Automatic material optimization method, device and equipment based on integration algorithm
By integrating algorithms to construct multiple material property prediction models, the problem of inaccurate optimization results from single models in existing technologies is solved, achieving high efficiency and accuracy in material optimization and providing an intelligent material design tool.
Patent Information
- Application Number
- CN202511626328.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-07
- Publication Date
- 2026-02-06
AI Technical Summary
Existing material optimization methods rely on only one machine learning fitting model, resulting in inaccurate optimization results, low efficiency, and high cost.
An automatic material optimization method based on ensemble algorithms is adopted. Multiple material performance prediction sub-models are established through a machine learning regression algorithm library, an ensemble model for material performance prediction is constructed, the optimal model is determined by the correlation coefficient, and the material composition and process are optimized by combining the optimization algorithm.
It improves the accuracy and efficiency of material optimization, reduces manual operations, and provides intelligent material design tools.
Smart Images
Figure CN121483451A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of material design, and particularly relates to a material automatic optimization method, device and equipment based on an integrated algorithm. BACKGROUND
[0002] In the field of materials, it is an eternal topic to pursue higher performance indicators of materials by optimizing the component ratio and process parameters of the materials.
[0003] In the traditional material optimization process, the trial-and-error method is mainly used. The component ratio and process parameters of the materials are manually adjusted to formulate an experimental scheme, and then experiments are performed to obtain the performance indicators of the materials under the experimental scheme. However, due to the limitation of the experience of experts, the optimization result also has limitations, and experiments can only be performed near the optimal region range in the experience. The development cycle is long, the efficiency is low, and the cost is very high.
[0004] Then, an automatic material optimization method, called automatic optimization, is used. First, a data set, i.e., the correspondence between the material performance and the component ratio and process parameters, is collected. Then, a machine learning fitting model is obtained by modeling the relationship between the material performance and the component ratio and process parameters. Then, the genetic algorithm and other automatic optimization algorithms are used to obtain the optimal performance component ratio and process parameters. However, this method also has limitations. The genetic algorithm only uses one kind of machine learning fitting model for optimization, but each machine learning fitting model algorithm has its special properties. Using only one algorithm leads to inaccurate optimization results. SUMMARY
[0005] Therefore, the present application provides a material automatic optimization method, device and equipment based on an integrated algorithm to overcome the problem that the current material optimization only uses one algorithm, leading to inaccurate optimization results.
[0006] To achieve the above purpose, the present application adopts the following technical solutions: In a first aspect, the present application provides a material automatic optimization method based on an integrated algorithm, comprising: Step S1, obtaining original data, and automatically and randomly dividing the original data into a training data set and a test data set according to a preset division ratio and a division number M; Step S2, according to N algorithms covered by a machine learning regression algorithm library, establishing a material performance prediction sub-model set F(x i ) for M training data sets, and calculating the correlation coefficient R i 2 of each material performance prediction sub-model, respectively; wherein the number of the material performance prediction sub-models is MxN; Step S3, constructing a material performance prediction integrated model set U(x i ) according to the material performance prediction sub-model set; Step S4, training each material performance prediction integrated model through the training data set; Step S5, verifying the trained material performance prediction integrated model through the test data set, and determining the optimal material performance prediction integrated model through the correlation coefficient R i 2 of each material performance prediction integrated model; Step S6, obtaining material data and predicting through the optimal material performance prediction integrated model; Step S7, according to the prediction result and the optimal material performance prediction integrated model, taking the optimal material performance prediction integrated model as the objective function of the optimization algorithm, and using m kinds of optimization algorithms to perform material composition and process optimization calculation to obtain m kinds of optimization results of composition and process.
[0007] Further, the above method, the original data includes: material characteristic data and material performance index data.
[0008] Further, the above method further comprises, before the step S2: performing data preprocessing on the original data through a normalization method and a principal component analysis method.
[0009] Further, the above method, the step S3, comprises: constructing a first material performance prediction integrated model according to the material performance prediction sub-model set whose correlation coefficient R i 2 is greater than 0.5; the correlation coefficient R i 2 as the weight term of the material performance prediction sub-model integrated calculation parameter in the first material performance prediction integrated model; the error function of the first material performance prediction integrated model is: ; constructing a second material performance prediction integrated model according to the material performance prediction sub-model set whose correlation coefficient R i 2 is greater than 0.8; the correlation coefficient R i 2 as the weight term of the material performance prediction sub-model integrated calculation parameter in the second material performance prediction integrated model; The error function of the second material performance prediction integrated model is: ; The correlation coefficient R of each material performance prediction sub-model in the material performance prediction sub-model set i 2 is compared, and the correlation coefficient R i 2 The highest and the second highest material performance prediction sub-models are used to construct a third material performance prediction integrated model; The error function of the third material performance prediction integrated model is: .
[0010] Further, the above method, the step S5, comprises: The trained material performance prediction integrated model is verified by the test data set; The correlation coefficient R of each material performance prediction integrated model is calculated 2 The correlation coefficient R of each material performance prediction integrated model is calculated 2 ; wherein the correlation coefficient R 2 The calculation formula is: ; Wherein, R j 2 is the correlation coefficient R of the jth material performance prediction integrated model 2 , is the average value of the true value; The material performance prediction integrated model with the highest correlation coefficient R 2 is determined as the optimal material performance prediction integrated model.
[0011] Further, the above method, before the step S6, further comprises: The model effect of the optimal material performance prediction integrated model is calculated; According to the model effect, it is judged whether the material performance prediction integrated model set needs to be retrained.
[0012] Secondly, the application provides a material automatic optimization device based on an integrated algorithm, comprising: A data acquisition module is configured to acquire original data and material data, and automatically and randomly divide the original data into training data sets and test data sets according to a preset division ratio and a division number M; A model creation module is configured to establish a material performance prediction sub-model set F(x according to N algorithms covered by a machine learning regression algorithm library, for M training data sets i), and the correlation coefficient R of each material performance prediction sub-model is calculated i 2 , the number of the material performance prediction sub-models is MxN, and a material performance prediction integrated model set U(x i ) is constructed according to the material performance prediction sub-model set, each material performance prediction integrated model is trained through the training data set, the trained material performance prediction integrated model is verified through the test data set, and the correlation coefficient R i 2 of each material performance prediction integrated model is determined to determine the optimal material performance prediction integrated model; An automatic optimization module is configured to perform prediction through the optimal material performance prediction integrated model, use the optimal material performance prediction integrated model as a target function of an optimization algorithm according to a prediction result and the optimal material performance prediction integrated model, and perform material composition and process optimization calculation using m optimization algorithms to obtain m optimization results of the composition and the process.
[0013] In a third aspect, the present application provides a material automatic optimization equipment based on an integrated algorithm, including a processor and a memory, the processor being connected with the memory: The processor is configured to call and execute a program stored in the memory. The memory is configured to store the program, and the program is used to execute the material automatic optimization method based on the integrated algorithm.
[0014] The present application has the following advantages: Firstly, the present application acquires original data, automatically and randomly divides the original data into a training data set and a test data set according to a preset division proportion and a division number M, then establishes a material performance prediction sub-model set F(x i ) according to N algorithms covered by a machine learning regression algorithm library, and calculates the correlation coefficient R of each material performance prediction sub-model i 2 , and constructs a material performance prediction integrated model set U(x i ) according to the material performance prediction sub-model set, trains each material performance prediction integrated model through the training data set, verifies the trained material performance prediction integrated model through the test data set, and determines the correlation coefficient R i 2, determine the optimal material performance prediction integrated model, finally, obtain the material data, predict through the optimal material performance prediction integrated model, according to the prediction result and the optimal material performance prediction integrated model, the optimal material performance prediction integrated model is used as the objective function of the optimization algorithm, and m kinds of optimization algorithms are used for material composition and process optimization calculation, and m kinds of optimization results of composition and process are obtained. Therefore, the problem that the optimization result is inaccurate due to the use of only one algorithm in the current material optimization is solved. The present application has automatic optimization workflow, can automatically clean the material design data set, divide data, train integrated calculation model, automatically optimize composition and process multi-objective and generate optimization scheme report, and no programming and personnel operation are needed in the process, so that the material design efficiency can be greatly improved, and an efficient intelligent tool is provided for material designers. BRIEF DESCRIPTION OF DRAWINGS
[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.
[0016] Figure 1 is a flow chart provided by an embodiment of the present application; Figure 2 is a structural schematic diagram of an embodiment of the material automatic optimization device based on the integrated algorithm provided by the present application; Figure 3 is a structural schematic diagram of an embodiment of the material automatic optimization device based on the integrated algorithm provided by the present application; Figure 4 is a first prediction value and true value comparison chart provided by an embodiment of the material automatic optimization method based on the integrated algorithm of the present application; Figure 5 is a second prediction value and true value comparison chart provided by an embodiment of the material automatic optimization method based on the integrated algorithm of the present application; Figure 6 is a third prediction value and true value comparison chart provided by an embodiment of the material automatic optimization method based on the integrated algorithm of the present application. DETAILED DESCRIPTION
[0017] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described in detail below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the scope of protection of the present application.
[0018] Figure 1 is a flowchart provided by an embodiment of the present application. Please refer to Figure 1 , the embodiment can include the following steps: Step S1, obtaining original data, and automatically and randomly dividing the original data into training data sets and test data sets according to a preset division ratio and a division number M; Step S2, according to N algorithms covered by a machine learning regression algorithm library, establishing a material performance prediction sub-model set F(x i ) for M training data sets, and respectively calculating the correlation coefficients R i 2 of each material performance prediction sub-model; wherein the number of material performance prediction sub-models is MxN; Step S3, constructing a material performance prediction integrated model set U(x i ) according to the material performance prediction sub-model set; Step S4, training each material performance prediction integrated model through the training data set; Step S5, verifying the trained material performance prediction integrated model through the test data set, and determining the optimal material performance prediction integrated model through the correlation coefficient R i 2 of each material performance prediction integrated model; Step S6, obtaining material data, and predicting through the optimal material performance prediction integrated model; Step S7, according to the prediction result and the optimal material performance prediction integrated model, taking the optimal material performance prediction integrated model as an objective function of an optimization algorithm, and using m optimization algorithms to perform material composition and process optimization calculation to obtain m optimization results of the composition and the process.
[0019] It can be understood that, first, the embodiment obtains original data, and automatically and randomly divides the original data into training data sets and test data sets according to a preset division ratio and a division number M, then, according to N algorithms covered by a machine learning regression algorithm library, establishes a material performance prediction sub-model set F(x i ) for M training data sets, and respectively calculates the correlation coefficients R i 2, according to the material performance prediction sub-model set, a material performance prediction integrated model set U(x i ), each material performance prediction integrated model is trained through a training data set, the trained material performance prediction integrated model is verified through a test data set, and the correlation coefficient R i 2 of each material performance prediction integrated model is determined, finally, material data is obtained, the optimal material performance prediction integrated model is used for prediction, and the optimal material performance prediction integrated model is used as an objective function of an optimization algorithm according to the prediction result and the optimal material performance prediction integrated model, and m kinds of optimization algorithms are used for material composition and process optimization calculation to obtain the optimization results of m kinds of compositions and processes. Thus, the problem that the optimization result is inaccurate due to the use of only one algorithm for material optimization is solved. The present application has an automatic optimization workflow, can automatically perform material design data set cleaning, data division, integrated calculation model training, composition and process multi-objective automatic optimization, and generate an optimization scheme report, and no programming and personnel operation are required in the process, which can greatly improve the material design efficiency and provide a highly efficient intelligent tool for material designers.
[0020] It should be noted that material optimization refers to a process of improving the performance indicators of the finally produced material by adjusting the composition ratio of the material and the process parameters used during production. For example, in metal material optimization, the composition ratio is the percentage of iron, chromium, manganese, and niobium in the total mass, and the performance indicators of the material can be directly affected by adjusting the composition ratio; the process parameters are solid solution strengthening temperature and grain size, etc., and by adjusting the process parameters, the metal crystal morphology can be affected under the same composition ratio, which will ultimately affect the performance indicators of the material. The performance indicators usually refer to the strength of the specific material.
[0021] Genetic algorithm (Genetic Algorithm, GA) is an optimization algorithm based on the principles of natural selection and genetics, and is widely used in the optimization and search of complex problems. It simulates the process of biological evolution, and through selection, crossover (recombination) and mutation operations, it constantly evolves solutions with higher fitness, in order to find the optimal or sub-optimal solution to the problem.
[0022] Preferably, the original data includes material characteristic data and material performance indicator data.
[0023] It can be understood that the material characteristic data refers to the composition ratio and the process parameters. The material performance indicator data is the material performance indicator obtained through experiment or computer simulation calculation.
[0024] Preferably, before step S2, it further includes: The original data is preprocessed by normalization method and principal component analysis method.
[0025] It can be understood that the data preprocessing is to process the features by normalization and principal component analysis method. The normalization scales the scale of each feature to 0-1, which can effectively eliminate the scale difference between different features; the principal component analysis is a dimension reduction method, after principal component analysis of all features, the original features can be converted into a new set of features, the first few features of the new features contain most of the information of the original features, and the effect of dimension reduction can be achieved.
[0026] Preferably, step S3 comprises: According to the correlation coefficient R of the material performance prediction sub-model set i 2 The material performance prediction sub-model with a correlation coefficient R greater than 0.5 is used to construct a first material performance prediction integrated model; Correlation coefficient R i 2 As a weight term of the material performance prediction sub-model integrated calculation parameter in the first material performance prediction integrated model; The error function of the first material performance prediction integrated model is: ; According to the correlation coefficient R of the material performance prediction sub-model set i 2 The material performance prediction sub-model with a correlation coefficient R greater than 0.8 is used to construct a second material performance prediction integrated model; Correlation coefficient R i 2 As a weight term of the material performance prediction sub-model integrated calculation parameter in the second material performance prediction integrated model; The error function of the second material performance prediction integrated model is: ; The correlation coefficient R of each material performance prediction sub-model in the material performance prediction sub-model set i 2 is compared, and according to the correlation coefficient R i 2 The highest and the second highest material performance prediction sub-model are used to construct a third material performance prediction integrated model; The error function of the third material performance prediction integrated model is: .
[0027] It can be understood that the integrated model is an algorithm that uses multiple sub-models by weighted average of the output of the sub-models to obtain the output of the integrated model, so as to utilize the advantages of multiple sub-models and obtain better results.
[0028] The output of the integrated model is:
[0029] wherein is the output of the jth sub-model, is the correlation coefficient indicator of the jth sub-model, is the output of the integrated model.
[0030] Preferably, step S5 comprises: verifying the trained material performance prediction integrated model by a test data set; calculating the correlation coefficient R 2 of each material performance prediction integrated model by the formula 2 ; wherein the correlation coefficient R 2 is calculated by the formula: ; wherein R j 2 is the correlation coefficient R 2 of the jth material performance prediction integrated model, is the average value of the true value; determining the material performance prediction integrated model with the highest correlation coefficient R 2 as the optimal material performance prediction integrated model.
[0031] It can be understood that the correlation coefficient is an index used to measure the ability of a model to explain the variation of a variable in regression analysis. It reflects the degree of explanation of the model to the target variable, and the value range is between 0 and 1, and the specific calculation method is as follows
[0032] wherein: is the average value of the actual value; characteristics are: R 2 =1: the model can perfectly explain the variation of the target variable, and all predicted values are completely consistent with the actual values.
[0033] R 2 =0: the model cannot explain the variation of the target variable, and its performance is the same as that of a simple mean prediction model.
[0034] R 2 <0: the model performs worse than the mean prediction model, which usually means that the fitting effect of the model is very poor.
[0035] Preferably, before step S6, it further comprises: Calculate the model effect of the optimal material performance prediction integrated model; According to the model effect, whether the material performance prediction integrated model set needs to be retrained.
[0036] In specific practice, using the three-component iron, niobium and cobalt alloy calculation dataset for optimization, first, a model needs to be built. The K-nearest neighbor algorithm selects 5 nearest neighbors, and all neighbors have equal weights; the Gaussian process regression uses a constant kernel, and the noise intensity is 1e-10.
[0037] In terms of algorithm indicators, the mean square error indicator is improved from 0.0036 of the K-nearest neighbor and 0.0037 of the Gaussian process regression to 0.0023 of the integrated algorithm, and the closer the mean square error indicator is to 0, the better; the correlation coefficient indicator is improved from 0.9894 of the K-nearest neighbor and 0.9889 of the Gaussian process regression to 0.9930 of the integrated algorithm, and the closer the correlation coefficient indicator is to 1, the better.
[0038] In terms of comparison between algorithm prediction values and true values, as shown in the horizontal coordinate of the test set sample in Figures 4-6 , the sample is arranged in ascending order of actual value; the vertical coordinate is the value, the green one is the actual value, and the red one is the predicted value. The closer the two values of the horizontal coordinate are, the more accurate the prediction is, and the smaller the deviation is. From Figures 4-6 , it can be known that the K-nearest neighbor model has a large deviation in the test set sequence number 75 and the value in the vicinity of 375; the Gaussian process regression model has a large deviation in the test set sequence number 215 and the value in the vicinity of 500; and the integrated model has a deviation in the two regions within an acceptable range, achieving good results.
[0039] Finally, the genetic algorithm is used to obtain the optimization result, the population size of the genetic algorithm is 50, the iteration evolution is 50 times, the real number coding is used, and the replacement probability is 0.001. The final optimization result of the mass percentages of iron, niobium and cobalt is 50.24350728, 2.24417832 and 47.5123144 respectively, and the pre-performance indicator is 721.41239791.
[0040] The application also provides a material automatic optimization device based on an integrated algorithm, which is used to realize the above method embodiment. Figure 2 is a structural schematic diagram of an embodiment of the material automatic optimization device based on the integrated algorithm provided by the application. As shown in Figure 2 , it comprises: A data acquisition module 1 is used to acquire original data and material data, and automatically randomly divide the original data into a training data set and a test data set according to a preset division proportion and a division number M; A model creation module 2 is used to establish a material performance prediction sub-model set F(x) according to N algorithms covered by a machine learning regression algorithm library and M training data sets;i ), and the correlation coefficient R of each material performance prediction sub-model is calculated i 2 , wherein the number of material performance prediction sub-models is MxN, and the material performance prediction integrated model set U(x is constructed according to the material performance prediction sub-model set i ), each material performance prediction integrated model is trained through the training data set, the trained material performance prediction integrated model is verified through the test data set, and the correlation coefficient R of each material performance prediction integrated model is calculated i 2 , to determine the optimal material performance prediction integrated model; The automatic optimization module 3 is configured to perform prediction through the optimal material performance prediction integrated model, use the optimal material performance prediction integrated model as a target function of an optimization algorithm according to the prediction result and the optimal material performance prediction integrated model, and perform material composition and process optimization calculation using m optimization algorithms to obtain m optimization results of the composition and the process.
[0041] As to the device in the above-mentioned embodiments, the specific manner in which each module performs an operation has been described in detail in the embodiments related to the method, and thus will not be described in detail here.
[0042] The application further provides a material automatic optimization device based on an integrated algorithm, which is used to implement the above-mentioned method embodiments. Figure 3 is a structural schematic diagram provided by an embodiment of the material automatic optimization device based on an integrated algorithm. As shown in Figure 3 The material automatic optimization device based on an integrated algorithm includes a processor 21 and a memory 22, and the processor 21 is connected to the memory 22. The processor 21 is configured to call and execute a program stored in the memory 22, and the memory 22 is configured to store the program, which is used to at least execute the material automatic optimization method based on an integrated algorithm in the above-mentioned embodiments.
[0043] The specific implementation of the material automatic optimization device based on an integrated algorithm provided by the embodiments of the application can refer to the implementation of the material automatic optimization method based on an integrated algorithm in any of the above-mentioned embodiments, and thus will not be described in detail here.
[0044] It can be understood that the same or similar parts in the above-mentioned embodiments can be mutually referred to, and the content not described in detail in some embodiments can be referred to the same or similar content in other embodiments.
[0045] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.
[0046] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.
[0047] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0048] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.
[0049] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.
[0050] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.
[0051] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0052] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.
Claims
1. A method for automatic material optimization based on ensemble algorithms, characterized in that, include: Step S1: Obtain the raw data and automatically and randomly divide the raw data into training dataset and test dataset according to the preset division ratio and division number M. Step S2: Based on the N algorithms covered by the machine learning regression algorithm library, establish a material property prediction sub-model set F(x) for the M training datasets. i ), and calculate the correlation coefficient R for each material property prediction sub-model. i 2 The number of material property prediction sub-models is M×N. Step S3: Construct an integrated material performance prediction model set U(x) based on the material performance prediction sub-model set. i ); Step S4: Train each material property prediction ensemble model using the training dataset; Step S5: Validate the trained material property prediction ensemble model using the test dataset, and use the correlation coefficient R of each material property prediction ensemble model. i 2 Determine the optimal integrated model for predicting the material properties; Step S6: Obtain material data and make predictions using the optimal integrated model for predicting material properties; Step S7: Based on the prediction results and the optimal integrated model for predicting material properties, the optimal integrated model for predicting material properties is used as the objective function of the optimization algorithm. m optimization algorithms are used to perform material composition and process optimization calculations to obtain the optimization results of m compositions and processes.
2. The method according to claim 1, characterized in that, The raw data includes: material characteristic data and material performance index data.
3. The method according to claim 2, characterized in that, Before step S2, the following is also included: The original data were preprocessed using normalization and principal component analysis.
4. The method according to claim 3, characterized in that, Step S3 includes: Based on the correlation coefficient R in the material property prediction sub-model set i 2 The material property prediction sub-models with a value greater than 0.5 are used to construct the first integrated material property prediction model; The correlation coefficient R i 2 As a weight term in the integrated calculation parameters of the material performance prediction sub-model in the first integrated material performance prediction model; The error function of the first material property prediction ensemble model is: ； Based on the correlation coefficient R in the material property prediction sub-model set i 2 For the material property prediction sub-model with a value greater than 0.8, a second integrated material property prediction model is constructed. The correlation coefficient R i 2 As a weight term in the integrated calculation parameters of the material performance prediction sub-model in the second integrated material performance prediction model; The error function of the second material property prediction ensemble model is: ； The correlation coefficient R of each material property prediction sub-model in the set of material property prediction sub-models i 2 Comparison based on the correlation coefficient R i 2 The highest and second highest material property prediction sub-models are used to construct a third integrated material property prediction model. The error function of the third material property prediction integrated model is: 。 5. The method according to claim 4, characterized in that, Step S5 includes: The trained material property prediction ensemble model was validated using the test dataset. Through the correlation coefficient R 2 The calculation formula calculates the correlation coefficient R for each of the material property prediction integrated models. 2 Wherein, the correlation coefficient R 2 The calculation formula is: ； Among them, R j 2 The correlation coefficient R of the j-th material property prediction ensemble model 2 , The average of the true values; The correlation coefficient R 2 The highest-performing integrated model for predicting material properties is determined to be the optimal integrated model for predicting material properties.
6. The method according to claim 5, characterized in that, Before step S6, the procedure also includes: Calculate the model performance of the optimal integrated model for predicting the material properties; Based on the model's performance, determine whether the integrated model set for predicting material properties needs to be retrained.
7. A material automatic optimization device based on an integrated algorithm, characterized in that, include: The data acquisition module is used to acquire raw data and material data, and automatically and randomly divide the raw data into training datasets and test datasets according to a preset division ratio and division number M. The model creation module is used to build a set of material property prediction sub-models F(x) based on N algorithms covered by the machine learning regression algorithm library for M training datasets. i ), and calculate the correlation coefficient R for each material property prediction sub-model. i 2 The number of material property prediction sub-models is M×N, and an integrated material property prediction model set U(x) is constructed based on the set of material property prediction sub-models. i Each material property prediction ensemble model is trained using the training dataset, and validated using the test dataset. The correlation coefficient R of each material property prediction ensemble model is then used to determine its validity. i 2 Determine the optimal integrated model for predicting the material properties; An automatic optimization module is used to make predictions using the optimal material performance prediction ensemble model. Based on the prediction results and the optimal material performance prediction ensemble model, the module uses the optimal material performance prediction ensemble model as the objective function of the optimization algorithm and uses m optimization algorithms to perform material composition and process optimization calculations to obtain the optimization results of m compositions and processes.
8. An automatic material optimization device based on an integrated algorithm, characterized in that, It includes a processor and a memory, wherein the processor is connected to the memory: The processor is used to call and execute the program stored in the memory; The memory is used to store the program, which is at least used to execute the material automatic optimization method based on the integrated algorithm as described in any one of claims 1-6.