Design support device, design support method, and design support program

Abstract

This design support device is provided with: a data acquisition unit that acquires a plurality of pieces of performance data comprising observation values of a plurality of characteristic terms and a design parameter set for a product or the like; a model construction unit that constructs, on the basis of the actual performance data, a prediction model that predicts an observed value of the characteristic item on the basis of the design parameter group; a target value acquisition unit that acquires a target value relating to the observed value of each characteristic item; a scoring function construction unit that constructs a scoring function including at least a target achievement probability term that represents an overall achievement probability of probabilities for achieving all characteristic term target values, said achievement probability term being calculated using the design parameter group as variables on the basis of the prediction model; a target achievement probability acquisition unit that acquires an overall achievement probability by optimizing the scoring function; and an output unit that outputs the overall achievement probability.

Description

Technical Field

[0001] One aspect of the present invention relates to a design support device, a design support method, and a design support program. Background Art

[0002] A product design using machine learning is being studied. As a field of product design, for example, in the design of functional materials, a model for inferring material properties is constructed by machine learning using learning data consisting of pairs of raw material ratios and properties related to experimental and manufactured materials, and properties of untested raw material ratios are predicted. By formulating an experimental plan based on such property predictions, parameters such as material properties and raw material ratios can be effectively optimized, thereby achieving improved development efficiency. In addition, as a method for such optimization, Bayesian optimization is known to be effective, and a design device that uses Bayesian optimization to output design values ​​is known (for example, refer to Patent Document 1).

[0003] Previous technical literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Publication No. 2020-52737 Summary of the invention

[0006] Technical issues to be solved by the invention

[0007] On the other hand, in product development of materials, etc., target values ​​are set for observed values ​​of characteristic items that represent characteristics of products, etc. In product development, setting appropriate target values ​​is important for efficiently developing products, determining whether to continue or stop development, making decisions such as changing development policies, etc. However, when products, etc. have many characteristic items and the number of target values ​​for the characteristic items to be set is large, it is difficult to set appropriate target values.

[0008] Therefore, the present invention is completed in view of the above-mentioned problems, and its purpose is to easily set the target value of the characteristic item constituting the target variable during the production process of products, work-in-progress, semi-finished products, parts or prototypes.

[0009] Means for solving technical problems

[0010] The design support device involved in the first aspect of the present invention supports the setting of target values ​​related to the observed values ​​of multiple characteristic items, wherein the multiple characteristic items represent the characteristics of a product, work-in-progress, semi-finished product, part or prototype manufactured based on a design parameter group composed of multiple design parameters, wherein the design support device comprises: a data acquisition unit, which acquires multiple performance data related to the manufactured product, work-in-progress, semi-finished product, part or prototype, which are composed of a design parameter group and the observed values ​​of each of the multiple characteristic items; a model construction unit, which constructs a prediction model based on the performance data, wherein the prediction model predicts the observed values ​​of the characteristic items as a probability distribution or its approximate or alternative index based on the design parameter group; a target value acquisition unit, which acquires the target value related to the observed values ​​of each characteristic item; a scoring function construction unit, which constructs a scoring function including at least a target achievement probability item, wherein the target achievement probability item represents the overall achievement probability calculated based on the prediction model using the design parameter group as a variable and is the probability of achieving the target values ​​of all characteristic items; a target achievement probability acquisition unit, which acquires the overall achievement probability by optimizing the scoring function; and an output unit, which outputs the overall achievement probability.

[0011] The design support method involved in the first aspect of the present invention is a design support method in a design support device that supports the setting of target values ​​related to observed values ​​of multiple characteristic items, wherein the multiple characteristic items represent characteristics of a product, work-in-progress, semi-finished product, part or prototype manufactured based on a design parameter group composed of multiple design parameters, wherein the design support method has the following steps: a data acquisition step, acquiring multiple performance data related to the manufactured product, work-in-progress, semi-finished product, part or prototype, which are composed of a design parameter group and the observed values ​​of each of the multiple characteristic items; a model construction step, constructing a prediction model based on the performance data, wherein the prediction model predicts the observed values ​​of the characteristic items as a probability distribution or its approximate or alternative index based on the design parameter group; a target value acquisition step, acquiring the target value related to the observed values ​​of each characteristic item; a scoring function construction step, constructing a scoring function that at least includes a target achievement probability term, wherein the target achievement probability term represents the overall achievement probability calculated based on the prediction model using the design parameter group as a variable and is the probability of achieving the target values ​​of all characteristic items; a target achievement probability acquisition step, acquiring the overall achievement probability by optimizing the scoring function; and an output step, outputting the overall achievement probability.

[0012] The design support program involved in the first aspect of the present invention is used to enable a computer to function as a design support device that supports the setting of target values ​​related to observed values ​​of multiple characteristic items, wherein the multiple characteristic items represent characteristics of a product, work-in-progress, semi-finished product, part or prototype manufactured based on a design parameter group composed of multiple design parameters, wherein the design support program enables the computer to implement the following functions: a data acquisition function for acquiring multiple performance data related to the manufactured product, work-in-progress, semi-finished product, part or prototype, which are composed of a design parameter group and the observed values ​​of each of the multiple characteristic items; a model construction function for constructing a prediction model based on the performance data, wherein the prediction model predicts the observed values ​​of the characteristic items as a probability distribution or its approximate or alternative index based on the design parameter group; a target value acquisition function for acquiring target values ​​related to the observed values ​​of each characteristic item; a scoring function construction function for constructing a scoring function that at least includes a target achievement probability term, wherein the target achievement probability term represents the overall achievement probability calculated based on the prediction model using the design parameter group as a variable and is the probability of achieving the target values ​​of all characteristic items; a target achievement probability acquisition function for acquiring the overall achievement probability by optimizing the scoring function; and an output function for outputting the overall achievement probability.

[0013] According to this aspect, a prediction model is constructed based on actual performance data and a target value related to each characteristic item is obtained. A scoring function including an overall achievement probability item is constructed based on the obtained target value and the prediction model. Furthermore, the overall achievement probability is obtained and output by optimizing the scoring function. Therefore, by referring to the output overall achievement probability, it is possible to determine whether the target value is appropriate, so that setting of an appropriate target value becomes easy.

[0014] The design support apparatus according to the second aspect may be the design support apparatus according to the first aspect, wherein the target value acquisition unit acquires the target value related to at least one characteristic item among the plurality of characteristic items based on a user input.

[0015] According to this aspect, the user can easily set the target value of each characteristic item. Therefore, by changing the input target value and repeating the input of the target value and the output of the overall achievement probability, an appropriate target value can be easily set.

[0016] In the design support device according to the third aspect, in the design support device according to the first or second aspect, the target value acquisition unit may acquire the target value related to at least one characteristic item among the plurality of characteristic items from a predetermined database storing characteristic item target values ​​in advance.

[0017] According to this aspect, the overall achievement probability can be output using the target values ​​of each characteristic item stored in advance. Thus, for example, when the number of characteristic items is large and the acquisition of the target value is based on the user's input, the operation can be labor-saving when the operation related to the input becomes cumbersome. And, for example, when a target value group with actual performance as an appropriate target value is pre-stored in a database, by obtaining the target value group as an initial target value and correcting a part of the target value, the setting of an appropriate target value becomes easy. And, by pre-storing the actual performance value with actual performance as the observed value of each characteristic item in a database, and using the actual performance value as the initial target value, the setting of an appropriate target becomes easy.

[0018] The design support device according to the fourth aspect may be the design support device according to any one of the first to third aspects, wherein the output unit outputs the target value acquired by the target value acquisition unit together with the overall achievement probability.

[0019] According to this aspect, the target values ​​of the characteristic items in the optimization process of the scoring function and the overall achievement probability obtained by the optimization are output together, for example, by display, so that the correspondence between the target value group and the overall achievement probability can be easily recognized.

[0020] In the design support device involved in the fifth aspect, it can be set that in the design support device involved in the fourth aspect, the target value acquisition unit acquires a range of target values ​​related to the observed value of at least one characteristic item among multiple characteristic items, the scoring function construction unit constructs a scoring function corresponding to each target value within the range of target values, the target achievement probability acquisition unit acquires the overall achievement probability corresponding to each target value within the acquired target value range by optimizing each scoring function, and the output unit associates each target value within the set target value range with the corresponding overall achievement probability and outputs it.

[0021] According to this aspect, by referring to the correlation between each target value and the overall achievement probability, it is possible to recognize a change in the overall achievement probability corresponding to a change in the target value of the characteristic item. Therefore, it becomes easy to set an appropriate target value.

[0022] In the design support device involved in aspect 6, it can be set that in the design support device involved in any one of aspects 1 to 5, the target achievement probability acquisition unit acquires the achievement probability of the target value of each characteristic item, and the output unit outputs the achievement probability of the target value of each characteristic item together with the overall achievement probability.

[0023] According to this aspect, the achievement probability of the target value of each characteristic item is outputted, for example, by displaying it, thereby identifying the target value of the characteristic item that is the bottleneck for achieving a high overall achievement probability. Therefore, by adjusting the target value, it becomes easy to improve the overall achievement probability.

[0024] The design support device involved in aspect 7 can be set as the design support device involved in any one of aspects 1 to 6, further comprising an optimization range acquisition unit, which acquires an optimization range of at least one design parameter among each design parameter included in the design parameter group as a search range in the optimization process of the scoring function, in which the optimization range of the design parameter is specified by at least one of the upper limit value, the lower limit value, the upper and lower limits and the fixed value of the design parameter value, and the target achievement probability acquisition unit optimizes the scoring function based on the acquired optimization range.

[0025] According to this aspect, the overall achievement probability within the optimization range of the acquired design parameters can be obtained. Therefore, for example, by confirming the overall achievement probability and resetting the optimization range multiple times, the optimal design parameter range in the development process of the product can be obtained.

[0026] The design support device according to claim 8 may be the design support device according to any one of claims 1 to 7, wherein the model construction unit constructs a prediction model for at least one characteristic item among the plurality of characteristic items based on input of parameters constituting the prediction model.

[0027] According to this aspect, part of the calculation of the probability of achieving the target value of the characteristic item does not depend on the actual performance data. Therefore, when there is insufficient actual performance data, for example, it is possible to prevent a decrease in the accuracy of the overall achievement probability due to a deterioration in the estimation accuracy of the characteristic item observation value distribution in the prediction model.

[0028] In the design support device involved in aspect 9, it can be set as a design support device involved in any one of aspects 1 to 8, wherein the target achievement probability acquisition unit obtains the design parameter group that maximizes the overall achievement probability by optimizing the scoring function, and the output unit outputs the design parameter group that maximizes the overall achievement probability.

[0029] According to this aspect, by referring to the output design parameter group, it is possible to confirm whether the design parameter group that can achieve the output overall achievement probability is appropriate. In addition, by referring to the output design parameter group, it is possible to manufacture and test products, etc.

[0030] Effects of the Invention

[0031] According to an aspect of the present invention, in the manufacturing process of a product, a work-in-progress, a semi-finished product, a part, or a prototype, target values ​​of characteristic items constituting a target variable can be easily set. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1It is a diagram showing an overview of a material design process to which the design support device according to the embodiment is applied.

[0033] Figure 2 This is a block diagram showing an example of a functional configuration of a design support device according to an embodiment.

[0034] Figure 3 This is a hardware block diagram of a design support device according to an embodiment.

[0035] Figure 4 It is a diagram showing an example of a design parameter group related to a manufactured material.

[0036] Figure 5 This is a diagram showing an example of observed values ​​of characteristic items related to manufactured materials.

[0037] Figure 6 : is a diagram showing an example of the target values ​​of the acquired characteristic items.

[0038] Figure 7 This is a diagram schematically showing an example of a display screen example (user interface) for inputting a target value and outputting an overall achievement probability.

[0039] Figure 8 It is a diagram showing an example of the acquired optimization range of the design parameters.

[0040] Fig. 9 This is a flowchart showing an example of the content of the design support method in the design support device according to the embodiment.

[0041] Fig.10 This is a diagram schematically showing an example of a display screen example (user interface) in which a target value and an overall achievement probability are displayed in association with each other.

[0042] Fig.11 This is a diagram showing the structure of a design support program. DETAILED DESCRIPTION

[0043] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or equivalent elements are denoted by the same reference numerals, and repeated descriptions are omitted.

[0044] Figure 1This is a diagram showing an overview of the design and development process of a material as an example of the design and development process of a product, work-in-progress, semi-finished product, part or prototype to which the design support device involved in the embodiment is applied. In addition, hereinafter, "product, work-in-progress, semi-finished product, part or prototype" is recorded as "product, etc." The design support device 10 of this embodiment can be applied to the design and development process of all products, etc. having multiple characteristic items representing the characteristics of the product, etc. and target values ​​of each characteristic item. Specifically, the design support device 10 can be applied to the setting of the preferred target value of the characteristic item in a process consisting of determining the design parameters (explanatory variables), making the product, etc. based on the determined design parameters, and obtaining the observed values ​​(target variables) of the characteristic items of the product, etc. In addition to the development / design of materials, the design support device 10 can also be applied to the design of products such as automobiles and drugs, the optimization of the molecular structure of drugs, etc. In addition, the design support device 10 of this embodiment can be applied to the use as a communication tool. Specifically, the target value of the characteristic item output by the design support device 10 and the probability of achieving the target value can be referred to between the manufacturer of the product etc. and the client (customer) of the product etc., and can be provided for forming a consensus on the setting of the target value to be achieved in the product etc. As described above, in this embodiment, the design support processing based on the design support device 10 is described by taking the example of the design development of materials as an example of the design development of the product etc.

[0045] like Figure 1 As shown, as an example, the design support processing by the design support device 10 is applied to the production and experiment of materials in the factory and laboratory A. That is, according to the set design parameter group x, the material is produced in the factory and laboratory A, and based on the produced material, the observed values ​​y of multiple characteristic items representing the material characteristics are obtained. In addition, the material production and experiment in the factory and laboratory A may also be simulated.

[0046] The design support device 10 supports the setting of target values ​​based on performance data, the performance data consisting of a design parameter group x and observation values ​​y of a plurality of characteristic items of a material produced based on the design parameter group x. Specifically, the design support device 10 constructs a prediction model for predicting the observation values ​​of the characteristic items based on the performance data consisting of the design parameter group x and the observation values ​​y, constructs a scoring function including an overall achievement probability term, and acquires and outputs the overall achievement probability by optimizing the scoring function, the overall achievement probability term indicating the overall achievement probability calculated based on the prediction model and achieving the target values ​​of all characteristic items.

[0047] For example, as an example of design and development of material products, when a certain material is produced by mixing a plurality of polymers and additives, the design support device 10 uses a design parameter group such as the compounding amount of each polymer and additive as a design variable, and uses the observed values ​​of the elastic modulus and thermal expansion coefficient as characteristic items as target variables for setting target values ​​of a plurality of characteristic items.

[0048] Figure 2 1 is a block diagram showing an example of the functional structure of the design support device involved in the embodiment. The design support device 10 of the present embodiment is a device that supports the setting of target values ​​related to the observed values ​​of a plurality of characteristic items, wherein the plurality of characteristic items represent the characteristics of a material produced based on a design parameter group composed of a plurality of design parameters. Figure 2 As shown, the design support device 10 may include a functional unit configured as a processor 101, a design parameter storage unit 21, an observation value storage unit 22, a target value storage unit 23, and an optimization range storage unit 24. Each functional unit will be described later.

[0049] Figure 3 1 is a diagram showing an example of a hardware configuration of a computer 100 constituting the design support apparatus 10 according to the embodiment. The computer 100 can constitute the design support apparatus 10 .

[0050] As an example, the computer 100 includes a processor 101, a main storage device 102, an auxiliary storage device 103, and a communication control device 104 as hardware components. The computer 100 constituting the design support apparatus 10 may further include an input device 105 such as a keyboard, a touch panel, or a mouse, and an output device 106 such as a display.

[0051] The processor 101 is a computing device that executes an operating system and application programs. Examples of processors include a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit), but the type of processor 101 is not limited to these. For example, the processor 101 may also be a combination of a sensor and a dedicated circuit. The dedicated circuit may be a programmable circuit such as an FPGA (Field-Programmable Gate Array), or other types of circuits.

[0052] The main storage device 102 is a device that stores programs for implementing the design support device 10, etc., calculation results output from the processor 101, etc. The main storage device 102 is composed of at least one of a ROM (Read Only Memory) and a RAM (Random Access Memory), for example.

[0053] The auxiliary storage device 103 is generally a device capable of storing a larger amount of data than the main storage device 102. The auxiliary storage device 103 is composed of a non-volatile storage medium such as a hard disk or a flash memory. The auxiliary storage device 103 stores a design support program P1 and various data for making the computer 100 function as the design support device 10 or the like.

[0054] The communication control device 104 is a device that performs data communication with other computers via a communication network. The communication control device 104 is composed of, for example, a network card or a wireless communication module.

[0055] Each functional element of the design support device 10 is realized by reading the corresponding program P1 on the processor 101 or the main storage device 102 and causing the processor 101 to execute the program. The program P1 includes codes for realizing each functional element of the corresponding server. The processor 101 operates the communication control device 104 according to the program P1 and executes the reading and writing of data in the main storage device 102 or the auxiliary storage device 103. Through such processing, each functional element of the corresponding server is realized.

[0056] The program P1 may be provided after being fixedly recorded on a tangible recording medium such as a CD-ROM, a DVD-ROM, or a semiconductor memory. Alternatively, at least one of the programs may be provided as a data signal superimposed on a carrier wave via a communication network.

[0057] Reference again Figure 2 The design support device 10 includes a data acquisition unit 11, a model construction unit 12, a target value acquisition unit 13, an optimization range acquisition unit 14, a scoring function construction unit 15, a target achievement probability acquisition unit 16 and an output unit 17. Figure 2 As shown, the design parameter storage unit 21 , the observation value storage unit 22 , the target value storage unit 23 , and the optimization range storage unit 24 may be configured as the design support device 10 , or may be configured as other devices accessible by the design support device 10 .

[0058] The data acquisition unit 11 acquires a plurality of performance data related to the manufactured material. The performance data is composed of a pair of a design parameter group and respective observation values ​​of a plurality of characteristic items. The design parameter storage unit 21 is a storage unit for storing the design parameter group in the performance data, and can be constituted by, for example, a main storage device 102 and an auxiliary storage device 103. The observation value storage unit 22 is a storage unit for storing the observation values ​​in the performance data.

[0059] Figure 4 2 is a diagram showing an example of a design parameter group stored in the design parameter storage unit 21. Figure 4 As shown, the design parameter storage unit 21 stores the design parameter set x in the t-th material production (t=1 to T). t The design parameter set x is, for example, composed of P design parameters v from 1 to P. p As an example, the design parameter group x may include the blending amount v1 of the raw material A, the blending amount v2 of the raw material B, ..., the design parameter v P , and vector data of dimension P corresponding to the number of design parameters can be formed. In addition to the exemplified data, the design parameters may also be non-vector data such as molecular structures and images. Furthermore, the design parameters may be data representing options among a plurality of molecules.

[0060] Figure 5 2 is a diagram showing an example of the observed value y stored in the observed value storage unit 22. Figure 5 As shown, the observation value storage unit 22 stores observation values ​​y of a plurality of characteristic items m (m=1 to M) representing the characteristics of the material in the t-th material production (t=1 to T). m,t As an example, the characteristic item m may include the glass transition temperature y 1,t , Adhesion 2,t And the characteristic item M(y M,t ). Design parameter group x t and the observed value y m,t The pairs constitute performance data.

[0061] The model construction unit 12 constructs a prediction model based on the actual performance data. The prediction model is based on the design parameter group x and the observed value y of the characteristic item m. m A model that predicts as a probability distribution or its approximation or alternative indicator. m The type of model used for prediction as a probability distribution or its approximation or alternative indicator is not limited. mPrediction models that make predictions as surrogate indicators for probability distribution use, for example, the distribution of predicted values ​​of predictors constituting an ensemble (random forest), the distribution obtained by Monte Carlo Dropout (neural network), the distribution of predictions of multiple predictors constructed under different conditions (arbitrary machine learning methods), etc. as surrogate indicators to predict the probability distribution of observed values.

[0062] For example, the prediction model can also take the design parameter set x as input and the observation value y m The prediction model is a regression model that uses the probability distribution of the prediction model as the output. When the prediction model is a regression model, the prediction model can be composed of any one of the regression models such as Gaussian process regression, random forest and neural network. The model construction unit 12 can construct the prediction model by a well-known machine learning method using the performance data. The model construction unit 12 can also construct the prediction model by a machine learning method that applies the performance data to the prediction model to update the parameters of the prediction model.

[0063] Furthermore, the prediction model can be a machine learning model that predicts the probability distribution of the observed value or its approximation or alternative index using any one of the posterior distribution of the predicted value based on Bayes' theorem, the distribution of the predicted value of the predictor constituting the ensemble, the theoretical formula of the prediction interval and the reliable interval of the regression model, the Monte Carlo discarding method, and the predicted distribution of multiple predictors constructed under different conditions. The prediction of the probability distribution of the observed value or its alternative index can be obtained by a method inherent to the model. Regarding the probability distribution of the observed value or its approximation or alternative index, if it is Gaussian process regression and Bayesian neural network, it can be obtained based on the posterior distribution of the predicted value, if it is a random forest, it can be obtained based on the predicted distribution of the predictor constituting the ensemble, if it is a linear regression, it can be obtained based on the prediction interval and the reliable interval, and if it is a neural network, it can be obtained based on the Monte Carlo discarding method. However, the calculation method of the distribution of the observed value or its alternative index for each machine learning model is not limited to the above method.

[0064] Furthermore, any model can be extended to a model that can predict the probability distribution of observed values ​​or their substitutes. For example, as an example, the following model can be cited: multiple data sets are constructed by bootstrapping methods, etc., and the distribution of predicted values ​​of each model obtained by constructing a prediction model for each data set is used as a substitute for the probability distribution of the observed value. However, the method of extending the machine learning model to a model that can predict the probability distribution of observed values ​​or their substitutes is not limited to the above method.

[0065] In addition, the prediction model can be constructed through linear regression, PLS regression, Gaussian process regression, bagging ensemble learning such as random forest, boosting ensemble learning such as gradient boosting, support vector machine and neural network.

[0066] In the prediction model constructed as Gaussian process regression, the probability distribution of the observation value is predicted by inputting the design parameter group x in the performance data of the explanatory variables constituting the training data, the observation value y constituting the target variable, and the design parameter x of the prediction object into the model.

[0067] Furthermore, the model building unit 12 may also adjust the hyperparameters of the prediction model by a known hyperparameter adjustment method. That is, the model building unit 12 may update the hyperparameters of the prediction model constructed by Gaussian process regression by using the maximum likelihood estimation of the vector representing the design parameter group x as the explanatory variable in the performance data and the observed value y as the target variable.

[0068] Furthermore, the prediction model may be constructed by a classification model. When the prediction model is a classification model, the model construction unit 12 can construct the prediction model by a machine learning method that can evaluate a known probability distribution using actual performance data.

[0069] In this way, by constructing a prediction model using a predetermined regression model or classification model, the model construction unit 12 can obtain a probability distribution of the observed values ​​of the characteristic items based on an arbitrary design parameter group x.

[0070] Furthermore, the prediction model may also be a single-task model that predicts the observed value of one characteristic item as a probability distribution or its approximation or an alternative indicator, or a multi-task model that predicts the observed values ​​of multiple characteristic items as a probability distribution or its approximation or an alternative indicator. In this way, by constructing a prediction model with a multi-task model or a single-task model appropriately constructed according to the properties of the characteristic items, the prediction accuracy of the observed values ​​based on the prediction model can be improved.

[0071] The model construction unit 12 may construct a prediction model of at least one characteristic item among the plurality of characteristic items based on the input of parameters constituting the prediction model. Specifically, the model construction unit 12 may construct the prediction model based on the input of parameters constituting the prediction model and contributing to probability calculation.

[0072] For example, when the prediction model is constructed as a model for predicting the observation value of a characteristic item that obeys a Gaussian distribution, the model building unit 12 can obtain the parameters representing the distribution of the observation values, i.e., the standard deviation or variance of the Gaussian distribution, based on user input, etc., and construct a prediction model for the characteristic item based on the obtained standard deviation or variance.

[0073] In this way, by inputting parameters that constitute the prediction model and contribute to the probability calculation, part of the probability calculation no longer depends on the actual performance data. Therefore, when there is no sufficient amount and accuracy of actual performance data, for example, it is possible to prevent the accuracy of the overall achievement probability from decreasing due to the deterioration of the estimation accuracy of the observed values ​​of the characteristic items in the prediction model.

[0074] The target value acquisition unit 13 acquires a target value related to the observed value of each characteristic item. Figure 6 is the target value y of each characteristic item stored in the target value storage unit 23. m(target) The target value storage unit 23 is a storage means for storing a target value input by a user via a predetermined user interface or a target value set in advance.

[0075] like Figure 6 As shown, the target value storage unit 23 stores the target value y of the characteristic item m (m=1 to M). m(target) As an example, the target value y m(target) Can include target value y for glass transition temperature 1(target) , target value of adhesion y 2(target) and the target value y of the characteristic item M M(target) .

[0076] The target value acquisition section 13 may acquire a target value related to at least one characteristic item among the plurality of characteristic items based on a user's input.

[0077] Figure 7 Schematically shows an example of a display screen example (user interface) for inputting a target value and outputting an overall achievement probability. Figure 7 As shown, the design support device 10 can display a user interface G on a display or the like as one form of the output device 106. The user interface G includes a user interface for inputting a target value y. m(target) Specifically, the user interface G includes a plurality of target values ​​y for inputting characteristic items 1, 2, 3, ..., M. 1(target) ,y 2(target) ,y 3(target) ,……,y M(target) The target value acquisition unit 13 acquires the target value of the characteristic item based on the operation input of the operation object gt by the user.

[0078] According to this method of acquiring the target value, the user can easily set the target value of each characteristic item. Therefore, by changing the input target value and repeatedly inputting the target value and outputting the overall achievement probability, it is easy to set an appropriate target value. In addition, the objects ga and gb included in the user interface G will be described later.

[0079] Furthermore, the target value acquisition unit 13 may acquire the target value associated with at least one of the plurality of characteristic items from a predetermined database in which target values ​​of characteristic items are pre-stored.

[0080] For example, a target value group having actual performance as appropriate target values ​​may be pre-stored in the target value storage unit 23, and the target value acquisition unit 13 may acquire the target value group having actual performance from the target value storage unit 23. In this way, by pre-storing an appropriate target value group, it is possible to easily acquire an initial target value group for setting target values ​​and studying.

[0081] Furthermore, actual performance values ​​with actual performance as observed values ​​of each characteristic item may be stored in advance in the target value storage unit 23, and the target value acquisition unit 13 acquires the actual performance value of each characteristic item from the target value storage unit 23. Thus, by applying the actual performance value to the initial target value, appropriate target setting becomes easy.

[0082] Furthermore, by using the pre-stored target values ​​of each characteristic item as the initial target value, for example, when the number of characteristic items is large, if the acquisition of the target value is based on the user's input, the operation related to the input becomes cumbersome, and the operation can be simplified.

[0083] The optimization range acquisition unit 14 acquires the optimization range of at least one design parameter among the design parameters included in the design parameter group. The optimization range is a search range of each design parameter as an explanatory variable in the optimization process of the score function.

[0084] The optimization range storage unit 24 stores the design parameters v p A storage mechanism for optimizing the range of. Figure 8 is the value of each design parameter v stored in the optimization range storage unit 24. p Figure 1 shows an example of the optimization range for . Figure 8 As shown, the optimization range storage unit 24 stores each design parameter v p The optimal range of design parameters v p The optimization range of can be set in the case where the possible values ​​of the design parameters as explanatory variables are limited to a plurality of limited candidates.

[0085] Design parameters p The optimization range of is determined by at least one of the upper limit, lower limit, upper and lower limits, and fixed value of the design parameter value, or may be determined by multiple of them. Figure 8 As shown in the figure, the optimization range of design parameter v1 is defined by upper and lower limits, which are set to "0-10". The optimization range of design parameter v2 is defined by a fixed value, which is set to "5". The optimization range of design parameter v3 is defined by two upper and lower limits, which are set to "0-10, 30-49".P The optimization range of is specified by 2 fixed values, set to "2 or 3".

[0086] When there is a desired value as the value of the design parameter, the design parameter v is specified by a fixed value. p The optimization range of the design parameter v p Since the optimization is performed while being fixed, it is possible to obtain the overall achievement probability within the range of the desired design parameters fixed as a result of the optimization.

[0087] As a more specific example, by setting the design parameter v representing the amount of expensive material p The optimization range of is set to "0", and the optimization can be performed without using the material as a condition. p When the optimization range is set to a fixed value to implement the optimization result and a lower target achievement probability is obtained, the design parameter v p By changing the optimization range to, for example, upper and lower limit values ​​having a certain degree of width, and performing the optimization again to evaluate the target achievement probability, it is possible to easily conduct a study on a preferred optimization range.

[0088] The optimization range storage unit 24 can store the design parameters v input by the user via the input device 105. p The optimization range acquisition unit 14 can acquire the design parameters v input by the user. p In this case, the user can easily set and change the design parameters v p Optimization range.

[0089] In this way, by optimizing the scoring function within the optimization range of the acquired design parameters, the overall achievement probability within the specified optimization range can be obtained. Therefore, for example, by setting the range that can be easily adopted as the setting design parameter as the optimization range, the overall achievement probability within the range of the easily adopted design parameters can be obtained. In addition, when the overall achievement probability within the range of the easily adopted design parameters is low, by repeatedly changing the range of the design parameters and confirming the overall achievement probability, the possibility of achieving the target can be increased, and the range of the adoptable design parameters can be obtained relatively easily.

[0090] The scoring function construction unit 15 constructs a scoring function including at least an overall achievement probability term. The overall achievement probability term represents the overall achievement probability as the probability of achieving the target values ​​of all characteristic items. The overall achievement probability is calculated based on the prediction model using the design parameter group as a variable. As long as the scoring function is a function including the overall achievement probability term, its structure is not limited.

[0091] Specifically, the scoring function construction unit 15 constructs a scoring function A(x) as shown in the following equation (1).

[0092] A(x)=g(P(x))…(1)

[0093] In formula (1), g(P(x)) is the target achievement probability term. That is, the scoring function A(x) at least includes the target achievement probability term g(P(x)). In addition, the scoring function of formula (1) is an example, and the scoring function may include terms other than the target achievement probability term.

[0094] The target achievement probability term can be configured to include an overall achievement probability P(x). For example, if the target achievement events of each characteristic item are independent of each other, the overall achievement probability P(x) can be defined as the following formula (2).

[0095] P(x)=Π 1≤m≤M Pm(x)…(2)

[0096] That is, the overall achievement probability P(x) is the product of the achievement probabilities Pm(x) of each characteristic item m (m=1~M). The prediction model can predict the probability distribution of the observed values ​​of the characteristic items based on the design parameter group x, so the achievement probability Pm(x) of each characteristic item can be expressed as a function using the design parameter group x of the prediction model of each characteristic item as an input variable. In addition, the overall achievement probability P(x) can be expressed as a function using the design parameter group x of the prediction model based on all characteristic items as an input variable without calculating the achievement probability of each characteristic item.

[0097] For example, the target achievement probability term g(P(x)) can be composed of the overall achievement probability P(x) as shown in formula (3), or can be composed of the logarithm of the overall achievement probability P(x) as shown in formula (4).

[0098] g(P(x))=P(x)…(3)

[0099] g(P(x))=log(P(x))…(4)

[0100] Furthermore, the target achievement probability term may be a term obtained by further multiplying the overall achievement probability P(x) or the logarithm of the overall achievement probability P(x) by a coefficient, and may further include terms for adding other factors. In an example of the scoring function constructed in this way, its optimization can be regarded as a maximization problem.

[0101] Thus, in one example of the present embodiment, the prediction model is configured to output the probability distribution of the observed value, so that the achievement probability Pm(x) of the target value of each characteristic item m corresponding to the design parameter group x can be obtained. Furthermore, the overall achievement probability calculated by multiplying the achievement probability Pm(x) of the target value of each characteristic item m is included in the target achievement probability term of the scoring function, so the overall achievement probability is appropriately reflected in the scoring function that is the optimization object.

[0102] The target achievement probability acquisition unit 16 acquires the overall achievement probability by optimizing the scoring function. In this embodiment, based on the above-mentioned equations (1) and (3), the scoring function A(x) is constructed as shown in equation (4).

[0103] A(x)=P(x)…(4)

[0104] Then, by optimizing (maximizing) the scoring function A(x), the target achievement probability acquisition unit 16 can acquire the overall achievement probability as the output of the scoring function A(x).

[0105] In the example of formula (4), the scoring function outputs the overall achievement probability, but when the output of the scoring function is the overall achievement probability multiplied by or added to the value of other factors, the target achievement probability acquisition unit 16 can obtain the overall achievement probability P(x) based on the design parameter group x and the target value of each characteristic item obtained by optimizing the scoring function.

[0106] The output unit 17 outputs the overall achievement probability acquired by the target achievement probability acquisition unit 16. The output method is not limited, and the output unit 17 can display the overall achievement probability on a display as one form of the output device 106. Figure 7 As shown, the output unit 17 may display the overall achievement probability acquired by the target achievement probability acquisition unit 16 in the overall achievement probability column ga of the user interface G.

[0107] Furthermore, the output unit 17 may output the target value acquired by the target value acquisition unit 13 together with the overall achievement probability. Figure 7 The user interface G shown includes a m(target) Since the operation object gt is an operation object gt, the output unit 17 can output and display the set target value and the overall achievement probability together by displaying the operation object gt and the overall achievement probability column ga on the display.

[0108] In this way, the target values ​​of the characteristic items in the optimization process of the scoring function and the overall achievement probability obtained by the optimization are output together, for example, by displaying them, so that the correspondence between the target value group and the overall achievement probability can be easily recognized.

[0109] Furthermore, the output unit 17 can output the target value achievement probability (Pm(x)) of each characteristic item m that can be obtained by the target achievement probability acquisition unit 16 in the process of obtaining the overall achievement probability together with the overall achievement probability. For example, by displaying the target value achievement probability of each characteristic item m on a display or the like, the target value of the characteristic item that becomes a bottleneck for achieving a high overall achievement probability can be identified, and thus by adjusting the target value such as relaxing the target value, it becomes easy to improve the overall achievement probability.

[0110] Furthermore, the target achievement probability acquisition unit 16 can further acquire a design parameter group that maximizes the overall achievement probability by optimizing the scoring function. Furthermore, the output unit 17 can further output the design parameter group that maximizes the overall achievement probability acquired by the target achievement probability acquisition unit 16. By referring to the output design parameter group, it is possible to confirm whether the design parameter group that can achieve the output overall achievement probability is appropriate. Furthermore, by referring to the output design parameter group, it is possible to implement the production and experiment of products, etc.

[0111] And, if Figure 7 As shown, the user interface G may include an instruction button gb for receiving an instruction to execute the optimization process based on the set target value. If the design support device 10 receives the user's instruction input to the instruction button gb, the scoring function based on the set target value is constructed and optimized, and the overall achievement probability is displayed in the overall achievement probability column ga. In this way, the setting of the target value, the construction and optimization of the scoring function, and the output of the overall achievement probability can be repeatedly implemented by referring to and operating the user interface G, and the study of the appropriate target value can be implemented.

[0112] Next, the design support process by the design support device 10 according to the present embodiment will be described in detail. Fig. 9 1 is a flowchart showing an example of the content of the design support method in the design support device 10 according to the embodiment. The design support method is executed by reading the design support program P1 in the processor 101 and executing the program, thereby realizing the functional units 11 to 17 .

[0113] In step S1, the data acquisition unit 11 acquires the design parameter set x t (t is an integer from 1 to T) and the observed value y of each characteristic item m,t The performance data composed of pairs.

[0114] In step S2 , the model construction unit 12 constructs each prediction model of the characteristic item m based on the actual performance data.

[0115] In step S3 , the target value acquisition unit 13 acquires the target value associated with the observed value of each characteristic item m.

[0116] In step S4 , the scoring function construction unit 15 constructs a scoring function based on the prediction model constructed in step S2 and the target value of each characteristic item acquired in step S3 .

[0117] In step S5, the target achievement probability acquisition unit 16 optimizes the scoring function. Here, the optimization range acquisition unit 14 acquires the design parameters v p The target achievement probability acquisition unit 16 optimizes the scoring function based on the acquired optimization range.

[0118] In step S6 , the target achievement probability acquisition unit 16 acquires the overall achievement probability obtainable based on the optimization of the scoring function A.

[0119] In step S7 , the output unit 17 outputs (displays) the overall achievement probability acquired by the target achievement probability acquisition unit 16 in step S6 .

[0120] Next, refer to Fig.10 , other examples of the output of the overall achievement probability are explained. Fig.10 It is a diagram schematically showing an example of a display screen (user interface) that displays a target value and an overall achievement probability in association with each other.

[0121] The target value acquisition unit 13 may acquire a range of target values ​​associated with an observed value of at least one characteristic item among the plurality of characteristic items. That is, the target value acquisition unit 13 may acquire a numerical range instead of a single value representing the target value. Fig.10 In the example shown, for convenience of explanation, the target value acquisition unit 13 acquires the numerical range (y) of the target values ​​of two characteristic items a and b among at least two characteristic items based on the input of the user. a(target) :1.0~4.0,y b(target) :10~40) when.

[0122] The scoring function construction unit 15 constructs a scoring function corresponding to each target value within the target value range. Specifically, for example, the scoring function construction unit 15 constructs a scoring function for each combination of points divided at predetermined intervals within the target value range of each characteristic item.

[0123] The target achievement probability acquisition unit 16 acquires the overall achievement probability corresponding to the combination of each target value within the acquired target value range by optimizing each scoring function. The output unit 17 associates each target value within the set target value range with the corresponding overall achievement probability and outputs it. Fig.10 In the example shown, the target value y a(target) Set on the horizontal axis, target value y b(target) The output unit 17 outputs the target value ya(target) and the target value y b(target) The overall achievement probability corresponding to the combination of , and the contour lines based on the output overall achievement probability are drawn and output.

[0124] By referring to the association between the combination of each target value and the overall achievement probability through the output of such a display screen, it is possible to recognize the change in the overall achievement probability corresponding to the change in the target value of the characteristic item. Therefore, it becomes easy to set an appropriate target value.

[0125] Next, a design support program for causing a computer to function as the design support device 10 according to the present embodiment will be described. Fig.11 This is a diagram showing the structure of a design support program.

[0126] The design support program P1 is composed of a main module m10 that centrally controls the design support processing in the design support device 10, a data acquisition module m11, a model construction module m12, a target value acquisition module m13, an optimization range acquisition module m14, a scoring function construction module m15, a target achievement probability acquisition module m16, and an output module m17. In addition, through each module m11 to m17, each function for the data acquisition unit 11, the model construction unit 12, the target value acquisition unit 13, the optimization range acquisition unit 14, the scoring function construction unit 15, the target achievement probability acquisition unit 16, and the output unit 17 can be realized.

[0127] In addition, the design support program P1 may be transmitted via a transmission medium such as a communication line, or may be transmitted as follows: Fig.11 As shown, this is the manner of storage in the recording medium M1.

[0128] According to the design support device 10, the design support method, and the design support program P1 of the present embodiment described above, a prediction model is constructed based on actual performance data, and a target value related to each characteristic item is obtained, and a scoring function including a target achievement probability item is constructed based on the obtained target value and the prediction model. Furthermore, by optimizing the scoring function, the overall achievement probability is obtained and output. Therefore, by referring to the output overall achievement probability, it is possible to determine whether the target value is appropriate, so it is easy to set an appropriate target value.

[0129] The present invention has been described in detail based on its embodiments. However, the present invention is not limited to the above-mentioned embodiments. The present invention can be modified in various ways without departing from the gist of the present invention.

[0130] Explanation of symbols

[0131] 10-design support device, 11-data acquisition unit, 12-model construction unit, 13-target value acquisition unit, 14-optimization range acquisition unit, 15-scoring function construction unit, 16-target achievement probability acquisition unit, 17-output unit, 21-design parameter storage unit, 22-observation value storage unit, 23-target value storage unit, 24-optimization range storage unit, M1-recording medium, m10-main module, m11-data acquisition module, m12-model construction module, m13-target value acquisition module, m14-optimization range acquisition module, m15-scoring function construction module, m16-target achievement probability acquisition module, m17-output module, P1-design support program.

Claims

1. A design support device that supports setting of target values ​​associated with observed values ​​of a plurality of characteristic items, wherein the plurality of characteristic items represent characteristics of a product, work-in-progress, semi-finished product, part, or prototype manufactured based on a design parameter group consisting of a plurality of design parameters, wherein: The design support device comprises: A data acquisition unit, which acquires a plurality of performance data related to the manufactured product, the work-in-progress, the semi-finished product, the part or the trial product, and is composed of the design parameter group and the observed values ​​of each of the plurality of characteristic items; A model building unit, which builds a prediction model based on the performance data, wherein the prediction model uses the observed value of the characteristic item as a probability distribution or an approximate or alternative indicator thereof to perform prediction based on the design parameter group; A target value acquisition unit, which acquires a target value related to the observed value of each characteristic item; a scoring function construction unit, which constructs a scoring function including at least a target achievement probability term, wherein the target achievement probability term represents an overall achievement probability calculated based on the prediction model using the design parameter group as a variable and is a probability of achieving the target values ​​of all characteristic items; a target achievement probability acquisition unit, which acquires the overall achievement probability by optimizing the scoring function; and An output unit outputs the overall achievement probability.

2. The design support device according to claim 1, wherein: The target value acquisition section acquires the target value associated with at least one characteristic item among the plurality of characteristic items based on a user input.

3. The design support device according to claim 1, wherein: The target value acquisition unit acquires a target value associated with at least one of the plurality of characteristic items from a prescribed database in which target values ​​of the characteristic items are pre-stored.

4. The design support device according to claim 1, wherein: The output unit outputs the target value acquired by the target value acquisition unit together with the overall achievement probability.

5. The design support device according to claim 4, wherein: The target value acquisition unit acquires a range of the target value related to an observed value of at least one characteristic item among the plurality of characteristic items, The scoring function construction unit constructs the scoring function corresponding to each target value within the target value range. The target achievement probability acquisition unit acquires the overall achievement probability corresponding to each target value within the acquired range of the target value by optimizing each scoring function. The output unit associates each target value within the set range of the target value with the corresponding overall achievement probability and outputs the resultant target value.

6. The design support device according to claim 1, wherein: The target achievement probability acquisition unit acquires the probability of achieving the target value of each characteristic item, The output unit outputs the achievement probability of the target value of each characteristic item together with the overall achievement probability.

7. The design support device according to claim 1, further comprising: an optimization range acquisition unit, which acquires an optimization range of at least one design parameter among the design parameters included in the design parameter group as a search range in the optimization process of the scoring function, wherein the optimization range of the design parameter is specified by at least one of an upper limit value, a lower limit value, upper and lower limits, and a fixed value of the design parameter value; The target achievement probability acquisition unit optimizes the scoring function based on the acquired optimization range.

8. The design support device according to claim 1, wherein: The model construction section constructs the prediction model of at least one characteristic item among a plurality of characteristic items based on input of parameters constituting the prediction model.

9. The design support device according to claim 1, wherein: The target achievement probability acquisition unit acquires the design parameter group that maximizes the overall achievement probability by optimizing the scoring function. The output unit outputs the design parameter group that maximizes the overall achievement probability.

10. A design support method in a design support device, which supports setting of target values ​​associated with observed values ​​of a plurality of characteristic items, wherein the plurality of characteristic items represent characteristics of a product, work-in-progress, semi-finished product, part, or prototype manufactured based on a design parameter group consisting of a plurality of design parameters, wherein: The design support method comprises the following steps: A data acquisition step of acquiring a plurality of performance data related to the finished product, the work-in-progress, the semi-finished product, the part or the trial product, which are composed of the design parameter group and the observed values ​​of each of the plurality of characteristic items; A model building step, building a prediction model based on the performance data, wherein the prediction model uses the observed values ​​of the characteristic items as probability distribution or its approximation or alternative index to perform prediction based on the design parameter group; A target value acquisition step, acquiring a target value related to the observed value of each characteristic item; A scoring function construction step, constructing a scoring function including at least a target achievement probability term, wherein the target achievement probability term represents an overall achievement probability calculated based on the prediction model using the design parameter group as a variable and is a probability of achieving the target values ​​of all characteristic items; a target achievement probability acquisition step, obtaining the overall achievement probability by optimizing the scoring function; and The output step outputs the overall achievement probability.

11. A design support program for causing a computer to function as a design support device for supporting setting of target values ​​associated with observed values ​​of a plurality of characteristic items, wherein the plurality of characteristic items represent characteristics of a product, work-in-progress, semi-finished product, part, or prototype manufactured based on a design parameter group consisting of a plurality of design parameters, wherein: The design support program enables the computer to implement the following functions: A data acquisition function for acquiring a plurality of performance data related to the finished product, the work-in-progress, the semi-finished product, the part or the trial product, which are composed of the design parameter group and the observed values ​​of each of the plurality of characteristic items; A model building function, which builds a prediction model based on the performance data, wherein the prediction model uses the observed values ​​of the characteristic items as probability distribution or its approximation or alternative index to make predictions based on the design parameter group; Target value acquisition function, to obtain the target value related to the observed value of each characteristic item; A scoring function construction function is used to construct a scoring function including at least a target achievement probability term, wherein the target achievement probability term represents an overall achievement probability calculated based on the prediction model using the design parameter group as a variable and is a probability of achieving the target values ​​of all characteristic items; A target achievement probability acquisition function, which acquires the overall achievement probability by optimizing the scoring function; and An output function is provided to output the overall achievement probability.

Images