 executes the data assimilation algorithm will be described.
Algorithm C executing section 46C executes a data assimilation algorithm as Algorithm C. FIG. Algorithm C then determines the parameters for Simulation D (weather forecast simulation) or Simulation E (production control simulation). In the following description, an example in which the algorithm C execution unit 46C determines parameters related to the simulation E (production control simulation) will be described, but the simulation D is the same. Algorithm C may determine parameters for Simulation A (structural analysis simulation), Simulation B (fluid simulation), or Simulation C (electromagnetic field simulation).

The simulator 20 executes a production management simulation as simulation E (S100). At this time, the function representing the model used in the production control simulation is simply Y=aX+b for clarity of explanation. Here, the input value X is the input (explanatory variable) of the simulation E, and the output value Y is the simulation result (objective variable). The input values for the simulation E may be design parameters such as the order of production processes or the number of facilities. The output value of the simulation E can be operating time, operating rate, or the like. In the simulation E, the model conceptually represents the relationship between the input value X and the output value Y for convenience of explanation. In this case, simulation E does not necessarily acquire the model explicitly. Also, the model is not necessarily linear and may be non-linear.

The simulation data is transmitted to the data management device 100 (S112). The data management device 100 manages the simulation data according to the data type by the method described above (S120). Here, in layer #3 illustrated in FIG. 9, as parameters related to simulation E, input value X and output value Y can be managed according to data types. That is, simulation data includes input values and output values. Here, the data types of the input value and the output value are data types suitable for Algorithm C (data assimilation algorithm). In this case, the data management device 100 can also acquire actual data (x, y) from the network edge 70 and manage it in the same manner as the simulation data. Note that the input value x and the output value y are actual data corresponding to the input value X and the output value Y, respectively.

When the algorithm C execution unit 46C executes the data assimilation algorithm to determine the parameters for the simulation E, the simulation data acquisition unit 48 requests data corresponding to the simulation E from the data management device 100 (S160). The data management device 100 (data transmission unit 152) transmits the simulation data (X, Y) to the calculation device 40 as parameters related to the simulation E (S162). Furthermore, the data management device 100 (data transmission unit 152) also transmits the actual data (x, y) to the arithmetic device 40.

Then, the algorithm C execution unit 46C executes the data assimilation algorithm (S170). Specifically, the algorithm C execution unit 46C executes a data assimilation algorithm using a Kalman filter or the like to more appropriately convert the actual data from the error between the simulation data (X, Y) and the actual data (x, y). Determine the parameters a and b of the function Y=aX+b to be reproduced. Then, the arithmetic unit 40 (algorithm execution result transmission unit 50) transmits the parameters a and b as the algorithm execution result to the simulator 20 (S180). The simulator 20 again performs a production management simulation using a model represented by a function using parameters a and b (S100). In this way, the simulation E executing section 26B can efficiently estimate the production control parameters that satisfy the necessary conditions.

(Modification)
It should be noted that the present invention is not limited to the above embodiments, and can be modified as appropriate without departing from the scope of the invention. For example, in the flowcharts described above, the order of each process (step) can be changed as appropriate. Also, one or more of a plurality of processes (steps) may be omitted.

For example, the processing of S124 to S128 shown in FIG. 7 may be omitted. In other words, the data type determination process (S130) may be performed regardless of the data size. Further, if the processing of S140 can be performed without the processing of S138 shown in FIG. 8, the processing of S138 may be omitted.

Also, the data removal device 60, the network edge 70 and the data monitoring device 80 of the devices shown in FIG. 2 may be omitted. On the other hand, a device that uses data managed by data management device 100 is not limited to arithmetic device 40 . For example, the simulator 20 may use data managed by the data management device 100 . Alternatively, network edge 70 or data monitoring device 80 may use data managed by data management device 100 .

Moreover, at least two of the devices constituting the simulation system 1 may be configured by the same device. In particular, the simulator 20 and the computing device 40 may be implemented as a physically integrated device. On the other hand, at least one of the devices that constitute the simulation system 1 may be composed of a plurality of devices.

Further, in the above-described embodiment, the data type determination unit 120 determines the data type of simulation data by the method illustrated in FIG. 8, but the configuration is not limited to this. For example, the data type determination unit 120 may determine the data type using some machine learning algorithm or the like, which is different from the processing illustrated in FIG.

In the above examples, the programs can be stored and delivered to computers using various types of non-transitory computer readable media. Non-transitory computer-readable media include various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic recording media (eg, flexible discs, magnetic tapes, hard disk drives), magneto-optical recording media (eg, magneto-optical discs), CD-ROMs (Read Only Memory), CD-Rs, CD-R/W, semiconductor memory (eg, mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, RAM (Random Access Memory)). The program may also be delivered to the computer on various types of transitory computer readable medium. Examples of transitory computer-readable media include electrical signals, optical signals, and electromagnetic waves. Transitory computer-readable media can deliver the program to the computer via wired channels, such as wires and optical fibers, or wireless channels.

Some or all of the above-described embodiments can also be described in the following supplementary remarks, but are not limited to the following.
(Appendix 1)
data acquisition means for acquiring first data relating to the simulation from at least one simulator;
data type determination means for determining a data type of the first data based on the character string related to the first data acquired by the data acquisition means;
and data management means for managing the first data by a management method according to the data type determined by the data type determination means.
(Appendix 2)
At least one computing device that determines parameters relating to the simulation executed by the simulator using a predetermined algorithm, of the first data managed by the data management means, of a data type suitable for the algorithm 1. The data management device according to appendix 1, further comprising transmission means for transmitting the first data.
(Appendix 3)
3. The data management device according to appendix 1 or 2, wherein the data type determination means uses an extension of the first data to determine the data type of the first data.
(Appendix 4)
The data type determination means determines the data type of the first data according to whether or not a character string constituting the first data can be treated as a numerical value. data management equipment.
(Appendix 5)
data size determination means for determining whether the size of the first data is equal to or greater than a predetermined threshold;
further comprising metadata extracting means for extracting metadata of the first data when the size of the first data is not equal to or greater than the threshold;
The data management device according to any one of attachments 1 to 4, wherein the data management means manages the first data whose size is not equal to or larger than the threshold by storing the metadata.
(Appendix 6)
at least one simulator for performing simulations;
a data management device that manages first data relating to simulation by the simulator;
a computing device for determining parameters related to the simulation;
The data management device
data acquisition means for acquiring the first data from the simulator;
data type determination means for determining a data type of the first data based on the character string related to the first data acquired by the data acquisition means;
data management means for managing the first data by a management method according to the data type determined by the data type determination means;
The computing device determines the parameter using the first data managed by the data management means,
A simulation system in which the simulator executes the simulation again using the parameters determined by the computing device.
(Appendix 7)
The simulation system according to appendix 6, wherein the data type determination means uses an extension of the first data to determine the data type of the first data.
(Appendix 8)
8. The simulation system according to appendix 6 or 7, wherein the data type determination means determines the data type of the first data according to whether or not a character string forming the first data can be treated as a numerical value.
(Appendix 9)
data size determination means for determining whether the size of the first data is equal to or greater than a predetermined threshold;
further comprising metadata extracting means for extracting metadata of the first data when the size of the first data is not equal to or greater than the threshold;
The simulation system according to any one of Appendices 6 to 8, wherein the data management means manages the first data whose size is not equal to or larger than the threshold by storing the metadata.
(Appendix 10)
obtaining first data about the simulation from at least one simulator;
Determining a data type of the first data based on the acquired character string related to the first data;
A data management method for managing the first data by a management method according to the determined data type.
(Appendix 11)
At least one computing device that determines parameters relating to a simulation executed on the simulator using a predetermined algorithm stores the first data of a data type suitable for the algorithm among the managed first data. 11. The data management method according to appendix 10, wherein the data is transmitted.
(Appendix 12)
12. The data management method according to appendix 10 or 11, wherein the extension of the first data is used to determine the data type of the first data.
(Appendix 13)
13. The data management method according to any one of Appendices 10 to 12, wherein a data type of the first data is determined according to whether or not a character string forming the first data can be treated as a numerical value.
(Appendix 14)
Determining whether the size of the first data is equal to or greater than a predetermined threshold,
extracting metadata of the first data if the size of the first data is not greater than or equal to the threshold;
14. The data management method according to any one of appendices 10 to 13, wherein the first data whose size is not equal to or larger than the threshold is managed by storing the metadata.
(Appendix 15)
obtaining first data about the simulation from at least one simulator;
determining a data type of the first data based on the obtained character string related to the first data;
A program for causing a computer to execute a step of managing the first data by a management method according to the determined data type.

1 simulation system 10 data management device 12 data acquisition unit 14 data type determination unit 16 data management unit 20 simulator 26 simulation execution unit 28 simulation data transmission unit 30 algorithm execution result reception unit 40 arithmetic unit 46 algorithm execution unit 48 simulation data acquisition unit 50 Algorithm execution result transmission unit 100 Data management device 110 Data storage unit 112 Data acquisition unit 114 Data size determination unit 116 Metadata extraction unit 120 Data type determination unit 130 Management method determination unit 140 Data management unit 150 Data request reception unit 152 Data transmission unit

Claims (9)

data acquisition means for acquiring first data relating to the simulation from at least one simulator;
data type determination means for determining a data type of the first data based on the character string related to the first data acquired by the data acquisition means;
data management means for managing the first data by a management method according to the data type determined by the data type determination means ;
At least one computing device that determines parameters relating to the simulation executed by the simulator using a predetermined algorithm, of the first data managed by the data management means, of a data type suitable for the algorithm a transmitting means for transmitting the first data;
A data management device having data acquisition means for acquiring first data relating to the simulation from at least one simulator;
data type determination means for determining a data type of the first data based on the character string related to the first data acquired by the data acquisition means;
data management means for managing the first data by a management method according to the data type determined by the data type determination means;
data size determination means for determining whether the size of the first data is equal to or greater than a predetermined threshold;
metadata extraction means for extracting metadata of the first data when the size of the first data is not equal to or greater than the threshold value;
has
The data management means manages the first data whose size is not equal to or larger than the threshold by storing the metadata.
Data management device. 3. The data management device according to claim 1, wherein said data type determining means determines the data type of said first data using an extension of said first data. 4. The data type determining means according to any one of claims 1 to 3, wherein the data type determining means determines the data type of the first data according to whether or not a character string forming the first data can be treated as a numerical value. Data management device as described. at least one simulator for performing simulations;
a data management device that manages first data relating to simulation by the simulator;
a computing device for determining parameters related to the simulation;
The data management device
data acquisition means for acquiring the first data from the simulator;
data type determination means for determining a data type of the first data based on the character string related to the first data acquired by the data acquisition means;
data management means for managing the first data by a management method according to the data type determined by the data type determination means;
The computing device determines the parameter using the first data managed by the data management means,
A simulation system in which the simulator executes the simulation again using the parameters determined by the computing device. obtaining first data about the simulation from at least one simulator;
Determining a data type of the first data based on the acquired character string related to the first data;
managing the first data by a management method according to the determined data type ;
At least one computing device that determines parameters relating to a simulation executed on the simulator using a predetermined algorithm stores the first data of a data type suitable for the algorithm among the managed first data. send data
Data management method. obtaining first data about the simulation from at least one simulator;
Determining whether the size of the acquired first data is equal to or greater than a predetermined threshold,
When the size of the first data is not equal to or greater than the threshold,
extracting metadata for the first data;
managing the first data whose size is not equal to or larger than the threshold by storing the metadata;
When the size of the first data is equal to or greater than the threshold,
Determining a data type of the first data based on a character string related to the first data;
managing the first data by a management method according to the determined data type
Data management method. obtaining first data about the simulation from at least one simulator;
determining a data type of the first data based on the acquired character string related to the first data;
managing the first data by a management method according to the determined data type ;
At least one computing device that determines parameters relating to a simulation executed on the simulator using a predetermined algorithm stores the first data of a data type suitable for the algorithm among the managed first data. sending data; and
A program that makes a computer run obtaining first data about the simulation from at least one simulator;
determining whether the size of the obtained first data is equal to or greater than a predetermined threshold;
When the size of the first data is less than or equal to the threshold, metadata of the first data is extracted and stored to manage the first data whose size is less than or equal to the threshold. a step;
determining a data type of the first data based on the acquired character string related to the first data when the size of the first data is equal to or larger than the threshold, and determining the determined data type managing the first data in a management method according to
A program that makes a computer run

Images