Highway guardrail simulation experiment multi-target optimization method

Abstract

The invention provides a road guardrail simulation experiment multi-target optimization method, which comprises the following steps: determining different level values of optimization variables based on a road guardrail simulation experiment, and collecting performance data corresponding to guardrail performance evaluation indexes under each key factor combination corresponding to the optimization variables through the simulation experiment; constructing and training a radial basis model to approximately represent a nonlinear relationship between the key factors and the guardrail performance evaluation indexes by using data obtained by a simulation experiment, and obtaining an agent model of a mathematical relationship between the guardrail key factors and an optimization target; and taking an optimization target value of each key factor combination predicted by the agent model within a constraint condition range as a target function, searching an optimal combination of key factors by adopting a multi-target particle swarm optimization algorithm, and searching a non-dominated solution meeting all optimization target requirements through a Pareto frontier solution set, and selecting an optimal factor level combination.

Description

Technical Field

[0001] The invention belongs to the technical field of design and optimization of traffic engineering safety facilities, and specifically relates to a multi-objective optimization method for highway guardrail simulation experiments. Background Art

[0002] With the continuous increase in traffic demand and the improvement of vehicle performance, the optimization design of the size of highway guardrails, as a key element to ensure road safety, has become increasingly important. Traditional design optimization methods mostly rely on tedious trial-and-error processes and the accumulation of expert experience, which is not only inefficient, but also difficult to fully capture and analyze the complex relationship between various factors that affect guardrail performance. Therefore, exploring an efficient and accurate quantitative optimization strategy to quickly and accurately determine the optimal size combination under multiple objectives has become a technical problem that needs to be overcome in the field of highway traffic safety facility design and optimization. Summary of the invention

[0003] In order to effectively, quickly and accurately find the optimal parameter combination for the design optimization of traffic engineering safety facilities, a multi-objective particle swarm optimization algorithm based on the radial basis model for the optimal combination of multiple factors in highway guardrail simulation is proposed, thereby realizing a multi-objective optimization method for highway guardrail simulation experiments.

[0004] In this scheme, according to the highway guardrail simulation experiment design, different levels of key factors are determined. Through simulation experiments, performance data such as the maximum composite acceleration of the center of mass and the maximum lateral dynamic deformation of the guardrail under various combinations of factors are collected; and a radial basis function model (RBF) is constructed to approximate the complex nonlinear relationship between these factors and the guardrail performance evaluation indicators; thus, the multi-objective particle swarm optimization algorithm (MOPSO) is used to find the optimal combination of experimental factors, and the non-dominated solution that meets all objective requirements is found by introducing the Pareto frontier solution set, and the optimal factor level combination is selected. Finally, the accurate quantification and scientific optimization of the optimal combination of highway guardrail simulation experiments under the influence of complex multiple factors are achieved, which improves the efficiency and effect of guardrail design.

[0005] During the design process of the scheme of the present invention, it was found that the method based on RBF neural network training, with its powerful nonlinear mapping ability and self-learning ability, has shown great potential in the modeling and optimization of complex systems. However, when using RBF neural network alone for size optimization, it may face challenges such as a large search space and easy to fall into local optimal solutions. To solve these problems, the present invention adopts a multi-objective particle swarm optimization algorithm with its parallel search, information sharing and fast convergence characteristics to provide a new idea for the optimization of highway guardrail size. However, when dealing with high-dimensional and nonlinear problems, the traditional multi-objective particle swarm optimization algorithm may also have problems such as slow convergence speed and uneven distribution of solution sets. In this context, an optimization strategy is proposed, which integrates a multi-objective particle swarm optimization algorithm based on RBF neural network training. First, the relationship between the size and performance of the highway guardrail is accurately modeled using RBF neural network, and secondly, the multi-objective particle swarm optimization algorithm is used to efficiently explore in a complex search space and quickly locate the optimal size combination that meets multiple performance indicators. This strategy can not only comprehensively evaluate the impact of various size parameters and their interactions on the performance of the guardrail under limited computing resources, but also significantly improve the optimization efficiency, providing a strong scientific basis and technical support for the design and optimization of highway traffic safety facilities.

[0006] The technical solution specifically adopted by the present invention to solve the technical problem is:

[0007] A highway guardrail simulation experiment multi-objective optimization method: based on highway guardrail simulation experiments, different level values ​​of optimization variables are determined, and through simulation experiments, performance data corresponding to guardrail performance evaluation indicators under various key factor combinations corresponding to the optimization variables are collected; using the data obtained from the simulation experiments, a radial basis model is constructed and trained to approximately represent the nonlinear relationship between key factors and guardrail performance evaluation indicators, and a proxy model of the mathematical relationship between guardrail key factors and optimization objectives is obtained; the optimization target value of each key factor combination within the constraint range predicted by the proxy model is used as an objective function, and a multi-objective particle swarm optimization algorithm is used to find the optimal combination of key factors, and a non-dominated solution that meets all optimization objectives is found through a Pareto front solution set to select the optimal factor level combination.

[0008] Furthermore, the optimization variables are determined by selecting several key factors that may have a significant impact on the simulation experiment results and do not affect each other according to the experimental purpose, and each selected key factor corresponds to a range of constraint conditions.

[0009] Furthermore, the guardrail performance evaluation indicators include data reflecting the guardrail performance, such as the maximum composite acceleration of the center of mass and the maximum lateral dynamic deformation of the guardrail.

[0010] The determination of evaluation indicators should clarify the purpose of the experiment, and consider the buffering, guiding and other functions of the guardrail design, as well as the economy and durability of the guardrail, such as the acceleration of the center of mass in the X and Y directions, the maximum lateral dynamic deformation of the guardrail, etc. Then, based on the economic and performance indicators of the highway guardrail, determine the evaluation indicators of the simulation experiment, that is, the optimization target, such as the maximum synthetic acceleration of the center of mass, the maximum lateral dynamic deformation of the guardrail, etc., and prepare a detailed experimental plan based on this, conduct simulation experiments according to the experimental plan, and record the experimental results.

[0011] Furthermore, the simulation experiment is a vehicle-guardrail coupling collision simulation experiment, and the simulation experiment environment needs to be set up in accordance with the highway traffic safety facility design specification JTG / D81-2017.

[0012] Furthermore, the scope of the constraint conditions is determined according to: Highway Traffic Safety Facilities Design Specifications JTG / D81-2017 and GB / T31439-2015 Corrugated Beam Steel Guardrail.

[0013] Furthermore, the key factors include: guardrail beam thickness, column thickness and barrier block thickness; the guardrail performance evaluation indicators include: vehicle center of mass acceleration, maximum dynamic lateral deformation of the guardrail and guardrail weight per unit length.

[0014] Furthermore, the proxy model uses a data set constructed using data obtained from simulation experiments, which is obtained after multiple iterations of training and convergence of the RBF neural network.

[0015] Furthermore, the multi-objective particle swarm optimization algorithm searches for the best balance point that satisfies multiple objectives in the solution space through information sharing and co-evolution among particles, and repeatedly runs the algorithm to obtain several Pareto frontier solution sets for selecting the optimal factor level combination.

[0016] And, an electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the multi-objective optimization method for a highway guardrail simulation experiment as described above are implemented.

[0017] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a multi-objective optimization method for a highway guardrail simulation experiment as described above.

[0018] In view of the defects and shortcomings of the prior art, the present invention and its preferred solution provide an intelligent optimization solution for the optimal combination of multiple factors in the simulation experiment of highway guardrail research and development and design, and through scientific means, accurately and efficiently solve the problem of optimizing the design parameters of highway guardrail dimensions.

[0019] A multi-objective particle swarm optimization algorithm (MOPSO) based on the radial basis function model (RBF) for the optimal combination of multiple factors in highway guardrail simulation experiments is provided. According to the design of highway guardrail simulation experiments, different levels of key factors are determined. Through simulation experiments, performance data such as the maximum synthetic acceleration of the center of mass and the maximum lateral dynamic deformation of the guardrail under various factor combinations are collected to construct the experimental space of factors and level values; and a radial basis model is constructed to approximate the complex nonlinear relationship between these factors and the performance evaluation index of the guardrail, that is, the radial basis proxy model, so as to adopt the multi-objective particle swarm optimization algorithm, and at the same time, the non-dominated solution that meets all the objective requirements is found by introducing the Pareto frontier solution set. Finally, by comprehensively analyzing the level combinations of each factor in the Pareto optimal solution set, the optimal level combination of factors is selected based on its comprehensive performance score. This process realizes the accurate quantification and scientific optimization of the optimal combination of highway guardrail simulation experiments under the influence of complex multiple factors, and improves the efficiency and effect of guardrail design. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0021] Figure 1 This is a schematic diagram of the RBF neural network training process according to an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of the MOPSO algorithm flow in an embodiment of the present invention;

[0023] Figure 3 A schematic diagram of part of the code for constructing a target proxy model according to an embodiment of the present invention;

[0024] Figure 4 This is a comparison chart of the prediction results of the training set and the test set of the embodiment of the present invention;

[0025] Figure 5 It is a schematic diagram of the overall flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the features and advantages of the present invention more clearly understood, the following embodiments are specifically described in detail as follows:

[0027] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0028] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0029] like Figure 5 As shown, the embodiment of the present invention provides a multi-factor optimal combination multi-objective optimization scheme for a highway guardrail simulation experiment, and its implementation process includes the following steps:

[0030] Step 1: According to the purpose of the experiment, select several key factors that may have a significant impact on the simulation experiment results and do not affect each other, that is, optimization variables, and determine different constraint ranges for each selected factor. Then, based on the economic and performance indicators of the highway guardrail, determine the evaluation indicators of the simulation experiment, that is, the optimization target, such as the maximum composite acceleration of the center of mass, the maximum lateral dynamic deformation of the guardrail, etc., and prepare a detailed experimental plan based on this, conduct simulation experiments according to the experimental plan, and record the experimental results.

[0031] Step 2: Organize the experimental results into an Excel file as a data set, and use Matlab software to train the data set with an RBF neural network. After multiple iterations of training, a stable RBF neural network model is finally obtained, which is a proxy model of the relationship between the key factors of the guardrail and the optimization target.

[0032] Step 3: Determine the optimization variables, constraints, and optimization goals. In order to efficiently explore the optimal combination of optimization variables under constraints, the RBF proxy model is used to predict the optimization target values ​​of various factor level combinations within the constraints, and it is used as the objective function. The multi-objective particle swarm optimization algorithm is implemented in Matlab software. Through information sharing and collaborative evolution among particles, the best balance point that meets multiple goals is found in the solution space, and several Pareto frontier solution sets are obtained by repeated operation. The average value is calculated to select the optimal factor level combination, and the optimal solution is verified by computer simulation.

[0033] As a preferred solution of this embodiment, the simulation experiment in step 1 is a vehicle-guardrail coupling collision simulation: after formulating the experimental plan, use 3D modeling software such as Solidworks and Caita to construct the guardrail and vehicle models, set keywords such as materials, loads, boundary conditions, etc. in finite element simulation software such as hypermesh, ansys, abaqus, etc., import the LS-dyna solver to obtain the simulation data, and finally use post-processing software such as LS-prepost and hyperview to collect and analyze the data results.

[0034] As a preferred solution of this embodiment, the RBF neural network training in step 2 uses its unique radial basis function as the activation function of the hidden layer neurons, which can effectively capture and approximate complex nonlinear relationships and continuously adjust network parameters so that the network output can be as close as possible to the true relationship between the key factors of the guardrail and the optimization target in the experimental data.

[0035] As a preferred solution of this embodiment, the RBF neural network model in step 2 utilizes the nonlinear approximation ability of the RBF neural network to approximate the complex functional relationship as a linear combination of a series of radial basis functions. By optimizing the parameters of the RBF neural network (such as the position of the center point, the width parameter, and the weight of the output layer), the proxy model can accurately approximate the output of the objective function, simulate the mathematical relationship between the optimization variable and the optimization target, and facilitate the subsequent optimization design.

[0036] As a preferred solution of this embodiment, the multi-objective particle swarm optimization algorithm in step 3 has two core iterative steps, namely, the particle update speed formula and the particle position update formula are as follows:

[0037] u i,j (t+1)=wu i,j (t)+c1r1(xp i.j (t)-x i.j (t))+c2r2(xg j (t)-x i.j (t)) Formula 1

[0038] x i,j (t+1)=x i,j (t)+v i,j (t+1) Formula 2

[0039] Where t is the number of iterations; w is the inertia weight; c1, c2 are learning factors; r1, r2 are random numbers that obey the uniform distribution U(0,1); is the best position that the particle itself has passed so far;

[0040] is the best position that the neighborhood particle has passed so far; j = 1, 2, 3…, n, where n is the dimension; i = 1, 2, 3…, N, where N is the particle size.

[0041] As a preferred solution of this embodiment, the Pareto front solution set in step 3, also called the Pareto front or non-inferior solution set, refers to a set of solutions that exist in a multi-objective optimization problem, and these solution sets achieve the best compromise between multiple objective functions. In other words, for any solution in the Pareto front solution set, there is no other solution that can make all objectives better without making at least one other objective worse. Therefore, these solutions form the boundary of an optimal solution in the objective space, namely the Pareto front.

[0042] like Figure 1 As shown, the complete steps of RBF neural network training are as follows:

[0043] 1) Data preprocessing: Preprocess the input data, including data cleaning, normalization and other operations, to ensure the accuracy and availability of the data.

[0044] 2) Initialize the hidden layer parameters: Determine the number of neurons in the hidden layer (i.e., the number of center points), which can be determined by random sampling, clustering (such as K-means algorithm), etc.

[0045] 3) Initialize the center and width (variance) of the radial basis function. These parameters determine the shape and coverage of the radial basis function.

[0046] 4) Feature extraction stage: Calculate the distance between the input data and each hidden layer neuron (center point). Substitute the distance into the radial basis function (such as Gaussian function) to obtain the output of the hidden layer.

[0047] 5) Output adjustment stage: Use the least squares method or other optimization algorithms to adjust the weight of the output layer according to the output of the hidden layer and the target output. Repeat this process until the output of the network is as close to the actual result as possible. When the preset number of iterations is reached or the convergence condition is met, the training process ends. At this time, the RBF neural network has learned the mapping relationship between the input data and the output, and the trained RBF neural network model is output.

[0048] like Figure 2 As shown, the complete steps of the MOPSO algorithm are as follows:

[0049] 1) Set algorithm-related parameters, including population size N, maximum number of iterations, inertia weight w, local and global acceleration constants c1 and c2, position and speed range limits, etc., initialize N populations, and determine the initial position;

[0050] ① Calculate the fitness values ​​corresponding to all particles and update the individual optimal position; ② According to the fitness value, store all particles with non-dominated relationships in the population into an external archive; ③ Each population selects its global optimal position.

[0051] 2) Each particle in the population uses equations 1 and 2 to update its position and velocity and calculate its fitness value.

[0052] 3) Select the individual optimal position and the global optimal position.

[0053] 4) Mutate the particles in the external archive formed in the previous generation.

[0054] 5) Calculate the new fitness value of the mutated particle, select the non-dominated solution together with the particles of the current generation, and update the external archive set.

[0055] 6) Determine whether the termination condition is met. If it is met, output the external archive as the final Pareto optimal solution set and exit. If it is not met, return to step 3).

[0056] In order to better understand the technical content of the present invention, the specific implementation scheme of the present method in practical engineering applications is described in detail below in combination with an engineering simulation example and accompanying drawings.

[0057] ①Simulation experiment stage

[0058] In the simulation experiment design, the highway guardrail simulation experiment selected the guardrail beam thickness, column thickness, and barrier thickness as key influencing factors, and conducted a total of 19 groups of simulation experiments to ensure the efficiency and representativeness of the experiment. At the same time, three evaluation indicators were selected: vehicle center of mass acceleration, maximum dynamic lateral deformation of the guardrail, and guardrail weight per unit length. The specific simulation experiment implementation plan is shown in Table 1.

[0059] Table 1 Implementation plan for the new guardrail collision simulation experiment

[0060]

[0061]

[0062] Through experimental simulation and data collection, the simulation experimental data of small passenger cars under different guardrail material combinations are shown in Table 2.

[0063] Table 2 Data of simulation experiment of new guardrail colliding with small passenger car

[0064]

[0065]

[0066] ②Constructing the target agent model

[0067] Import the data set into Matlab software, divide the first 16 data into training data sets, and the last 3 data as validation data sets, and normalize the data; use the newrb function in Matlab's Neural Network Toolbox to create an RBF neural network object, initialize the center and width parameters of the hidden layer nodes, and set the radial basis function expansion speed to 100. Part of the code is as follows Figure 3 shown.

[0068] Use the trained RBF neural network to make random predictions within the variable range and evaluate the performance of the model. Use the determination coefficient (the range is 0-1, the closer it is to 1, the better the model fit) and the mean square error to evaluate the model. Compare the prediction results of the training set and the test set. Figure 4 shown.

[0069] When the RBF neural network is trained and optimized to meet the requirements, it can be used as a proxy model to predict new input data.

[0070] ③Multi-objective particle swarm optimization stage

[0071] In Matlab software, a multi-objective particle swarm algorithm is used to perform multi-objective optimization on the optimization variables. The size of the particle swarm (i.e., population size) is set to 100, the position and velocity of each particle are randomly initialized, the individual optimal position and fitness value of each particle (i.e., individual historical optimal solution) are set, the global optimal solution set (i.e., Pareto frontier) is initialized, the inertia weight w=0.8, the learning factors c1, c2=1.5, and the number of optimization iterations T=500.

[0072] The RBF neural network model is used as the objective function (fitness function), and the optimization variable range and constraints are set according to the Highway Traffic Safety Facilities Design Specification (JTG / D81-2017). The Pareto dominance relationship is used to determine whether each particle is better than other particles or solutions in the global optimal solution set. The global optimal solution set is updated: the global optimal solution set is updated using the Pareto dominance relationship and the crowding distance. If the position of the current particle dominates some solutions in the global optimal solution set, these solutions are removed from the solution set and the current particle is added to the solution set. If the current particle is non-dominated with the solution in the global optimal solution set, the crowding distance strategy is used to decide whether to replace the solution in the solution set to maintain diversity.

[0073] Then update the speed and position: According to the speed and position update formula of the particle swarm optimization algorithm (Equation 1 and Equation 2), update the speed and position of each particle. The speed and position update may include the influence of inertia weight, cognitive and social components.

[0074] And perform boundary processing to ensure that the position and velocity of the particle are within the set search space. Check whether the termination conditions are met, such as reaching the maximum number of iterations and the global optimal solution set no longer changing. Finally, output the global optimal solution set, namely the Pareto frontier. The data set is shown in Table 3.

[0075] Table 3 Pareto frontier solution set after optimization

[0076] serial number <![CDATA[Maximum resultant acceleration / m·s -2 > Maximum dynamic deformation of guardrail / m <![CDATA[Weight of guardrail per unit length / kg·km -1 > 1 25.323155682 947.832765412 34.567281943 2 29.482716357 1063.529874361 35.298476531 3 22.953684201 765.419287365 42.783196452 4 21.674928530 981.653274189 33.124765839 5 24.197543862 1025.784369152 39.456827310 6 27.521087634 917.368542971 33.987653124 7 34.369875120 1092.486137259 35.623847519 8 30.092643785 953.276483195 41.237654821 9 28.857321946 796.541876329 33.451963872 10 23.410976583 974.825361947 44.876321459 11 51.276543189 1048.396725841 35.892764310 12 20.753216948 836.158472963 33.123965487 13 25.187439267 1018.743692857 36.548732169 14 21.349872651 995.632741895 33.765438291 15 29.963185472 823.965417283 40.983276541 16 37.816493275 1079.483629175 34.276531894 17 22.528741369 902.578364192 36.123876432 18 44.631789504 752.369184725 43.598763124 19 26.295473816 1039.287643159 33.847653129 20 28.409637258 968.417283695 41.653218764

[0077] According to the purpose of the experiment, the performance requirements, economy, durability, etc. of the guardrail are comprehensively considered to select the optimal solution. The tenth data is the best. While ensuring the minimum acceleration of the vehicle's center of mass, the maximum lateral dynamic deformation of the guardrail meets the specification value, and the dimensions of each part of the guardrail corresponding to the unit length weight can meet the requirements of GB / T 31439-2015 (corrugated beam steel guardrail); and its reliability is verified by the guardrail-vehicle coupling collision, that is, the safety performance, guidance performance, and buffering performance requirements of the guardrail installation and use can be met, and the importance of economy in engineering practice can be fully considered; thus, the optimal design of the multi-factor parameter combination of highway guardrails is finally successfully realized, which provides a theoretical basis for the design and optimization of highway guardrails, and provides a quantitative analysis method that can be used as a reference for similar multi-factor optimization problems.

[0078] Based on the same inventive concept, the present invention also provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the program instructions stored in the memory. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is used to implement one or more instructions, specifically for loading and executing one or more instructions in a computer storage medium to implement the above method.

[0079] It needs to be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium, on which a computer program is stored, and the computer program is executed by a processor to execute the above method. The storage medium can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0080] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0081] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

[0082] The present invention is not limited to the above-mentioned optimal implementation mode. Anyone can derive other forms of a multi-objective optimization method for highway guardrail simulation experiment under the inspiration of the present invention. All equal changes and modifications made according to the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A multi-objective optimization method for highway guardrail simulation experiment, characterized by: Based on the highway guardrail simulation experiment, different level values ​​of optimization variables are determined. Through the simulation experiment, the performance data corresponding to the guardrail performance evaluation index under each key factor combination corresponding to the optimization variable are collected; using the data obtained from the simulation experiment, a radial basis model is constructed and trained to approximate the nonlinear relationship between the key factors and the guardrail performance evaluation index, and a proxy model of the mathematical relationship between the guardrail key factors and the optimization target is obtained; the optimization target value of each key factor combination within the constraint condition predicted by the proxy model is used as the objective function, and a multi-objective particle swarm optimization algorithm is used to find the optimal combination of key factors, and the non-dominated solution that meets all optimization target requirements is found through the Pareto front solution set to select the optimal factor level combination.

2. A highway guardrail simulation experiment multi-objective optimization method according to claim 1, characterized in that: The optimization variables are determined by selecting several key factors that may have a significant impact on the simulation experiment results and do not affect each other according to the experimental purpose, and each selected key factor corresponds to a range of constraint conditions.

3. A highway guardrail simulation experiment multi-objective optimization method according to claim 1, characterized in that: The guardrail performance evaluation index is data reflecting the performance of the guardrail.

4. The multi-objective optimization method for highway guardrail simulation experiment according to claim 1 is characterized by: The simulation experiment is a vehicle-guardrail coupling collision simulation experiment, and the simulation experiment environment is set up in accordance with the highway traffic safety facility design specification JTG / D81-2017.

5. The multi-objective optimization method for highway guardrail simulation experiment according to claim 2 is characterized by: The scope of the constraint conditions is determined according to: Highway Traffic Safety Facilities Design Specification JTG / D81-2017 and GB / T 31439-2015 Corrugated Beam Steel Guardrail.

6. A highway guardrail simulation experiment multi-objective optimization method according to claim 1, characterized in that: The key factors include: the thickness of the guardrail beam, the thickness of the column and the thickness of the barrier block; the guardrail performance evaluation indicators include: the combined acceleration of the vehicle's center of mass, the maximum dynamic lateral deformation of the guardrail and the weight of the guardrail per unit length.

7. The multi-objective optimization method for highway guardrail simulation experiment according to claim 1 is characterized by: The proxy model uses a data set constructed using data obtained from simulation experiments, and is obtained after the RBF neural network is trained and converged through multiple iterations.

8. The multi-objective optimization method for highway guardrail simulation experiment according to claim 1 is characterized by: The multi-objective particle swarm optimization algorithm searches for the best balance point that satisfies multiple objectives in the solution space through information sharing and co-evolution among particles, and repeatedly runs the algorithm to obtain several Pareto frontier solution sets to select the optimal factor level combination.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of a multi-objective optimization method for a highway guardrail simulation experiment as described in any one of claims 1-8 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a multi-objective optimization method for a highway guardrail simulation experiment as described in any one of claims 1 to 8.

Images