KR20210045205A - Computing system implementing and operating models describing subject system and subject system operation prediction method therewith - Google Patents

Abstract

According to one embodiment of the present invention, a computing system for implementing and operating a model describing a target system includes: a communication interface for receiving or calling acquisition data obtained from the target system; a system model including a plurality of sub-models; and at least one processor. Input data and output data of each of the sub-models in the target system are defined based on structural information of the target system obtained by using knowledge of the target system. The structural information includes information on a data connection relation between a first sub-model and a second sub-model among the sub-models, and the data connection relation includes information on whether the output data of the first sub-model is able to be used as the input data of the second sub-model. Accordingly, a hybrid model that operates by combining advantages of a physically based model and a data-based model is provided.

Description

A computing system that implements and operates a model describing the target system, and a method of predicting the operation of the target system using the same {COMPUTING SYSTEM IMPLEMENTING AND OPERATING MODELS DESCRIBING SUBJECT SYSTEM AND SUBJECT SYSTEM OPERATION PREDICTION METHOD THEREWITH}

The present invention relates to a technology for modeling and simulating a physical complex system, and more particularly, to a technology for implementing and operating a digital model describing a target system using a computing system. In addition, the present invention relates to a technology for advancing a digital model by using a conventional theory-based model and a data-based model based on data analysis together in the process of implementing and operating a digital model.

The present invention is derived from research conducted as part of the technology transfer commercialization project of the Ministry of Science and ICT and the Foundation for Promotion of Special Research and Development Zones [Project management number: 2019-DD-RD-0056-01-101, project name: machine learning Built-in digital twin modeling simulation platform development].

The demand to improve productivity, economy, safety, etc. is spreading in industrial sites. In recent years, technologies such as Internet of Things (IoT), big data, artificial intelligence, and cyber physical systems (CPS) are widely used, and as a technology trend in which these technologies are concentrated, digital twin technology is paying attention. Are receiving.

A digital twin is a digital replica (twin) of physical objects (assets, processes and systems, etc.), which maintains the properties/states of target object elements throughout their lifecycle and describes the dynamic nature of how they behave. It can be said to be a model of.

The digital twin implemented in the computing system is interlocked with the target object (physical asset) to reflect the real situation and predict the situation that may occur in the real world or inform the conditions for optimizing the operation. It is recognized as a means to strengthen competitiveness.

Internet of Things (IoT) technology and digital twin technology are closely related. The advancement of IoT platform technology enables smart services such as machine learning/artificial intelligence-based prediction, fault diagnosis, optimization, and predictive diagnosis after collecting sensor data of the operating system in real time. The digital twin technology started on the premise that it is used in industrial sites, but in recent years, efforts have been made to expand the area to which it is applied, such as cyber city, to more diverse smart services. It is virtually impossible to install all of the sensors for collecting all the data required for various smart services into an operating system due to physical constraints (location and number of sensors), and/or economic constraints (number of sensors).

As a means to partially solve this problem, in US Patent Publication No. 2017/0286572 "Digital twin of twinned physical system", the actual sensor operation is simulated using the digital twin of the actual sensor, and the actual sensor does not operate. When it turned out, a technology was proposed to replace the operation of the real sensor by collecting data from a virtual sensor corresponding to the real sensor in the digital twin model.

In Korean Patent Laid-Open Publication No. 10-2019-0013610 "System, method, and control unit for controlling the operation of a technical system", virtual sensor data is generated using a digital twin of an actual sensor, and when an abnormality of the actual sensor is detected. A technology has been proposed to replace the abnormal sensor with virtual sensor data.

In general, in order to perform motion/performance analysis or prediction for a real-world system, an abstracted model is created and executed to measure/observe measures of interest aspects such as motion/performance. In order to ensure the reliability of the analysis/prediction results of the target system through such a model, how well (accurately) the system is modeled becomes an important key.

As a method of modeling, there is a method of creating an abstraction model using knowledge of the physical laws or rules of operation included in the target system. This is modeling and simulation that can express the causal relationship between the controlled input and the corresponding output as a model. It is a (M&S) based method. This method has the limitation that detailed information about the system to be modeled must be available. In addition, when a model is built in a modeling and simulation (M&S)-based method, it is essential to verify the model with real data on how well the actual system is reflected in order to ensure the validity of the model. When it is difficult to obtain data from a real-world system, since the validity of the model cannot be verified, there may be a problem that the reliability of the analysis/prediction results based on the model cannot be guaranteed.

As another modeling method for predicting and analyzing a system, there is a data modeling method in which the rules/patterns/functions contained in the corresponding system are derived by analyzing a large number of data acquired through operation and observation of the target system. The machine learning-based model, which is a representative method of data modeling, is a method that represents the correlation between one set of data and another data set. In the era of big data, more effective machine learning has become possible by utilizing a large number of data. . However, the data model built through machine learning can be predicted under the premise that the system will operate without any change in the future, but if the composition or operating laws of the system change, the future cannot be predicted with the previously learned model. There are limitations.

The term'big data' is widely spread as a means of predicting a diversified society. Big data refers to a set of data that exceeds the ability of common software tools to collect, manage, and process within acceptable elapsed time. Because such large amounts of data provide more insight than existing limited data, it is receiving a lot of interest in research in various fields such as science, engineering, defense, business, medicine, and politics. For this reason, modeling using big data is becoming an essential and important issue in the era of big data.

A model created by the laws of physics or motion rules is called a physics-driven model, and a model that analyzes data and derives rules hidden in the data is called a data-driven model. The physical-based model is not limited to the measured data and has the advantage of being able to expand the scope of application, but it has the disadvantage that it requires verification by the measured data, and the data-based model has the disadvantage that the verification of the measured data is performed first in the analysis process. There is a limit that it cannot be applied to a range outside the measured data. Recently, due to the development of artificial neural networks and deep learning techniques, cases in which data-based models have superior performance to physical-based models have been reported. However, even if data-based models have excellent performance, they cannot be applied to a range outside the range of measured data. Still exists. In addition, the data-based model can only analyze the correlation between data, and has a limitation in that it cannot explain the causal relationship which data is the cause and which data is the result.

There are attempts to complement the advantages and disadvantages of such a data-based model and a physical-based model. For example, US Patent Publication US 2018/0357343 "OPTIMIZATION METHODS FOR PHSICAL MODELS" proposes a hybrid model that combines the advantages and disadvantages of a physics-driven model having high fidelity and a data-driven model having low fidelity. In US 2018/0357343, the physics-driven model is executed when the calculated competence area of the data-driven model is out of the critical area, and the discrepancy between the data-driven model and the physics-driven model is expressed in the form of a function. After calibration, a hybrid model that combines this relationship is used. At this time, the competence of the data-driven model is calculated based on the accuracy of the sample data of the data-driven model.

In the case of a single system in which the input data and output data of the target system are clearly defined, a hybrid model that combines the merits of a data-driven model and a physics-driven model can be used by applying the US 2018/0357343. However, if the target system is complicated and there are many variables to be dealt with, it is necessary to apply more advanced data to digitize the target system and implement it as a digital twin model.

US Patent Publication No. 2017/0286572 "Digital twin of twinned physical system" (October 5, 2017) Korean Patent Laid-Open Patent No. 10-2019-0013610 "System, method and control unit for controlling the operation of a technical system" (February 11, 2019) US Patent Publication No. 2019/0033850 "Controlling operation of a technical system" (January 31, 2019) US Patent Publication No. 2019/0068618 "Using virtual sensors to accommodate industrial asset control systems during cyber attacks" (February 28, 2019) US Patent Publication US 2018/0357343 "OPTIMIZATION METHODS FOR PHSICAL MODELS" (December 13, 2018)

In the prior art and the above-described prior art, a real-world operation or state is simplified and simulated into a single model. At this time, the use of a hybrid model that combines a physics-based model capable of deductive inference and a data-based model based on data learning was initiated by the preceding document, US 2018/0357343, but the target system in the real world is very complex and operates with only a single model. It is not easy to determine the condition.

In addition, as the target system to be handled becomes larger and more complex, the limitations of attempts to describe the target system by a single model are revealed.

The present invention is an invention to solve the problems of the prior art and the prior art, by modeling a causal relationship between a number of causes and effects describing a target system, and combining sub-models capable of partially describing the target system. We propose a system model that can describe the entire target system in an integrated manner. In this case, a coupling relationship and/or a data connection relationship between sub-models in the system model is defined based on structural information, and the structural information can be obtained using knowledge of the target system.

The present invention aims to propose a hybrid model that combines the merits of a physical-based model and a data-based model to operate, and extends the application range of the hybrid model to apply the hybrid model to the system model described by the composite model. The purpose of this study is to propose a system modeling technique.

The present invention proposes a mutually collaborative alternative that can overcome the limitations of each approach by complementarily utilizing the advantages of the two modeling methods, a physical model and a data-based model to enable robust analysis/prediction support. It aims to do. At this time, the object is to propose a technique that can implement a complex model by combining a physically-based model and a data-based model more strongly than the prior art or prior art in which a physically-based model or a data-based model is selectively applied to the issue. It is done.

An object of the present invention is to propose a system model that utilizes the accuracy of a data-based model to the fullest, but allows inference on data outside the data range, which is a limitation of the data-based model.

The present invention proposes a system model capable of describing the operation or state of a target system due to a complex cause by using a data-based model, and the interrelationship between data in the data-based model can be explained based on structural information. It aims to generate descriptive information.

An object of the present invention is to propose a system model capable of applying a normative analysis, which was difficult to implement by a data-based model while using a data-based model.

The present invention aims to implement a hybrid model that combines the advantages of a data-based model and a physical-based model as a complex system model, and to apply the complex system model to a digital twin for describing a very complex target system. It is done.

The present invention is a configuration derived to achieve the above object, and a computing system according to an embodiment of the present invention is a computing system that implements and operates a model describing a target system. A communication interface capable of receiving or calling; A system model including a plurality of sub-models; And at least one or more processors. Each of the plurality of sub-models is a model capable of receiving part of the acquired data and inferring or predicting at least part of the remaining data as an output, and input data and output data of each of the plurality of sub-models in the system model Is defined based on structural information of the target system that can be obtained using knowledge of the target system, and the structural information is a data connection relationship between a first sub-model and a second sub-model among the plurality of sub-models. And the data connection relationship includes information on whether output data of the first sub-model can be used as input data of the second sub-model.

The at least one processor selects the first data from among new input data for the target system based on the structural information and provides the first data as the input of the first sub-model, and provides the output data of the first sub-model to the second As data, the system model outputs third data that is generated based on the output data of the second sub-model or is generated based on the output data of the second sub-model, and is provided as an input of the second sub-model based on the structural information. Control the first sub-model so that the first sub-model can infer the operation of the target system based on the first data, and the second sub-model is based on the second data Thus, the second sub-model is controlled so that the operation of the target system can be inferred.

At this time, each of the plurality of sub-models is defined based on obtainable domain knowledge, experience, and theory related to the target system, and any of a theory-based model capable of deductive inference or a data-based model that is learned based on the acquired data It may be one, or a model in which the theory-based model and the data-based model are combined to complement each other. The data-based model may be a model using an artificial neural network (ANN).

At this time, the structural information is a data connection relationship between parameters in the target system constituting a theory-based primitive model that can be acquired using knowledge of the target system, and whether each of the parameters can be included in the acquired data. Can be defined based on whether or not.

At this time, the structural information may be selected as input data or output data of each of the plurality of sub-models among parameters in the target system constituting a theory-based original model that can be obtained using knowledge of the target system. Information on related parameters may be included, and each of the related parameters may include sub-model data structure information defined as input data and output data of each of the plurality of sub-models based on the theory-based original model.

The computing system further includes a user interface for receiving a user query regarding the target system and transmitting it to the at least one processor, and the at least one processor interprets the user query to execute instructions that can be executed in the computing system. Can be converted to an order.

When the user query includes a query for at least two of the first data, the second data, and the third data, the at least one processor may perform the Generate a corresponding search result by searching for at least one of a condition variable, a control variable, or a design variable of the target system corresponding to at least two or more of the first data, the second data, and the third data, and the A response to the user query may be generated based on a corresponding search result, and the response to the user query may be delivered to the user interface.

When the user query includes a prediction of the operation of the target system when the first data is out of the acquired data domain covered by the acquired data, the at least one processor By inputting the first data, the first sub-model may be controlled to predict the operation of the target system in response to the first data. The at least one processor applies one of a theory-based model capable of deductive inference as the first sub-model or a data-based model that is learned based on the acquired data, or the theory-based model and the data-based model They may be combined to complement each other and applied to the first sub-model.

When the user query includes a condition in which the distribution of the third data is adjusted, the at least one processor determines the first data applicable to the first data from among the acquired data based on the structural information and the system model. A first distribution, a second distribution of second data related to the first distribution, and a third distribution of the third data related to the second distribution may be generated. The at least one processor provides a condition for limiting the range of the first distribution in response to the user query based on the first distribution, the second distribution, the third distribution, and the structural information, or the target system It is possible to provide the user with the possibility of changing at least one or more of a condition variable, a control variable, or a design variable.

A computing system according to an embodiment of the present invention is a computing system that implements and operates a model describing a target system, comprising: a communication interface capable of receiving or calling acquisition data obtained from the target system; A system model including a plurality of sub-models; And at least one or more processors. Each of the plurality of sub-models is a model capable of receiving part of the acquired data and inferring or predicting at least part of the remaining data as an output, and input data and output data of each of the plurality of sub-models in the system model Is defined based on structural information of the target system that can be obtained using knowledge of the target system, and the structural information is a data connection relationship between a first sub-model and a second sub-model among the plurality of sub-models. And the data connection relationship includes information on whether output data of the first sub-model can be used as input data of the second sub-model.

In this case, the at least one processor provides first data of the acquired data as an input of the first sub-model, and provides second data defined as an output of the first sub-model based on the structural information. Provided as an input of a model, and control the first sub-model so that the first sub-model can learn the relationship between the first data and the second data, or the second sub-model is used as the second data and the second data The second sub-model may be controlled to learn a relationship between third data defined as an output of the second sub-model based on structural information.

According to an embodiment of the present invention, a method of operating a system model executed by a computing system that implements and operates a model describing a target system, wherein the communication interface receives or calls new input data for the target system. step; Selecting, by the at least one processor, the first data based on the structural information among the new input data and providing the first data as an input of the first sub-model; Controlling, by the at least one processor, the first sub-model so that the first sub-model can infer an operation of the target system based on the first data; Providing, by the at least one processor, output data of the first sub-model as second data as an input of the second sub-model based on the structural information; Controlling, by the at least one processor, the second sub-model so that the second sub-model can infer an operation of the target system based on the second data; And providing, by the at least one processor, third data, which is generated based on the output data of the second sub-model or is output data of the second sub-model, as an output of the system model.

In this case, the method of operating the system model of the present invention may further include converting, by the at least one processor, a user query regarding a target system into an instruction command that can be executed in the computing system.

According to an embodiment of the present invention, a method of learning a system model executed by a computing system that implements and operates a model describing a target system includes the steps of: receiving or calling the acquired data by the communication interface; Providing, by the at least one processor, first data of the acquired data as an input of the first sub-model; Providing, by the at least one processor, second data, which is defined as an output of the first sub-model based on the structural information, as an input of the second sub-model; And the at least one processor controlling the first sub-model so that the first sub-model can learn a relationship between the first data and the second data, or the second sub-model is the second data and And controlling the second sub-model to learn a relationship between third data defined as an output of the second sub-model based on the structural information.

According to the present invention, a system model capable of integrally describing the entire target system by modeling a causal relationship between a plurality of causes and effects describing a target system, and combining sub-models capable of partially describing the target system is provided. Can provide In this case, a coupling relationship and/or a data connection relationship between sub-models in the system model is defined based on structural information, and the structural information can be obtained using knowledge of the target system.

According to the present invention, it is possible to provide a hybrid model that operates by combining the advantages of a physical-based model and a data-based model. At this time, a system model that can apply the hybrid model to the system model described by the composite model can be implemented by expanding the application range of the hybrid model.

According to the present invention, it is possible to implement a system model that utilizes the accuracy of a data-based model to the maximum, but allows inference on data outside the data range, which is a limitation of the data-based model.

According to the present invention, it is possible to provide a system model capable of describing an operation or state of a target system due to a complex cause by using a data-based model. In addition, it is possible to generate descriptive information that can describe interrelationships between data in a data-based model based on structural information.

According to the present invention, while using a data-based model, it is possible to implement a system model to which a normative analysis, which was difficult to implement by the data-based model, can be applied.

According to the present invention, a hybrid model that combines the advantages of a data-based model and a physical-based model can be implemented as a complex system model, and a complex system model can be applied to a digital twin for describing a very complex target system. have.

According to the present invention, a system model with a built-in data-based model embeds machine learning contents using acquired data acquired through operation and observation of a target system, so that the model itself becomes a verified model in the domain of the acquired data. You can overcome limitations that may be encountered when analyzing or predicting a target system using only the base model or only the physically based model.

By embedding machine learning into the system model, it reduces the degree of prior knowledge required for the target system and achieves the effect of model verification (verification with real data), while changing the structure/rules within the target system that could not be done with machine learning alone. Can analyze and predict the behavior of the target system according to.

According to the present invention, it can have the following advantages over the existing data-based model. In general, a data-based model exhibits very high accuracy for static data, but it is known that it is not easy to apply variables that affect behavior to various dynamic analyses. According to the present invention, a data-based model can also be applied to a target system having a space variant, a time variant, and a stochastic feature.

In addition, according to the present invention, non-linear characteristics of the target system can be reflected. Since many systems in the real world have nonlinear features without the overlapping theorem between inputs and outputs, it is a great advantage that the state transition function can take into account nonlinear features. In addition, since additional situational information such as weather conditions can be expressed through data obtained from the actual system, it can be applied as a robust modeling technique to the realization of complex digital models such as digital twin models that mimic the real world.

1 is a diagram illustrating a computing system according to an embodiment of the present invention.
2 is a diagram illustrating an embodiment of a system model included in a computing system according to an embodiment of the present invention.
3 is a diagram illustrating an example of a process of deriving structural information used in a computing system according to an embodiment of the present invention.
4 is a diagram illustrating an example of a domain of acquired data and a theoretically expandable data domain used by a computing system according to an embodiment of the present invention.
5 is a diagram illustrating an embodiment of a sub-model included in a system model according to an embodiment of the present invention.
6 is a diagram illustrating a computing system according to an embodiment of the present invention.
7 is a diagram illustrating an example of an operation process of a system model according to an embodiment of the present invention.
8 is a diagram illustrating an embodiment of a theory-based primitive model on which a system model according to an embodiment of the present invention is based.
9 is a diagram illustrating an embodiment of a system model generated based on the theory-based primitive model of FIG. 8.
10 is a flowchart illustrating a method of learning a system model executed by a computing system according to an embodiment of the present invention.
11 is an operation flow diagram illustrating a method of operating a system model such that the system model is executed by the computing system according to an embodiment of the present invention and infers a target system.

In addition to the above objects, other objects and features of the present invention will become apparent through the description of the embodiments with reference to the accompanying drawings.

A preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the subject matter of the present invention, a detailed description thereof will be omitted.

1 is a diagram illustrating a computing system 100 according to an embodiment of the present invention.

The computing system 100 includes a system model 120 that describes a target system, and may implement and operate the system model 120. Specifically, the computing system 100 includes at least one processor 110 and a communication interface 130 capable of receiving or calling acquisition data obtained from a target system.

The target system refers to a real-world system to be analyzed, inferred, and predicted, and may be a system that operates according to physical laws or a system that operates according to an operation rule. The laws on which the target system depends are not limited to physical laws, and according to embodiments, the target system may be described by social, psychological, and economic rules or theories. The target system is closely related to the problem to be solved. In the case of analyzing and predicting the spreading process of forest fires, the distribution of objects such as forest topography, trees, rocks, etc., where forest fires may occur, can constitute the target system. In the case of analyzing and predicting traffic flow, the target system can be composed of road networks, signal systems, capacity of each road, and vehicles located on the road. In this process, in the case of analyzing the flow of traffic conditions when an unexpected situation occurs, behavior changes due to human psychological behavior patterns can be added as factors to be identified on the target system. In the case of analyzing and predicting the occurrence, spread, or healing process of a disease, the composition of cells, blood vessels, skeletons, organs, or blood flow can constitute the target system.

The target system may be a large and complex system such as a digital twin of a smart city. When there are very many variables affecting the target system, they can be combined and implemented as a target system, and the elements to be analyzed can be abstracted. For example, in the case of a smart city, national public data (population statistics, maps, etc.) can be collected or utilized, and the collected data can be linked using IoT infrastructure in the city. Using such data, it is possible to simulate changes in the behavior of citizens due to changes in policies within the city, or simulate economic activities and floating population activities according to disasters or geographic environments.

The processor 110 controls the system model 120 so that the system model 120 can learn acquired data obtained from a target system in order to implement the system model 120. When a new input for the target system is given after the learning of the system model 120 is finished, the processor 110 responds to the new input to predict the action to be taken by the target system, and the system model 120 is based on the new input. The system model 120 may be controlled to infer or predict the operation of the target system.

The system model 120 may include a theory-driven model and a data-driven model. The data-driven model may include an artificial neural network (ANN). The system model 120 is generally a collection of data stored in a memory or storage, and parameters analyzed for correlations between acquired data that have become a learning target constitute the system model 120. In the process of the processor 110 controlling the system model 120, the processor 110 accesses the storage or memory in which the system model 120 is stored, and the processor 110 reads the system model 120 to perform other operations. Alternatively, the result of the operation may be used in a form of rewriting some of the parameters of the system model 120.

2 is a diagram illustrating an embodiment of a system model 120 included in the computing system 100 according to an embodiment of the present invention.

The system model 120 includes first sub-models 122a and 122b, and a second sub-model 124. The first sub-models 122a and 122b may receive the first data X1, X2, X3, and X4 as inputs and may output the second data Y1 and Y2. The second sub-model 124 may receive second data Y1 and Y2 as inputs and may output third data Z as an output. The system model 120 may use the third data Z as a final output, but the spirit of the present invention is not limited to the embodiment of FIG. 2.

Each of the plurality of sub-models 122a, 122b, and 124 included in the system model 120 is a model capable of receiving some of the acquired data and inferring or predicting at least some of the remaining data as an output. In the system model 120, the input data and output data of each of the plurality of sub-models 122a, 122b, 124 may be defined based on structural information of the target system that can be obtained using knowledge of the target system. I can. The structural information includes information on a data connection relationship between the first sub-models 122a and 122b and the second sub-model 124 among the plurality of sub-models 122a, 122b, and 124, and the data connection relationship is first Information on whether output data of the lower models 122a and 122b can be used as input data of the second lower model 124 may be included.

The first data (X1, X2, X3, X4), the second data (Y1, Y2), and the third data (Z) are data that can be actually measured or observed and obtained from a target system. The processor 110 includes first data (X1, X2, X3, X4) suitable for the first sub-models 122a and 122b, and the second sub-model 124 based on structural information among actual acquired data of the target system, The second data Y1 and Y2, and the third data Z may be selected.

The processor 110 provides the first data (X1, X2, X3, X4) as input of the first sub-models 122a and 122b based on structural information among the acquired data, and provides the first data among the acquired data based on the structural information. 2 The data Y1 and Y2 may be provided as outputs of the first sub-models Y1 and Y2 for learning the first sub-models 122a and 122b. The processor 110 may provide the second data Y1 and Y2 as an input of the second sub-model 124 based on the structural information. The processor 110 controls the first sub-model so that the first sub-models 122a and 122b can learn the relationship between the first data (X1, X2, X3, X4) and the second data (Y1, Y2). I can.

In addition, the processor 110 may provide the third data Z among acquired data as an output of the second sub-model 124 for learning of the second sub-model 124 based on the structural information. The processor 110 may control the second sub-model 124 so that the second sub-model 124 can learn a relationship between the second data Y1 and Y2 and the third data Z.

In a state in which the first sub-models 122a and 122b and/or the second sub-model 124 are trained, the first sub-models 122a and 122b and/or the second sub-model 124 are It can infer or predict the behavior of the target system.

In this case, the processor 110 may select new input data other than the measured data as the first data X1, X2, X3, X4, and provide the first data as inputs to the first sub-models 122a and 122b. In this case, data output by inference or prediction of the first sub-models 122a and 122b is designated as the second data Y1 and Y2, and the second data Y1 and Y2 are designated as the second sub-model 124. Can be provided as input. The process of providing the second data (Y1, Y2) output by inference or prediction of the first sub-models 122a and 122b by the processor 110 as input to the second sub-model 124 is based on structural information. It can be done by doing. In addition, a process of selecting the first data X1, X2, X3, and X4 among new input data by the processor 110 may be performed based on structural information.

The processor 110 is a first sub-model so that the first sub-models 122a and 122b can infer or predict the operation, state, or state change of the target system based on the first data (X1, X2, X3, X4). The second sub-model 124 controls (122a, 122b) and allows the second sub-model 124 to infer or predict the behavior, state, or state change of the target system based on the second data (Y1, Y2). ) Can be controlled. In this case, FIG. 2 shows an embodiment in which the output data of the second lower model 124 is provided to the user as the final output of the system model 120 as the third data Z, but according to the embodiment of the present invention, the second Additional operations may be performed based on the output data of the lower model 124, and thus data generated based on the output data of the second submodel 124 may be provided to the user as a final output of the system model 120. have.

At this time, each of the plurality of sub-models 122a, 122b, 124 is defined based on obtainable domain knowledge, experience, and theory related to the target system, and is a theory-based model capable of deductive inference or data learned based on acquired data. It may be any one of the base models, or a model in which the theory-based model and the data-based model are complementary to each other. In this case, the data-based model may be a model using an artificial neural network (ANN). The theory-based model may be defined by physical laws, rules of operation, and the like, and the obtainable domain knowledge, experience, and theory may include all theories based on physics, psychology, economics, and sociology.

A general data-based model has recently improved the accuracy of data analysis and prediction through rapid development, but there is a problem in that it is difficult for humans to understand the internal operation because the interior thereof is treated as a black box. Therefore, accurate prediction is possible only within the actually obtained data domain, and it is difficult to accurately predict inputs outside the data domain. On the other hand, a general theory-based model is capable of descriptive/deductive reasoning, and it is possible to reason about inputs outside the specified domain, but it is necessary to verify whether it predicts what is happening in the actual system by applying actual data.

The system model 120 according to an embodiment of the present invention can complementaryly combine a data-based model and a theory-based model, and a large frame is configured like a theory-based model using structural information, and each sub-model They 122a, 122b, and 124 may be configured as a data-based model. In this case, since the entire structure of the system model 120 is obtained by knowledge, theory, or experience, explanatory/deductive inference is possible, and inference can be made about inputs outside a predetermined domain. On the other hand, when each of the sub-models 122a, 122b, and 124 is a data-based model, since it is a model already obtained based on data in the acquired data domain, additional verification by actual measurement data is not required.

In addition, in the embodiment of the present invention, all of the sub-models 122a, 122b, 124 may be composed of a theory-based model, some of the sub-models 122a, 122b, and 124 are theory-based models, and the others are data-based models. It can also be configured. When all the sub-models 122a, 122b, 124 are composed of a theory-based model, consider a case in which each sub-models 122a, 122b, 124 undergo verification by actual data and exhibit the same accuracy as the data-based model. I will be able to.

For example, in the system model 120 of FIG. 2, only a specific sub-model 122a is a data-based model, and the other sub-models 122b and 124 predict the actual target system with an accuracy comparable to that of the data-based model. Can be assumed. At this time, the specific sub-model 122a is valid only within the domain of the given acquisition data, but the other sub-models 122b and 124 can operate in an extended range outside the domain of the acquired data. In this system model 120, when the user inputs X3 and X4 out of the existing range, it is possible to provide a solution that satisfies the user's needs when he/she wants to predict the operation of the target system.

That is, when a large part of the target system can be modeled as a theory-based model, when the performance of the data-based model is superior to that of the theory-based model only for a specific sub-model 122a, the data-based model is used as a specific sub-model 122a In addition, the rest of the system model 120 may use a theory-based model.

In general, data-driven models are known to have strength in predictive analysis that analyzes "what will happen?" based on descriptive analysis that analyzes "what happened?", but "how did that event occur?" Or "How do I get the results I want?" It is known that it is difficult to support the normative analysis that analyzes the back.

The system model 120 according to an embodiment of the present invention complementarily combines a data-based model and a theory-based model to support a normative analysis using a data-based model and structural information based on theory. .

In the above example, since the specific sub-model 122a is a black box, the operation rules of the specific sub-model 122a cannot be changed, but the operation rules of the other sub-models 122b and 124 may be changed. In addition, it is not possible to change the internal operation rule of the specific sub-model 122a, but by changing the operation rules of the other sub-models 122b and 124 in the vicinity connected to the specific sub-model 122a, the overall system model 120 ) May be treated as if a relative operation rule of a specific sub-model 122a is changed. Since the relative operation rules in the system model 120 can be changed, the system model 120 can perform a normative analysis (including a control policy, etc.) from a predictive analysis, and a cognitive analysis from the normative analysis. (Including planning, etc.) can be carried out.

If a single data-based model is treated as a black box, the theory-based model can be thought of as a white box, and a model that combines the data-based model and theory-based model can be thought of as a gray box. At this time, if only the input and output of the data-based model is observed in conjunction, the sub-model block may be considered a black box, but from the perspective of the entire system model 120, the input and output of the data-based model within the data domain Since the system operation can be completely reproduced by other theory-based models in the vicinity, including data, from the perspective of the entire system model 120, even when the data-based model is included as a sub-block, the white box goes beyond the gray box. It would be okay to recognize it as.

In addition, according to an embodiment of the present invention, a theory-based model and a data-based model may be used together in one sub-model. For example, there may be cases where a theory-based model presents a mathematical relationship, but does not accurately inform the coefficient or order of the relationship. In this case, coefficients or orders of the theory-based model can be obtained by data-based learning.

For convenience of explanation, as a simple example, it is assumed that the first sub-models 122a and 122b are data-based models, and the second sub-model 124 is a theory-based model. It may be noted that the second sub-model 124, which is a theory-based model, can be described by a simple linear relation Z = mY1 + nY2. However, it is assumed that the values of the coefficients m and n are not known only by theory. In this case, in order to obtain m and n, coefficients m and n may be obtained through data-based learning between Y1, Y2, and Z.

That is, one sub-model may be modeled as a data-based model, or although it is a theory-based model, its coefficient or order may be specified by data-based learning. Also, if the theoretical basis is clear, the theory-based model can be implemented as a very definitive mathematical model.

There may be various types of sub-models, and the system model 120 of the present invention is a feature that differentiates from the prior art in that the structure of the entire system model 120 is determined based on structural information. Each sub-model is connected based on, and although each sub-model corresponds to a black box, other sub-models connected to the sub-model can fully reproduce the system operation using the input and output of the black box. The system model 120 may provide a result such as a white box to a user based on structural information. The characteristic that operates like a white box of the system model 120 provides descriptive information about the association between data inside the system model 120 to the user even when the data-based model is used as a lower model. It provides a platform that can be used.

3 is a diagram illustrating an example of a process of deriving structural information used in a computing system according to an embodiment of the present invention.

Structural information defining the internal structure of the system model 120 may be obtained based on a theory-based primitive model that can be obtained using knowledge of a target system. The theory-based primitive model can be derived based on the data connection relationship and causal relationship between parameters in the target system. Whether each of the various parameters included in the theory-based primitive model can be included in the acquired data, that is, is it a parameter actually measured or observed for the target system, is identified, and the information identified for each parameter is the structural information for the target system. Is derived.

Structural information can be derived through the process of deriving a theory-based primitive model from knowledge, experience, experiments, observations, and theories, and identifying data that can be obtained from the target system. This process may be executed by the computing system 100 or a separate computing system, or may be executed by interacting with the computing system 100 by human intervention in some processes.

For example, setting the theory-based primitive model is whether each of the parameters constituting each node of the theory-based primitive model corresponds to data that can be actually obtained from the target system in a state executed by human actions. May be identified by a computing system.

At this time, the structural information may be selected as input data or output data of each of the plurality of sub-models among parameters in the target system constituting a theory-based original model that can be obtained using knowledge of the target system. Information on related parameters may be included, and each of the related parameters may include sub-model data structure information defined as input data and output data of each of the plurality of sub-models based on the theory-based original model.

4 is a diagram illustrating an example of a domain of acquired data used by a computing system according to an embodiment of the present invention and a theoretically expandable data domain.

When the first data (X1, X2, X3, X4) is obtained from the actual acquired data, the same temporal variation as the first data (X1, X2, X3, X4), or the second data (Y1, Y2) appearing in the spatial variation The distribution of is shown in the acquisition data domain 410. Meanwhile, the distribution of the third data Z appearing in the same temporal variation or spatial variation as the second data Y1 and Y2 obtained from the actual acquired data is also shown by the acquired data domain 430.

Data beyond the acquired data domains 410 and 430 may be provided as input to the system model 120 for the purpose of establishing a design or plan of the system model 120. For example, when the first data (X1, X2, X3, X4) is out of the acquisition data domain 410, the system model 120 is the first data (X1, X2, X3, The operation of the target system is inferred based on X4), and the distribution of the second data Y1 and Y2 appearing at this time may constitute the extended data domain 420.

Likewise, the system model 120 infers the operation of the target system based on the second data Y1 and Y2 out of the acquired data domain 430, and the distribution of the third data Z that appears at this time is the extended data domain ( 440).

The system model 120 of the present invention makes full use of the accuracy of the data-based model, but can infer or predict the operation of the target system even for data that deviates from the acquired data domains 410 and 430, which is a limitation of the data-based model.

5 is a diagram illustrating an embodiment of a sub-model 122a included in the system model 120 according to an embodiment of the present invention.

The sub-model 122a of FIG. 5 includes a data-based model 510 and a theory-based model 520 therein, and the logic module 530 is any one between the data-based model 510 and the theory-based model 520. One output is selected, or the output of the lower model 122a is generated by combining the outputs of the data-based model 510 and the theory-based model 520.

According to an embodiment of the present invention, the data-based model 510 and the theory-based model 520 may be competitively combined. For example, when input data within the range of the actual acquired data is given, the logic module 530 is based on the prediction accuracy of each of the data-based model 510 and the theory-based model 520 It can be designed to select any one between models 520.

According to an embodiment of the present invention, the data-based model 510 and the theory-based model 520 may be complementarily combined with each other. Based on whether the input data (X1, X2) given in FIG. 5 deviates from the acquisition data domain 410, the logic module 530 may apply either of the data-based model 510 and the theory-based model 520. You can choose. For example, the data-based model 510 within the acquired data domain 410 may be applied, and the theory-based model 520 may be applied to input data outside the acquired data domain 410.

In order to apply the theory-based model 520 in the extended data domain 420, a process for increasing the association/consistency between the theory-based model 520 and the data-based model 510 in the acquired data domain 410 may be required. have. This process may be performed in the form of, for example, a curve fitting for coefficients or orders of the theory-based model 520. In order to improve the consistency between the theory-based model 520 and the data-based model 510, a separate hybrid model having a form similar to that of the system model 120 of FIG. 2 may be applied.

Referring to FIGS. 4 and 5 together, an attempt to extend data to the extended data domains 420 and 440 is, for example, particularly when the distribution of data in the acquired data domains 410 and 430 is qualitatively asymmetric. It makes sense. In the case of a system model implemented using factory facilities as a target system, normal operation data will be significantly greater than abnormal operation data. Therefore, in the case of learning using data in an asymmetric state, there is a possibility that the system model is trained while being overfitted with the normal operation data. In addition, since it is not easy to acquire abnormal motion data in an actual situation, the distribution of asymmetric data can be compensated for a little by expanding the data assuming a virtual situation.

This expansion of data is necessary to ensure a balanced distribution of data when predicting a future operation from a current operation of a target system. In addition, not only normal/abnormal operation, but also in the case of design changes and planning to improve the reliability, stability, quality, or performance of the target system, analysis and prediction of the system model for the extended data domains (420, 440) Is necessary.

6 is a diagram illustrating a computing system 100 according to an embodiment of the present invention. Referring to FIG. 6, the computing system 100 may further include a user interface 140 to easily implement a service model for interacting with a user.

The computing system 100 further includes a user interface 140 for receiving a user query regarding a target system and transmitting it to the at least one processor 110, and the at least one processor 110 interprets the user query to provide a computing system. (It can be converted into an instruction instruction that can be executed within 1000.

In response to a user query, the processor 110 may interpret and express the operation and state of the target system based on structural information so that the user can easily understand it.

7 is a diagram illustrating an example of an operation process of the system model 120 according to an embodiment of the present invention.

Referring to FIG. 7, the lower model 122a has a condition variable C1, the lower model 122b has a condition variable C2, and the lower model 124 has a condition variable C3. C1, C2, and C3 may be condition variables, control variables, or design variables. Even if each of the sub-models 122a, 122b, 124 is a data-based model, the condition variable reflected in the process of combining the sub-models 122a, 122b, 124 with other sub-models in the system model 120 It can be determined based on structural information.

When there is a user query, the processor 110 uses structural information for the user query and conditional variable/control variable/design variable C1, C2, C3 to conduct a normative analysis, plan establishment, and design optimization for the target system. You can create an offer.

The at least one processor 110 includes a query for the association of at least two of the first data (X1, X2, X3, X4), the second data (Y1, Y2), and the third data (Z). In this case, the first data (X1, X2, X3, X4), the second data (Y1, Y2), and the third data ( Z) generates a response search result by searching for at least one of a condition variable, a control variable, or a design variable of a target system corresponding to at least two of the associations, and generates a response to a user query based on the corresponding search result. And, a response to a user query may be delivered to the user interface 140.

For example, if all of the sub-models 122a, 122b, 124 are data-based models, the system model 120 predicts the third data Z from the first data X1, X2, X3, X4. When a user query includes the process of deriving a result, the processor 110 connects the second data (Y1, Y2) and sub-models (122a, 122b, 124), which are intermediate results that are not directly revealed in the system model 120. Descriptive information on the system model 120 describing the target system based on the relationship may be provided. In addition, the processor 110 may verify the validity, reliability, and stability of the explanatory information based on actual measured data of the second data Y1 and Y2 as intermediate results.

User query, when the first data (X1, X2, X3, X4) out of the acquisition data domain 410 covered by the acquired data (intentional harsh environment or unknown operating conditions) to the operation of the target system In the case of including prediction, the at least one processor 110 inputs the first data (X1, X2, X3, X4) to the first sub-models 122a and 122b so that the first sub-models 122a and 122b In response to the first data (X1, X2, X3, X4), the operation of the target system may be predicted. The at least one processor 110 applies one of a theory-based model capable of deductive inference as the first sub-models 122a and 122b or a data-based model that is learned based on acquired data, or a theory-based model and data The base models may be combined to complement each other and applied to the first sub-models 122a and 122b.

When the user query includes a condition in which the distribution of the third data Z is adjusted (may include an intention to improve reliability, stability, performance, and quality of the target system), at least one processor 110 Is a first distribution of data applicable to the first data (X1, X2, X3, X4) among acquired data based on the structural information and the system model 120, and the second data related to the first distribution (Y1, Y2). A second distribution of) and a third distribution of the third data Z related to the second distribution may be generated. At this time, it is assumed that the characteristic information indicated by the distribution of the third data Z in the acquired data domain 430 can be evaluated as an indicator such as stability, reliability, quality, and performance of the target system. This process is, for example, by varying the set of condition variables {C1, C2, C3} and repeating inference/prediction repeatedly to obtain simulation results corresponding to the set of various condition variables {C1, C2, C3}. Can be understood. Alternatively, the processor 110 analyzes the relationship between the real distribution of each data in the acquired data and the target distribution of data representing the desired quality or performance, and a set of condition variables (C1, C2) for deriving the target distribution from the real distribution. , C3} can generate a proposal for a change. As such, the at least one processor 110 provides a limiting condition of the range of the first distribution in response to a user query based on the first distribution, the second distribution, the third distribution, and the structural information, or the The possibility of changing at least one or more of a condition variable, a control variable, or a design variable of the target system may be provided to the user.

For example, in FIG. 4, the data is uniformly distributed in the acquired data domains 410 and 430. In order to increase the stability and reliability of the target system, or to improve the quality and performance of the final output, the acquired data domain ( If the data is more concentrated in the center of 410 and 430, the target distribution may be set according to these needs. At this time, the user query may include complex needs. For example, the complex needs of improving the quality and performance of the final output without deteriorating the stability and reliability of the target system can be converted into target distribution and set. The processor 110 may generate a proposal to adjust the structural information and condition variables C1, C2, and C3 in the system model 120 or limit the range of input data in order to obtain a target distribution from a given real distribution.

8 is a diagram illustrating an embodiment of a theory-based primitive model 800 on which a system model is based according to an embodiment of the present invention.

9 is a diagram illustrating an embodiment of a system model 900 generated based on the theory-based primitive model 800 of FIG. 8.

For system modeling, a hypothetical model for the target system can be defined in the form of a gray box by first grasping the domain knowledge/experience and theories that can be acquired related to the target system. Since simulation is difficult with a hypothetical model itself, a process of constructing a complete model by securing information such as motion functions and parameters necessary for model completion is required. The information necessary for model completion is necessary to obtain big data by actually operating/observing the target system, and to generate the theory-based primitive model 800 by using a hypothetical model through machine learning for the acquired big data. Acquire. By learning big data acquired through actual operation and observation of the target system using a machine learning algorithm such as an artificial neural network, information necessary for the theory-based primitive model 800 can be learned. A theory-based primitive model 800 for a target system is generated by applying the learned and verified information based on actual data to a hypothetical model.

The theory-based primitive model 800 is implemented to have structural information by a number of parameters that can be represented in the target system. At this time, assuming that all parameters are secured, the theory-based primitive model 800 corresponds to a white box. However, in practice, not all parameters can be measured or observed from the target system, and thus editing is required to implement this as the system model 900.

In FIG. 8, it is assumed that y1' is a parameter that cannot be measured or observed. The sub-models 810 and 820 related to the parameter y1' cannot be applied to the system model 900 as they are, and modifications are required in order to be applied to the system model 900.

As shown in FIG. 9, the submodels 810 and 820 related to the parameter y1' may be merged into one submodel 910 and included in the system model 900. In this case, it is assumed that all of x1', x2', and w1' can be actually measured or observed from the target system.

Since only parameters that can be actually measured or observed from the target system are set as main nodes of the system model 900, sub-models of the system model 900 may be implemented as a data-based model or a theory-based model. . At this time, since the structural information includes information on a connection relationship and a coupling relationship between sub-models in the system model 900, it is possible to participate in the operation of each sub-model as Semantics of the system model 900.

In an embodiment of the present invention, sub-models may be partially data-based models, or all sub-models may be configured as a data-based model. However, since the connection relationship of each sub-model, that is, the internal structure of the system model 900 is given by structural information, the system model 900 can perform a normative analysis while using a data-based model and provides explanatory information. You can do it, and you can support design change and planning. If all of the parameters displayed in FIG. 9 are data that cannot be actually obtained in the target system, the system model 900 of FIG. 9 will be treated as one black box, so it should be interpreted that the present invention is not applied in this case. something to do.

Structural information of the present invention is set based on observation data required for each of the sub-models to learn by treating the behavior of sub-models inside the system model 900 as a black box. If there is no observation data of the parameters inside the system model 900, it is impossible to learn, and if there is no learning data, the internal behavior of the system model 900 cannot be determined.

Collected data (x1', x2', x3', x4', x5', x6', y2', y3', w1', w2', z') are input variable values, output variable values, or state It does not matter whether the value of a variable is a value, and whether it is a value that can be actually measured or observed with a sensor or the like is important.

10 is an operation flowchart illustrating a method of learning a system model 120 executed by the computing system 100 according to an embodiment of the present invention.

The operating method of the system model 120 and the learning method of the system model 120 executed in the computing system 100 according to an embodiment of the present invention include a plurality of sub-models 122a, 122b, 124 The system model 120 is executed on the computing system 100 including at least one or more processors 110.

The learning method of the system model 120 executed by the computing system 100 that implements and operates the system model 120 describing a target system according to an embodiment of the present invention is obtained by the communication interface 140 Receiving or calling data (S1010); Providing, by the at least one processor 110, first data (X1, X2, X3, X4) of the acquired data as an input of the first sub-models 122a and 122b (S1020); Providing, by at least one processor 110, second data (Y1, Y2) defined as outputs of the first sub-models 122a and 122b as input to the second sub-model 124 based on structural information (S1040); And at least one processor 110, the first sub-models (122a, 122b) to learn the relationship between the first data (X1, X2, X3, X4) and the second data (Y1, Y2) The third sub-models 122a and 122b are controlled (S1030), or the second sub-model 124 is defined as an output of the second sub-model 124 based on the second data Y1 and Y2 and structural information. And controlling the second sub-model 124 to learn the relationship between the data Z (S1050). When the first sub-models 122a and 122b are theory-based models and do not require training, step S1030 may be omitted, and when the second sub-model 124 is a theory-based model and does not require training, Step S1050 may be omitted. However, at least one of the first sub-models 122a and 122b and the second sub-model 1224 is a data-based model, and thus at least one of the steps S1030 and S1050 is Executed by

11 is an operational flow diagram illustrating a method of operating the system model 120 such that the system model 120 is executed by the computing system 100 and infers a target system according to an embodiment of the present invention.

According to an embodiment of the present invention, the operation method of the system model 120 executed by the computing system 100 that implements and operates the system model 120 describing a target system, the communication interface 140 Receiving or calling new input data for the system; Selecting, by at least one or more processors 110, first data (X1, X2, X3, X4) based on structural information among new input data and providing the first data as inputs of the first sub-models 122a and 122b; The first sub-model 122a, so that the at least one processor 110 can infer the operation of the target system based on the first data (X1, X2, X3, X4) of the first sub-models 122a and 122b. Controlling 122b); At least one processor 110 providing output data of the first sub-models 122a and 122b as second data Y1 and Y2 as an input of the second sub-model 124 based on structural information; Controlling, by at least one processor 110, the second sub-model 124 so that the second sub-model 124 can infer the operation of the target system based on the second data Y1 and Y2; And at least one processor 110 outputs the third data Z, which is generated based on the output data of the second sub-model 124 or is output data of the second sub-model 124, to the system model 120. And providing it to the user.

In this case, in the operating method of the system model 120 of the present invention, the at least one processor 110 interprets a user query regarding a target system and converts it into an instruction command that can be executed in the computing system 100 (shown Not) may be further included.

According to the present invention, it can have the following advantages over the existing data-based model. In general, a data-based model exhibits very high accuracy for static data, but it is known that it is not easy to apply variables that affect behavior to various dynamic analyses. According to the present invention, a data-based model can also be applied to a target system having a space variant, a time variant, and a stochastic feature.

As an example of the spatial transition, when the entire system model is expressed on a cell basis, the state transition function is not collectively expressed for all cells, but may be expressed differently according to individual cells. Therefore, the features of the terrain that vary depending on the location can be reflected for each cell. As an example of the time shift, the state transition function used to describe the system model may be set to be different according to the time of the morning, afternoon, or the like. For example, when there are mornings, afternoons, weekends, holidays, weekdays, and specific events in the traffic flow, a system model can be implemented by reflecting seasonal factors. Such variations can be further specified and optimized through learning of data-based models.

As an example of a probabilistic characteristic, in the case of a traffic simulation, a driver's behavior is not always determined, but may reflect a point that may appear probabilistically. Since the data-based model reflects only acquired data and learns, an event may occur with a specific probability under the condition learned by the data-based model. At this time, when the situation of the road network changes or a disaster occurs, the change of probability cannot be simulated in a pure data-based model, but in the system model of the present invention, the probabilistic system is based on the data connection relationship in the model and the structural information between the models. It can be simulated by adjusting or changing properties.

In addition, according to the present invention, non-linear characteristics of the target system can be reflected. Since many systems in the real world have nonlinear features without the overlapping theorem between inputs and outputs, it is a great advantage that the state transition function can take into account nonlinear features. In addition, additional situational information such as weather conditions can be expressed through data obtained from the actual system, making it robust to the realization of complex digital models such as digital twin models that mimic the real world.

The operating method of the system model and the learning method of the system model executed by the computing system according to an embodiment of the present invention are implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. have. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -A hardware device specially configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

However, the present invention is not limited or limited by the embodiments. The same reference numerals shown in each drawing indicate the same members. The length, height, size, width, etc. introduced in the embodiments and drawings of the present invention may be exaggerated to aid understanding.

As described above, in the present invention, specific matters such as specific components, etc., and limited embodiments and drawings have been described, but this is provided only to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. , If a person of ordinary skill in the field to which the present invention belongs, various modifications and variations are possible from these descriptions.

Accordingly, the spirit of the present invention is limited to the described embodiments and should not be defined, and all things that have equivalent or equivalent modifications to the claims as well as the claims to be described later belong to the scope of the spirit of the present invention. .

100: computing system
110: processor
120: system model
130: communication interface
140: user interface
122a, 122b: first sub-model
124: second sub-model

Claims (16)

In a computing system that implements and operates a model describing a target system,
The computing system
A communication interface capable of receiving or calling acquisition data obtained from the target system;
A system model including a plurality of sub-models; And
At least one processor; Including,
Each of the plurality of sub-models is a model capable of receiving part of the acquired data and inferring or predicting at least part of the remaining data as an output,
Input data and output data of each of the plurality of sub-models in the system model are defined based on structural information of the target system that can be obtained using knowledge of the target system,
The structural information includes information on a data connection relationship between a first sub-model and a second sub-model among the plurality of sub-models, and the data connection relationship includes output data of the first sub-model and the second sub-model. Contains information on whether it can be used as input data of,
The at least one or more processors
Among new input data for the target system, the first data is selected based on the structural information and provided as an input of the first sub-model, and output data of the first sub-model is provided as second data to the structural information. Based on the input of the second sub-model, and providing third data that is generated based on the output data of the second sub-model or that is the output data of the second sub-model as an output of the system model,
The first sub-model is controlled so that the first sub-model can infer the operation of the target system based on the first data, and the second sub-model operates the target system based on the second data. The computing system for controlling the second sub-model so that it can be inferred. The method of claim 1,
Each of the plurality of sub-models is defined based on obtainable domain knowledge, experience, and theory related to the target system, and any one of a theory-based model capable of deductive inference or a data-based model that is learned based on the acquired data Or, the theory-based model and the data-based model are combined to complement each other. The method of claim 1,
The above structural information is
Defined based on a data connection relationship between parameters in the target system constituting a theory-based primitive model that can be acquired using knowledge of the target system, and whether each of the parameters can be included in the acquired data Computing system. The method of claim 1,
The above structural information is
Information on a related parameter that can be selected as input data or output data of each of the plurality of sub-models among parameters in the target system constituting a theory-based original model that can be obtained using knowledge of the target system And wherein each of the related parameters includes sub-model data structure information defined as input data and output data of each of the plurality of sub-models based on the theory-based original model. The method of claim 1,
The computing system
A user interface for receiving a user query regarding the target system and transmitting it to the at least one processor;
Including more,
The at least one or more processors
A computing system that interprets the user query and converts it into instruction instructions that can be executed in the computing system. The method of claim 5,
The at least one or more processors
When the user query includes a query for at least two of the first data, the second data, and the third data, the first data and the first data are 2 data and at least one of a condition variable, a control variable, or a design variable of the target system corresponding to the association of at least two of the data and the third data,
A computing system for generating a response to the user query based on the corresponding search result and delivering the response to the user query to the user interface. The method of claim 5,
When the user query includes a prediction of the operation of the target system when the first data is out of the acquired data domain covered by the acquired data, the at least one processor Inputting the first data to control the first sub-model to predict the operation of the target system in response to the first data,
The at least one processor applies one of a theory-based model capable of deductive inference as the first sub-model or a data-based model that is learned based on the acquired data, or the theory-based model and the data-based model Computing systems that are combined to complement each other and applied to the first sub-model. The method of claim 5,
When the user query includes a condition in which the distribution of the third data is adjusted, the at least one processor
A first distribution of data applicable to the first data among the acquired data based on the structural information and the system model, a second distribution of second data related to the first distribution, and a second distribution related to the second distribution Generate a third distribution of the third data,
Based on the first distribution, the second distribution, the third distribution, and the structural information, a limiting condition of the range of the first distribution is provided in response to the user query, or a condition variable or a control variable of the target system Or, a computing system that provides the user with the possibility of changing at least one or more of the design variables. In a computing system that implements and operates a model describing a target system,
The computing system
A communication interface capable of receiving or calling acquisition data obtained from the target system;
A system model including a plurality of sub-models; And
At least one processor; Including,
Each of the plurality of sub-models is a model capable of receiving part of the acquired data and inferring or predicting at least part of the remaining data as an output,
Input data and output data of each of the plurality of sub-models in the system model are defined based on structural information of the target system that can be obtained using knowledge of the target system,
The structural information includes information on a data connection relationship between a first sub-model and a second sub-model among the plurality of sub-models, and the data connection relationship includes output data of the first sub-model and the second sub-model. Contains information on whether it can be used as input data of,
The at least one or more processors
Providing first data of the acquired data as an input of the first sub-model, and providing second data defined as an output of the first sub-model based on the structural information as an input of the second sub-model,
The first sub-model is controlled so that the first sub-model can learn a relationship between the first data and the second data, or the second sub-model is based on the second data and the structural information. 2 Computing system for controlling the second sub-model to learn a relationship between third data defined as an output of the sub-model. The method of claim 9,
Each of the plurality of sub-models is defined based on obtainable domain knowledge, experience, and theory related to the target system, and any one of a theory-based model capable of deductive inference or a data-based model that is learned based on the acquired data Or, the theory-based model and the data-based model are combined to complement each other. In the operating method of a system model executed by a computing system that implements and operates a model describing a target system,
The computing system
A communication interface capable of receiving or calling acquisition data obtained from the target system;
A system model including a plurality of sub-models; And
At least one processor; Including,
Each of the plurality of sub-models is a model capable of receiving part of the acquired data and inferring or predicting at least part of the remaining data as an output,
Input data and output data of each of the plurality of sub-models in the system model are defined based on structural information of the target system that can be obtained using knowledge of the target system,
The structural information includes information on a data connection relationship between a first sub-model and a second sub-model among the plurality of sub-models, and the data connection relationship includes output data of the first sub-model and the second sub-model. Contains information on whether it can be used as input data of,
The communication interface receiving or calling new input data for the target system;
Selecting, by the at least one processor, the first data based on the structural information among the new input data and providing the first data as an input of the first sub-model;
Controlling, by the at least one processor, the first sub-model so that the first sub-model can infer an operation of the target system based on the first data;
Providing, by the at least one processor, output data of the first sub-model as second data as an input of the second sub-model based on the structural information;
Controlling, by the at least one processor, the second sub-model so that the second sub-model can infer an operation of the target system based on the second data; And
Providing, by the at least one processor, third data generated based on output data of the second sub-model or output data of the second sub-model to a user as an output of the system model;
Operating method of the system model comprising a. The method of claim 11,
The computing system
A user interface for receiving a user query regarding the target system and transmitting it to the at least one processor;
Including more,
Converting, by the at least one processor, the user query into an instruction command that can be executed in the computing system;
The operating method of the system model further comprising. The method of claim 12,
When the user query includes a query for at least two relationships among the first data, the second data, and the third data, the at least one processor, based on the structural information and the data connection relationship, Generating a corresponding search result by searching for at least one or more of a condition variable, a control variable, or a design variable of the target system corresponding to at least two of the first data, the second data, and the third data. ;
Generating, by the at least one processor, a response to the user query based on the corresponding search result; And
Transmitting, by the at least one processor, a response to the user query to the user interface;
The method of operating a system model further comprising The method of claim 12,
When the user query includes a prediction of the operation of the target system when the first data is out of the acquired data domain covered by the acquired data, the at least one processor Controlling the first sub-model to predict the operation of the target system in response to the first data by inputting the first data; And
The at least one processor applies one of a theory-based model capable of deductive inference as the first sub-model, or a data-based model that is learned based on the acquired data, or the theory-based model and the data-based model Combining them to complement each other and applying them to the first sub-model;
The operating method of the system model further comprising. The method of claim 12,
When the user query includes a condition in which the distribution of the third data is adjusted, the at least one processor selects the data applicable to the first data from among the acquired data based on the structural information and the system model. Generating a first distribution, a second distribution of second data related to the first distribution, and a third distribution of the third data related to the second distribution; And
The at least one processor provides a limit condition of the range of the first distribution in response to the user query based on the first distribution, the second distribution, the third distribution, and the structural information, or the target system Providing the user with a possibility of changing at least one or more of a condition variable, a control variable, or a design variable;
The operating method of the system model further comprising. In the learning method of a system model executed by a computing system that implements and operates a model describing a target system,
The computing system
A communication interface capable of receiving or calling acquisition data obtained from the target system;
A system model including a plurality of sub-models; And
At least one processor; Including,
Each of the plurality of sub-models is a model capable of receiving part of the acquired data and inferring or predicting at least part of the remaining data as an output,
Input data and output data of each of the plurality of sub-models in the system model are defined based on structural information of the target system that can be obtained using knowledge of the target system,
The structural information includes information on a data connection relationship between a first sub-model and a second sub-model among the plurality of sub-models, and the data connection relationship includes output data of the first sub-model and the second sub-model. Contains information on whether it can be used as input data of,
Receiving or calling the acquisition data by the communication interface;
Providing, by the at least one processor, first data of the acquired data as an input of the first sub-model;
Providing, by the at least one processor, second data, which is defined as an output of the first sub-model based on the structural information, as an input of the second sub-model; And
The at least one processor controls the first sub-model so that the first sub-model can learn a relationship between the first data and the second data, or the second sub-model is the second data and the second data. Controlling the second sub-model to learn a relationship between third data defined as an output of the second sub-model based on structural information;
Learning method of a system model comprising a.

Images