An automobile structure optimization method, device and storage medium

Abstract

This invention discloses a method, apparatus, and storage medium for automotive structure optimization. The method includes: acquiring historical data for automotive structure optimization; wherein the historical data includes structural parameters and corresponding performance simulation test results; mathematically modeling the automotive structure optimization problem based on the historical data to determine the optimization objective of the model; wherein the optimization objective is to minimize vehicle weight and maximize safety; and using a hierarchical particle swarm optimization algorithm assisted by a classification model to optimize the mathematical model and obtain structural design parameters that minimize vehicle weight and maximize safety. This invention utilizes classification model prediction to replace most of the automotive structure performance simulation during the evolution process, and uses the classification results to drive hierarchical particle swarm evolution, solving the problems of poor automotive structure optimization results and low search efficiency caused by large search space and long evaluation time in existing technologies. This invention can be widely applied in two major fields: evolutionary computation and industrial automotive structure design.

Description

Technical Field

[0001] This invention relates to two major fields: evolutionary computation and industrial vehicle structural design, and in particular to a method, apparatus and storage medium for optimizing vehicle structure. Background Technology

[0002] In recent years, with the increasing emphasis on environmental protection and the diversification of customer needs, the automotive industry has faced numerous challenges and development opportunities. For example, in terms of automotive structure optimization, how to reduce vehicle weight while improving collision safety and fuel efficiency are urgent issues that need to be addressed. However, automotive structure optimization is a complex problem, characterized by a large search space, high optimization difficulty, and low efficiency, making it difficult to solve effectively using traditional optimization algorithms.

[0003] Evolutionary computation is a common and effective method for solving complex optimization problems, and it has been widely applied in various industrial optimization problems. For automotive structural optimization problems, it can better find the optimal solution from a global perspective. However, evolutionary computation involves performing tens of thousands, hundreds of thousands, or even millions of fitness value evaluations on candidate solutions to achieve a "survival of the fittest" optimization process. In automotive structural optimization design, structural performance requires simulation analysis using finite element analysis software, safety performance requires crash simulation analysis using collision simulation software, and dynamic performance requires computational fluid dynamics simulation analysis using computational fluid dynamics software. A single simulation analysis can take tens of minutes or even several hours, making the time consumption for evaluating the performance of a candidate solution enormous, and making it difficult to complete tens of thousands of fitness value evaluations within an acceptable timeframe. Therefore, how to utilize evolutionary computation methods to complete the global optimization design of automotive structures within an acceptable timeframe remains a problem to be solved. Summary of the Invention

[0004] In order to at least partially solve one of the technical problems existing in the prior art, the present invention aims to provide a method, apparatus and storage medium for optimizing automobile structure.

[0005] The technical solution adopted in this invention is:

[0006] A method for optimizing automobile structure includes the following steps:

[0007] Acquire historical data on automotive structure optimization; wherein, the historical data includes structural parameters and performance simulation test results of the corresponding structures;

[0008] Mathematical modeling of the vehicle structure optimization problem is performed based on historical data to determine the optimization objective of the model; among which, the minimum vehicle weight and the highest safety are taken as the optimization objectives.

[0009] A hierarchical particle swarm optimization algorithm assisted by a classification model is used to optimize the mathematical model, thereby obtaining structural design parameters that minimize the vehicle weight and maximize safety.

[0010] Furthermore, the classification model performs performance prediction on the generated structural parameters during the algorithm optimization process, replacing the performance simulation of the vehicle structural parameters during the evolution process, thereby saving algorithm optimization time;

[0011] The hierarchical particle swarm algorithm is used as an evolution operator to drive the evolution of the hierarchical particle swarm based on the performance results predicted by the classification model, and to perform global search optimization of the automotive structure design.

[0012] Furthermore, the classification model is trained using historical data on vehicle structure optimization:

[0013] The classification model is trained using historical data from vehicle structure optimization. Wherein, the structural parameters are inputs The output is the simulation test results of the corresponding vehicle weight and safety.

[0014] New structural parameters are generated during the optimization process. Then, a classification model is used to predict the performance of this structural parameter.

[0015] Select structural parameters with better predictive performance Perform automotive structural performance simulation and utilize and Update the training classification model

[0016] Furthermore, the hierarchical particle swarm optimization algorithm assisted by the classification model is used to optimize the mathematical model, including:

[0017] A1. Encode each individual based on all structural parameters and initialize the parameters of the database; the initialized parameters include decision variables and corresponding fitness values, as well as initial velocity; where each structural parameter corresponds to a dimension, and structural parameter refers to the value corresponding to the dimension;

[0018] A2. Select the NP individuals with the best fitness from the database to form a population, including decision variables. fitness initial velocity Where NP is the population size;

[0019] A3. Sort the populations from best to worst fitness and divide them into four layers; train a four-classifier using the four-layer populations.

[0020] A4. Update the learning process on individuals in the population to obtain intermediate learning individuals;

[0021] A5. Use a classifier to predict the category of intermediate learners;

[0022] A6. Perform local information development, and repeat steps A3 and A4 until all intermediate learning individuals are predicted to be of the first class.

[0023] A7. Among the intermediate learning individuals who are all predicted to be of the first category, potential individuals are selected using geometric relationships for automobile structural performance simulation analysis.

[0024] A8. To increase population diversity, in addition to potential individuals, an intermediate learning individual is randomly selected for automobile structural performance simulation analysis.

[0025] A9. Add the individuals from the real performance simulation analysis to the database;

[0026] A10. Repeat steps A2 to A7 until the algorithm reaches the termination condition, that is, the number of real performance simulation analysis attempts has been exhausted.

[0027] A11. Output the individual variable values ​​with the best fitness values ​​in the database as the optimization results of the vehicle structure parameters in the global optimization.

[0028] Further, step A4 specifically includes:

[0029] Keep the first layer of individuals in the population unchanged;

[0030] For each individual in the second layer of the population, two individuals are randomly selected from the first layer. And meet the fitness value Superior in, Two randomly selected individuals are labeled with indices k1 and k2, respectively.

[0031] For each individual in the third layer of the population, randomly select an individual from the first layer. Randomly select an individual from the second layer As can be seen from the stratification, these two individuals must satisfy the fitness value. Superior

[0032] For each individual in the fourth layer of the population, firstly, two learning layers are randomly selected, and then learners are selected from each of the two learning layers. And meet the fitness value Superior

[0033] Furthermore, the intermediate learning individuals are updated using the following formula:

[0034]

[0035]

[0036] Where t represents the evolutionary algebra, and i represents the index of the structural parameter to be updated. Let φ be a random number vector with values ​​in the space [0,1], and let φ be a weight control parameter. and This indicates the ratio of the prediction performance of the two classification models. Good structural parameters, i.e. This represents the structural parameters of the i-th individual in generation t. This represents the structural parameters of the i-th individual in generation t+1. This represents the velocity vector of the i-th individual in generation t. Let represent the velocity vector of the i-th individual in the (t+1)-th generation.

[0037] Furthermore, step A6 specifically includes:

[0038] The first layer of individuals in the marked population is Intermediate learners were predicted to be in the first category.

[0039] Calculate the Euclidean distance between intermediate learning individual i and individual j in the first layer of the population;

[0040] Label the differences between intermediate learner individual i and the first layer of the population;

[0041] The intermediate learning individual with the smallest difference was selected as the potential individual for automotive structural performance simulation analysis.

[0042] Furthermore, the formula for calculating the Euclidean distance is:

[0043]

[0044] The following method is used to label the differences between intermediate learning individual i and the first layer of the population:

[0045]

[0046] in, This represents the i-th intermediate learner. This represents an individual in the first layer of the j-th population.

[0047] Another technical solution adopted in this invention is:

[0048] A vehicle structure optimization device, comprising:

[0049] At least one processor;

[0050] At least one memory for storing at least one program;

[0051] When the at least one program is executed by the at least one processor, the at least one processor implements the method described above.

[0052] Another technical solution adopted in this invention is:

[0053] A computer-readable storage medium storing a processor-executable program, which, when executed by a processor, performs the method described above.

[0054] The beneficial effects of this invention are: This invention uses classification model prediction to replace most of the simulation of automobile structure performance during the evolution process, and uses the classification results to drive hierarchical particle swarm evolution, which solves the problems of poor automobile structure optimization results and low search efficiency caused by large search space and long evaluation time in the prior art. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following description is provided with accompanying drawings of the relevant technical solutions in the embodiments of the present invention or the prior art. It should be understood that the accompanying drawings described below are only for the purpose of clearly illustrating some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart illustrating the steps of a vehicle structure optimization method according to an embodiment of the present invention;

[0057] Figure 2 This is an algorithm flowchart of a vehicle structure optimization method based on a classification model and a hierarchical particle swarm algorithm in an embodiment of the present invention.

[0058] Figure 3 This is a schematic diagram of a hierarchical particle swarm optimization algorithm for a vehicle structure optimization method based on a classification model and a hierarchical particle swarm algorithm, as described in an embodiment of the present invention.

[0059] Figure 4 This is a schematic diagram of local information development in an embodiment of the present invention for a vehicle structure optimization method based on a classification model and a hierarchical particle swarm optimization algorithm;

[0060] Figure 5 This is a schematic diagram illustrating the geometric relationship selection of potential individuals in an automotive structure optimization method based on a classification model and a hierarchical particle swarm optimization algorithm, as described in an embodiment of the present invention. Detailed Implementation

[0061] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0062] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0063] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0064] In the description of this invention, unless otherwise explicitly defined, terms such as "set up," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0065] like Figure 1 As shown, this embodiment provides a method for optimizing vehicle structure. This method employs a hierarchical particle swarm optimization algorithm as the evolution operator, utilizes historical data from vehicle structure optimization to train a classification model as an aid, and replaces most of the vehicle structure performance simulation during the evolution process with classification prediction, saving performance simulation analysis time. The classification results drive the hierarchical particle swarm evolution to perform a global search optimization of the vehicle structure design. The method includes the following steps:

[0066] S1. Obtain historical data on vehicle structure optimization; wherein, the historical data includes structural parameters and performance simulation test results of the corresponding structure.

[0067] S2. Based on historical data, mathematical model the automobile structure optimization problem and determine the optimization objective of the model; among them, the minimum automobile weight and the highest safety are taken as the optimization objectives.

[0068] The structural optimization information is used to mathematically model the automotive structural optimization problem, resulting in decision variables with structural parameters as variables and objective functions with automotive weight and safety as objectives.

[0069] S3. The mathematical model is optimized using a hierarchical particle swarm optimization algorithm assisted by a classification model to obtain structural design parameters that minimize the car's weight and maximize its safety.

[0070] A hierarchical particle swarm optimization algorithm assisted by a classification model is used to calculate the structural design parameters that minimize the vehicle weight and maximize safety, with the optimization objectives being the minimum vehicle weight and the highest safety, provided that all structural parameters satisfy the feasible region.

[0071] Specifically, the classification model utilizes historical data from automotive structure optimization for training as an aid, replacing most of the automotive structure performance simulation during the evolution process with classification prediction, thus saving performance simulation and analysis time. The hierarchical particle swarm optimization algorithm serves as the evolution operator, driving the hierarchical particle swarm evolution with the performance prediction results from the classification model, and performing global search optimization of the automotive structure design.

[0072] As an optional implementation, a classification model is trained using historical data from vehicle structure optimization. During algorithm optimization, performance prediction is performed on the generated structural parameters, replacing performance simulation of most vehicle structural parameters during evolution, thus saving algorithm optimization time. This includes steps B1-B3:

[0073] B1. Training a classification model using historical data on automobile structure optimization. Wherein, the structural parameters are inputs The output is the simulation test results of the corresponding vehicle weight and safety.

[0074] B2. New structural parameters are generated during the optimization process. Then, a classification model is used to predict the performance of this structural parameter.

[0075] B3. Select structural parameters with better predictive performance. Perform automotive structural performance simulation and utilize and Update the training classification model

[0076] As an optional implementation, the hierarchical particle swarm evolution is driven by the prediction results of the classification model, and the structural design parameters that minimize the vehicle weight and maximize safety are calculated with the goal of minimizing the vehicle weight and maximizing safety, provided that all structural parameters satisfy the feasible region.

[0077] The hierarchical particle swarm optimization strategy driven by the classification model is as follows:

[0078]

[0079]

[0080] Where t represents the evolutionary algebra, and i represents the index of the structural parameter to be updated. Let be a random number vector with values ​​in the space [0,1], and φ be a weight control parameter set to 0.4. and This indicates the ratio of the prediction performance of the two classification models. Good structural parameters, i.e.

[0081] As an optional implementation, step S3 specifically includes the following steps S31-S311:

[0082] S31: Encode each individual with all structural parameters as individuals, each structural parameter corresponding to a dimension, and the structural parameter refers to the value corresponding to the dimension. Initialize the database, including decision variables and corresponding fitness values, i.e. target values, and initial velocities.

[0083] S32: Select the NP individuals with the best fitness from the database to form a population, including decision variables. fitness speed Where NP is the population size.

[0084] S33: Sort the populations from best to worst fitness and divide them into four layers; correspondingly, train a four-classifier using the four-layer populations.

[0085] S34: Update the learning of individuals in the population to obtain intermediate learning individuals.

[0086] Specifically, step S34 includes:

[0087] For each individual in the second layer of the population, two individuals are randomly selected from the first layer. And meet the fitness value Superior Where k1 and k2 mark the indexes of the selected individuals;

[0088] For each individual in the third layer of the population, randomly select an individual from the first layer. Randomly select an individual from the second layer As can be seen from the stratification, these two individuals must satisfy the fitness value. Superior

[0089] For each individual in the fourth layer of the population, firstly, two learning layers are randomly selected, and then learners are selected from each of the two learning layers. And meet the fitness value Superior

[0090] S35: Use a classifier to predict the category of intermediate learners.

[0091] S36: Local information development, repeat steps S33 and S34 until all intermediate learning individuals are predicted to be of the first class.

[0092] Step S36 specifically includes:

[0093] The first layer of individuals in the marked population is Intermediate learners were predicted to be in the first category.

[0094] The Euclidean distance between intermediate learning individual i and first-level individual j in the population is calculated as follows:

[0095]

[0096] The differences between the labeled intermediate learner individual i and the first layer of the population are as follows:

[0097]

[0098] Select the middle learner with the smallest difference. As a potential individual, conduct automotive structural performance simulation analysis.

[0099] S37: Among the intermediate learning individuals who are all predicted to be of the first class, potential individuals are selected using geometric relationships for automotive structural performance simulation analysis.

[0100] S38: To increase population diversity, in addition to potential individuals, an intermediate learning individual is randomly selected for vehicle structural performance simulation analysis.

[0101] S39: Add individuals from the real performance simulation analysis to the database.

[0102] S310: Repeat steps S32 to S37 until the algorithm reaches the termination condition, that is, the number of times the real performance simulation analysis has been consumed.

[0103] S311: Output the individual variable value with the best fitness value in the database as the result of global optimization of the vehicle structure parameters.

[0104] The above method will be explained in detail below with reference to the accompanying drawings and specific embodiments.

[0105] This invention provides a method for optimizing vehicle structure based on a classification model and a hierarchical particle swarm optimization algorithm, comprising the following steps:

[0106] S101: Obtain historical data on automotive structure optimization, including structural parameters and performance simulation test results of the corresponding structures.

[0107] S102: Using the structural optimization information, mathematical modeling is performed on the automotive structural optimization problem to obtain decision variables with structural parameters as variables and objective functions with automotive weight and safety as objectives.

[0108] S103: Using a hierarchical particle swarm optimization algorithm assisted by a classification model, with the goal of minimizing the vehicle weight and maximizing safety, the structural design parameters that minimize the vehicle weight and maximize safety are calculated under the condition that all structural parameters satisfy the feasible region.

[0109] This invention provides a vehicle structure optimization method based on a classification model and a hierarchical particle swarm optimization algorithm. This method can be applied to different types of vehicles, such as passenger cars and commercial vehicles, with the optimization goal of reducing vehicle weight and improving collision safety performance. In this invention, a hierarchical particle swarm optimization algorithm based on a classification model is implemented. Through iterative search and population evolution, a set of relatively optimized vehicle structure parameters is found.

[0110] See Figure 2 , Figure 2This invention provides an algorithm flowchart for a vehicle structure optimization method based on a classification model and a hierarchical particle swarm optimization algorithm. First, the database is initialized; then, NP individuals with the best fitness values ​​are selected from the database to form a population; the population is divided into four layers from best to worst based on fitness values; a four-classifier is trained using the hierarchical population based on the layered results; the population is updated using a hierarchical particle swarm optimization algorithm to generate intermediate learning individuals; the trained four-classifier is used to predict the class of the intermediate learning individuals; if any intermediate learning individual is not predicted to be class 1, the hierarchical particle swarm optimization algorithm is repeated until all intermediate learning individuals are predicted to be class 1; if all intermediate learning individuals are predicted to be class 1, potential individuals are selected using geometric relationships; vehicle structure parameters are simulated and analyzed on the potential individuals; the structural parameters and simulation results are added to the database as known solutions; if the algorithm has not reached the termination condition (i.e., the number of simulation analyses of the vehicle structure parameters has not been fully consumed), the population selection and evolution process is repeated; if the algorithm reaches the termination condition, the algorithm stops and the individual with the best fitness value in the database is output as the globally optimized vehicle structure parameters. By combining a classification model with a hierarchical particle swarm optimization algorithm, this method can significantly improve the global search capability and search efficiency for automotive structure optimization problems.

[0111] Specifically, the hierarchical particle swarm optimization algorithm based on a classification model provided in this embodiment of the invention includes the following steps:

[0112] A1. Treat all structural parameters as individuals, with each structural parameter corresponding to a dimension. Structural parameters refer to the values ​​corresponding to the dimensions. Encode the individuals and initialize the database, including decision variables and corresponding fitness values, i.e., target values, and initial velocities.

[0113] 1) Database initialization:

[0114] This embodiment uses Latin hypercube sampling to initialize the database in order to obtain initial individuals that are relatively evenly distributed in the search space. Where i represents the individual index with a total of N, j represents the dimension index with a dimension of D, and each dimension corresponds to a car structure design parameter.

[0115] A2. Select the NP individuals with the best fitness from the database to form a population, including decision variables. fitness initial velocity

[0116] 2) Population selection:

[0117] In this embodiment of the invention, NP individuals with the best fitness values ​​are selected from the database DB to form a population.

[0118] A3. Sort the populations from best to worst fitness and divide them into four layers; correspondingly, train a four-classifier using the four-layer populations.

[0119] 3) Population stratification:

[0120] In this embodiment of the invention, NP individuals are divided into four layers according to their fitness values ​​from best to worst, i.e., the layers with the best fitness values ​​are... Individuals are in the first layer, ranked by fitness value. The individuals in the second layer are ranked by fitness value. The individuals are in the third layer, ranked by fitness value. The individuals are in the fourth layer.

[0121] 4) Classifier training:

[0122] This invention utilizes the hierarchical population correspondence to train a four-class classifier, with the classifier input for training. Keeping the vehicle structure parameters unchanged, the input training y is changed to the individual level, i.e. {1,2,3,4}.

[0123] A4. Update the learning of individuals in the population to obtain intermediate learning individuals.

[0124] 5) Population evolution:

[0125] In this embodiment of the invention, the first layer of individuals in the population remains unchanged; for each individual in the second layer of the population, two individuals are randomly selected from the first layer. And meet the fitness value Superior For each individual in the third layer of the population, randomly select an individual from the first layer. Randomly select an individual from the second layer As can be seen from the stratification, these two individuals must satisfy the fitness value. Superior For each individual in the fourth layer of the population, firstly, two learning layers are randomly selected, and then learners are selected from each of the two learning layers. And meet the fitness value Superior

[0126] The intermediate learning individuals are obtained using the following update formula:

[0127]

[0128]

[0129] Where t represents the evolutionary generation, and i represents the index of the individual to be updated from the second to the fourth level. φ is a random number vector with values ​​in the space [0,1], and φ is a weight control parameter set to 0.4.

[0130] A5. Use a classifier to predict the category of intermediate learners.

[0131] 6) Classifier prediction:

[0132] In this embodiment of the invention, a pre-trained classifier is used to predict the category of intermediate learning individuals. The prediction structure corresponds to the population hierarchy, that is, category 1 corresponds to the first layer of the population, category 2 corresponds to the second layer of the population, category 3 corresponds to the third layer of the population, and category 4 corresponds to the fourth layer of the population.

[0133] A6. Local information development: Repeat steps A3 and A4 until all intermediate learning individuals are predicted to be of the first class.

[0134] 7) Local information development:

[0135] In this embodiment of the invention, a classification model is used to drive a hierarchical particle swarm algorithm for local development. If there are intermediate learning individuals that are not predicted as class 1, the steps 5) population evolution and 6) classifier prediction are repeated until all intermediate learning individuals are predicted as class 1.

[0136] A7. Among the intermediate learning individuals who are all predicted to be of the first category, potential individuals are selected using geometric relationships for automobile structural performance simulation analysis.

[0137] 8) Selection of geometric relations:

[0138] This invention utilizes the positional geometric relationship between intermediate learning individuals and the first-layer individuals of the population to perform long-range individual selection, marking the first-layer individuals of the population as... Intermediate learners were predicted to be in the first category. The Euclidean distance between intermediate learning individual i and first-level individual j in the population is calculated as follows:

[0139]

[0140] The differences between the labeled intermediate learner individual i and the first layer of the population are as follows:

[0141]

[0142] A8. To increase population diversity, in addition to potential individuals, an intermediate learning individual is randomly selected for automobile structural performance simulation analysis.

[0143] 9) Automobile structure simulation analysis:

[0144] Select the middle learner with the smallest difference. As potential individuals, we conduct automotive structural performance simulation analysis; in addition, to improve population diversity, we randomly select an intermediate learning individual predicted to be of class 1 for automotive structural performance simulation analysis.

[0145] A9. Add the individuals from the real performance simulation analysis to the database.

[0146] 10) Database update:

[0147] Add the individual data and analysis values ​​from the vehicle structural performance simulation analysis to the database and update the database.

[0148] A10. Repeat steps A2 to A7 until the algorithm reaches the termination condition, that is, the number of times the real performance simulation analysis has been consumed.

[0149] 11) Algorithm termination determination

[0150] If the algorithm reaches the termination condition, that is, the number of times the vehicle structure performance simulation analysis is exhausted, the algorithm terminates and outputs the individual with the best fitness value in the database as the vehicle structure design parameters for global optimization; if the algorithm does not reach the termination condition, that is, the number of times the vehicle structure performance model analysis is not exhausted, the algorithm jumps to step 2) population selection to continue evolution.

[0151] A11. Output the individual variable values ​​with the best fitness values ​​in the database as the optimization results of the vehicle structure parameters in the global optimization.

[0152] See Figure 3 , Figure 3 This diagram illustrates a hierarchical particle swarm optimization (PSO) algorithm based on a classification model and a hierarchical PSO algorithm, as provided in an embodiment of the present invention. First, the population is divided into four layers (L1, L2, L3, L4) according to their fitness values ​​from best to worst. Individuals in layer L1 have the best fitness values, and individuals in layer L4 have the worst fitness values. The number of individuals in each layer is... Then, keeping the L1 individuals unchanged, L2 individuals learn from L1 individuals, L3 individuals learn from both L1 and L2 individuals, and L3 individuals learn from L1, L2, and L3 individuals. This approach preserves elite solutions and improves population diversity. The specific learning strategy is as follows:

[0153] Keep the individuals in the first layer of the population unchanged; for each individual in the second layer of the population, randomly select two individuals from the first layer. And meet the fitness value Superior For each individual in the third layer of the population, randomly select an individual from the first layer. Randomly select an individual from the second layer As can be seen from the stratification, these two individuals must satisfy the fitness value. Superior For each individual in the fourth layer of the population, firstly, two learning layers are randomly selected, and then learners are selected from each of the two learning layers. And meet the fitness value Superior

[0154] The intermediate learning individuals are obtained using the following update formula:

[0155]

[0156]

[0157] See Figure 4 , Figure 4 This diagram illustrates the local information development of a vehicle structure optimization method based on a classification model and a hierarchical particle swarm optimization algorithm, as provided in an embodiment of the present invention. After one hierarchical particle swarm update learning, the number of intermediate learning individuals predicted as class 1 increases, while the number of intermediate learning individuals predicted as class 4 decreases. This indicates that the intermediate learning individuals are converging towards the L1 layer of the population, but the L1 layer information has not been fully developed. To fully develop local information, the present invention uses the prediction results of the classification model to drive the hierarchical particle swarm update learning. After multiple update learnings, all intermediate learning individuals are predicted as class 1, indicating that the intermediate learning individuals have converged to the L1 layer of the population, thus completing the local information development.

[0158] See Figure 5 , Figure 5 This diagram illustrates the geometric relationship selection of potential individuals in a vehicle structure optimization method based on a classification model and hierarchical particle swarm optimization algorithm, as provided in this embodiment of the invention. Black circles represent fitness contour lines, blue dots represent the outer points of individuals in the L1 layer of the population, purple dots represent intermediate learning individuals predicted as class 1, red squares represent the centers of the outer points of individuals in the L1 layer of the population, and yellow asterisks represent individuals selected based on geometric relationships. When the vehicle structure optimization problem is a unimodal optimization problem, such as... Figure 5 As shown in (a), individuals selected based on geometric relationships are located near the peak of the contour lines, indicating that these individuals have the potential for better fitness, which can improve local exploitation capabilities and help the population converge; when the automotive structure optimization problem is a multimodal optimization problem, such as Figure 5 As shown in (b), individuals selected based on geometric relationships are located near the saddles of contour lines. These individuals are more conducive to helping the model fit the relationships between different peaks and improving classification accuracy.

[0159] In summary, this embodiment has at least the following advantages and beneficial effects compared to the prior art:

[0160] (1) The vehicle structure optimization method based on classification model and hierarchical particle swarm optimization algorithm provided in this embodiment of the invention uses the acquired vehicle structure optimization data as the mathematical model for establishing vehicle structure optimization, quantitatively calculates vehicle structure optimization design, takes structural parameters as decision variables and vehicle weight and safety as objective functions, and adopts hierarchical particle swarm optimization algorithm to improve the global search capability of variable space, thus solving the problem of poor globality in solving vehicle structure optimization problems in the prior art.

[0161] (2) The vehicle structure optimization method based on classification model and hierarchical particle swarm algorithm provided in this embodiment of the invention uses classification model as an aid, and uses historical data of vehicle structure optimization to train classification model as an aid. It uses classification prediction to replace most of the vehicle structure performance simulation in the evolution process, saving performance simulation analysis time and solving the problem of huge time consumption in solving the vehicle structure optimization problem in the existing technology.

[0162] (3) The vehicle structure optimization method based on classification model and hierarchical particle swarm algorithm in this embodiment combines the classification characteristics of classification model with the hierarchical characteristics of hierarchical particle swarm algorithm. The classification result drives the evolution of hierarchical particle swarm, which improves the global search and development capability for vehicle structure optimization design. It solves the problems of poor optimization results and low search efficiency in the existing technology when solving vehicle structure optimization problems.

[0163] This embodiment also provides an automotive structure optimization device, including:

[0164] At least one processor;

[0165] At least one memory for storing at least one program;

[0166] When the at least one program is executed by the at least one processor, the at least one processor implements Figure 1 The method is described above.

[0167] This embodiment of the vehicle structure optimization device can execute the vehicle structure optimization method provided in the method embodiment of the present invention, and can execute any combination of the implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.

[0168] This application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform... Figure 1 The method shown.

[0169] This embodiment also provides a storage medium storing instructions or programs that can execute the vehicle structure optimization method provided in the method embodiment of the present invention. When the instructions or programs are run, any combination of implementation steps of the method embodiment can be executed, and the method has the corresponding functions and beneficial effects.

[0170] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0171] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0172] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0173] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0174] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0175] 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.

[0176] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, 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.

[0177] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0178] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for optimizing automobile structure, characterized in that, Includes the following steps: Acquire historical data on automotive structure optimization; wherein, the historical data includes structural parameters and performance simulation test results of the corresponding structures; Mathematical modeling of the vehicle structure optimization problem is performed based on historical data to determine the optimization objective of the model; among which, the minimum vehicle weight and the highest safety are taken as the optimization objectives. A hierarchical particle swarm optimization algorithm assisted by a classification model is used to optimize the mathematical model, thereby obtaining structural design parameters that minimize the vehicle weight and maximize safety. The hierarchical particle swarm optimization algorithm assisted by a classification model is used to optimize the mathematical model, including: A1. Encode each individual based on all structural parameters and initialize the parameters of the database; the initialized parameters include decision variables and corresponding fitness values, as well as initial velocity. A2. Select the optimal fitness value from the database. NP Individuals make up a population, which includes decision variables. fitness value initial velocity ,in NP It refers to population size; A3. Sort the populations from best to worst fitness and divide them into four layers; train a four-classifier using the four-layer populations. A4. Update the learning process on individuals in the population to obtain intermediate learning individuals; A5. Use a classifier to predict the category of intermediate learners; A6. Perform local information development, and repeat steps A3-A4 until all intermediate learning individuals are predicted to be of the first class. A7. Among the intermediate learning individuals who are all predicted to be of the first category, potential individuals are selected using geometric relationships for automobile structural performance simulation analysis. A8. To increase population diversity, in addition to potential individuals, an intermediate learning individual is randomly selected for automobile structural performance simulation analysis. A9. Add the individuals from the real performance simulation analysis to the database; A10. Repeat steps A2 to A7 until the algorithm reaches the termination condition; A11. Output the individual variable values ​​with the best fitness values ​​in the database as the optimization results of the vehicle structure parameters in the global optimization.

2. The method for optimizing automobile structure according to claim 1, characterized in that, The classification model performs performance prediction on the generated structural parameters during the algorithm optimization process, replacing the performance simulation of the vehicle structural parameters during the evolution process, thereby saving algorithm optimization time. The hierarchical particle swarm optimization algorithm, as an evolution operator, drives the evolution of the hierarchical particle swarm based on the performance results predicted by the classification model, and performs global search optimization for the automotive structure design.

3. The method for optimizing automobile structure according to claim 1, characterized in that, The classification model is trained using historical data on vehicle structure optimization. The classification model was trained using historical data from vehicle structure optimization. The structural parameters are the input parameters. The output is the simulation test results of the corresponding vehicle weight and safety. ; New structural parameters are generated during the optimization process. Then, a classification model is used to predict the performance of this structural parameter. ; Selecting structural parameters with better predictive performance Conduct automotive structural performance simulation and utilize and Update the training classification model .

4. The method for optimizing automobile structure according to claim 1, characterized in that, Step A4 specifically includes: Keep the first layer of individuals in the population unchanged; For each individual in the second layer of the population, two individuals are randomly selected from the first layer. And satisfy the fitness value Superior ,in To label the selected individual index; For each individual in the third layer of the population, randomly select an individual from the first layer. Randomly select an individual from the second layer. As can be seen from the stratification, these two individuals must satisfy the fitness value. Superior ; For each individual in the fourth layer of the population, firstly, two learning layers are randomly selected, and then learners are selected from each of the two learning layers. And meet the fitness value Superior .

5. The method for optimizing automobile structure according to claim 4, characterized in that, The intermediate learning individuals are updated using the following formula: in, Represents evolutionary algebra, This indicates the index of the structure parameter to be updated. Let be a vector of random numbers taking values ​​in the space [0,1]. For weight control parameters, and This indicates the ratio of the prediction performance of the two classification models. Good structural parameters, i.e. , ; Indicates the first The individual in the first The structural parameters of the generation, Indicates the first The individual in the first The structural parameters of the generation, Indicates the first The individual in the first The velocity vector of the generation, Indicates the first The individual in the first The velocity vector of the generation.

6. The method for optimizing automobile structure according to claim 1, characterized in that, Step A6 specifically includes: The first layer of individuals in the marked population is Intermediate learners were predicted to be in the first category. ; Computing intermediate learning individuals With individuals in the first layer of the population Euclidean distance; Label intermediate learners Differences from the first layer of the population; The intermediate learning individual with the smallest difference was selected as the potential individual for automotive structural performance simulation analysis.

7. The method for optimizing automobile structure according to claim 6, characterized in that, The formula for calculating the Euclidean distance is: The following method is used to label intermediate learning individuals. Differences from the first layer of the population: in, Indicates the first An intermediate learning individual, Indicates the first The first layer of individuals in a population.

8. A vehicle structure optimization device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method of any one of claims 1-7.

9. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to perform the method as described in any one of claims 1-7.

Images