Search device, search method, and program

Abstract

This search device (100) comprises: a first data set update unit (111) that updates a first data set (121) indicating an objective variable y and a first explanatory variable in association with each other; a first prediction model generation unit (130) that generates a first prediction model (123) using the first data set (121); a search condition derivation unit (160) that derives a search condition for obtaining a target value of the objective variable y using the first prediction model (123); and a prediction unit (150) that predicts whether or not a characteristic value indicated by the objective variable y deviates from the target value when raw materials are processed continuously according to the search condition. The first data set update unit (111) adds a second explanatory variable to the first data set (121) if the characteristic value is predicted to deviate from the target value, and the first prediction model generation unit (130) updates the first prediction model (123) using the first data set (121) to which the second explanatory variable is added.

Description

Searching device, searching method, and program

[0001] The present disclosure relates to a searching device for searching conditions used in processing raw materials.

[0002] Material processing equipment, such as chemical reaction equipment and semiconductor processing equipment, produces products by synthesizing and processing raw materials according to processing conditions. The characteristic values ​​of the products vary depending on the various processing conditions of the material processing equipment. For example, characteristic values ​​of the products are obtained for each of a plurality of different processing conditions. To obtain a product having a target characteristic value, it is necessary to search for input parameters that will produce output parameters that indicate the target characteristic value from a plurality of input parameters that each represent a different value as a processing condition. However, it is not practical to actually measure or evaluate the output parameters for each of the input parameters. Therefore, machine learning techniques, such as neural networks, are used to search for optimal solutions for the input parameters.

[0003] Patent Literature 1 discloses a search device that generates a prediction model showing the relationship between input parameters and output parameters through machine learning and searches for an optimal solution based on this prediction model. Specifically, the search device of Patent Literature 1 calculates, based on the prediction model, predicted values ​​of input parameters that yield output parameters showing target characteristic values, and determines whether the actual measured values ​​corresponding to these predicted values ​​reach the target. The actual measured values ​​are the actual measured characteristic values ​​of a product obtained by processing a semiconductor processing device according to the predicted values. If the actual measured values ​​do not reach the target, the search device adds the predicted and actual measured values ​​to the training data as new input and output parameters, thereby updating the prediction model. This improves search efficiency. Furthermore, if the actual measured values ​​are not close to the target, the search device of Patent Literature 1 excludes from the search area the region where the combination of the actual measured value and the predicted value exists. This improves the accuracy of reaching the optimal solution.

[0004] JP 2019-40984 A

[0005] a search condition derivation unit that performs a search using the first prediction model to derive, as search conditions, values ​​of the first explanatory variables for obtaining a target value of the dependent variable; and a prediction unit that predicts whether the characteristic value indicated by the dependent variable will deviate from the target value at a time point later than the current time point when the raw materials are continuously processed according to the search conditions derived by the search condition derivation unit. When the prediction unit predicts that the characteristic value will deviate from the target value, the first dataset update unit updates the first dataset by adding a second explanatory variable to the first dataset, the second explanatory variable indicating a state of the raw materials. The first prediction model generation unit updates the first prediction model using the first dataset updated by adding the second explanatory variable.

[0006] These comprehensive or specific aspects may be realized as a system, a method, an integrated circuit, a computer program, or a computer-readable recording medium such as a CD-ROM, or may be realized as any combination of a system, a method, an integrated circuit, a computer program, and a recording medium. The recording medium may also be a non-transitory recording medium.

[0007] Further advantages and effects of one aspect of the present disclosure will become apparent from the specification and drawings. Such advantages and / or effects are provided by some of the embodiments and configurations described in the specification and drawings, but not necessarily all of the configurations.

[0008] FIG. 1 is a diagram showing an example of a processing system according to the first embodiment. FIG. 2 is a block diagram showing an example of the configuration of a material processing apparatus according to the first embodiment. FIG. 3 is a block diagram showing an example of the configuration of a searching apparatus according to the first embodiment. FIG. 4 is a diagram showing examples of a first dataset and a second dataset according to the first embodiment. FIG. 5 is a diagram showing an example of a first dataset updated by adding a second explanatory variable according to the first embodiment. FIG. 6 is a flowchart showing an example of the processing operation of the searching apparatus according to the first embodiment. FIG. 7 is a block diagram showing an example of the configuration of a material processing apparatus according to the second embodiment. FIG. 8 is a block diagram showing an example of the configuration of the searching apparatus according to the second embodiment. FIG. 9 is a diagram showing an example of a first dataset updated by adding a second explanatory variable and a third explanatory variable according to the second embodiment. FIG. 10 is a flowchart showing an example of the processing operation of the searching apparatus according to the second embodiment.

[0009] (Findings that form the basis of the present invention) The present inventor has found that the above-mentioned Patent Document 1, described in the "Background Art" section, has the following problem of possibly reducing search accuracy.

[0010] The search device of Patent Document 1 searches for an optimal solution without considering changes in raw materials over time. Therefore, when raw materials that change significantly over time are used to produce a product, the search device of Patent Document 1 decreases the accuracy of the search for the optimal solution over time.

[0011] In other words, even if the search device of Patent Document 1 searches for processing conditions (i.e., input parameters) that are considered to be optimal solutions, if the material processing device continuously produces a product under those processing conditions using raw materials that change significantly over time, the characteristic values ​​of the product produced at that time may deviate significantly from the target as time passes. In other words, in a continuous process in which the production of a product is continuously performed, if raw materials that change significantly over time are used, the search accuracy for the optimal solution decreases.

[0012] Microreactors, which have a flow channel, are known as material processing devices for continuous processing. Microreactors involve contacting and mixing mutually soluble liquids (i.e., reactants) to react with each other and produce a desired product. Because products are produced in a small reaction space, microreactors tend to produce high-quality products with minimal variation. However, when producing large quantities of products, microreactors require the raw materials to be used for a long period of time. Consequently, if the physical properties of the raw materials change significantly over time, the properties of the resulting products may also change significantly over time. Therefore, even if a continuous process is initiated under optimal processing conditions, it is difficult to produce a large quantity of a product with the desired properties.

[0013] The present disclosure provides a search device that can improve search accuracy. a search condition derivation unit that performs a search using the first prediction model to derive, as search conditions, values ​​of the first explanatory variables for obtaining a target value of the dependent variable; and a prediction unit that predicts whether the characteristic value indicated by the dependent variable will deviate from the target value at a time point later than the current time point when the raw materials are continuously processed according to the search conditions derived by the search condition derivation unit. When the prediction unit predicts that the characteristic value will deviate from the target value, the first dataset update unit updates the first dataset by adding a second explanatory variable to the first dataset, the second explanatory variable indicating a state of the raw materials. The first prediction model generation unit updates the first prediction model using the first dataset updated by adding the second explanatory variable. The first explanatory variables and search conditions are also called processing conditions.

[0014] The search device disclosed herein can improve search accuracy. Specifically, when search conditions are derived as an optimal solution and raw materials are continuously processed according to the search conditions, i.e., during the continuous process stage, it is predicted whether the characteristic value of the raw materials will deviate from the target value at a later time point. Therefore, if the raw materials change over time, it is predicted that the characteristic value will deviate from the target value. In such a case, the first prediction model is updated using a first dataset to which a second explanatory variable indicating the state of the raw materials has been added. As a result, if a search is performed again using the updated first prediction model, it is possible to perform a search that takes the second explanatory variable into account, i.e., a search that takes into account the change in the raw materials over time. Therefore, if raw materials are continuously processed according to the new search conditions derived by the search, it is possible to increase the likelihood that the characteristic value will remain at the target value even at a later time point. This improves the search accuracy for an optimal solution, even when raw materials that change significantly over time are used to produce a product, and enables continuous mass production of a product having the target characteristic value. The optimal solution is a search condition, a processing condition, or a value of the first explanatory variable for generating a product having a characteristic value that is a target value.

[0015] The search device according to a second aspect may further include a second dataset updater that acquires the second explanatory variables and the dependent variable when processing of the raw materials is continuously performed according to the search conditions derived by the search condition derivation unit, and generates or updates a second dataset that associates the acquired second explanatory variables with the dependent variable; and a second prediction model generator that uses the second dataset to generate a second prediction model that indicates a temporal relationship between the second explanatory variables and the dependent variable. The prediction unit may use the second prediction model to predict whether the characteristic value deviates from the target value. For example, the second dataset associates an actually measured second explanatory variable with the dependent variable. Note that the second aspect may be dependent on the first aspect.

[0016] This makes it possible to predict with high accuracy whether or not the characteristic value will deviate from the target value at a time point later than the present time.

[0017] In the searching device according to a third aspect, the first dataset update unit may, in adding the second explanatory variable to the first dataset, integrate the second dataset into the first dataset to update the first dataset to a dataset indicating the first explanatory variable and the second explanatory variable in association with the objective variable. Note that the third aspect may be dependent on the second aspect.

[0018] This allows the first explanatory variable and the second explanatory variable to be appropriately associated with the dependent variable, and the second explanatory variable to be effectively added to the first dataset.

[0019] In addition, in the searching device according to a fourth aspect, when the prediction unit predicts that the characteristic value will deviate from the target value when the state of the raw material indicated by the second explanatory variable is in a predetermined state, the search condition derivation unit may perform a search using the updated first prediction model to derive, as a new search condition, a value of the first explanatory variable for obtaining the target value of the objective variable when the state of the raw material is in the predetermined state. Note that the fourth aspect may be dependent on the second or third aspect.

[0020] This means that even if the raw materials change over time and reach a predetermined state, a product with the target characteristic values ​​can be produced by having the material processing device perform processing according to the new search conditions.

[0021] In a fifth aspect of the present invention, the first dataset update unit may, when adding the second explanatory variable to the first dataset, update the first dataset by adding a third explanatory variable to the first dataset together with the second explanatory variable for adjusting a state of the raw materials, and the first prediction model generation unit may update the first prediction model using the first dataset updated by adding the second explanatory variable and the third explanatory variable. Note that the fifth aspect may be dependent on any one of the second to fourth aspects.

[0022] For example, in a material processing apparatus, the state of the raw material is adjusted according to the third explanatory variable. Thus, by performing a search using the first prediction model, it is possible to perform a search that takes into account the time-varying state of the raw material and adjustments to that state. As a result, it is possible to derive with high accuracy an optimal solution for the search conditions, including the values ​​of the first explanatory variable and the third explanatory variable, for producing a product having a target characteristic value.

[0023] Hereinafter, the embodiments will be specifically described with reference to the drawings.

[0024] The embodiments described below are all comprehensive or specific examples. The numerical values, shapes, materials, components, component placement and connection configurations, steps, and step order shown in the following embodiments are merely examples and are not intended to limit the present disclosure. Furthermore, among the components in the following embodiments, components that are not described in the independent claims that represent the highest concepts are described as optional components. Furthermore, each drawing is a schematic diagram and is not necessarily an exact illustration. Furthermore, the same components are designated by the same reference numerals in each drawing.

[0025] First Embodiment FIG. 1 is a diagram showing an example of a processing system according to this embodiment.

[0026] The processing system 1000 of this embodiment can improve the accuracy of searching for an optimal solution, even when raw materials that change significantly over time are used to produce a product, and can continuously produce a large amount of a product having target characteristic values. Furthermore, the processing phase of the processing system 1000 can be switched between a search phase and a continuous process phase. The search phase is a phase in which processing conditions (i.e., search conditions) for processing raw materials are searched for, and the continuous process phase is a phase in which raw materials are continuously processed according to the search conditions, i.e., a continuous process is performed. Such a processing system 1000 includes a search device 100 and a material processing device 200.

[0027] The material processing apparatus 200 is an apparatus that processes raw materials to produce a product. For example, the material processing apparatus 200 performs flow synthesis by flowing multiple types of liquids (e.g., a first liquid and a second liquid) that are raw materials through a flow path and sequentially reacting the liquids within the flow path. The liquids flowing within the flow path may be heated or cooled. A product is produced by such flow synthesis. The material processing apparatus 200 in this embodiment may also be referred to as a reactor or a chemical reaction apparatus. The material processing apparatus 200 in this embodiment is not limited to the example shown in FIG. 1 , and may be an apparatus having another configuration, such as a semiconductor processing apparatus, as long as it processes raw materials as a continuous process.

[0028] In the search stage, the search device 100 searches for processing conditions for the material processing apparatus 200 to obtain target characteristic values ​​(also called physical property values) of the product. That is, the search device 100 searches for optimal solutions for the processing conditions. The processing conditions are expressed as first explanatory variables or input parameters, and the characteristic values ​​of the product are expressed as response variables or output parameters. The search device 100 inputs or sets first explanatory variables to the material processing apparatus 200 and acquires, from the material processing apparatus 200, response variables obtained by processing raw materials according to the first explanatory variables. The search device 100 then generates a first prediction model using the known first explanatory variables and response variables, and uses the first prediction model to derive processing conditions for obtaining the target characteristic values, i.e., the value of the first explanatory variable, as an optimal solution. The value of the first explanatory variable and the processing conditions are also called search conditions. Then, in the continuous process stage, the search device 100 sets the value of the first explanatory variable (i.e., the processing conditions) derived as the optimal solution in the material processing device 200, and causes the material processing device 200 to perform processing on the raw materials according to the value of the first explanatory variable as a continuous process.

[0029] Furthermore, during the continuous process phase in which the material processing apparatus 200 is performing a continuous process, the searching apparatus 100 continuously acquires from the material processing apparatus 200 a response variable obtained by processing raw materials according to the value of the first explanatory variable and a second explanatory variable indicating the state of the raw materials. The searching apparatus 100 then uses the known second explanatory variable and response variable to predict whether the characteristic value of the product indicated by the response variable will deviate from the target value at a later time point than the current time point. If the searching apparatus 100 predicts that the characteristic value of the product will deviate from the target value, the continuous process phase is interrupted, and the processing phase of the processing system 1000 transitions from the continuous process phase to the search phase. During the search phase, the searching apparatus 100 then performs a search for processing conditions using the second explanatory variable.

[0030] In a specific example, the search device 100 is configured as a personal computer or a server, and is connected to input devices such as a keyboard and a mouse, and a display device such as a display.

[0031] FIG. 2 is a block diagram showing an example of the configuration of a material processing apparatus 200 according to this embodiment.

[0032] The material processing apparatus 200 in this embodiment is a reaction apparatus that produces a liquid mixture as a product by mixing and synthesizing a first liquid and a second liquid that are soluble in each other. The material processing apparatus 200 includes a first liquid supply unit 210, a second liquid supply unit 220, a mixing channel unit 230, a first explanatory variable setting unit 241, a second explanatory variable acquisition unit 242, a target variable acquisition unit 250, and a recovery unit 260. In FIG. 2, the flow of the liquid is indicated by solid arrows.

[0033] The first liquid supply unit 210 supplies the first liquid to the mixing channel unit 230 via the first liquid channel 213. Specifically, the first liquid supply unit 210 has a first liquid container 211 and a first liquid pump 212. The first liquid container 211 is a container that stores the first liquid. The first liquid pump 212 pumps out the first liquid from the first liquid container 211 and pumps it to the mixing channel unit 230 via the first liquid channel 213.

[0034] The second liquid supply unit 220 supplies the second liquid to the mixing channel unit 230 via the second liquid channel 223. Specifically, the second liquid supply unit 220 has a second liquid container 221 and a second liquid pump 222. The second liquid container 221 is a container that stores the second liquid. The second liquid pump 222 pumps out the second liquid from the second liquid container 221 and pumps it to the mixing channel unit 230 via the second liquid channel 223.

[0035] The mixing channel section 230 has a mixing channel for sequentially mixing and flowing the first liquid flowing in the first liquid channel 213 and the second liquid flowing in the second liquid channel 223. The first liquid and the second liquid are mixed to produce a mixed liquid, which flows through the mixing channel. The mixed liquid is also called a synthesized product, a product, or a product. The mixing channel section 230 may have a channel configured by, for example, a microreactor and a micromixer. In this case, the diameter of the channel is several μm to several mm, and the material processing apparatus 200 can be said to be a microflow reaction apparatus for performing microflow synthesis.

[0036] The first explanatory variable setting unit 241 sets the value of the first explanatory variable derived by the searching device 100 to at least one of the first liquid supply unit 210, the second liquid supply unit 220, the first liquid flow path 213, the second liquid flow path 223, and the mixing flow path unit 230. As a result, the material processing apparatus 200 processes the raw materials by mixing the first liquid and the second liquid according to the first explanatory variable. The first explanatory variable indicates, for example, the heating temperature, the raw material concentration, the raw material flow rate, etc. Note that the first explanatory variable is not limited to these and may indicate other numerical values. The heating temperature is, for example, at least one of the set temperature for heating the first liquid by the first liquid supply unit 210 and the set temperature for heating the second liquid by the second liquid supply unit 220. In this case, the first liquid supply unit 210 and the second liquid supply unit 220 may be equipped with a thermostatic bath or the like capable of controlling the temperatures of the first liquid and the second liquid according to the first explanatory variable. Alternatively, the heating temperature may be, for example, a set temperature for heating the mixed liquid. In this case, the mixing channel section 230 may be equipped with a thermostatic bath or the like capable of controlling the temperature of the mixed liquid in accordance with the first explanatory variable. The concentration of the raw material may be, for example, the concentration of at least one of the first liquid and the second liquid. In this case, the first liquid supply section 210 and the second liquid supply section 220 may have a mechanism capable of adjusting the concentration in accordance with the first explanatory variable. The flow rate of the raw material may be, for example, the flow rate of at least one of the first liquid and the second liquid. In this case, the first liquid supply section 210 and the second liquid supply section 220 may adjust the flow rate of the supplied liquid in accordance with the first explanatory variable. The mixing ratio of the first liquid and the second liquid may be adjusted by adjusting the concentration or flow rate of the raw material.

[0037] The second explanatory variable acquisition unit 242 has a sensor and acquires second explanatory variables indicating the state of the raw materials by using the sensor. Specifically, the second explanatory variable acquisition unit 242 acquires second explanatory variables by measuring the molecular weight, pH, viscosity, temperature, etc. of the first liquid and the second liquid contained in the first liquid container 211 and the second liquid container 221. Note that the second explanatory variables are not limited to these and may indicate other numerical values.

[0038] The objective variable acquisition unit 250 has a sensor and acquires objective variables indicating characteristic values ​​of the product, which is a mixed liquid, by using the sensor. For example, the characteristic values ​​may be the particle size, particle size distribution, absorbance, etc. of the product, but are not limited to these and may be other numerical values. The acquired objective variables indicate values ​​actually measured by the above-mentioned sensor.

[0039] The recovery section 260 recovers the product, which is a mixed liquid.

[0040] 2 mixes two types of liquids, the number of types of liquids to be mixed is not limited to two and may be three or more. For example, if the number of types of liquids is three or more, the material processing apparatus 200 is provided with liquid supply units equal to the number of types of liquids. These liquid supply units have the same configuration as the first liquid supply unit 210 and the second liquid supply unit 220, and supply the type of liquid corresponding to the liquid supply unit to the mixing channel unit 230.

[0041] In addition, although the main component of the first liquid and the second liquid in this embodiment is water, the first liquid and the second liquid may be either water-soluble or water-insoluble as long as they are soluble in each other. For example, the first liquid and the second liquid may be water or an aqueous solution, or may be an organic solvent or an oil-based liquid.

[0042] FIG. 3 is a block diagram showing an example of the configuration of the searching device 100 according to this embodiment.

[0043] The search device 100 includes a first dataset update unit 111, a second dataset update unit 112, a storage unit 120, a first prediction model generation unit 130, a second prediction model generation unit 140, a prediction unit 150, a search condition derivation unit 160, and a device control unit 170.

[0044] In the search stage, the first dataset update unit 111 acquires first explanatory variables from the apparatus control unit 170 and acquires, from the objective variable acquisition unit 250 of the material processing apparatus 200, objective variables obtained by mixing according to the first explanatory variables. That is, the first dataset update unit 111 acquires combinations including values ​​indicated by the first explanatory variables and values ​​indicated by the objective variables. The values ​​indicated by the first explanatory variables are also referred to as values ​​of the first explanatory variables, and the values ​​indicated by the objective variables are also referred to as values ​​of the objective variables. The first dataset update unit 111 then generates or updates a first dataset 121 including the combination. That is, if the first dataset 121 is not stored in the storage unit 120, the first dataset update unit 111 generates a first dataset 121 including the combination and stores it in the storage unit 120. On the other hand, if the first dataset 121 is already stored in the storage unit 120, the dataset update unit 110 reads the existing first dataset 121 from the storage unit 120 and updates the first dataset 121 stored in the storage unit 120 by including the acquired combination in the existing first dataset 121. Such a first dataset 121 includes one or more of the above-mentioned combinations, thereby indicating the first explanatory variable and the target variable in association with each other.

[0045] In this way, the first dataset update unit 111 in this embodiment generates or updates the first dataset 121 that associates the first explanatory variables with the objective variables that indicate the characteristic values ​​of the products obtained by processing raw materials by the material processing apparatus 200 according to the first explanatory variables.

[0046] The storage unit 120 is a recording medium for storing the first data set 121 and the like. The storage unit 120 in this embodiment has a storage capacity for storing not only the first data set 121 but also the second data set 122, the first prediction model 123, and the second prediction model 124. For example, the storage unit 120 is a hard disk drive, a random access memory (RAM), a read-only memory (ROM), or a semiconductor memory. Note that such a storage unit 120 may be volatile or non-volatile.

[0047] In the search stage, the first prediction model generation unit 130 reads the first data set 121 from the storage unit 120 and uses the first data set 121 to generate a first prediction model 123 that indicates a relationship between a first explanatory variable and a dependent variable. The first prediction model 123 is, for example, a model generated by machine learning so that, in response to an input value of a dependent variable, the first prediction model 123 outputs a value of the first explanatory variable that results in the input value of the dependent variable. In other words, the first prediction model generation unit 130 generates the first prediction model 123 by machine learning using the first data set 121. The first prediction model 123 is, for example, a neural network. The first prediction model generation unit 130 stores the generated first prediction model 123 in the storage unit 120.

[0048] In the search stage, the search condition derivation unit 160 reads the first prediction model 123 from the storage unit 120 and performs a search using the first prediction model 123 to derive, as search conditions, values ​​of the first explanatory variables for obtaining a target value of the objective variable. The search condition derivation unit 160 then outputs the derived values ​​of the first explanatory variables, which are the search conditions, to the apparatus control unit 170. The search conditions are used as conditions for processing by the material processing apparatus 200, i.e., the processing conditions described above.

[0049] The apparatus control unit 170 controls the material processing apparatus 200. Specifically, when the apparatus control unit 170 acquires the value of the first explanatory variable from the search condition derivation unit 160 as a processing condition, it outputs the value of the first explanatory variable to the material processing apparatus 200 and causes the material processing apparatus 200 to perform mixing according to the value of the first explanatory variable. Note that the value of the first explanatory variable, which is the processing condition, is also output from the apparatus control unit 170 to the first dataset update unit 111.

[0050] As a result, the material processing apparatus 200 outputs a new value of the response variable obtained according to the derived value of the first explanatory variable to the first data set update unit 111. Specifically, the first explanatory variable setting unit 241 of the material processing apparatus 200 sets the value of the first explanatory variable. Then, the response variable acquisition unit 250 of the material processing apparatus 200 outputs a new value of the response variable obtained by processing according to the value of the first explanatory variable to the first data set update unit 111.

[0051] The first dataset update unit 111 updates the first dataset 121 using the new value of the objective variable in the same manner as described above. If the new value of the objective variable reaches the target value, the search for the search conditions is interrupted. That is, as a result of the search, the value of the explanatory variable derived immediately before is determined as the optimal solution. This stops the search phase, and the processing system 1000 transitions from the search phase to the continuous process phase. For example, the device control unit 170 causes the material processing device 200 to start the continuous process.

[0052] The second dataset update unit 112 continuously acquires the second explanatory variables and the response variable from the second explanatory variable acquisition unit 242 and the response variable acquisition unit 250 of the material processing apparatus 200 during the continuous process. That is, the second dataset update unit 112 acquires combinations including values ​​indicated by the second explanatory variables and values ​​indicated by the response variables. The values ​​indicated by the second explanatory variables are also referred to as values ​​of the second explanatory variables. The second dataset update unit 112 may periodically acquire the combinations. Then, the second dataset update unit 112 generates or updates a second dataset 122 including the combinations. That is, if the second dataset 122 is not stored in the storage unit 120, the second dataset update unit 112 generates a second dataset 122 including the combination and stores it in the storage unit 120. On the other hand, if the second dataset 122 is already stored in the storage unit 120, the dataset update unit 110 reads the existing second dataset 122 from the storage unit 120 and updates the second dataset 122 stored in the storage unit 120 by including the acquired combination in the existing second dataset 122. Such a second dataset 122 includes one or more of the above-mentioned combinations, thereby indicating the association between the second explanatory variable and the target variable.

[0053] In this way, the second dataset update unit 112 in this embodiment acquires the second explanatory variables and the dependent variable when processing of raw materials is continuously performed according to the search conditions derived by the search condition derivation unit 160, i.e., during the continuous process stage, and generates or updates the second dataset 122 that shows the acquired second explanatory variables and dependent variable in association with each other.

[0054] The second prediction model generation unit 140 reads the second data set 122 from the storage unit 120 in the continuous process stage and uses the second data set 122 to generate a second prediction model 124 that indicates the temporal relationship between the second explanatory variable and the dependent variable. The second prediction model 124 is a model generated by machine learning, for example, so that, in response to an input of a time later than the current time, the second prediction model 124 outputs the respective values ​​of the second explanatory variable and the dependent variable at that time. In other words, the second prediction model generation unit 140 generates the second prediction model 124 by machine learning using the second data set 122. The second prediction model 124 is, for example, a neural network. The second prediction model generation unit 140 stores the generated second prediction model 124 in the storage unit 120.

[0055] During the continuous process stage, the prediction unit 150 reads the second prediction model 124 from the storage unit 120 and uses the second prediction model 124 to predict the values ​​of the second explanatory variable and the dependent variable at a time point later than the current time. The prediction unit 150 then predicts whether the value of the dependent variable (i.e., the characteristic value) at a time point later than the current time will deviate from the target value. That is, during the continuous process stage, when raw materials are continuously processed according to the search conditions derived by the search condition derivation unit 160, the prediction unit 150 in this embodiment predicts whether the characteristic value indicated by the dependent variable will deviate from the target value at a time point later than the current time. At this time, the prediction unit 150 uses the second prediction model 124 to predict whether the characteristic value will deviate from the target value.

[0056] When the prediction unit 150 predicts that the characteristic value will deviate from the target value, it causes the material processing apparatus 200 to stop the continuous process via the apparatus control unit 170, and causes the first data set update unit 111 to update the first data set 121. As a result, the processing stage of the processing system 1000 transitions from the continuous process stage to the search stage.

[0057] The first dataset update unit 111 updates the first dataset 121 by adding a second explanatory variable to the first dataset 121 in accordance with the prediction result by the prediction unit 150. That is, when the prediction unit 150 predicts that the characteristic value will deviate from the target value, the first dataset update unit 111 updates the first dataset 121 by adding a second explanatory variable indicating the state of the raw materials to the first dataset 121. Then, the first prediction model generation unit 130 updates the first prediction model 123 using the first dataset 121 updated by adding the second explanatory variable.

[0058] As a result, the first prediction model 123 is updated to a prediction model that indicates the relationship between the first and second explanatory variables and the dependent variable. As a result, the search condition derivation unit 160 performs a new search using the updated first prediction model 123, thereby deriving, as a new search condition, the value of the first explanatory variable for obtaining the target value of the dependent variable.

[0059] FIG. 4 is a diagram showing an example of the first data set 121 and the second data set 122. As shown in FIG.

[0060] 4A, the first data set 121 shows an explanatory variable X1, an explanatory variable X2, and a response variable y in association with one another. The explanatory variables X1 and X2 are each a first explanatory variable. In a specific example, the first data set 121 shows a value "X1a" of the explanatory variable X1, a value "X2a" of the explanatory variable X2, and a value "30" of the response variable y in association with one another.

[0061] For example, the search condition derivation unit 160 derives the value "X1b" of the explanatory variable X1 and the value "X2b" of the explanatory variable X2 as search conditions. In this case, the first dataset update unit 111 adds the value "X1b" of the explanatory variable X1 and the value "X2b" of the explanatory variable X2 to the first dataset 121. Then, the material processing apparatus 200 processes raw materials according to the search conditions expressed as the value "X1b" of the explanatory variable X1 and the value "X2b" of the explanatory variable X2, and as a result, the value "65" of the objective variable y is obtained as an actual measured value. When the first dataset update unit 111 acquires the value "65" of the objective variable y from the material processing apparatus 200, it adds the value "65" of the objective variable y to the first dataset 121 in association with the value "X1b" of the explanatory variable X1 and the value "X2b" of the explanatory variable X2. As a result, a combination including the value “X1b” of the explanatory variable X1, the value “X2b” of the explanatory variable X2, and the value “65” of the objective variable y is added to the first data set 121.

[0062] Furthermore, when the target value is equal to or greater than a threshold, the first dataset update unit 111 compares the value of the objective variable y with the threshold, and adds the comparison result as a judgment result to the first dataset 121 in association with the value of the objective variable y. In the example of FIG. 4A, the judgment result "good" indicates that the value of the objective variable y is equal to or greater than the threshold, and the judgment result "poor" indicates that the value of the objective variable y is less than the threshold. In this embodiment, the threshold is, for example, 70.

[0063] For example, each time a combination described above is added to the first data set 121, the first prediction model generation unit 130 generates or updates the first prediction model 123 using all combinations included in the first data set 121. The search condition derivation unit 160 uses the first prediction model 123 thus generated or updated to derive, as the next search conditions, values ​​of the explanatory variables X1 and X2 for obtaining a target value of the objective variable y. That is, the search condition derivation unit 160 derives, as the next search conditions, values ​​of the explanatory variables X1 and X2 that cause the value of the objective variable y to reach the target value. This derivation of search conditions is repeatedly performed. As a result, for example, a value of "80" for the objective variable y that is equal to or greater than the threshold is obtained for a search condition represented by the value "X1o" of the explanatory variable X1 and the value "X2o" of the explanatory variable X2. A judgment result of "good" is associated with the value "80" of the objective variable y. Furthermore, for the search conditions expressed by the value "X1p" of explanatory variable X1 and the value "X2p" of explanatory variable X2, a value of "85" of objective variable y, which is equal to or greater than the threshold, is obtained. This value of objective variable y, "85," is also associated with the judgment result "good."

[0064] Here, the search condition derivation unit 160 notifies the apparatus control unit 170 that the search for search conditions will be suspended because, for example, at time t1, a value of "85" for the objective variable y, which is equal to or greater than "threshold + 15," has been added to the first data set 121. In response to this notification, the apparatus control unit 170 causes the material processing apparatus 200 to execute a continuous process according to the search conditions associated with the value of the objective variable y, "85," i.e., a continuous process according to the value of the explanatory variable X1, "X1p," and the value of the explanatory variable X2, "X2p." Note that in this embodiment, when a value of the objective variable y equal to or greater than "threshold + 15," is added to the first data set 121, the search is suspended. However, the value added to the threshold is not limited to 15 and may be any other value equal to or greater than 0.

[0065] 4B, the second data set 122 shows the explanatory variable A1, the explanatory variable A2, and the objective variable y in association with each other. The explanatory variables A1 and A2 are each a second explanatory variable. In a specific example, the second data set 122 shows the value "A1a" of the explanatory variable A1, the value "A2a" of the explanatory variable A2, and the value "85" of the objective variable y in association with each other.

[0066] For example, in the continuous process stage, i.e., when the material processing apparatus 200 starts the continuous process, the second dataset update unit 112 acquires from the material processing apparatus 200 a combination including the values ​​of the explanatory variables A1 and A2 and the value of the response variable y. The value of the response variable y is a characteristic value of a product obtained by processing the material processing apparatus 200 in accordance with the value "X1p" of the explanatory variables X1 and the value "X2p" of the explanatory variables X2, with respect to the state of the raw materials represented by the values ​​of the explanatory variables A1 and A2. For example, the second dataset update unit 112 adds to the second dataset 122 a combination including the value "A1a" of the explanatory variables A1 and the value "A2a" of the explanatory variables A2, and the value "85" of the response variable y.

[0067] Furthermore, if the target value is equal to or greater than the threshold value, the second dataset update unit 112 compares the value of the objective variable y with the threshold value, and adds the comparison result as a determination result, associated with the value of the objective variable y, to the second dataset 122. In this embodiment, the threshold value is, for example, 70. Each time the second dataset update unit 112 acquires a combination, it adds the combination to the second dataset 122 together with the determination result.

[0068] For example, each time a combination described above is added to the second data set 122, the second prediction model generation unit 140 generates or updates the second prediction model 124 using all of the combinations included in the second data set 122. The prediction unit 150 uses the second prediction model 124 generated or updated in this manner to predict the values ​​of the explanatory variable A1 and the explanatory variable A2 and the value of the dependent variable at a time point later than the current time point.

[0069] For example, if the current time is time tn, the prediction unit 150 predicts the values ​​of the explanatory variables A1 and A2 and the value of the dependent variable at times t(n+1), t(n+2), and t(t+3), which are later than the current time. The prediction unit 150 then determines whether the predicted values ​​of the dependent variable at those times are equal to or greater than a threshold. The prediction unit 150 may add combinations including the predicted values ​​of the explanatory variables A1 and A2 and the dependent variable values, as well as the determination results for the dependent variable values, to the second data set 122.

[0070] Furthermore, when the prediction unit 150 determines that the predicted value of the objective variable is less than the threshold, it notifies the apparatus control unit 170 of the determination result. For example, the prediction unit 150 notifies the apparatus control unit 170 that the predicted value of the objective variable at time t (t+3), "68," is less than the threshold value, "70." Upon receiving the notification from the prediction unit 150, the apparatus control unit 170 causes the material processing apparatus 200 to suspend the continuous process. Then, the apparatus control unit 170 instructs the first dataset update unit 111 to update the first dataset 121 using the second explanatory variable.

[0071] FIG. 5 is a diagram showing an example of the first data set 121 updated by adding the second explanatory variable.

[0072] The first dataset 121 updated by adding the second explanatory variable shows the first explanatory variables X1 and X2, the second explanatory variables A1 and A2, and the dependent variable y in association with each other, as shown in FIG. 5 . When adding the second explanatory variable to the first dataset 121, the first dataset update unit 111 integrates the second dataset 122 into the first dataset 121, thereby updating the first dataset 121 to a dataset in which the first and second explanatory variables are associated with the dependent variable. The first dataset 121 updated in this manner is hereinafter also referred to as an integrated dataset.

[0073] In a specific example, the integrated data set represents combinations at times t1 to tn in a continuous process stage and combinations at times t(n+1) to t(n+3) in a search stage following the continuous process stage, each of which includes a value of a first explanatory variable, a value of a second explanatory variable, a value of a response variable, and a determination result for the value of the response variable.

[0074] When the first dataset update unit 111 is instructed by the device control unit 170 to update the first dataset 121 using the second explanatory variable, it first adds each combination of times t1 to tn to the first dataset 121. In other words, by integrating the second dataset 122 into the first dataset 121, the second explanatory variable is added to the first dataset 121. As a result, the first dataset 121 is updated to an integrated dataset.

[0075] The first prediction model generation unit 130 generates a first prediction model 123 using the integrated dataset, which is the updated first dataset 121. Furthermore, the search condition derivation unit 160 performs a search using the first prediction model 123 to derive the value "X1q" of the explanatory variable X1 and the value "X2q" of the explanatory variable X2 as search conditions at time t(n+1). The first dataset update unit 111 acquires, from the material processing apparatus 200, the value "75" of the objective variable y obtained by processing the raw materials according to the search conditions, as well as the value "A1i" of the explanatory variable A1 and the value "A2i" of the explanatory variable A2, which indicate the state of the raw materials at that time. The first dataset update unit 111 then determines that the value "75" of the objective variable y is equal to or greater than a threshold. The first dataset update unit 111 then adds the combination at time t(n+1) to the first dataset 121. The combination includes the value "X1q" of explanatory variable X1 and the value "X2q" of explanatory variable X2, the value "A1i" of explanatory variable A1 and the value "A2i" of explanatory variable A2, the value "75" of objective variable y, and the judgment result "good" for the value "75" of objective variable y.

[0076] Addition of such combinations to the first data set 121 is repeatedly performed until, for example, a value equal to or greater than "threshold + 15" is obtained as the value of the objective variable y. In this embodiment, the threshold is, for example, 70. Here, as shown in FIG. 4B, at time t(n+3), the state of the raw materials changes to a state represented by the value "A1k" of the explanatory variable A1 and the value "A2k" of the explanatory variable A2, and as a result, it was predicted that the value of the objective variable y would be less than the threshold. However, in this embodiment, even if the state of the raw materials changes to a state represented by the value "A1k" of the explanatory variable A1 and the value "A2k" of the explanatory variable A2 at time t(n+3), the search condition derivation unit 160 can increase the likelihood of deriving search conditions that result in a value of the objective variable y that is equal to or greater than the threshold. In the example of FIG. 5, the value of the objective variable y is "85" at time t(n+3).

[0077] When the first dataset updater 111 adds a value of "85" for the objective variable y that is equal to or greater than "threshold value + 15" to the first dataset 121, the search condition deriver 160 notifies the device control unit 170 that the search for the search conditions will be suspended. In response to this notification, the device control unit 170 causes the material processing apparatus 200 to execute a continuous process according to the search conditions associated with the value of the objective variable y, "85," i.e., a continuous process according to the value of the explanatory variable X1, "X1s," and the value of the explanatory variable X2, "X2s." This improves the search accuracy and enables the continuous mass production of a product having the characteristic values ​​that are the target values.

[0078] FIG. 6 is a flowchart showing an example of the processing operation of the searching device 100 according to this embodiment.

[0079] First, the search condition derivation unit 160 of the search device 100 acquires a target value for the objective variable y and search ranges for each of the explanatory variables X1 and X2 (step S1). For example, the search condition derivation unit 160 may acquire the target value and search range in response to an input operation on an input device such as a keyboard, or, if the target value and search range are stored in the storage unit 120, may acquire them by reading them from the storage unit 120. Alternatively, if the search device 100 has a communication interface, the search condition derivation unit 160 may acquire the target value and search range from outside the search device 100 via the communication interface.

[0080] Next, the first dataset update unit 111 acquires explanatory variables X1 and X2, which are first explanatory variables, from the apparatus control unit 170 through an initial experiment using the material processing apparatus 200, and acquires from the material processing apparatus 200 a response variable y obtained by processing according to the explanatory variables X1 and X2. That is, the first dataset update unit 111 acquires a combination including the value of the explanatory variable X1, the value of the explanatory variable X2, and the value of the response variable y. The first dataset update unit 111 then generates a first dataset 121 including the combination and stores it in the storage unit 120 (step S2). Note that the processing of step S2 may be performed before the processing of step S1 or may be performed simultaneously with the processing of step S1.

[0081] In the initial experiment, the apparatus control unit 170 may randomly select values ​​for the explanatory variables X1 and X2 from the search range acquired in step S1 and set these values ​​in the material processing apparatus 200. Alternatively, the apparatus control unit 170 may select values ​​for the explanatory variables X1 and X2 using a statistical method such as experimental design. In step S2, the first dataset update unit 111 may acquire two or more of the above-described combinations and generate a first dataset 121 including these combinations. The number of combinations is not particularly limited and may be one or two. However, the greater the number of combinations, the higher the prediction accuracy of the first prediction model 123 generated from the first dataset 121.

[0082] Next, the first prediction model generation unit 130 generates a first prediction model 123 by machine learning using the first dataset 121 generated in step S2 (step S3).Then, the search condition derivation unit 160 uses the first prediction model 123 generated in step S3 to derive, as search conditions, values ​​of explanatory variables X1 and X2 for obtaining a target value of the objective variable y from the search range acquired in step S1 (step S4).For example, the target value is a value equal to or greater than a threshold.

[0083] Next, the apparatus control unit 170 sets the values ​​of the first explanatory variable X1 and the second explanatory variable X2, which are the search conditions derived in step S4, in the material processing apparatus 200, and causes the material processing apparatus 200 to perform processing according to the search conditions, i.e., mixing of the first liquid and the second liquid. As a result, the first dataset update unit 111 obtains the values ​​of the explanatory variables X1 and X2 from the apparatus control unit 170 and obtains the value of the response variable y obtained by mixing according to the values ​​of the explanatory variables X1 and X2 from the material processing apparatus 200. That is, the first dataset update unit 111 obtains a combination including the values ​​of the explanatory variables X1, X2, and y as an actual measurement value (step S5). The first dataset update unit 111 then updates the first dataset 121 stored in the storage unit 120 (step S6). That is, the first dataset update unit 111 adds the combination including the value of the explanatory variable X1, the value of the explanatory variable X2, and the value of the objective variable y acquired in step S5 to the first dataset 121. At this time, the first dataset update unit 111 determines whether the value of the objective variable y is equal to or greater than a threshold, and adds the determination result to the first dataset 121.

[0084] Next, the search condition derivation unit 160 determines whether the value of the objective variable y most recently added to the first dataset 121 reaches the target value acquired in step S1 (step S7). For example, if the value of the objective variable y is equal to or greater than "threshold value + 15," the search condition derivation unit 160 determines that the value of the objective variable y has reached the target value. Here, if it is determined that the value of the objective variable y has not reached the target value (No in step S7), the search device 100 repeats the processing from step S3. That is, the first prediction model generation unit 130 generates or updates the first prediction model 123 using the first dataset 121 updated in step S6.

[0085] On the other hand, if it is determined in step S7 that the value of the objective variable y has reached the target value (Yes in step S7), the equipment control unit 170 causes the material processing apparatus 200 to start the continuous process (step S8). That is, the equipment control unit 170 causes the material processing apparatus 200 to start the continuous process according to the search conditions included in the combination acquired in the most recent step S5, i.e., the values ​​of the explanatory variables X1 and X2.

[0086] When the continuous process starts, the second dataset update unit 112 acquires the second explanatory variables, explanatory variables A1 and A2, and the response variable y from the material processing apparatus 200. That is, the second dataset update unit 112 acquires a combination including the value of explanatory variable A1, the value of explanatory variable A2, and the value of response variable y. At this time, the second dataset update unit 112 may acquire multiple combinations. Then, the second dataset update unit 112 generates a second dataset 122 including the combination and stores it in the storage unit 120 (step S9).

[0087] Next, the second prediction model generation unit 140 generates or updates the second prediction model 124 using the second data set 122 generated in step S9 (step S10). The prediction unit 150 then predicts the values ​​of the explanatory variables A1 and A2 and the value of the dependent variable at a time point later than the current time. The prediction unit 150 then determines whether the predicted value, which is the predicted value of the dependent variable, deviates from the target value (step S11). If the target value is equal to or greater than a threshold, the prediction unit 150 determines that the predicted value deviates from the target value if the predicted value is less than the threshold. If it is determined that the predicted value does not deviate from the target value (No in step S11), the apparatus control unit 170 determines whether the termination condition is satisfied (step S12). For example, the apparatus control unit 170 determines that the termination condition is satisfied if the recovery unit 260 of the material processing apparatus 200 has recovered a predetermined amount of product (e.g., a mixed liquid) or more. If the device control unit 170 determines that the termination condition is satisfied (Yes in step S12), it terminates the continuous process by the material processing apparatus 200 and also terminates the search for the search conditions. On the other hand, if the device control unit 170 determines that the termination condition is not satisfied (No in step S12), it continues the continuous process by the material processing apparatus 200 and causes the second dataset update unit 112 to update the second dataset 122 (step S13). In this case, the second dataset update unit 112 obtains a combination including new values ​​of the explanatory variable A1, the explanatory variable A2, and the value of the objective variable y, and adds the combination to the second dataset 122, thereby updating the second dataset 122.

[0088] On the other hand, if it is determined in step S11 that the predicted value deviates from the target value (Yes in step S11), the first dataset update unit 111 integrates the second dataset 122 stored in the storage unit 120 into the first dataset 121 (step S14). At this time, the device control unit 170 terminates the continuous process by the material processing apparatus 200. Accordingly, the continuous process phase ends, and the processing phase of the processing system 1000 transitions from the continuous process phase to the search phase. Through the integration in step S14, explanatory variables A1 and A2, which are second explanatory variables, are added to the first dataset 121, and the first dataset 121 is updated to an integrated dataset. The first prediction model generation unit 130 then generates or updates the first prediction model 123 using the integrated dataset, i.e., the first dataset 121 to which the second explanatory variables have been added. In other words, the derivation of new search conditions begins.

[0089] As described above, in this embodiment, when search conditions are derived as an optimal solution and raw materials are continuously processed according to the search conditions, i.e., in the continuous process stage, it is predicted whether the characteristic value of the raw materials will deviate from the target value at a later time point. Therefore, if the raw materials change over time, it is predicted that the characteristic value will deviate from the target value. In such a case, the first prediction model is updated using the first dataset to which the second explanatory variable indicating the state of the raw materials has been added. As a result, a search is performed again using the updated first prediction model, taking into account the second explanatory variable, i.e., a search that takes into account the change in the raw materials over time. Therefore, raw materials can be processed according to the new search conditions derived by the search as a continuous process, thereby increasing the likelihood that the characteristic value will remain at the target value even at a later time point. This improves the accuracy of the search for an optimal solution, even when raw materials that change significantly over time are used to produce a product, and enables continuous mass production of a product having the target characteristic value.

[0090] Furthermore, in the continuous process stage of this embodiment, a second data set 122 is generated or updated, and a second prediction model 124 generated from the second data set 122 is used to predict whether the characteristic value will deviate from the target value. This makes it possible to predict with high accuracy whether the characteristic value will deviate from the target value at a time point later than the current time point.

[0091] Furthermore, in this embodiment, by integrating the second data set 122 into the first data set 121, the first data set 121 is updated to a data set that indicates the first and second explanatory variables and the dependent variable in association with each other. This makes it possible to appropriately associate the first and second explanatory variables with the dependent variable, and effectively add the second explanatory variable to the first data set 121.

[0092] Furthermore, in this embodiment, if the prediction unit 150 predicts that the characteristic value will deviate from the target value when the state of the raw materials indicated by the second explanatory variable is in a predetermined state, the search condition derivation unit 160 performs a search using the updated first prediction model 123 to derive, as new search conditions, values ​​of the first explanatory variable for obtaining the target value of the objective variable y when the raw materials are in that predetermined state. For example, as shown in FIG. 4B, the predetermined state is represented by the value "A1k" of the explanatory variable A1 and the value "A2k" of the explanatory variable A2. When the raw materials are in this state, it is predicted that the characteristic value will deviate from the target value. Then, as shown in FIG. 5B, values ​​of the first explanatory variables for obtaining the target value of the objective variable y when the raw materials are in that predetermined state, i.e., the value "X1s" of the explanatory variable X1 and the value "X2s" of the explanatory variable X2, are derived as new search conditions. This means that even if the raw material changes over time and reaches a predetermined state, the material processing device 200 can be made to perform processing according to the new search conditions, thereby producing a product with the target characteristic values.

[0093] Second Embodiment The material processing apparatus 200 in this embodiment has a configuration for adjusting the state of raw materials. The searching apparatus 100 then sets a third explanatory variable for adjusting the state of the raw materials in the material processing apparatus 200 together with the first explanatory variable. Furthermore, the searching apparatus 100 adds the third explanatory variable together with the second explanatory variable to the first data set 121. This makes it possible to easily mass-produce a product having target characteristic values, even when raw materials that change significantly over time are used.

[0094] Of the components in this embodiment, the same components as those in embodiment 1 are denoted by the same reference numerals as those in embodiment 1, and detailed description thereof will be omitted. The following will mainly describe in detail the configuration and processing operations of this embodiment that are different from those in embodiment 1.

[0095] FIG. 7 is a block diagram showing an example of the configuration of a material processing apparatus 200 according to this embodiment.

[0096] The material processing apparatus 200 in this embodiment has the same configuration as that in the first embodiment, and further includes a first liquid adjuster 214 , a second liquid adjuster 224 , and a third explanatory variable setting unit 243 .

[0097] The third explanatory variable setting unit 243 sets the value of the third explanatory variable derived by the searching device 100 in the first liquid adjuster 214 and the second liquid adjuster 224 .

[0098] The first liquid adjuster 214 adjusts the state of the first liquid contained in the first liquid container 211. For example, the first liquid adjuster 214 adjusts the state of the first liquid by adding an adjuster such as water, acid, or base to the first liquid in an amount corresponding to the value of the third explanatory variable set by the third explanatory variable setting unit 243. In other words, the second explanatory variable of the first liquid is adjusted by the third explanatory variable.

[0099] The second liquid adjuster 224 adjusts the state of the second liquid contained in the second liquid container 221. For example, similar to the first liquid adjuster 214, the second liquid adjuster 224 adjusts the state of the second liquid by adding an adjuster such as water, acid, or base to the second liquid in an amount corresponding to the value of the third explanatory variable set by the third explanatory variable setting unit 243. In other words, the second explanatory variable of the second liquid is adjusted by the third explanatory variable.

[0100] FIG. 8 is a block diagram showing an example of the configuration of a searching device 100 according to this embodiment.

[0101] The search device 100 in this embodiment has a configuration similar to that of embodiment 1, but the first dataset update unit 111, the first prediction model generation unit 130, the search condition derivation unit 160, and the device control unit 170 in this embodiment perform additional operations compared to embodiment 1.

[0102] Specifically, in this embodiment, if the predicted value of the dependent variable y deviates from the target value, the first dataset updater 111 adds the second explanatory variable and the third explanatory variable to the first dataset 121 to update the first dataset 121 to an integrated dataset. The first prediction model generator 130 then uses the integrated dataset to generate a first prediction model 123 that indicates the relationship between the dependent variable and the first, second, and third explanatory variables. The search condition derivation unit 160 performs a search using the first prediction model 123 to derive, as search conditions, values ​​of the first explanatory variable and the third explanatory variable for obtaining the target value of the dependent variable. The search condition derivation unit 160 outputs the derived search conditions, the values ​​of the first explanatory variable and the third explanatory variable, to the device control unit 170. When the device control unit 170 acquires the values ​​of the first explanatory variable and the third explanatory variable as search conditions from the search condition derivation unit 160, it outputs those values ​​to the material processing apparatus 200 and causes the material processing apparatus 200 to perform mixing according to those values. The values ​​of the first explanatory variable and the third explanatory variable are also output from the device control unit 170 to the first dataset update unit 111.

[0103] The first explanatory variable setting unit 241 of the material processing apparatus 200 acquires the value of the first explanatory variable from the apparatus control unit 170 and sets it in, for example, the mixing channel unit 230. Furthermore, the third explanatory variable setting unit 243 of the material processing apparatus 200 acquires the value of the third explanatory variable from the apparatus control unit 170 and sets it in the first liquid adjustment unit 214 and the second liquid adjustment unit 224. Then, the response variable acquisition unit 250 of the material processing apparatus 200 outputs a new value of the response variable obtained by processing according to the value of the first explanatory variable and the value of the third explanatory variable to the first dataset update unit 111.

[0104] 9 is a diagram showing an example of the first data set 121 updated by adding the second and third explanatory variables. In this embodiment, as in the example of the first embodiment, the first data set 121 shown in FIG. 4A is generated in the initial search stage, and the second data set 122 shown in FIG. 4B is generated in the next continuous process stage.

[0105] The first dataset 121 updated by adding the second and third explanatory variables shows, for example, as shown in FIG. 9 , explanatory variables X1 and X2, which are first explanatory variables, explanatory variables A1 and A2, which are second explanatory variables, explanatory variables B1 and B2, which are third explanatory variables, and a dependent variable y, all associated with one another. The first dataset update unit 111 integrates the second dataset 122 into the first dataset 121 and further adds a third explanatory variable, thereby updating the first dataset 121 to a dataset that associates the first, second, and third explanatory variables with the dependent variable. The first dataset 121 updated in this manner is also referred to as an integrated dataset.

[0106] In a specific example, the integrated data set represents combinations at times t1 to tn in a continuous process stage and combinations at times t(n+1) to t(n+3) in an exploration stage following the continuous process stage, each of which includes values ​​of a first explanatory variable, a second explanatory variable, a third explanatory variable, a response variable, and a determination result for the response variable value.

[0107] When the first dataset update unit 111 is instructed by the device control unit 170 to update the first dataset 121 using the second explanatory variable and the third explanatory variable, the first dataset update unit 111 first adds each combination of times t1 to tn to the first dataset 121. That is, the first dataset update unit 111 integrates the second dataset 122 into the first dataset 121, and further adds the initial value (e.g., 0) of the third explanatory variable at each of times t1 to tn to the first dataset 121. As a result, the second explanatory variable and the third explanatory variable are added to the first dataset 121. As a result, the first dataset 121 is updated to an integrated dataset.

[0108] The first prediction model generation unit 130 generates a first prediction model 123 using the integrated dataset, which is the updated first dataset 121. Furthermore, the search condition derivation unit 160 performs a search using the first prediction model 123 to derive, as search conditions at time t(n+1), the value "X1q" of explanatory variable X1, the value "X2q" of explanatory variable X2, and the value "B1a" of explanatory variable B1 and the value "B2a" of explanatory variable B2. The first dataset update unit 111 acquires, from the material processing apparatus 200, the value "75" of objective variable y obtained by processing the raw materials according to the search conditions, as well as the value "A1i'" of explanatory variable A1 and the value "A2i'" of explanatory variable A2, which indicate the state of the raw materials at that time. The first dataset update unit 111 then determines that the value "75" of objective variable y is equal to or greater than a threshold. The first dataset update unit 111 then adds the combination at time t(n+1) to the first dataset 121. The combination includes the value "X1q" of explanatory variable X1 and the value "X2q" of explanatory variable X2, the value "A1i'" of explanatory variable A1 and the value "A2i'" of explanatory variable A2, the value "B1a" of explanatory variable B1 and the value "B2a" of explanatory variable B2, the value "75" of objective variable y, and the judgment result "good" for the value "75" of objective variable y.

[0109] Such combinations are repeatedly added to the first data set 121 until a value equal to or greater than "threshold + 15" is obtained as the value of the objective variable y. In this embodiment, the threshold is, for example, 70. Note that the value added to the threshold is not limited to 15 and may be any other value equal to or greater than 0. Here, as shown in FIG. 4B, at time t(n+3), the state of the raw materials changes to a state represented by the value "A1k" of the explanatory variable A1 and the value "A2k" of the explanatory variable A2, and as a result, it is predicted that the value of the objective variable y will be less than the threshold. However, in this embodiment, the likelihood that the state of the raw materials will change to a state represented by the value "A1k" of the explanatory variable A1 and the value "A2k" of the explanatory variable A2 at time t(n+3) is reduced. Furthermore, the search condition derivation unit 160 can further increase the likelihood of deriving search conditions that result in a value of the objective variable y that is equal to or greater than the threshold.

[0110] When the first dataset updater 111 adds a value of "85" for the objective variable y that is equal to or greater than "threshold value + 15" to the first dataset 121, the search condition deriver 160 notifies the device control unit 170 that the search for the search conditions will be suspended. In response to this notification, the device control unit 170 causes the material processing apparatus 200 to execute a continuous process according to the search conditions associated with the value of the objective variable y, "85," i.e., a continuous process according to the value of the explanatory variable X1, "X1s," and the value of the explanatory variable X2, "X2s." This improves the search accuracy and enables the continuous mass production of a product having the characteristic values ​​that are the target values.

[0111] FIG. 10 is a flowchart showing an example of the processing operation of the searching device 100 in this embodiment.

[0112] The searching device 100 in this embodiment executes each step included in the flowchart of FIG. 6, and further executes the process of step S15.

[0113] In step S15, the first dataset update unit 111 adds a third explanatory variable to the first dataset 121 integrated in step S14. In step S15, explanatory variables B1 and B2, which are third explanatory variables, are added to the first dataset 121, and the first dataset 121 is updated to an integrated dataset. The first prediction model generation unit 130 then generates or updates the first prediction model 123 using the integrated dataset, i.e., the first dataset 121 to which the third explanatory variable has been added. In other words, derivation of new search conditions begins.

[0114] As described above, in this embodiment, when the first dataset update unit 111 adds a second explanatory variable to the first dataset 121, the first dataset update unit 111 updates the first dataset by adding, together with the second explanatory variable, a third explanatory variable for adjusting the state of the raw materials to the first dataset 121. Then, the first prediction model generation unit 130 updates the first prediction model 123 using the first dataset 121 updated by adding the second and third explanatory variables.

[0115] This allows a search to be performed that takes into account the state of the raw materials that changes over time and adjustments to that state, by performing a search using the first prediction model 123. As a result, it is possible to derive with high accuracy an optimal solution for the search conditions that include the values ​​of the first explanatory variable and the third explanatory variable and are used to generate a product having characteristic values ​​that are set as target values.

[0116] (Other Modifications) The searching device 100 according to one aspect of the present disclosure has been described above based on the above-described embodiments, but the present disclosure is not limited to those embodiments. As long as the modifications do not deviate from the spirit of the present disclosure, various modifications that would occur to those skilled in the art may also be included in the present disclosure.

[0117] For example, in the first or second embodiment, the number of each of the first explanatory variable, the second explanatory variable, and the third explanatory variable is two, but the number is not limited to two and may be one, or three or more. Also, in the first and second embodiments, the number of the objective variable is one, but the number is not limited to one and may be two or more.

[0118] In the first embodiment, in step S9 of FIG. 6 , a second data set 122 is generated that includes explanatory variables A1 and A2 as second explanatory variables. However, if the process of step S9 is executed again, a second data set 122 that includes second explanatory variables different from explanatory variables A1 and A2 may be generated. Even if the process of step S9 of FIG. 10 in the second embodiment is executed again, a second data set 122 that includes second explanatory variables different from explanatory variables A1 and A2 may be generated. Furthermore, in the second embodiment, in step S15 of FIG. 10 , a first data set 121 that includes explanatory variables B1 and B2 as third explanatory variables is generated. However, if the process of step S15 is executed again, a first data set 121 that includes a third explanatory variable different from explanatory variables B1 and B2 may be generated.

[0119] In addition, although the target value is a value equal to or greater than the threshold in the above-described first and second embodiments, it may be a value equal to or less than the threshold. In this case, the search condition derivation unit 160 determines that the value of the objective variable y has reached the target value if the value of the objective variable y is equal to or less than "threshold - 15."

[0120] Furthermore, the first prediction model 123 and the second prediction model 124 in the first and second embodiments may be any model that indicates the relationship, i.e., the correlation, between the explanatory variables and the dependent variable.

[0121] Furthermore, the third explanatory variable in the above embodiment may be any variable other than the first explanatory variable and the second explanatory variable.

[0122] In each of the above embodiments, each component may be configured with dedicated hardware, or may be realized by executing a software program suitable for each component. Each component may be realized by a program execution unit such as a CPU or processor reading and executing a software program recorded on a recording medium such as a hard disk or semiconductor memory. Here, the software that realizes the search device 100 of each of the above embodiments is, for example, a program that causes a computer to execute each step of the flowchart shown in FIG. 6 or FIG. 10.

[0123] The following cases are also included in this disclosure:

[0124] (1) The at least one device is specifically a computer system comprising a microprocessor, ROM (Read Only Memory), RAM (Random Access Memory), a hard disk unit, a display unit, a keyboard, a mouse, etc. A computer program is stored in the RAM or hard disk unit. The at least one device achieves its function when the microprocessor operates in accordance with the computer program. Here, the computer program is composed of a combination of multiple instruction codes that indicate commands to the computer to achieve a predetermined function.

[0125] (2) Some or all of the components constituting at least one of the above devices may be configured as a single system LSI (Large Scale Integration). A system LSI is an ultra-multifunctional LSI manufactured by integrating multiple components on a single chip, and specifically, is a computer system configured to include a microprocessor, ROM, RAM, etc. A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor operating in accordance with the computer program.

[0126] (3) Some or all of the components constituting at least one of the above devices may be configured as an IC card or a standalone module that can be attached to or detached from the device. The IC card or module is a computer system configured with a microprocessor, ROM, RAM, etc. The IC card or module may include the above-mentioned ultra-multifunctional LSI. The IC card or module achieves its functions when the microprocessor operates in accordance with a computer program. This IC card or module may be tamper-resistant.

[0127] (4) The present disclosure may be embodied as the methods described above, a computer program that implements these methods on a computer, or a digital signal that includes the computer program.

[0128] The present disclosure may also be a computer program or a digital signal recorded on a computer-readable recording medium, such as a flexible disk, a hard disk, a CD (Compact Disc)-ROM, a DVD, a DVD-ROM, a DVD-RAM, a BD (Blu-ray (registered trademark) Disc), a semiconductor memory, etc. Alternatively, the present disclosure may be a digital signal recorded on such a recording medium.

[0129] The present disclosure may also be applied to transmitting a computer program or digital signal via a telecommunications line, a wireless or wired communication line, a network such as the Internet, data broadcasting, or the like.

[0130] Furthermore, the program or digital signal may be recorded on a recording medium and transferred, or the program or digital signal may be transferred via a network or the like, so that the program or digital signal may be implemented by another independent computer system.

[0131] The searching device of the present disclosure can be applied to, for example, devices and systems that search for processing conditions for producing a product having desired characteristic values ​​as search conditions.

[0132] 100 Searching device 111 First dataset updating unit 112 Second dataset updating unit 120 Storage unit 121 First dataset 122 Second dataset 123 First prediction model 124 Second prediction model 130 First prediction model generating unit 140 Second prediction model generating unit 150 Prediction unit 160 Search condition derivation unit 170 Device control unit 200 Material processing device 210 First liquid supply unit 211 First liquid container 212 First liquid pump 213 First liquid flow path 214 First liquid adjustment unit 220 Second liquid supply unit 221 Second liquid container 222 Second liquid pump 223 Second liquid flow path 224 Second liquid adjustment unit 230 Mixing flow path unit 241 First explanatory variable setting unit 242 Second explanatory variable acquisition unit 243 Third explanatory variable setting unit 250 Objective variable acquisition unit 260 Recovery section 1000 Processing system

Claims

1. A search device comprising: a first dataset updating unit that generates or updates a first dataset indicating an association between a first explanatory variable and an objective variable indicating a characteristic value of a product obtained by processing raw materials by a material processing device in accordance with a first explanatory variable; a first prediction model generating unit that generates a first prediction model indicating a relationship between the first explanatory variable and the objective variable using the first dataset; a search condition derivation unit that performs a search using the first prediction model to derive a value of the first explanatory variable as a search condition for obtaining a target value of the objective variable; and a prediction unit that predicts whether the characteristic value indicated by the objective variable will deviate from the target value at a time point later than a current time when processing of the raw materials in accordance with the search conditions derived by the search condition derivation unit is continuously performed, wherein the first dataset updating unit updates the first dataset by adding a second explanatory variable indicating a state of the raw materials to the first dataset when the prediction unit predicts that the characteristic value will deviate from the target value, and the first prediction model generating unit updates the first prediction model using the first dataset updated by adding the second explanatory variable.

2. The search device according to claim 1, further comprising: a second dataset update unit that acquires the second explanatory variable and the dependent variable when processing of the raw materials is continuously performed in accordance with the search conditions derived by the search condition derivation unit, and generates or updates a second dataset that correlates the acquired second explanatory variable with the dependent variable; and a second prediction model generation unit that uses the second dataset to generate a second prediction model that indicates the relationship over time between the second explanatory variable and the dependent variable, wherein the prediction unit uses the second prediction model to predict whether the characteristic value will deviate from the target value.

3. The exploration device described in claim 2, wherein, when adding the second explanatory variable to the first dataset, the first dataset update unit integrates the second dataset into the first dataset, thereby updating the first dataset to a dataset that indicates the first explanatory variable and the second explanatory variable in association with the target variable.

4. The search device described in claim 2, wherein when the prediction unit predicts that the characteristic value will deviate from the target value when the state of the raw materials indicated by the second explanatory variable is in a specified state, the search condition derivation unit derives, as a new search condition, the value of the first explanatory variable for obtaining the target value of the objective variable when the state of the raw materials is in the specified state by performing a search using the updated first prediction model.

5. The exploration device described in claim 2, wherein the first dataset update unit, when adding the second explanatory variable to the first dataset, updates the first dataset by adding a third explanatory variable for adjusting the state of the raw materials to the first dataset together with the second explanatory variable, and the first prediction model generation unit updates the first prediction model using the first dataset updated by adding the second explanatory variable and the third explanatory variable.

6. A search method using a computer, comprising: generating or updating a first data set indicating an association between a first explanatory variable and a response variable indicating a characteristic value of a product obtained by processing raw materials by a material processing device in accordance with a first explanatory variable; generating a first prediction model using the first data set indicating the relationship between the first explanatory variable and the response variable; deriving a value of the first explanatory variable for obtaining a target value of the response variable as a search condition by performing a search using the first prediction model; predicting whether the characteristic value indicated by the response variable will deviate from the target value at a time point after the current time when processing of the raw materials in accordance with the derived search conditions is continuously performed; updating the first data set by adding a second explanatory variable indicating a state of the raw materials to the first data set when it is predicted that the characteristic value will deviate from the target value; and updating the first prediction model using the first data set updated by adding the second explanatory variable.

7. A program causing a computer to execute the following steps: generate or update a first dataset indicating an association between a first explanatory variable and an objective variable indicating a characteristic value of a product obtained by processing raw materials by a material processing device in accordance with a first explanatory variable; generate a first prediction model indicating a relationship between the first explanatory variable and the objective variable using the first dataset; derive a value of the first explanatory variable for obtaining a target value of the objective variable as a search condition by performing a search using the first prediction model; predict whether the characteristic value indicated by the objective variable will deviate from the target value at a time point after the current time when processing of the raw materials in accordance with the derived search conditions is continuously performed; update the first dataset by adding a second explanatory variable indicating a state of the raw materials to the first dataset when it is predicted that the characteristic value will deviate from the target value; and update the first prediction model using the first dataset updated by adding the second explanatory variable.

Images