WO2018142507A1

WO2018142507A1 - Simulation method, system, and program

Info

Publication number: WO2018142507A1
Application number: PCT/JP2017/003602
Authority: WO
Inventors: 正啓間瀬; 宇▲しん▼ 梁; 恵木　正史
Original assignee: 株式会社日立製作所
Priority date: 2017-02-01
Filing date: 2017-02-01
Publication date: 2018-08-09
Also published as: JPWO2018142507A1; JP6726312B2

Abstract

For a simulation method in a system which executes simulations with different granularities, the system includes: a step for generating second data, which is used in a second simulation, by using first data which is used in a first simulation; a step for executing the first simulation and outputting a first simulation result on the basis of the first data and a model of the first simulation; a step for executing the second simulation and outputting a second simulation result on the basis of the second data and a model of the second simulation; and a step for calibrating at least any one among the first simulation model and the second simulation model on the basis of the first simulation result and the second simulation result.

Description

Simulation method, system, and program

The present invention relates to a method for correcting a simulation model.

In various fields such as transportation, finance, distribution, and industry, with the progress of IoT, analysis and visualization of the current situation and statistical prediction using statistical analysis processing or agent simulation using data collected from devices Etc. are done.

The simulation model, which is the definition information for simulation, depends on the type and characteristics of the data to be acquired. Therefore, it is necessary to correct the model according to changes in purpose and environment. In particular, in a simulation that requires real-time analysis, it is necessary to dynamically correct the simulation model according to environmental changes.

The technique of Patent Document 1 is known as a technique for correcting a simulation model. Patent Document 1 includes a “simulation model storage unit that stores a simulation model, a simulation unit that receives the simulation model and performs simulation, a real world statistical information storage unit that stores statistical information in the real world, and the simulation. A simulation model correction method creation unit for creating a simulation model correction method based on the model and the statistical information, a simulation model correction unit for correcting the simulation model based on the simulation model correction method and the simulation result, Is provided. "

US Pat. No. 6,772,363

In the technique described in Patent Document 1, since the simulation model is corrected so that the statistical information and the simulation result coincide with each other, the technique cannot be applied to a real-time simulation and a simulation considering future prediction.

The present invention aims to provide a method, a system, and a program for correcting a simulation model based on a current situation or a future prediction.

A typical example of the invention disclosed in the present application is as follows. That is, a simulation method in a system for executing a simulation with different granularities, the system including at least one computer having a processor, a memory connected to the processor, and a network interface connected to the processor, At least one computer includes model management information for managing a model, which is definition information of a first simulation and a second simulation having a larger granularity than the first simulation, and a first simulation used for the first simulation. History management information for managing a history of data, and the simulation method uses the first data acquired from the history management information by the at least one computer for the second simulation. Second A first step of generating data; and the at least one computer executes the second simulation based on the second simulation and the model of the second simulation acquired from the model management information. A second step of outputting a second simulation result, and the at least one computer includes the first data acquired from the history management information and the model of the first simulation acquired from the model management information. And a third step of executing the first simulation and outputting a first simulation result, and the at least one computer is based on the first simulation result and the second simulation result. , The first simulation model and the second simulation model Characterized in that it comprises a fourth step of correcting at least one of emissions model, a.

According to the present invention, a real-time simulation and a simulation considering future prediction can be realized. Problems, configurations, and effects other than those described above will become apparent from the following description of embodiments.

1 is a diagram illustrating a hardware configuration of a simulation system according to Embodiment 1. FIG. 3 is a diagram illustrating a software configuration of the simulation system according to the first embodiment. It is a figure which shows operation | movement of the simulation system in the correction process of the simulation model of Example 1. FIG. 6 is a flowchart illustrating an example of processing executed by an integrated processing server according to the first embodiment. 6 is a flowchart illustrating an example of processing executed by a micro simulation execution server according to the first embodiment. 6 is a flowchart illustrating an example of processing executed by a micro simulation execution server according to the first embodiment. 6 is a flowchart illustrating an example of simulation model correction processing executed by the integrated processing server according to the first embodiment. It is a figure which shows an example of the history management information of Example 2. It is a figure which shows an example of the space definition information of Example 2. It is a figure which shows an example of the observation data of Example 2. It is a figure which shows an example of the statistical data of Example 2. It is a figure which shows an example of the prediction data of Example 2. It is a figure which shows an example of the prediction data of Example 2. FIG. 10 is a diagram illustrating an example of a statistic calculated from a processing result of a micro simulation according to the second embodiment. FIG. 10 is a diagram illustrating an example of a statistic calculated from a processing result of a micro simulation according to the second embodiment. It is a figure which shows an example of the internal information which the macro simulation execution server of Example 2 outputs. It is a figure which shows an example of the space definition information of Example 3.

Embodiments according to the present invention will be described below with reference to the accompanying drawings. In each figure, the same reference numerals are given to common configurations.

FIG. 1 is a diagram illustrating a hardware configuration of the simulation system according to the first embodiment.

The simulation system shown in FIG. 1 includes an integrated processing server 100, a database server 110, a macro simulation execution server 120, a micro simulation execution server 130, and a user terminal 140. Each device is connected via a network 190. In addition, as a kind of the network 190, WAN (Wide Area Network), LAN (Local Area Network), etc. can be considered. The connection method of the network 190 may be either wired or wireless.

The integrated processing server 100 corrects the simulation model which is the simulation definition information. The simulation model includes parameters that characterize the simulation, variables that are substituted with input values, and the like. The integrated processing server 100 includes a CPU 101, a memory 102, a storage device 103, and a network interface 104 as hardware. Each hardware is connected via an internal bus or the like.

CPU 101 executes a program stored in memory 102. When the CPU 101 executes processing in accordance with a program, it operates as a module having a predetermined function. In the following description, when a process is described with a module as a subject, it indicates that the CPU 101 is executing a program that realizes the module.

The memory 102 stores a program executed by the CPU 101 and information necessary for the program. The memory 102 also includes a work area that is temporarily used by the program.

Storage device 103 stores data permanently. The storage device 103 may be a storage medium such as HDD (Hard Disk Drive) and SSD (Solid State Drive), or a non-volatile memory. Note that the program and information stored in the memory 102 may be stored in the storage device 103. In this case, the CPU 101 reads out the program and information from the storage device 103, loads the program and information into the memory 102, and executes the program loaded into the memory 102.

The network interface 104 connects to other devices via a network.

The database server 110 manages various data. The database server 110 includes a CPU 111, a memory 112, a storage device 113, and a network interface 114 as hardware. Each hardware is connected via an internal bus or the like.

The CPU 111, the memory 112, the storage device 113, and the network interface 114 are the same hardware as the CPU 101, the memory 102, the storage device 103, and the network interface 104.

The macro simulation execution server 120 executes macro simulation. For example, as a macro simulation related to traffic, a simulation for predicting traffic volume, traffic jams, and the like can be considered. The macro simulation execution server 120 includes a CPU 121, a memory 122, a storage device 123, and a network interface 124 as hardware. Each hardware is connected via an internal bus or the like.

The CPU 121, the memory 122, the storage device 123, and the network interface 124 are the same hardware as the CPU 101, the memory 102, the storage device 103, and the network interface 104.

The micro simulation execution server 130 executes a micro simulation having a smaller granularity than the macro simulation. For example, as a micro simulation related to traffic, a simulation for predicting the optimum route of each vehicle can be considered. The micro simulation execution server 130 includes a CPU 131, a memory 132, a storage device 133, and a network interface 134 as hardware. Each hardware is connected via an internal bus or the like.

The CPU 131, the memory 132, the storage device 133, and the network interface 134 are the same hardware as the CPU 101, the memory 102, the storage device 103, and the network interface 104.

The user terminal 140 is a terminal used by the user. The user uses the user terminal 140 to instruct execution of simulation and correction of the simulation model. The user terminal 140 includes a CPU 141, a memory 142, a storage device 143, a network interface 144, an input device 145, and an output device 146 as hardware. Each hardware is connected via an internal bus or the like.

The CPU 141, the memory 142, the storage device 143, and the network interface 144 are hardware similar to the CPU 101, the memory 102, the storage device 103, and the network interface 104.

The input device 145 is a device for inputting data and the like, and includes a keyboard, a mouse, a touch panel, and the like.

The output device 146 is a device for outputting data and the like, and includes a display and a touch panel.

FIG. 2 is a diagram illustrating a software configuration of the simulation system according to the first embodiment.

The database server 110 holds simulation model information 211, history management information 212, and space definition information 213.

The simulation model information 211 is information for managing the simulation model. The simulation model information 211 stores one or more simulation models for macro simulation and one or more simulation models for micro simulation.

The history management information 212 stores various data acquired from the agent in the micro simulation.

The space definition information 213 is information for generating data (statistical data) used for the macro simulation from the data stored in the history management information 212. For example, in the micro simulation for analyzing the movement of the moving body, information such as a space in which the moving body moves and a time range of data to be acquired is stored in the space definition information 213.

The integrated processing server 100 holds programs for realizing the preprocessing module 201, the statistical data generation module 202, and the model correction module 203.

The preprocessing module 201 generates observation data 301 (see FIG. 3) used for micro simulation from the history management information 212.

The statistical data generation module 202 generates statistical data 303 (see FIG. 3) used for the micro simulation based on the observation data 301 and the space definition information 213.

The model correction module 203 corrects the simulation model based on the processing result of the macro simulation and the processing result of the micro simulation.

The macro simulation execution server 120 holds a program that realizes the macro simulation execution module 221 that executes the macro simulation.

The micro simulation execution server 130 holds a program for realizing the micro simulation execution module 231 for executing the micro simulation.

The user terminal 140 holds a program that realizes the application 241 that performs operations on the integrated processing server 100 and the like.

FIG. 3 is a diagram illustrating the operation of the simulation system in the correction process of the simulation model of the first embodiment.

The integrated processing server 100 calls the preprocessing module 201 when receiving a simulation model correction request from the application 241 of the user terminal 140.

The preprocessing module 201 accesses the database server 110 to acquire the space definition information 213, and acquires data in a predetermined time range from the history management information 212. The preprocessing module 201 performs processing such as data conversion processing and data extraction processing using data acquired from the space definition information 213 and the history management information 212 to generate observation data 301.

Next, the integrated processing server 100 calls the statistical data generation module 202. At this time, the observation data 301 and the space definition information 213 are input to the statistical data generation module 202.

The statistical data generation module 202 generates statistical data 303 using the observation data 301 and the space definition information 213.

Next, the integrated processing server 100 transmits a macro simulation execution instruction including the statistical data 303 to the macro simulation execution server 120, and transmits a micro simulation execution instruction including the observation data 301 to the micro simulation execution server 130. .

The macro simulation execution server 120 calls the macro simulation execution module 221 when receiving an execution instruction.

The macro simulation execution module 221 accesses the database server 110 and acquires a simulation model of macro simulation from the simulation model information 211. The macro simulation execution module 221 executes the macro simulation using the simulation model and the statistical data 303 and transmits the processing result as the prediction data 304 to the integrated processing server 100.

The micro simulation execution server 130 calls the micro simulation execution module 231 when an execution instruction is received.

The micro simulation execution module 231 accesses the database server 110 and acquires a simulation model for micro simulation from the simulation model information 211. The micro simulation execution module 231 executes the micro simulation using the simulation model and the observation data 301 and transmits the processing result as the prediction data 302 to the integrated processing server 100.

The integrated processing server 100 calls the model correction module 203 when receiving the

prediction data

302 and 304 from the macro simulation execution server 120 and the micro simulation execution server 130. At this time, the

prediction data

302 and 304 are input to the model correction module 203.

The model correction module 203 accesses the database server 110 and acquires a simulation model for macro simulation and a simulation model for micro simulation from the simulation model information 211. The model correction module 203 corrects at least one of the simulation model for micro simulation and the simulation model for macro simulation using the

prediction data

302 and 304. In this embodiment, it is assumed that the simulation model of micro simulation is corrected. The model correction module 203 registers the corrected simulation model of the micro simulation in the simulation model information 211 as the simulation model of the micro simulation to be executed.

As described above, the system of the present embodiment executes a micro simulation, executes a macro simulation using the statistical data 303 generated from data used for the micro simulation, and sets a simulation model so that each processing result is consistent. to correct.

Conventionally, since the simulation model was corrected so as to match the history, the simulation corresponding to the state change could not always be realized. On the other hand, in the present embodiment, the results of the simulations are compared, and the simulation models are corrected based on the comparison results. Since the simulation is a process for predicting various situation changes, the simulation corresponding to the state change such as the environmental change can be performed by performing the correction as in the present embodiment.

FIG. 4 is a flowchart illustrating an example of processing executed by the integrated processing server 100 according to the first embodiment.

The integrated processing server 100 receives a simulation model correction request from the user terminal 140 (step S401).

Next, the integrated processing server 100 acquires the space definition information 213 and acquires data in a predetermined time range from the history management information 212 (step S402).

Specifically, the preprocessing module 201 accesses the database server 110 and acquires data in a predetermined time range from the history management information 212. The time range may be specified by the user or may be set in advance. Further, the preprocessing module 201 may determine a time range with reference to a simulation model of macro simulation.

Next, the integrated processing server 100 generates observation data 301 (step S403).

Specifically, the preprocessing module 201 generates observation data 301 using data acquired from the space definition information 213 and the history management information 212.

Next, the integrated processing server 100 generates statistical data 303 (step S404).

Specifically, the statistical data generation module 202 generates the statistical data 303 using the space definition information 213 and the observation data 301.

Next, the integrated processing server 100 transmits a macro simulation execution instruction including the statistical data 303 to the macro simulation execution server 120 (step S405).

Next, the integrated processing server 100 receives the prediction data 304 from the macro simulation execution server 120 (step S406). At this time, the integrated processing server 100 sets the received prediction data 304 as an index for determining whether or not the simulation model needs to be corrected.

Next, the integrated processing server 100 transmits a micro simulation execution instruction including the observation data 301 to the micro simulation execution server 130 (step S407). Note that the integrated processing server 100 may transmit a micro simulation execution instruction before receiving the prediction data 304.

Next, the integrated processing server 100 executes a simulation model correction process (step S408). Details of the correction process of the simulation model will be described with reference to FIG.

In addition, a present Example is not limited to the processing content of a macro simulation and a micro simulation.

5A and 5B are flowcharts illustrating an example of processing executed by the micro simulation execution server 130 according to the first embodiment.

FIG. 5A shows processing when there is one simulation model for micro simulation stored in the simulation model information 211.

When receiving the execution instruction, the micro simulation execution module 231 accesses the database server 110 and acquires a simulation model from the simulation model information 211 (step S501).

The micro simulation execution module 231 executes a micro simulation using the acquired simulation model and the observation data 301 included in the execution instruction (step S502).

The micro simulation execution module 231 transmits the prediction data 302 that is the processing result of the micro simulation to the integrated processing server 100 (step S503) and transmits a completion notification (step S504).

FIG. 5B shows processing when there are two or more simulation models for micro simulation stored in the simulation model information 211.

When the micro simulation execution module 231 receives the execution instruction, the micro simulation execution module 231 accesses the database server 110 and acquires simulation models of all the micro simulations from the simulation model information 211 (step S511).

The micro simulation execution module 231 selects one target simulation model from among a plurality of simulation models (step S512).

The micro simulation execution module 231 executes the micro simulation using the observation data 301 included in the target simulation model and the execution instruction (step S513).

The micro simulation execution module 231 transmits the prediction data 302 to the integrated processing server 100 (step S514). At this time, the micro simulation execution module 231 transmits the prediction data 302 together with the identification information of the target simulation model.

The micro simulation execution module 231 determines whether or not the processing has been completed for all the acquired simulation models (step S515).

If it is determined that the processing has not been completed for all the acquired simulation models, the micro simulation execution module 231 returns to step S512 and executes the same processing.

When it is determined that the processing has been completed for all the acquired simulation models, the micro simulation execution module 231 transmits a completion notification to the integrated processing server 100 (step S516).

Note that the micro simulation execution module 231 may present internal information indicating the state of the agent after simulation in response to a request from the integrated processing server 100 or the user terminal 140.

FIG. 6 is a flowchart illustrating an example of simulation model correction processing executed by the integrated processing server 100 according to the first embodiment.

The integrated processing server 100 calls the model correction module 203 when receiving a response to the execution instruction from the micro simulation execution server 130.

The model correction module 203 determines whether or not the received response is a completion notification (step S601).

When it is determined that the received response is the prediction data 302, the model correction module 203 stores the received prediction data 302 in the work area of the memory 102 (step S602). After that, the model correction module 203 shifts to a waiting state and returns to step S601 when a response is received.

When the prediction data 302 and the simulation model identification information are included, the model correction module 203 associates the prediction data 302 and the simulation model identification information and stores them in the work area of the memory 102.

If it is determined that the received response is a completion notification, the model correction module 203 determines whether or not the prediction data 302 has been received a plurality of times (step S603).

Specifically, the model correction module 203 refers to the work area of the memory 102 and determines whether or not two or more prediction data 302 are stored.

When it is determined that the prediction data 302 has been received a plurality of times, the model correction module 203 calculates a statistic of each simulation model (step S604).

Specifically, the model correction module 203 selects one prediction data 302 and uses the prediction data 302 to calculate a statistic that can be compared with the processing result of the macro simulation. Assume that the calculation method of the statistic is set in advance. Note that the statistic calculation method may be managed in association with the simulation model. The model correction module 203 executes the same process for all prediction data 302 stored in the work area of the memory 102.

Next, the model correction module 203 calculates a difference between the statistic and the determination index of each simulation model, and selects a simulation model to be actually used based on the difference (step S605). At this time, the model correction module 203 gives a flag indicating that the selected simulation model is a simulation model to be used. Thereafter, the model correction module 203 ends the simulation model correction processing.

The micro simulation execution module 231 executes the micro simulation using the simulation model to which the flag is assigned when executing the micro simulation for the purpose of normal analysis.

Specifically, the model correction module 203 calculates a difference between the statistic and the determination index, and selects a simulation model corresponding to the statistic having the smallest difference.

If it is determined in step S603 that the prediction data 302 has not been received a plurality of times, the model correction module 203 calculates a statistic of the simulation model (step S606). Note that the method for calculating the statistic is the same as that in step S604.

Next, the model correction module 203 determines whether or not the difference between the statistics of the simulation model and the determination index is equal to or less than the threshold value m (step S607).

When it is determined that the difference is equal to or less than the threshold value m, the model correction module 203 ends the simulation model correction process.

If it is determined that the difference is larger than the threshold value m, the model correction module 203 corrects the simulation model (step S608), and registers the corrected simulation model in the simulation model information 211. For example, the model correction module 203 can be a method of correcting parameters included in the simulation model.

The micro simulation execution module 231 executes the micro simulation using the corrected simulation model when executing the micro simulation for the purpose of normal analysis.

Next, the model correction module 203 transmits a micro simulation execution instruction to the micro simulation execution server 130 (step S609). Thereafter, the model correction module 203 ends the simulation model correction processing. Therefore, the simulation model is corrected until the difference between the statistic of the simulation model and the determination index is equal to or less than the threshold value m.

As described above, in the first embodiment, the simulation model of the micro simulation is corrected so as to match the processing result of the macro simulation.

For example, by generating statistical data using data between a predetermined time and the current time, it is possible to correct to a simulation model reflecting a global state change or the like. Further, it is not necessary to prepare statistical data used for macro simulation in advance.

In the first embodiment, the example of correcting the simulation model of the micro simulation is shown. However, the simulation model of the macro simulation may be corrected.

In this embodiment, each server is realized by using a physical computer, but each server may be realized by using a virtualization technology. Further, the functions possessed by each server may be integrated into one computer. For example, the integrated processing server 100, the macro simulation execution server 120, and the micro simulation execution server 130 may be integrated into one computer. Further, the integrated processing server 100 and the database server 110 may be integrated into one computer.

In Example 2, the simulation system of Example 1 will be described using specific values. The simulation system of the second embodiment simulates the operation of a ship. Since the system configuration and software configuration have been described in the first embodiment, a description thereof will be omitted. The agent of the micro simulation of the second embodiment is a ship.

FIG. 7 is a diagram illustrating an example of the history management information 212 according to the second embodiment.

The history management information 212 stores ship state information. Specifically, the history management information 212 includes an entry including a time 701, a moving body ID 702, a longitude 703, a latitude 704, and a destination 705.

Time 701 is the time when the state information is acquired. The moving body ID 702 is identification information for uniquely identifying a ship. The longitude 703 and the latitude 704 are information indicating the detailed position of the ship. The destination 705 is information indicating a place where the ship is to go.

FIG. 8 is a diagram illustrating an example of the space definition information 213 according to the second embodiment.

The space definition information 213 includes information 800 that defines the space (XY plane) in which the ship moves and information 810 that defines the history time.

The information 800 includes an entry composed of a space dimension 801 and a space mesh 802. The space dimension 801 is identification information of coordinates that define a space. The space mesh 802 is a unit for dividing the space in the coordinate direction.

Information 810 includes an entry composed of a time dimension 811 and a time step 812. The time dimension 811 is information indicating a unit time for acquiring a history. The time step 812 is information indicating a time unit for collecting state information.

FIG. 9 is a diagram illustrating an example of the observation data 301 of the second embodiment.

The observation data 301 of Example 2 is data used by the micro simulation. Specifically, the input data includes an entry including a time 901, a mobile object ID 902, a position (X) 903, a position (Y) 904, a speed (X) 905, a speed (Y) 906, and a destination 907. Including.

Time 901 is the same as time 701. The mobile object ID 902 is the same as the mobile object ID 702. A position (X) 903 indicates the position of the X axis of the ship in space. A position (Y) 904 indicates the position of the Y axis of the ship in space. Speed (X) 905 indicates the speed of the ship in the X-axis direction. Speed (Y) 906 indicates the speed of the ship in the Y-axis direction. The destination 907 is the same as the destination 705.

FIG. 10 is a diagram illustrating an example of the statistical data 303 of the second embodiment.

The statistical data 303 of Example 2 indicates the density of ships in space at a certain time. Specifically, the statistical data 303 includes an entry including a time 1001, a spot ID 1002, and a density 1003.

Time 1001 indicates a time that is a starting point for generating statistical data. The point ID 1002 is identification information of a certain point (region) in space. The density 1003 is the density of the ship at the point corresponding to the point ID 1002.

FIG. 11 is a diagram illustrating an example of the prediction data 304 of the second embodiment.

The prediction data 304 of the second embodiment has the same data structure as the statistical data 303. The prediction data 304 stores information indicating the density of each point after the macro simulation is executed. Note that the time after the macro simulation is set as the time 1101.

In Example 2, the density of each point is set as a determination index.

FIG. 12 is a diagram illustrating an example of the prediction data 302 according to the second embodiment.

In the prediction data 302 of the second embodiment, information indicating the state of each ship after the micro simulation is executed is stored. Specifically, the prediction data 302 includes an entry including a time 1201, a mobile object ID 1202, a position (X) 1203, a position (Y) 1204, and a destination 1205.

Time 1201 is the time after the micro simulation. The mobile object ID 1202 and the destination 1205 are the same as the mobile object ID 902 and the destination 907. A position (X) 1203 and a position (Y) 1204 are the positions of the X axis and the Y axis of the ship after the micro simulation.

FIG. 13A and FIG. 13B are diagrams illustrating an example of statistics calculated from the processing results of the micro simulation of the second embodiment.

In Example 2, based on the prediction data 302, the density of the ships at each point is calculated as a statistic. The statistic includes an entry including a time 1301, a spot ID 1302, a density 1303, a determination index 1304, and a difference 1305.

Time 1301, point 1302, and density 1303 are the same as time 1001, point ID 1002, and density 1003. The determination index 1304 is the same as the density 1103. The difference 1305 is the difference between the density 1303 and the determination index 1304.

FIG. 14 is a diagram illustrating an example of internal information output by the macro simulation execution server 120 according to the second embodiment.

The internal information stores information indicating the state of the ship after the micro simulation. Specifically, the internal information includes an entry including a time 1401, a mobile object ID 1402, a speed 1403, and a traveling direction 1404.

Time 1401 is the time after simulation. The mobile object ID 1402 is the same as the mobile object ID 702. The speed 1403 is an absolute value of the speed of the ship. The traveling direction 1404 is the traveling direction of the ship.

Here, the processing flow of the second embodiment will be described with reference to FIG.

The pre-processing module 201 generates observation data 301 shown in FIG. 9 using the history management information 212 shown in FIG. 7 and the space definition information 213 shown in FIG.

The statistical data generation module 202 generates the statistical data 303 shown in FIG. 10 using the space definition information 213 shown in FIG. 8 and the observation data 301 shown in FIG.

The macro simulation execution module 221 executes a macro simulation using the statistical data 303 shown in FIG. 10 and calculates the density at each point after 5 minutes. Specifically, prediction data 304 shown in FIG. 11 is output.

The micro simulation execution module 231 executes the micro simulation using the observation data 301 shown in FIG. 9 and calculates the position of each ship after 5 minutes. Specifically, prediction data 302 shown in FIG. 12 is output.

Here, it is assumed that the micro simulation shown in FIG. 5A has been executed.

The model correction module 203 sets the prediction data 304 as a determination index for each point, and calculates a statistic using the prediction data 302. The model correction module 203 calculates the difference between the determination index and the statistic for each point. As the above processing results, results as shown in FIGS. 13A and 13B are output.

The model correction module 203 determines whether or not the total difference of the points is equal to or less than the threshold value m. When it is determined that the total difference of the points is larger than the threshold value m, the simulation model is corrected.

As described above, by correcting the simulation model of the micro simulation so that the processing result of the macro simulation and the processing result of the micro simulation match, it is possible to perform a simulation corresponding to a future state change.

In Example 3, the simulation system of Example 1 will be described using specific values. The simulation system according to the third embodiment simulates sales of merchandise. Since the system configuration and software configuration have been described in the first embodiment, a description thereof will be omitted. The agent of the micro simulation of the third embodiment is a customer. Accordingly, the history management information 212 stores purchase histories of customer products in chronological order.

The pre-processing module 201 uses the history management information 212 and the space definition information 213, and includes observation data including entries composed of fields indicating purchase product identification information, the number of purchases, the purchase amount, and presence / absence of advertisement recognition, etc. 301 is generated.

Here, the space definition information 213 according to the third embodiment will be described. FIG. 15 is a diagram illustrating an example of the space definition information 213 according to the third embodiment.

The space definition information 213 of the third embodiment has the same data structure as the space definition information 213 of the second embodiment. In Example 3, a product space is assumed instead of a physical space. The space includes an axis indicating a product category and an axis indicating a product grade. The space mesh 802 stores information indicating the division of the product category axis and information indicating the division of the product category axis.

Further, as shown in the information 810, the history management information 212 stores a purchase history in units of dates, and indicates that the statistical data is generated by aggregating purchase histories in units of weeks.

Returning to the explanation of FIG.

The statistical data generation module 202 generates statistical data 303 using the observation data 301 and the space definition information 213 shown in FIG. The statistical data includes an entry including a group of fields indicating the sales amount, sales number, and advertisement recognition rate after one week for each product category.

The macro simulation execution module 221 executes the macro simulation using the statistical data 303 and outputs the prediction data 304. The prediction data 304 includes an entry made up of a field group indicating the sales amount, the sales number, and the advertisement recognition rate of each product category after one week.

The micro simulation execution module 231 executes a micro simulation using the observation data 301 and outputs prediction data 302. The prediction data 302 includes an entry made up of a group of fields indicating the purchased products, the number of purchases, the purchase amount, and the presence / absence of advertisement recognition after one week.

The model correction module 203 sets the prediction data 304 using the sales amount of each product category as a determination index, and uses the prediction data 302 to calculate a statistic indicating the sales amount of each product category. The model correction module 203 calculates the difference between the determination index and the statistic for each product category.

The model correction module 203 determines whether or not the total difference value of each product category is equal to or less than the threshold value m. When it is determined that the total value of the differences between the product categories is larger than the threshold value m, the simulation model is corrected.

In addition, this invention is not limited to the above-mentioned Example, Various modifications are included. Further, for example, the above-described embodiments are described in detail for easy understanding of the present invention, and are not necessarily limited to those provided with all the described configurations. In addition, a part of the configuration of the embodiment can be added to, deleted from, or replaced with another configuration.

In addition, each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. The present invention can also be realized by software program codes that implement the functions of the embodiments. In this case, a storage medium in which the program code is recorded is provided to the computer, and a CPU included in the computer reads the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the program code itself and the storage medium storing it constitute the present invention. Examples of storage media for supplying such program codes include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs (Solid State Drives), optical disks, magneto-optical disks, CD-Rs, magnetic tapes, A non-volatile memory card, ROM, or the like is used.

Further, the program code for realizing the functions described in this embodiment can be implemented by a wide range of programs or script languages such as assembler, C / C ++, Perl, Shell, PHP, Java, and the like.

Furthermore, by distributing the program code of the software that realizes the functions of the embodiments via a network, the program code is stored in a storage means such as a hard disk or memory of a computer or a storage medium such as a CD-RW or CD-R. The CPU included in the computer may read and execute the program code stored in the storage unit or the storage medium.

In the above-described embodiments, the control lines and information lines indicate those that are considered necessary for the explanation, and do not necessarily indicate all the control lines and information lines on the product. All the components may be connected to each other.

Claims

A simulation method in a system for executing a simulation with different granularity,
The system includes at least one computer having a processor, a memory connected to the processor, and a network interface connected to the processor,
The at least one computer includes model management information for managing a model which is definition information of a first simulation and a second simulation having a granularity larger than that of the first simulation, and a first used for the first simulation. History management information for managing the history of data, and
The simulation method includes:
A first step in which the at least one computer uses the first data acquired from the history management information to generate second data used for the second simulation;
The at least one computer executes the second simulation based on the second simulation model acquired from the second data and the model management information, and outputs a second simulation result. Two steps,
The at least one computer executes the first simulation based on the first data acquired from the history management information and the model of the first simulation acquired from the model management information. A third step of outputting the simulation result of
A fourth step in which the at least one computer corrects at least one of the first simulation model and the second simulation model based on the first simulation result and the second simulation result; The simulation method characterized by including these.
The simulation method according to claim 1,
The fourth step includes
A fifth step of using the first simulation result to calculate a statistic that can be compared with the second simulation result;
And a sixth step of correcting a model of the first simulation based on a comparison result between the statistic and the second simulation result.
The simulation method according to claim 2,
The sixth step includes
Calculating a difference between the statistic and the second simulation result;
Correcting the model of the first simulation so that the difference becomes small.
The simulation method according to claim 2,
The model management information stores a plurality of first simulation models,
In the third step, the first simulation result is calculated for each of the plurality of first simulation models,
In the fifth step, the statistic is calculated for each of the plurality of first simulation results,
The sixth step includes
Calculating a difference between each of the plurality of statistics and the second simulation result;
Determining a first simulation model corresponding to a statistic having the smallest difference as a model to be used for the first simulation.
A system that executes simulations with different granularities,
The system includes at least one computer having a processor, a memory, and a network interface;
A model management unit that manages a model that is definition information of the first simulation and the second simulation having a larger granularity than the first simulation;
A history management unit that manages the history of the first data used in the first simulation;
Based on the first data managed by the history management unit, a generation unit that generates second data used for the second simulation;
A first simulation execution unit that executes the first simulation based on the first data acquired from the history management unit and the model of the first simulation, and outputs a first simulation result;
A second simulation execution unit that executes the second simulation based on the second data generated by the generation unit and the model of the second simulation, and outputs a second simulation result;
A system comprising: a model correction unit that corrects the first simulation model and the second simulation model based on the first simulation result and the second simulation result.
6. The system according to claim 5, wherein
The model correction unit is
Using the first simulation result, calculate a statistic that can be compared with the second simulation result;
A system for correcting a model of the first simulation based on a comparison result between the statistic and the second simulation result.
The system according to claim 6, comprising:
The model correction unit is
Calculating a difference between the statistic and the second simulation result;
A system that corrects the model of the first simulation so that the difference becomes small.
The system according to claim 6, comprising:
The model management unit manages a plurality of first simulation models,
The first simulation execution unit calculates the first simulation result for each of the plurality of first simulation models;
The model correction unit is
Calculating the statistic for each of the plurality of first simulation results;
Calculating a difference between each of the plurality of statistics and the second simulation result;
A system that determines a model of a first simulation corresponding to a statistic having the smallest difference as a model to be used for the first simulation.
A program executed by a computer that executes a simulation with different granularity,
The calculator is
A processor, a memory connected to the processor, and at least one computer having a network interface connected to the processor;
Model management information for managing a model, which is definition information of a first simulation and a second simulation having a granularity larger than that of the first simulation, and a history of first data used for the first simulation are managed. History management information, and the program
A first procedure for generating second data used for the second simulation using the first data acquired from the history management information;
A second procedure for executing the second simulation based on the second data and the model of the second simulation acquired from the model management information and outputting a second simulation result;
Based on the first data acquired from the history management information and the model of the first simulation acquired from the model management information, the first simulation is executed, and a first simulation result is output. 3 steps,
Based on the first simulation result and the second simulation result, the computer executes a fourth procedure for correcting at least one of the first simulation model and the second simulation model. A program characterized by letting
The program according to claim 9, wherein
The fourth procedure includes:
A fifth procedure that uses the first simulation result to calculate a statistic comparable to the second simulation result;
And a sixth procedure for correcting a model of the first simulation based on a comparison result between the statistic and the second simulation result.
The program according to claim 10,
The sixth procedure includes:
Calculating a difference between the statistic and the second simulation result;
And a procedure for correcting the model of the first simulation so that the difference becomes small.
The program according to claim 10,
The model management information stores a plurality of first simulation models,
In the third procedure, the first simulation result is calculated for each of the plurality of first simulation models,
In the fifth procedure, the statistic is calculated for each of the plurality of first simulation results,
The sixth procedure includes:
Calculating a difference between each of the plurality of statistics and the second simulation result;
And a procedure for determining a model of a first simulation corresponding to a statistic having the smallest difference as a model to be used for the first simulation.