Glass elastic modulus prediction method based on spider bee optimization algorithm and machine learning algorithm

By combining the spider-bee optimization algorithm and machine learning algorithm, a glass elastic modulus database was constructed and the parameters of the random forest model were optimized, which solved the problem of inaccurate glass elastic modulus prediction and achieved fast and accurate prediction results.

CN118969144AInactive Publication Date: 2024-11-15CNBM RESEARCH INSTITUTE FOR ADVANCED GLASS MATERIALS GROUP CO LTD +1

Patent Information

Application Number
CN202410981786.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-19
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing technologies cannot accurately and quickly predict the elastic modulus of glass.

Method used

A method for predicting the elastic modulus of glass based on the spider-bee optimization algorithm and machine learning algorithm is adopted. By constructing a glass elastic modulus database, the parameters are optimized using the random forest model and the spider-bee optimization algorithm to establish the optimal random forest model and predict the elastic modulus of glass components.

Benefits of technology

It improves the accuracy and speed of glass elastic modulus prediction, simplifies the model structure, and enhances convergence speed and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969144A_ABST
    Figure CN118969144A_ABST
Patent Text Reader

Abstract

The invention discloses a glass elastic modulus prediction method based on a spider bee optimization algorithm and a machine learning algorithm. The method comprises the following steps: constructing an elastic modulus database of glass materials with different components; constructing a feature descriptor; providing a training set sample, and initializing parameters of a spider bee search algorithm; setting spider bee optimization algorithm optimization parameters, optimizing random forest algorithm parameters by using the spider bee optimization algorithm optimization parameters to obtain optimal random forest algorithm parameters, and establishing a glass elastic modulus prediction model; inputting test sample data, and predicting the elasticity modulus of the glass. According to the method, the spider bee optimization algorithm is used for optimizing the parameters of the random forest algorithm, the structure is simple, the convergence speed and precision are improved, the optimal random forest algorithm parameters obtained through optimization can obviously improve the performance of the random forest algorithm, and the method has practical significance for improving the accuracy of predicting the glass elastic modulus.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a glass elastic modulus prediction method, and in particular to a glass elastic modulus prediction method based on a spider bee optimization algorithm and a machine learning algorithm. Background Art

[0002] Glass is a non-equilibrium, amorphous material that can spontaneously relax to a supercooled liquid state. Unlike crystals, glass does not need to meet strict stoichiometric rules and can be considered as a continuous solution of chemical elements. Therefore, a large number of elements may become components of glass materials. 80 chemical elements can produce 10 52 There are 10 possible glass compositions. However, the number of reported inorganic glasses is only 10 6 This means there is huge room for exploring new glass-forming compositions with special properties.

[0003] As the basic mechanical property of solid materials, the accurate and fast measurement method of elastic modulus is very basic and very important. There are four types of measurement methods for the elastic modulus of solid materials, which are divided into quasi-static method, low-frequency method, resonance method and wave propagation method. The quasi-static method is to determine the Young's modulus of the material by measuring the relationship between force and deformation under quasi-static loading. The dynamic method is divided into low-frequency method, resonance method and wave propagation method according to the measurement frequency from low to high. At present, the more mature solid elastic modulus test methods are free beam resonance method and pulse excitation method. The essence of these two methods is to measure the elastic modulus by exciting the bending vibration of long strip samples. The free beam resonance method generally uses two suspension wires to suspend the sample and is used to excite and receive vibration signals at the same time. By changing the suspension mode of the suspension wire, the bending vibration or torsional vibration can be selected. This test method is easy to cause the sample to fall off and damage the sample, and it also requires a long experimental preparation time. Molecular dynamics simulation can simulate the performance of glass quickly, but there is a certain error. There is no prediction method that can effectively solve the problem of inaccurate prediction of the elastic modulus of glass. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the glass elastic modulus cannot be accurately and quickly predicted in the prior art, and to propose a glass elastic modulus prediction method based on spider bee optimization algorithm and machine learning algorithm.

[0005] The technical solution adopted by the present invention is:

[0006] The glass elastic modulus prediction method based on spider bee optimization algorithm and machine learning algorithm is characterized by comprising the following steps:

[0007] Step 1: Collect elastic modulus data of oxide glasses of different components and construct a glass elastic modulus database, wherein the glass elastic modulus database includes one-to-one mapping of glass components and their corresponding elastic moduli;

[0008] Step 2: Descriptors with element molar content and preparation process parameters as input parameters, that is, the molar content of each component of the glass is used as a group of descriptors, and the descriptors are constructed with glass preparation process parameters, including heating rate, melting temperature, and holding time;

[0009] Step 3: Using the descriptor constructed in step 2 as the input of the random forest model and the elastic modulus database constructed in step 1 as the output of the random forest model, construct a training set and a test set, and establish a random forest model;

[0010] Step 4: Introduce the spider bee optimization algorithm to optimize the parameters of the selected random forest model;

[0011] Step 5: Establish the best performing random forest model based on the optimized parameters;

[0012] Step 6: For the glass component to be predicted, use the optimal random forest model to predict the glass elastic modulus of the glass component.

[0013] On the basis of the above technical solutions, there are the following further technical solutions:

[0014] The step 4 includes the following sub-steps:

[0015] Step 4-1: Select the parameters to be optimized for the random forest algorithm. The position coordinates of the zebra are the parameters to be optimized. Define the population size N of the spider bee optimization algorithm and search for the upper bound. Search lower bound

[0016]

[0017] Among them, SW Pop is the initial population of spider bees, and the following formula can generate any solution in the search space:

[0018]

[0019] Where t is the generated index; i is the midbody index, i = 1, 2, 3, ..., N; r is a vector consisting of D-dimensional random initialization numbers between 0 and 1;

[0020] Step 4-2: In the search phase, the spider bee randomly explores the search space with a constant step size, as mentioned above, to find spiders that are suitable for its offspring. This behavior is modeled as shown in formula (3). The current position of each spider bee is updated with a constant motion at each generation t to simulate the exploration behavior of the spider bee.

[0021]

[0022] Among them, a and b are two indicators randomly selected from the population to determine the exploration direction, followed by the spider bee, which uses μ1 to determine the constant movement through the current direction, and the formula is as follows:

[0023] μ1=|rn|×r1 (4)

[0024] Where r1 is a randomly generated number between 0 and 1, and rn is also a random number, but generated using a normal distribution;

[0025] Step 4-3: Spider bees sometimes cannot find the spider that fell from the ball, therefore, they search the entire area around the exact spot where the spider fell. According to this behavior, another equation under different exploration methods is constructed. In order to enable the proposed algorithm to explore the area around the discarded spider with a small step size different from that of formula (3), this equation is also based on updating the current spider bee in each generation with a constant movement, based on a position of a spider bee randomly selected from the population representing the location of the dropped spider. The equation is described as follows:

[0026]

[0027] μ2=B×cos(2πl) (6)

[0028] Among them, c is an indicator randomly selected from the population, l is a randomly generated number between 1 and -2,

[0029] Formulas (3) and (5) generate the next position of the female spider bee randomly as follows:

[0030]

[0031] Among them, r3 and r4 are two random numbers in [0,1];

[0032] Step 4-4: Following and escaping phase: The process of the spider bee chasing the spider is simulated, in which the distance between the spider as the prey and the spider bee is initially small and may increase or decrease according to the speed of the spider bee and the prey. Under this trend, we proposed a mathematical model to simulate two situations: (1) the spider bee is faster than the spider (prey) C>0.5, (2) the prey is faster than the spider bee C<0.5, where C is the distance control factor that determines the speed of the spider bee, starting from speed 2 and linearly decreasing to 0. In this case, the spider bee moves faster than the prey, simulating the spider as the prey escaping from the spider bee. The distance between the spider bee and the spider increases. Therefore, when the spider bee moves at a speed lower than 0.5, the rate of change of its position is so small that it cannot reach the prey. The formula is as follows:

[0033]

[0034] In the formula, a is an indicator randomly selected from the population; t and t max Represent the current and maximum values ​​respectively; is a vector that represents a randomly generated value in the interval [0,1]; r6 is a random number in the interval [0,1];

[0035] When the prey spider escapes from the spider bee, the distance between the spider bee and the spider gradually increases. This stage is the initial development stage. As the distance increases, this behavior is simulated using the following formula:

[0036]

[0037] in is a vector generated between k and -k according to the normal distribution. Based on this, k is generated using formula (11) to gradually increase the distance between the spider bee and the spider.

[0038]

[0039] The trade-off between these two tendencies is achieved stochastically as shown in the following equation:

[0040]

[0041] The exchange between the search phase and the following agencies is adjusted according to the following formula:

[0042]

[0043] Where p is a random number in [0,1];

[0044] Step 4-5: Utilization phase, the first equation is based on pulling the spiders towards the area containing the most fit spiders and considering it as the best place to build a nest in order to place the paralyzed spider in its abdomen and lay eggs. The first equation is described as follows:

[0045]

[0046] Among them, SW * Representing the best solution so far, the second equation will build a nest at a randomly selected spider bee location from the population, using an extra step to avoid building two nests at the same location. The equation is designed as follows:

[0047]

[0048] Among them, r3 is a random number generated in the interval [0,1]; γ is a number generated according to the levy flight; a, b, c are the indicators of three solutions randomly selected from the population; is a binary vector that determines when to apply a step size to avoid building two nests at the same location; The allocation formula is as follows:

[0049]

[0050] in, and are two vectors representing random values ​​in the interval [0,1]. Equations (14) and (15) are randomly swapped as follows:

[0051]

[0052] The spider wasp will pull the appropriate spider into a pre-prepared nest.

[0053]

[0054] Step 4-6: Matching phase: Each spider bee represents a possible solution of the current generation, and the spider bee egg represents the newly generated potential solution of this generation. The new solution / spider bee egg is generated by the following formula:

[0055]

[0056] Among them, Crossover is a unified crossover operator SW applied between solutions i t and With Cross Rate (CR); SW i t and are two vectors representing male and spider bees. The male spider bees generated in our proposed algorithm are different from the female spider bees according to the following formula:

[0057]

[0058] Where β and β1 are two numbers randomly generated according to normal distribution, e is an exponential constant, and Generated by the following formula:

[0059]

[0060] In the formula, a, b, and c are the indices of three solutions randomly selected from the population, such that a≠i≠b≠c. Crossover is the process of recombining the genetic material of the parental spider bees to produce offspring eggs with the characteristics of both parents.

[0061] Steps 4-7: During the iteration process, some spider bees in the population will be terminated to provide more function evaluations for other spider bees, which will also reduce the diversity of the population and speed up the convergence to the optimal solution. At each evaluation of the entire function, the length of the new population will be updated using the following formula:

[0062] N=N min +(NN min )×k (23)

[0063] Step 4-8: Output the position and accuracy of the spider bee. The position of the spider bee is the parameter value of the random forest.

[0064] The beneficial effects of the present invention are:

[0065] The present invention uses the spider bee optimization algorithm to optimize the random forest algorithm parameter optimization, has a simple structure, improves the convergence speed and accuracy, and the optimal random forest algorithm parameters obtained by optimization can significantly improve the performance of the random forest algorithm, which has practical significance for improving the accuracy of predicting the glass elastic modulus. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the attached embodiment of the present invention Figure 1 , the technical solutions in the embodiments of the present invention are described clearly and completely. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0068] Embodiment 1:

[0069] Step 1: Collect elastic modulus data of oxide glasses with different components and build a glass elastic modulus database. Specifically, collect elastic modulus data of 800 groups of SiO2-Na2O-CaO-Al2O3 glasses with different contents and build a glass elastic modulus database.

[0070] Step 2: Descriptors with element molar content and preparation process parameters as input parameters;

[0071] Step 2-1: Take the molar contents of elements of 800 groups of SiO2-Na2O-CaO-Al2O3 glasses as a set of descriptors;

[0072] Step 2-2: construct a descriptor based on the glass preparation process parameters (heating rate, melting temperature, and holding time); Step 3: use the descriptor constructed in step 2 as the input of the model, and use the elastic modulus database constructed in step 1 as the output of the model to construct 640 sets of data for training and 160 sets of data for testing, and establish a random forest model;

[0073] Step 4: Introduce the spider bee optimization algorithm to optimize the parameters of the selected random forest model;

[0074] Step 4-1: The hyperparameters of the random forest algorithm (number of decision trees, maximum depth of decision trees, minimum number of separation samples) are the parameters that need to be optimized, the position coordinates of the spider bee are the parameters that need to be optimized, and the population size of the spider bee optimization algorithm is defined as 50, and the maximum number of iterations is 1000;

[0075] Step 4-2: In the search phase, female wasps randomly explore the search space with a constant step size, as mentioned above, to find spiders that are suitable for their offspring. This behavior is modeled as shown in formula (3). The current position of each female wasp is updated with a constant movement at each generation t to simulate the exploration behavior of the female wasp.

[0076]

[0077] Among them, a and b are two indicators randomly selected from the population to determine the exploration direction, followed by the female bee; μ1 is used to determine the constant movement through the current direction, and the formula is as follows:

[0078] μ1=|rn|×r1 (4)

[0079] Where r1 is a randomly generated number between 0 and 1, and rn is also a random number, but generated using a normal distribution. Step 4-3: Spider wasps sometimes fail to find the spider that fell from the ball; therefore, they search the entire area around the exact spot where the spider fell. Based on this behavior, another equation under different exploration methods is constructed in order to enable the proposed algorithm to explore the area around the discarded spider with a small step size different from that of equation (3). This equation is also based on updating the current spider wasp in each generation with a constant movement, based on a position of a female wasp randomly selected from the population representing the location of the dropped spider. The equation is described as follows:

[0080]

[0081] μ2=B×cos(2πl) (6)

[0082] Among them, c is an indicator randomly selected from the population, and l is a randomly generated number between 1 and -2.

[0083] Formulas (3) and (5) generate the next position of the spider bee randomly as follows:

[0084]

[0085] Among them, r3 and r4 are two random numbers in [0,1].

[0086] Step 4-4: Following and escaping phase: The process of the spider bee chasing the spider is simulated, in which the distance between the prey and the spider bee is initially small and may increase or decrease depending on the speed of the spider bee and the prey. Under this trend, we proposed a mathematical model to simulate two cases: (1) the spider bee is faster than the spider (prey) C>0.5 and (2) the prey is faster than the spider bee C<0.5, where C is the distance control factor that determines the speed of the spider bee, starting from speed 2 and linearly decreasing to 0. In this case, the spider bee moves faster than the prey, simulating the spider escaping from the spider bee, and the distance between the spider bee and the spider increases. Therefore, when the spider bee moves at a speed lower than 0.5, the rate of change of its position is so small that it cannot reach the prey. The formula is as follows:

[0087]

[0088]

[0089] In the formula, a is an indicator randomly selected from the population; t and t max Respectively represent the current and maximum values is a vector representing randomly generated values ​​in the interval [0,1]; r6 is a random number in the interval [0,1].

[0090] As the spider escapes from the spider bee, the distance between the spider bee and the spider gradually increases. This stage is the initial development stage. As the distance increases, this behavior is simulated using the following formula:

[0091]

[0092] in is a vector generated between k and -k according to the normal distribution. Based on this, k is generated using formula (11) to gradually increase the distance between the spider bee and the spider.

[0093]

[0094] The trade-off between these two tendencies is achieved stochastically as shown in the following equation:

[0095]

[0096] The exchange between the search phase and the following agencies is adjusted according to the following formula:

[0097]

[0098] Where p is a random number in [0,1].

[0099] Step 4-5: Utilization phase, the first equation is based on pulling the spiders towards the area containing the most fit spiders, considering it the best place to build a nest in order to place the paralyzed spider in its abdomen and lay eggs. The first equation is described as follows:

[0100]

[0101] Among them, SW * represents the best solution so far. The second equation will build a nest at a randomly selected spider bee location from the population, using an extra step to avoid building two nests at the same location. The equations are designed as follows:

[0102]

[0103] Among them, r3 is a random number generated in the interval [0,1]; γ is a number generated according to the levy flight; a, b, c are the indicators of three solutions randomly selected from the population; is a binary vector that determines when to apply a step size to avoid building two nests at the same location; The allocation formula is as follows:

[0104]

[0105] in, and are two vectors representing random values ​​in the interval [0,1]. Equations (14) and (15) are randomly swapped as follows:

[0106]

[0107] The spider wasp will pull suitable spiders into a pre-prepared nest.

[0108]

[0109] Step 4-6: Matching phase: Each spider bee represents a possible solution of the current generation, and the spider bee egg represents a new potential solution of the generation. The new solution / spider bee egg is generated by the following formula:

[0110]

[0111] Among them, Crossover is a unified crossover operator SW applied between solutions i t and With Cross Rate (CR); SW i t and are two vectors representing male and spider bees. The male spider bees generated in our proposed algorithm are different from the female bees according to the following formula:

[0112]

[0113] Where β and β1 are two numbers randomly generated according to normal distribution, e is an exponential constant, and Generated by the following formula:

[0114]

[0115] In the formula, a, b, and c are the indices of three solutions randomly selected from the population, such that a≠i≠b≠c. Crossover is the process of recombining the genetic material of the parent spider bees to produce offspring (eggs) with the characteristics of both parents.

[0116] Steps 4-7: During the iteration process, some spider bees in the population will be terminated to provide more function evaluations for other spider bees, which will also reduce the diversity of the population and speed up the convergence to the optimal solution. At each evaluation of the entire function, the length of the new population will be updated using the following formula:

[0117] N=N min +(NN min )×k (23)

[0118] Step 4-8: Output the position and accuracy of the spider bee. The position of the spider bee is the parameter value of the random forest.

[0119] Step 5: Establish the best performing random forest model based on the optimized parameters;

[0120] Step 6: For the glass component to be predicted, the optimal random forest model is used to predict the glass elastic modulus of the glass component.

[0121] Embodiment 2:

[0122] Step 1: Collect elastic modulus data of oxide glasses with different components and build a glass elastic modulus database. Specifically, collect 1,600 sets of elastic modulus data of SiO2-Na2O-CaO-Al2O3 glasses and build a glass elastic modulus database.

[0123] Step 2: Descriptors with element molar content and preparation process parameters as input parameters;

[0124] Step 2-1: Take the molar content of elements of 1600 groups of SiO2-Na2O-CaO-Al2O3 glasses as a set of descriptors; Step 2-2: Construct descriptors based on glass preparation process parameters (heating rate, melting temperature, and holding time). Step 3: Take the descriptors constructed in step 2 as the input of the model, and the elastic modulus database constructed in step 1 as the output of the model, construct 1280 groups of data to construct a training set, and 320 groups of data to construct a test set, and establish a random forest model;

[0125] Step 4: Introduce the spider bee optimization algorithm to optimize the parameters of the selected random forest model;

[0126] Step 4-1: The hyperparameters of the random forest algorithm (number of decision trees, maximum depth of decision trees, minimum number of separation samples) are the parameters that need to be optimized. The position coordinates of the spider bee are the parameters that need to be optimized. The population size of the spider bee optimization algorithm is defined as 100, and the maximum number of iterations is 2000;

[0127] Step 4-2: In the search phase, female wasps randomly explore the search space with a constant step size, as described above, to find spiders that are suitable for their offspring. This behavior is modeled as shown in formula (3), where the current position of each female wasp is updated with a constant movement at each generation t to simulate the exploration behavior of the female wasp.

[0128]

[0129] Among them, a and b are two indicators randomly selected from the population to determine the exploration direction, followed by the female bee; μ1 is used to determine the constant movement through the current direction, and the formula is as follows:

[0130] μ1=|rn|×r1 (4)

[0131] Where r1 is a randomly generated number between 0 and 1, and rn is also a random number, but generated using a normal distribution. Step 4-3: Spider wasps sometimes fail to find the spider that fell from the ball; therefore, they search the entire area around the exact spot where the spider fell. Based on this behavior, another equation under different exploration methods is constructed in order to enable the proposed algorithm to explore the area around the discarded spider with a small step size different from that of equation (3). This equation is also based on updating the current spider wasp in each generation with a constant movement, based on a position of a female wasp randomly selected from the population representing the location of the dropped spider. The equation is described as follows:

[0132]

[0133] μ2=B×cos(2πl) (6)

[0134] Among them, c is an indicator randomly selected from the population, and l is a randomly generated number between 1 and -2.

[0135] Formulas (3) and (5) generate the next position of the spider bee randomly as follows:

[0136]

[0137] Among them, r3 and r4 are two random numbers in [0,1].

[0138] Step 4-4: Following and escaping phase: The process of the spider bee chasing the spider is simulated, in which the distance between the prey and the spider bee is initially small and may increase or decrease depending on the speed of the spider bee and the prey. Under this trend, we proposed a mathematical model to simulate two cases: (1) the spider bee is faster than the spider (prey) C>0.5 and (2) the prey is faster than the spider bee C<0.5, where C is the distance control factor that determines the speed of the spider bee, starting from speed 2 and linearly decreasing to 0. In this case, the spider bee moves faster than the prey, simulating the spider escaping from the spider bee, and the distance between the spider bee and the spider increases. Therefore, when the spider bee moves at a speed lower than 0.5, the rate of change of its position is so small that it cannot reach the prey. The formula is as follows:

[0139]

[0140] In the formula, a is an indicator randomly selected from the population; t and t max Represent the current and maximum values ​​respectively; is a vector representing randomly generated values ​​in the interval [0,1]; r6 is a random number in the interval [0,1].

[0141] As the spider escapes from the spider bee, the distance between the spider bee and the spider gradually increases. This stage is the initial development stage. As the distance increases, this behavior is simulated using the following formula:

[0142]

[0143] in is a vector generated between k and -k according to the normal distribution. Based on this, k is generated using formula (11) to gradually increase the distance between the spider bee and the spider.

[0144]

[0145] The trade-off between these two tendencies is achieved stochastically as shown in the following equation:

[0146]

[0147] The exchange between the search phase and the following agencies is adjusted according to the following formula:

[0148]

[0149] Where p is a random number in [0,1].

[0150] Step 4-5: Utilization phase, the first equation is based on pulling the spiders towards the area containing the most fit spiders, considering it the best place to build a nest in order to place the paralyzed spider in its abdomen and lay eggs. The first equation is described as follows:

[0151]

[0152] Among them, SW * represents the best solution so far. The second equation will build a nest at a randomly selected spider bee location from the population, using an extra step to avoid building two nests at the same location. The equations are designed as follows:

[0153]

[0154] Among them, r3 is a random number generated in the interval [0,1]; γ is a number generated according to the levy flight; a, b, c are the indicators of three solutions randomly selected from the population; is a binary vector that determines when to apply a step size to avoid building two nests at the same location; The allocation formula is as follows:

[0155]

[0156] in, and are two vectors representing random values ​​in the interval [0,1]. Equations (14) and (15) are randomly swapped as follows:

[0157]

[0158] The spider wasp will pull suitable spiders into a pre-prepared nest.

[0159]

[0160] Step 4-6: Matching phase: Each spider bee represents a possible solution of the current generation, and the spider bee egg represents a new potential solution of the generation. The new solution / spider bee egg is generated by the following formula:

[0161]

[0162] Among them, Crossover is a unified crossover operator SW applied between solutions i t and With Cross Rate (CR); SW i t and are two vectors representing male and spider bees. The male spider bees generated in our proposed algorithm are different from the female bees according to the following formula:

[0163]

[0164] Where β and β1 are two numbers randomly generated according to normal distribution, e is an exponential constant, and Generated by the following formula:

[0165]

[0166] In the formula, a, b, and c are the indices of three solutions randomly selected from the population, such that a≠i≠b≠c. Crossover is the process of recombining the genetic material of the parent spider bees to produce offspring (eggs) with the characteristics of both parents.

[0167] Steps 4-7: During the iteration process, some spider bees in the population will be terminated to provide more function evaluations for other spider bees, which will also reduce the diversity of the population and speed up the convergence to the optimal solution. At each evaluation of the entire function, the length of the new population will be updated using the following formula:

[0168] N=N min +(NN min )×k (23)

[0169] Step 4-8: Output the position and accuracy of the spider bee. The position of the spider bee is the parameter value of the random forest.

[0170] Step 5: Establish the best performing random forest model based on the optimized parameters;

[0171] Step 6: For the glass component to be predicted, the optimal random forest model is used to predict the glass elastic modulus of the glass component.

[0172] The model performance of Example 1-2 is as follows:

[0173] Table 1: Performance table of the models of Examples 1 and 2

[0174] R2 MSE MAE Before optimization 0.91 20.19 2.61 Example 1 0.97 3.48 1.84 Example 2 0.96 4.56 1.97

[0175] In the table, R2 is between [0,1]. The larger the value, the more accurate the model prediction;

[0176] The smaller the values ​​of MSE and MAE, the smaller the prediction error of the model.

[0177] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A glass elastic modulus prediction method based on spider bee optimization algorithm and machine learning algorithm, characterized in that: The following steps are involved: Step 1: Collect elastic modulus data of oxide glasses of different components and construct a glass elastic modulus database, wherein the glass elastic modulus database includes one-to-one mapping of glass components and their corresponding elastic moduli; Step 2: Descriptors with element molar content and preparation process parameters as input parameters, that is, the molar content of each component of the glass is used as a group of descriptors, and the descriptors are constructed with glass preparation process parameters, including heating rate, melting temperature, and holding time; Step 3: Using the descriptor constructed in step 2 as the input of the random forest model and the elastic modulus database constructed in step 1 as the output of the random forest model, construct a training set and a test set, and establish a random forest model; Step 4: Introduce the spider bee optimization algorithm to optimize the parameters of the selected random forest model; Step 5: Establish the best performing random forest model based on the optimized parameters; Step 6: For the glass component to be predicted, use the optimal random forest model to predict the glass elastic modulus of the glass component.

2. The glass elastic modulus prediction method based on spider bee optimization algorithm and machine learning algorithm according to claim 1, characterized in that: The step 4 includes the following sub-steps: Step 4-1: Select the parameters to be optimized for the random forest algorithm. The position coordinates of the zebra are the parameters to be optimized. Define the population size N of the spider bee optimization algorithm and search for the upper bound. Search lower bound Among them, SW Pop is the initial population of spider bees, and the following formula can generate any solution in the search space: Where t is the generated index; i is the midbody index, i = 1, 2, 3, ..., N; r is a vector consisting of D-dimensional random initialization numbers between 0 and 1; Step 4-2: In the search phase, the spider bee randomly explores the search space with a constant step size, as mentioned above, to find spiders that are suitable for its offspring. This behavior is modeled as shown in formula (3). The current position of each spider bee is updated with a constant movement at each generation t to simulate the exploration behavior of the spider bee. Among them, a and b are two indicators randomly selected from the population to determine the exploration direction, followed by the spider bee, which uses μ1 to determine the constant movement through the current direction, and the formula is as follows: μ1=|rn|×r1 (4) Where r1 is a randomly generated number between 0 and 1, and rn is also a random number, but generated using a normal distribution; Step 4-3: Spider bees sometimes cannot find the spider that fell from the ball, therefore, they search the entire area around the exact spot where the spider fell. According to this behavior, another equation under different exploration methods is constructed. In order to enable the proposed algorithm to explore the area around the discarded spider with a small step size different from that of formula (3), this equation is also based on updating the current spider bee in each generation with a constant movement, based on a position of a spider bee randomly selected from the population representing the location of the dropped spider. The equation is described as follows: μ2=B×cos(2πl) (6) Where c is a randomly selected index from the population, l is a randomly generated number between 1 and -2, and formulas (3) and (5) generate the next position of the female spider bee randomly implemented as follows: Among them, r3 and r4 are two random numbers in [0,1]; Step 4-4: Following and escaping phase: The process of the spider bee chasing the spider is simulated, in which the distance between the spider as the prey and the spider bee is initially small and may increase or decrease according to the speed of the spider bee and the prey. Under this trend, we proposed a mathematical model to simulate two situations: (1) the spider bee is faster than the spider (prey) C>0.5, (2) the prey is faster than the spider bee C<0.5, where C is the distance control factor that determines the speed of the spider bee, starting from speed 2 and linearly decreasing to 0. In this case, the spider bee moves faster than the prey, simulating the spider as the prey escaping from the spider bee. The distance between the spider bee and the spider increases. Therefore, when the spider bee moves at a speed lower than 0.5, the rate of change of its position is so small that it cannot reach the prey. The formula is as follows: In the formula, a is an indicator randomly selected from the population; t and t max Represent the current and maximum values ​​respectively; r5 is a vector, which represents a randomly generated value in the interval [0,1]; r6 is a random number in the interval [0,1]; When the prey spider escapes from the spider bee, the distance between the spider bee and the spider gradually increases. This stage is the initial development stage. As the distance increases, this behavior is simulated using the following formula: in is a vector generated between k and -k according to the normal distribution. Based on this, k is generated using formula (11) to gradually increase the distance between the spider bee and the spider. The trade-off between these two tendencies is achieved stochastically as shown in the following equation: The exchange between the search phase and the following agencies is adjusted according to the following formula: Where p is a random number in [0,1]; Step 4-5: Utilization phase, the first equation is based on pulling the spiders towards the area containing the most fit spiders and considering it as the best place to build a nest in order to place the paralyzed spider in its abdomen and lay eggs. The first equation is described as follows: Among them, SW * Representing the best solution so far, the second equation will build a nest at a randomly selected spider bee location from the population, using an extra step to avoid building two nests at the same location. The equation is designed as follows: Among them, r3 is a random number generated in the interval [0,1]; γ is a number generated according to the levy flight; a, b, c are the indicators of three solutions randomly selected from the population; is a binary vector that determines when to apply a step size to avoid building two nests at the same location; The allocation formula is as follows: in, and are two vectors representing random values ​​in the interval [0,1]. Equations (14) and (15) are randomly swapped as follows: The spider wasp will pull the appropriate spider into a pre-prepared nest. Step 4-6: Matching phase: Each spider bee represents a possible solution of the current generation, and the spider bee egg represents the newly generated potential solution of this generation. The new solution / spider bee egg is generated by the following formula: Among them, Crossover is a unified crossover operator SW applied between solutions i t and With Cross Rate (CR); SW i t and are two vectors representing male and spider bees. The male spider bees generated in our proposed algorithm are different from the female spider bees according to the following formula: Where β and β1 are two numbers randomly generated according to normal distribution, e is an exponential constant, and Generated by the following formula: In the formula, a, b, and c are the indices of three solutions randomly selected from the population, such that a≠i≠b≠c. Crossover is the process of recombining the genetic material of the parental spider bees to produce offspring eggs with the characteristics of both parents. Steps 4-7: During the iteration process, some spider bees in the population will be terminated to provide more function evaluations for other spider bees, which will also reduce the diversity of the population and speed up the convergence to the optimal solution. At each evaluation of the entire function, the length of the new population will be updated using the following formula: N=N min +(N-N min )×k (23) Step 4-8: Output the position and accuracy of the spider bee. The position of the spider bee is the parameter value of the random forest.

Citation Information

Patent Citations

  • Durable concrete mix proportion optimization method based on RF-NSGA-II

    CN111986737A

  • Elastic modulus prediction method based on molecular dynamics and elastic network regression model

    CN113255083A

  • Glass hardness prediction method based on squirrel optimization algorithm and machine learning algorithm

    CN114373523A

  • Oxide glass performance prediction method based on interpretable high-dimensional space prediction model

    CN117497087A

  • Excitation vibration elasticity modulus measuring device

    CN214953240U

Cited By

  • Multi-system glass material modulus analysis method

    CN119274723A

  • A modulus analysis method for multi-system glass materials

    CN119274723B

  • High-specific-modulus glass fiber composition, glass fiber and composite material

    CN120364946A

  • Glass fiber molecular structure-modulus performance relation prediction method based on graph neural network

    CN121054112A

  • A method for predicting the molecular structure-modulus property relationship of glass fibers based on graph neural networks

    CN121054112B