JP2022035737A - Control system, control method, control device and program - Google Patents

Abstract

To provide a control system capable of obtaining a control parameter value of high explainability in the case of intervention to automatic control.SOLUTION: A control system 1 comprises: a creation unit 201 for creating a model representing a relation between a state of a controlled object 40 and a control parameter value through simulation learning on the basis of a history of the control parameter value in a case where an operator intervenes the controlled object; and a calculation unit 103 for calculating the control parameter value through the model in accordance with a state of the controlled object.SELECTED DRAWING: Figure 3

Description

The present invention relates to control systems, control methods, control devices and programs.

In various plants such as chemical plants, steelmaking plants, and energy plants, automatic control using PID (Proportional-Integral-Differential) control is widely performed. PID control is a simple but excellent automated control method, but it is known that depending on the state of the plant, a human operator often has to manually intervene in the control. For example, when it becomes difficult to bring the controlled object closer to the target by automatic control due to changes in the state of the plant or the influence of disturbance, the operator needs to manually intervene in the control as necessary while monitoring the sensor values. be.

Since an increase in operator intervention leads to an increase in work load and labor cost, it is desirable to reduce operator intervention. Therefore, in recent years, an automatic control method using reinforcement learning has attracted attention in order to reduce operator intervention. Reinforcement learning is an effective method for automatic control of complex systems, but it is difficult to apply it to an operating plant because random control is performed on the plant at the initial stage of learning, resulting in poor controllability. On the other hand, instead of performing automatic control, the optimum control parameter value when the plant is automatically controlled by reinforcement learning is learned, and the optimum control parameter value is given to the operator when intervention is required. It is also possible to make a suggestion.

Japanese Unexamined Patent Publication No. 2019-67238

However, since reinforcement learning models the optimal automatic control of the plant, the possibility of explaining the control parameter value proposed to the operator (that is, the possibility of explaining the judgment basis such as why the control parameter value was proposed). Was low. Therefore, it is difficult for the operator to determine whether the control parameter value is really the optimum value.

One embodiment of the present invention has been made in view of the above points, and an object of the present invention is to obtain control parameter values that are highly explainable when intervening in automatic control.

In order to achieve the above object, the control system according to the embodiment has the state of the controlled object and the control parameter value based on the history of the control parameter value when the operator intervenes in the controlled object. It has a creation unit that creates a model representing a relationship by imitation learning, and a calculation unit that calculates a control parameter value by the model according to the state of the control target.

It is possible to obtain control parameter values that are highly explainable when intervening in automatic control.

It is a figure which shows an example of the whole structure of the control system which concerns on this embodiment. It is a figure which shows an example of the hardware composition of the control device which concerns on this embodiment. It is a figure which shows an example of the functional structure of the control system which concerns on this embodiment. It is a flowchart which shows an example of the flow of the model creation process which concerns on this embodiment. It is a flowchart which shows an example of the flow of the control process which concerns on this embodiment.

Hereinafter, an embodiment of the present invention will be described. In this embodiment, a control system 1 capable of obtaining control parameter values that are highly explainable when intervening in automatic control of a controlled object (for example, various plants, various facilities, various devices, etc.) will be described. The control system 1 according to the present embodiment models the relationship between the state of the controlled object and the control parameter value when the operator intervenes in the past by imitation learning, which is one of the machine learning methods. This model (hereinafter, also referred to as “intervention model”) is used to obtain control parameter values at the time of intervention for automatic control. As a result, the same control parameter value as when the operator actually intervened in the past can be obtained, so that it is possible to obtain a control parameter value with high explainability. Therefore, the reliability of the control parameters proposed to the operator is ensured, and for example, it becomes possible to contribute to the stable operation of the plant.

The control parameter value is a parameter value for controlling the control target. For example, an operation amount (MV: Manipulative Variable) for the control target or a target value (SV: Set Variable) that affects the operation amount. And so on.

<Overall configuration>
First, the overall configuration of the control system 1 according to the present embodiment will be described with reference to FIG. FIG. 1 is a diagram showing an example of the overall configuration of the control system 1 according to the present embodiment.

As shown in FIG. 1, the control system 1 according to the present embodiment includes a control device 10, a server 20, an operator terminal 30, and a controlled object 40. The control device 10 and the server 20 are communicably connected via a communication network such as the Internet. Further, the control device 10, the operator terminal 30, and the controlled object 40 are communicably connected via a communication network such as a control network.

The control device 10 is a computer or a computer system that controls the controlled object 40. At this time, the control device 10 controls the control target 40 by an automatic control method such as PID control, which is one of the feedback controls. As the control device 10, for example, a PLC (Programmable Logic Controller), a DCS (Distributed Control System), or the like can be used.

Further, the control device 10 intervenes when the intervention of the operator is required (for example, when the state of the controlled object 40 (that is, the observed value (PV: Process Variable)) is about to deviate from the target). The control parameter value is calculated by the model and transmitted to the operator terminal 30. Thereby, the control parameter value is proposed to the operator who uses the operator terminal 30.

The server 20 is a computer or a computer system that creates an intervention model by imitation learning using the history of the operator's intervention in the past (hereinafter, also referred to as “intervention history”) and transmits it to the control device 10.

The operator terminal 30 is various terminals used by an operator who monitors or intervenes in the control of the controlled object 40. As the operator terminal 30, for example, a PC (personal computer), a tablet terminal, a smartphone, or the like can be used.

The control target 40 is various plants, various facilities, various devices, and the like controlled by the control device 10. The control target 40 is equipped with various sensors (for example, a temperature sensor, a flow meter, a pressure gauge, a densitometer, etc.), and observation values indicating the state of the control target 40 are transmitted to the control device 10 at each control cycle (for example, a temperature sensor, a flow meter, a pressure gauge, a densitometer, etc.). Feedback). The observed values are various sensor values (for example, temperature, flow rate, pressure, concentration of a specific component, etc.) representing the state of the controlled object 40, but other than these, the observed values include the state of the controlled object 40. Arbitrary information (for example, a photographed image of the controlled object 40, sound data obtained by recording the sound output from the controlled object 40, etc.) may be included.

The overall configuration of the control system 1 shown in FIG. 1 is an example, and may be another configuration. For example, the control system 1 does not include the server 20, and the control device 10 may create an intervention model.

<Hardware configuration>
Next, the hardware configuration of the control device 10 according to the present embodiment will be described with reference to FIG. FIG. 2 is a diagram showing an example of the hardware configuration of the control device 10 according to the present embodiment.

As shown in FIG. 2, the control device 10 according to the present embodiment is realized by a hardware configuration of a general computer or computer system, and includes an input device 11, a display device 12, an external I / F 13, and a communication I /. It has an F14, a processor 15, and a memory device 16. Each of these hardware is connected so as to be communicable via the bus 17.

The input device 11 is, for example, a keyboard, a mouse, a touch panel, or the like. The display device 12 is, for example, a display or the like. The control device 10 does not have to have at least one of the input device 11 and the display device 12.

The external I / F 13 is an interface with an external device. The external device includes a recording medium 13a and the like. The control device 10 can read or write the recording medium 13a via the external I / F 13. The recording medium 13a includes, for example, a CD (Compact Disc), a DVD (Digital Versatile Disk), an SD memory card (Secure Digital memory card), a USB (Universal Serial Bus) memory card, and the like.

The communication I / F 14 is an interface for connecting the control device 10 to the communication network. The processor 15 is, for example, various arithmetic units such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). The memory device 16 is, for example, various storage devices such as an HDD (Hard Disk Drive), an SSD (Solid State Drive), a RAM (Random Access Memory), a ROM (Read Only Memory), and a flash memory.

By having the hardware configuration shown in FIG. 2, the control device 10 according to the present embodiment can realize various processes described later. However, the hardware configuration shown in FIG. 2 is an example, and the control device 10 may have another hardware configuration. For example, the control device 10 may have a plurality of processors 15 or a plurality of memory devices 16.

Similarly, the server 20 and the operator terminal 30 are realized by the hardware configuration of a general computer or computer system, and the input device, the display device, the external I / F, the communication I / F, the processor, and the memory are realized. Has a device. However, the server 20 does not have to have at least one of an input device and a display device. Further, the server 20 and the operator terminal 30 may have a plurality of processors or may have a plurality of memory devices.

<Functional configuration>
Next, the functional configuration of the control system 1 according to the present embodiment will be described with reference to FIG. FIG. 3 is a diagram showing an example of the functional configuration of the control system 1 according to the present embodiment.

<< Control device 10 >>
As shown in FIG. 3, the control device 10 according to the present embodiment includes a control unit 101, an intervention determination unit 102, a calculation unit 103, a proposal unit 104, and a re-learning unit 105. Each of these parts is realized by a process of causing the processor 15 to execute one or more programs installed in the control device 10.

Further, the control device 10 according to the present embodiment has a storage unit 106. The storage unit 106 is realized by, for example, a memory device 16. The storage unit 106 may be realized by a storage device (for example, a database server or the like) connected to the control device 10 via a communication network.

The control unit 101 controls the control target 40 by an automatic control method such as PID control, or controls the control target 40 by the control parameter value calculated by the intervention model. That is, the control unit 101 transmits the operation amount calculated by the automatic control method using the observed value and the target value to the control target 40, and the operation amount based on the control parameter value calculated by the intervention model is sent to the control target 40. The control target 40 is controlled by transmitting or the like. Here, the manipulated variable based on the control parameter value is, for example, the manipulated variable itself when the control parameter value is the manipulated variable, and the observed value and the said when the control parameter value is the target value. It is the amount of operation calculated by the automatic control method using the target value.

The intervention determination unit 102 determines whether or not intervention is necessary for the automatic control. When intervention is required for automatic control, for example, the difference between the observed value indicating the state of the controlled object 40 and the target value exceeds a predetermined threshold value, or the observed value exceeds a predetermined threshold value (or). (Below), etc.

When the intervention determination unit 102 determines that intervention is necessary, the calculation unit 103 calculates the control parameter value by the intervention model using the observation value indicating the current state of the control target 40.

The proposal unit 104 transmits the control parameter value calculated by the calculation unit 103 to the operator terminal 30, and proposes the control parameter value to the operator.

The re-learning unit 105 determines the current state of the controlled object 40 according to whether or not the control parameter value proposed by the proposal unit 104 to the operator is adopted (that is, whether or not the intervention is performed with the control parameter value). The intervention model is relearned using the observed values indicating the above and the control parameter values.

The storage unit 106 stores the intervention model created by the server 20. In addition, the storage unit 106 also stores the intervention history when the operator intervened in the past. In addition, every time an intervention for automatic control is performed, the storage unit 106 stores an intervention history related to this intervention.

Here, each intervention history includes, for example, the date and time when the intervention was performed, the observed value indicating the state of the controlled object 40 at the time of the intervention, and the control parameter value at the time of the intervention. In addition to these, each intervention history may include, for example, the ID of the operator who performed the intervention (hereinafter, also referred to as “operator ID”), and information indicating the result of the intervention. (For example, the difference between the observed value and the target value in the next control cycle (or the control cycle after that)) may be included.

≪Server 20≫
As shown in FIG. 3, the server 20 according to the present embodiment has a model creation unit 201. The model creation unit 201 is realized, for example, by a process of causing a processor to execute one or more programs installed in the server 20.

Further, the server 20 according to the present embodiment has a storage unit 202. The storage unit 202 is realized by, for example, a memory device. The storage unit 202 may be realized by a storage device (for example, a database server or the like) connected to the server 20 via a communication network.

The model creation unit 201 creates (learns) an intervention model by imitation learning using a plurality of intervention histories stored in the storage unit 202. That is, the model creation unit 201 uses a plurality of intervention histories to model and observe the relationship between the observed value indicating the state of the controlled object 40 and the control parameter value of the intervention performed in the state by imitation learning. Create an intervention model that inputs values and outputs control parameter values. Then, the model creation unit 201 transmits the intervention model to the control device 10. It should be noted that imitation learning is one of the machine learning methods (particularly, one of the machine learning methods of the framework similar to the enhanced learning), and the environment (in the present embodiment) using the behavior history (intervention history in the present embodiment). Then, it is a method of learning the optimum behavior (control parameter value in this embodiment) for the observed value).

The storage unit 202 stores a plurality of intervention histories transmitted from the control device 10. These intervention histories are transmitted, for example, from the control device 10 when the intervention model is created.

The functional configuration of the control system 1 shown in FIG. 3 is an example, and may be another configuration. For example, when the intervention model is created by the control device 10, the control device 10 may have the model creation unit 201.

<Model creation process>
Next, the flow of the model creation process according to the present embodiment will be described with reference to FIG. FIG. 4 is a flowchart showing an example of the flow of the model creation process according to the present embodiment. The model creation process shown in FIG. 4 is executed before the control process described later. Hereinafter, it is assumed that the storage unit 202 of the server 20 stores a plurality of intervention histories transmitted from the control device 10.

The model creation unit 201 creates an intervention model by imitation learning using a plurality of intervention histories stored in the storage unit 202 (step S101). At this time, the model creation unit 201 may create an intervention model using all the intervention histories stored in the storage unit 202, or may select from a plurality of intervention histories stored in the storage unit 202. An intervention model may be created using some of the intervention histories. Here, when selecting the intervention history used for creating the intervention model, it may be selected by any method, but for example, it is conceivable to select by any of the following selection methods 1 to 3. Will be.

Selection method 1: The intervention history used for creating the intervention model is selected from the plurality of intervention histories stored in the storage unit 202 at the discretion of the operator (or the person in charge of creating the intervention model). This means that a human determines and selects "intervention history when a good intervention is performed" and "intervention history when a bad intervention is performed" from the past intervention history.

Selection method 2: After dividing a plurality of intervention histories stored in the storage unit 202 into predetermined periods, a predetermined statistical value (for example, an autocorrelation function value or a cross-correlation function value) is calculated in each period. , Select the intervention history used to create the intervention model based on these statistics. Specifically, for example, when the statistical value is an autocorrelation function value or a cross-correlation function value, an intervention history in which the autocorrelation function value or the cross-correlation function value is included in a period equal to or less than a predetermined threshold is selected. do it. This allows you to select an intervention history that is included in a period with (or without) correlation (autocorrelation or cross-correlation).

Selection method 3: Among a plurality of intervention histories stored in the storage unit 202, an intervention history including a specific operator ID is selected, or an intervention history other than an intervention history including a specific operator ID is selected. Or select. Specifically, for example, an intervention history including an operator ID of a skilled operator may be selected, or an intervention history other than an intervention history including an operator ID of an inexperienced operator may be selected. As a result, it is possible to select the intervention history of an operator who is good at determining the control parameter value at the time of intervention, or conversely, it is possible to exclude the intervention history of an operator who is not good at determining the control parameter value.

Then, the model creation unit 201 transmits the intervention model created in step S101 to the control device 10 (step S102). As a result, the intervention model is stored in the storage unit 106 of the control device 10.

As described above, the control system 1 according to the present embodiment uses the history of the actual intervention performed by the operator in the past, and the observed value indicating the state of the controlled object 40 at the time of the intervention and the control parameter at the time of the intervention. The relationship with the value is modeled by imitation learning. This makes it possible to model the same control rules as the operator at the time of intervention, and as will be described later, it is possible to reduce the burden on the operator when the need for intervention occurs, and it is highly explainable. It becomes possible to propose control parameter values to the operator.

<Control processing>
Next, the flow of the control process according to the present embodiment will be described with reference to FIG. FIG. 5 is a flowchart showing an example of the flow of control processing according to the present embodiment. The control process shown in FIG. 5 is repeatedly executed every control cycle. Hereinafter, the control process in a certain control cycle will be described. Further, from now on, it is assumed that the intervention model created by the server 20 is stored in the storage unit 106 of the control device 10.

The control unit 101 receives an observed value indicating the current state of the controlled object 40 (step S201).

The intervention determination unit 102 determines whether or not intervention is necessary from the observation value received in step S201 above (step S202). As described above, the case where intervention is required is, for example, the case where the difference between the observed value and the target value exceeds a predetermined threshold value or the case where the observed value exceeds (or falls below) a predetermined threshold value. And so on.

When it is determined in step S202 that intervention is not necessary, the control unit 101 transmits the manipulated variable calculated by the automatic control method using the observed value and the target value to the controlled object 40 (step S203). As a result, the control target 40 is controlled according to the operation amount.

On the other hand, when it is determined in step S202 that intervention is necessary, the calculation unit 103 calculates the control parameter value by the intervention model stored in the storage unit 106 (step S204). That is, the calculation unit 103 inputs the observation value received in step S201 to the intervention model, and calculates the control parameter value as its output.

Here, in addition to the above-mentioned control parameter value, the calculation unit 103 may create evidence information indicating the basis of the control parameter value. For example, the calculation unit 103 may create one or more of the following grounds information 1 to ground information 4.

Evidence information 1: Using the observed values indicating the current state of the controlled object 40 and the control parameter values, a plurality of intervention histories used for creating and relearning the intervention model are searched, and the search results are used as evidence information. Create as. As a result, for example, the operator ID included in the search result (that is, the operator ID that intervened with the same control parameter value when the control target 40 was in the same state in the past) or the like is given to the operator of the operator terminal 30. It will be possible to present. At this time, for example, when the intervention history includes information indicating the result of the intervention, the information indicating the result of the intervention can also be presented to the operator. The re-learning of the intervention model will be described later.

Evidence information 2: The information obtained by quantifying the operator ID (and the information indicating the result of the intervention) obtained in the above-mentioned evidence information 1 may be used as the evidence information. At this time, for example, depending on the skill level and experience of the operator, the higher the skill level or the more experience, the higher the value of the evidence information, and the lower the skill level or the less experience, the lower the value of the evidence information. good. Further, depending on the information indicating the result of the intervention, the value of the evidence information may be increased as the state of the controlled object 40 approaches the target, and the value of the evidence information may be decreased as the state of the controlled object 40 approaches the target.

Rationale information 3: The most recent N-1 (however, N is predetermined) among the observed value indicating the current state of the controlled object 40, the control parameter value, and the plurality of intervention histories stored in the storage unit 106. Using the natural number) intervention history, the cross-correlation function value with N intervention histories out of the multiple intervention histories used to create and relearn the intervention model was calculated as the degree of similarity, and then the most. The history of N interventions for which a high degree of similarity was obtained and the degree of similarity are used as evidence information. This makes it possible to present to the operator the past intervention history similar to the current state of the controlled object 40 and how similar it is. Instead of the above-mentioned cross-correlation function, the similarity may be calculated by the dynamic time warping method (DTW).

Evidence information 4: Using known factor visualization technology, information indicating which intervention history is the basis for judgment among a plurality of intervention histories used for creating and relearning an intervention model is created as evidence information. It should be noted that such a factor visualization technique is generally known as a technique for visualizing the judgment basis (factor) for the inference result of the machine learning model.

Following step S204, the proposal unit 104 transmits the control parameter value (and its basis information) calculated in step S204 to the operator terminal 30 (step S205). As a result, the control parameter value is proposed to the operator of the operator terminal 30. The operator terminal 30 that has received the control parameter value displays the control parameter value on the screen in an arbitrary form (for example, a numerical value, a graph, etc.), issues an alert, or blinks a warning light. You may. On the other hand, the operator operates the operator terminal 30 and returns to the control device 10 whether or not to adopt the control parameter value proposed by the control device 10. At this time, if the operator does not adopt the control parameter value, a new control parameter value different from the control parameter value is returned.

If the operator determines that intervention is unnecessary, the operator may operate the operator terminal 30 and return information indicating that intervention is unnecessary to the control device 10. In this case, the above step S203 is executed and automatic control is performed.

Next, when the operator terminal 30 returns the information indicating the adoption, the control unit 101 transmits the operation amount based on the control parameter value calculated in step S204 to the control target 40, and the operator terminal 30 fails. When the information indicating adoption and the new control parameter value are returned, the operation amount based on the new control parameter value is transmitted to the control target 40 (step S206). At this time, the control unit 101 has the observed value received in step S201 (that is, the observed value indicating the current state of the controlled object 40), the control parameter value calculated in step S204, or new control. An intervention history including parameter values is created and stored in the storage unit 106.

Subsequently, the re-learning unit 105 relearns the intervention model stored in the storage unit 106 according to the reply result (adopted or rejected) of the operator terminal 30 in step S205 (step S207). That is, the re-learning unit 105 relearns the intervention model by imitation learning using the observation value received in step S201 and the control parameter value calculated in step S204. At this time, the re-learning unit 105 relearns the intervention model so that a penalty is imposed if the reply result of the operator terminal 30 in step S205 is information indicating rejection. Such a penalty imposes a penalty on the evaluation of the objective function value when the operator terminal 30 returns information indicating rejection to the objective function used for creating and relearning the intervention model. (This is called a penalty item or a penalty item, etc.) can be added.

As described above, the control system 1 according to the present embodiment intervenes in the automatic control of the controlled object 40 by using the intervention model in which the history of the actual intervention performed by the operator in the past is modeled by imitation learning. Propose control parameter values to the operator when the need arises. At this time, the control system 1 according to the present embodiment can also present to the operator information indicating the basis for calculating the control parameter value by the intervention model. As a result, the burden on the operator can be reduced, and control parameter values that are highly explainable can be proposed to the operator.

<Modification example>
Hereinafter, a modification of the present embodiment will be described.

<< Modification 1 >>
In the present embodiment, the control parameter value calculated by the intervention model is proposed to the operator, but the operation amount based on the control parameter value may be transmitted to the control target 40 without proposing to the operator. That is, when it is determined that intervention is necessary for automatic control, the control target 40 may be controlled by the operation amount based on the control parameter value calculated by the intervention model.

Further, at this time, the value of the above-mentioned evidence information 2 and the degree of similarity of the above-mentioned evidence information 3 (these values and the degree of similarity may be referred to as “confidence” or the like) exceed a predetermined threshold value. The operation amount based on the control parameter value is transmitted to the control target 40 only in the case (that is, when the certainty is high and there is a high possibility that the control target 40 can be appropriately controlled by the control parameter value calculated by the intervention model). May be good.

<< Modification 2 >>
In the present embodiment, one intervention model is created and control parameter values are calculated by this intervention model. However, even if a plurality of intervention models are created and the intervention models are switched and used according to predetermined conditions. good. For example, an intervention model for nighttime and an intervention model for daytime may be created, and the intervention model may be switched according to the operating time zone of the controlled object 40. Similarly, for example, an intervention model may be created for each product type, and the intervention model may be switched according to the product manufactured by the controlled object 40. Further, for example, a plurality of intervention models may be created for each range (for example, a temperature range) in which the state of the control target 40 can be taken, and the intervention model may be switched according to the state of the control target 40.

The present invention is not limited to the above-described embodiment disclosed specifically, and various modifications and modifications, combinations with known techniques, and the like are possible without departing from the description of the scope of claims. be.

1 Control system 10 Control device 11 Input device 12 Display device 13 External I / F
13a Recording medium 14 Communication I / F
15 Processor 16 Memory device 17 Bus 20 Server 30 Operator terminal 40 Controlled object 101 Control unit 102 Intervention judgment unit 103 Calculation unit 104 Proposal unit 105 Re-learning unit 106 Storage unit 201 Model creation unit 202 Storage unit

Claims (11)

A creation unit that creates a model representing the relationship between the state of the controlled object and the control parameter value by imitation learning based on the history of the control parameter value when the operator intervenes in the controlled object.
A calculation unit that calculates control parameter values using the model according to the state of the control target,
Control system with. The control system according to claim 1, further comprising a proposal unit that proposes a control parameter value calculated by the calculation unit to the operator. The calculation unit creates evidence information representing the basis of the control parameter value calculated by the model.
The control system according to claim 2, wherein the proposal unit proposes the basis information to the operator in addition to the control parameter value. The second or third aspect of claim 2 or 3, further comprising a re-learning unit that relearns the model by imitation learning depending on whether or not an intervention is performed on the controlled object by the control parameter value proposed by the proposal unit. Control system. The control system according to claim 1, further comprising a control unit that controls the control target by the control parameter value calculated by the calculation unit. The calculation unit further calculates a predetermined index value related to the control parameter value calculated by the model.
The control system according to claim 5, wherein the control unit controls the control target by the control parameter value calculated by the calculation unit when the index value exceeds a predetermined threshold value. The creating unit creates a plurality of the models according to the time zone, the type of the product manufactured by the controlled object, or the range of values that the state of the controlled object can take.
The calculation unit calculates a control parameter value by one of a plurality of the models according to the state of the control target and the time zone or the type of the product manufactured by the control target. The control system according to any one of 1 to 6. The creation unit calculates a predetermined statistical value from the history at a predetermined period, selects a control parameter value to be used for creating the model from the calculated statistical value, and uses the selected control parameter value and the control parameter value. The control system according to any one of claims 1 to 7, wherein the model is created using the state of the controlled object at the time of intervention. A creation procedure for creating a model representing the relationship between the state of the controlled object and the control parameter value by imitation learning based on the history of the control parameter value when the operator intervenes in the controlled object.
A calculation procedure for calculating a control parameter value by the model according to the state of the controlled object, and
The control method that the computer performs. A creation unit that creates a model representing the relationship between the state of the controlled object and the control parameter value by imitation learning based on the history of the control parameter value when the operator intervenes in the controlled object.
A calculation unit that calculates control parameter values using the model according to the state of the control target,
Control device with. A creation procedure for creating a model representing the relationship between the state of the controlled object and the control parameter value by imitation learning based on the history of the control parameter value when the operator intervenes in the controlled object.
A calculation procedure for calculating a control parameter value by the model according to the state of the controlled object, and
A program that causes a computer to run.

Images