Design method and device for S-shaped front longitudinal beam parameters, electronic equipment and medium

Abstract

The invention relates to an S-shaped front longitudinal beam parameter design method and device, electronic equipment and a medium, and belongs to the technical field of intelligent optimization algorithms, and the method comprises the steps: obtaining a first population and a temporary population based on an initial population and a direction vector of a target model of an S-shaped front longitudinal beam; performing grid division on the first population to obtain first grid data; performing selection operation on the first population based on the temporary population according to the probability table to generate a second population; performing grid division on the second population to obtain second grid data of the second population; updating a probability table based on the first grid data and the second grid data, wherein the updated probability table is used for selection operation in the next iteration step; combining the first population and the second population to obtain a third population, and performing environment selection on the third population to obtain a fourth population; and the fourth population is used as an initial population for iteration, and the optimal parameters of the S-shaped front longitudinal beam are determined based on the population after iteration. According to the method, the design precision of the S-shaped front longitudinal beam parameters is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent optimization algorithms, and in particular to a design method, device and electronic equipment for parameters of an S-shaped front longitudinal beam. Background Art

[0002] With the vigorous development of the automobile industry, automobile sales in my country have also been rising year by year, and the passive safety of vehicles has received more and more attention from consumers and regulatory authorities.

[0003] According to research, the vast majority of automobile accidents are caused by frontal collisions. The S-shaped front longitudinal beam is used to connect the front and rear ends of the vehicle. It is the main energy-absorbing structure when the vehicle collides head-on, absorbing collision energy and reducing impact force. Therefore, how to enhance the vehicle's anti-collision performance in a head-on collision has become the focus of the design.

[0004] The problem with the existing technology for designing the parameters of the S-shaped front longitudinal beam is that the existing intelligent algorithms for solving such problems have low convergence accuracy, and the optimized results often fail to meet the accuracy requirements of the designer.

[0005] In summary, the prior art lacks an improved multi-objective evolutionary algorithm to improve the design accuracy of the S-type front longitudinal beam parameters. Summary of the invention

[0006] In view of this, it is necessary to provide a method, device, electronic equipment and medium for designing parameters of an S-shaped front longitudinal beam, so as to solve the problem of poor design accuracy of the parameters of the S-shaped front longitudinal beam.

[0007] In order to solve the above problems, the present invention provides a method for designing parameters of an S-shaped front longitudinal beam, comprising: S1: Construct the target model of the S-shaped front longitudinal beam, generate the direction vector in the target space, and generate a random initial population in the decision space; S2: Get the first population and temporary population based on the initial population and direction vector; S3: gridding the first population to obtain first grid data of the first population; S4: performing a selection operation on the first population based on the temporary population according to the probability table to generate a second population; S5: gridding the second population to obtain second grid data of the second population; S6: updating the probability table based on the first grid data and the second grid data, and the updated probability table is used for the selection operation in S4 in the next iteration step; S7: The first population and the second population are combined to obtain the third population, and the third population is subjected to environmental selection to obtain the fourth population; S8: using the fourth population as the initial population, repeating steps S2-S7, and determining the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

[0008] In a possible implementation, constructing a target model of the S-shaped front longitudinal beam, generating a direction vector in a target space, and generating a random initial population in a decision space include: Construct a target model of the S-shaped front longitudinal beam to maximize energy absorption, minimize impact force, and the width, height, and thickness of the S-shaped front longitudinal beam; According to the range of decision variables of the target model, random initialization is performed in the decision space to obtain the initial population; Generates uniformly distributed direction vectors in target space.

[0009] In a possible implementation, obtaining the first population and the temporary population based on the initial population and the direction vector includes: Determine the angle between individuals in the initial population and the direction vector; Associating the individuals in the initial population with the direction vector with the smallest angle to obtain a plurality of first associated subpopulations of the initial population; The first population and the temporary population are determined based on the number of the first associated sub-populations and the preset number of sub-populations of the initial population.

[0010] In a possible implementation, determining the first population and the temporary population based on the number of the first associated sub-population and the preset number of sub-populations of the initial population includes: When it is determined that the number of the first associated subpopulations is less than the preset number of subpopulations of the initial population, a preset number of subpopulations is selected from the initial population to be added to the first associated subpopulation to obtain a first population and a temporary population, wherein the preset number of subpopulations is the preset number of subpopulations of the initial population minus the number of the first associated subpopulations; After determining that the number of the associated sub-populations is not less than the preset number of sub-populations of the initial population, a first population and a temporary population are obtained based on a non-dominated sorting algorithm.

[0011] In a possible implementation, the first grid data includes a grid state and a grid dominance ranking, and the grid division of the first population to obtain the first grid data of the first population includes: Determining an upper boundary and a lower boundary of a first target space based on the preset dimension and the target model; Divide the target model into a preset number of parts to obtain a number of equally divided first target models; determining the coordinates of each individual in the first population based on the upper boundary and the lower boundary; determining a grid state of the first population based on the coordinates of each individual in the first population; A grid dominance ranking of the first population is determined based on the grid status and the preset number of shares of the first population.

[0012] In a possible implementation manner, updating the probability table based on the first grid data and the second grid data includes: The first individual in the first population generates a second individual as an offspring based on the strategy in the probability table, and the second individual is an individual in the second population; Calculate the mesh states of the first body and the second body; Determining a grid dominance ranking of the first individual based on a grid state of the first individual; determining a grid dominance ranking of the second individual based on the grid state of the second individual; When it is determined that the grid dominance ranking of the second individual is less than the grid dominance ranking of the first individual, determining the reward value to be a first preset value; When it is determined that the grid dominance ranking of the second individual is not less than the grid dominance ranking of the first individual, determining the reward value to be a second preset value; The probability value of the strategy in the probability table is determined based on the reward value and the current probability value of the strategy.

[0013] In a possible implementation, performing environmental selection on the third population to obtain the fourth population includes: Calculating the angle between each individual in the third population and the direction vector; Associating the individuals in the third population with the direction vector with the smallest angle to obtain a plurality of second associated subpopulations of the third population; When it is determined that the number of the second associated subpopulations is less than the preset number of subpopulations of the initial population, a preset number of subpopulations is selected from the three populations and added to the second associated subpopulation, where the preset number of subpopulations is the preset number of subpopulations of the initial population minus the number of the second associated subpopulations; After determining that the number of the second associated sub-populations is not less than the preset number of sub-populations of the initial population, a fourth population is obtained based on the angle and grid dominance ranking of each individual of the third population.

[0014] On the other hand, the present invention also provides a device for designing parameters of an S-shaped front longitudinal beam, comprising: Initialization parameter module, used to build the target model of the S-shaped front longitudinal beam, generate direction vectors in the target space, and generate random initial populations in the decision space; A first population acquisition module, used for obtaining a first population and a temporary population based on an initial population and a direction vector; A first grid data acquisition module is used to perform grid division on the first population to obtain first grid data of the first population; a second population acquisition module is used to perform a selection operation for the first population based on the temporary population according to the probability table to generate a second population; A second grid data acquisition module, used for performing grid division on the second population to obtain second grid data of the second population; A probability table updating module, used for updating the probability table based on the first grid data and the second grid data, wherein the updated probability table is used for the selection operation in the next iteration step; A population merging module, used for merging the first population and the second population to obtain a third population, and performing environmental selection on the third population to obtain a fourth population; The loop iteration module is used to repeat steps S2-S7 using the fourth population as the initial population, and determine the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

[0015] On the other hand, the present invention also provides an electronic device, including a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the method for designing parameters of an S-shaped front longitudinal beam described in any one of the above implementations.

[0016] On the other hand, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps in a method for designing parameters of an S-shaped front longitudinal beam described in any of the above-mentioned implementation methods.

[0017] The beneficial effects of the present invention are as follows: the present invention provides a method for designing parameters of an S-shaped front longitudinal beam, the method comprising: constructing and generating an initial population in a decision space and a direction vector in a target space based on a target model of the S-shaped front longitudinal beam, obtaining a first population and a temporary population based on the initial population and the direction vector, meshing the first population to obtain first grid data of the first population, thereby better evaluating the convergence and distribution of a solution set, performing a selection operation on the first population based on the temporary population according to a probability table to generate a second population, thereby selecting the most appropriate offspring generation strategy for individuals according to different evolutionary states, and dynamically inverting the grid state based on the grid state control. Feedback reward strategy, according to the change of the grid state transition dominance relationship between the parent generation, i.e., the first population individual and the offspring generation, i.e., the second population individual, a dynamic feedback reward is given to the strategy, thereby affecting the preference for the generation strategy, the second population is gridded to obtain the second grid data of the second population; based on the first grid data and the second grid data, the probability table is updated, and the updated probability table is used for the selection operation in the next iteration step, the first population and the second population are merged to obtain the third population, and the third population is subjected to environmental selection to obtain the fourth population, and the fourth population is used as the initial population for iteration, and the optimal parameters of the S-shaped front longitudinal beam are determined based on the iterated population. The present invention divides the population into grids and generates a dynamic feedback reward strategy based on the grid state dominance in the generation of the offspring sub-population, according to the change of the grid state transition dominance relationship between the parent generation, i.e., the first population individual and the offspring generation, i.e., the second population individual, a dynamic feedback reward is given to the strategy, thereby affecting the preference for the generation strategy, thereby improving the convergence accuracy of the algorithm, and further improving the accuracy of the parameters of the S-shaped front longitudinal beam. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flow chart of an embodiment of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Figure 2 A schematic structural diagram of an anti-collision box provided by the present invention; Figure 3 A schematic structural diagram of an S-shaped front longitudinal beam provided by the present invention; Figure 4 A schematic diagram of dividing the initial population in the decision space of a design method for parameters of an S-shaped front longitudinal beam provided by the present invention; Figure 5 For the present invention Figure 1 A schematic flow chart of an embodiment of S102; Figure 6 For the present invention Figure 1 A schematic flow chart of an embodiment of S103; Figure 7 A schematic diagram of each target grid division of a decision space in an embodiment of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Figure 8 A schematic diagram of decision space population grid division in an embodiment of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig. 9 Two method flow charts of embodiments of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig.10 A schematic diagram of method population division for two embodiments of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig.11 A method flow chart for generating offspring for two embodiments of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig.12 A schematic flow chart of a method for selecting an environment for a merged population in two embodiments of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig.13 The figure is a result of experimental comparison in two embodiments of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig.14 A schematic flow chart of an embodiment of a device for designing parameters of an S-shaped front longitudinal beam provided by the present invention; Fig.15 A schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0019] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0020] Before presenting the embodiments, the following terms are explained.

[0021] Evolutionary algorithm is designed and proposed based on the evolution law of organisms in nature. It is a computational model of biological evolution process that simulates the natural selection and genetic mechanism of Darwin's theory of biological evolution. It is a method of searching for the optimal solution by simulating the natural evolution process. The algorithm uses mathematical methods and computer simulation operations to convert the problem-solving process into a process similar to the crossover and mutation of chromosome genes in biological evolution. When solving more complex combinatorial optimization problems, it can usually obtain better optimization results faster than some conventional optimization algorithms.

[0022] Embodiment 1: The present invention provides a method, device and electronic equipment for designing parameters of an S-shaped front longitudinal beam, which are respectively described below.

[0023] Figure 1A schematic flow chart of an embodiment of a method for designing parameters of an S-shaped front longitudinal beam provided by the present invention is shown in FIG. Figure 1 As shown in FIG. 1 , the design method of the S-shaped front longitudinal beam parameters includes: S101: Construct a target model of the S-shaped front longitudinal beam, generate a direction vector in the target space, and generate a random initial population in the decision space; S102: obtaining a first population and a temporary population based on the initial population and the direction vector; S103: gridding the first population to obtain first grid data of the first population; S104: performing a selection operation on the first population based on the temporary population according to the probability table to generate a second population; S105: Grid-dividing the second population to obtain second grid data of the second population; S106: updating the probability table based on the first grid data and the second grid data, and the updated probability table is used for the selection operation in S104 in the next iteration step; S107: merging the first population and the second population to obtain a third population, and performing environmental selection on the third population to obtain a fourth population; S108: Repeat steps S102-S107 with the fourth population as the initial population, and determine the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

[0024] Compared with the prior art, the present embodiment provides a method for designing parameters of an S-shaped front longitudinal beam, the method comprising: constructing and generating an initial population in a decision space and a direction vector in a target space based on a target model of the S-shaped front longitudinal beam, obtaining a first population and a temporary population based on the initial population and the direction vector; performing grid division on the first population to obtain first grid data of the first population, so as to better evaluate the convergence and distribution of the solution set, performing a selection operation on the first population based on the temporary population according to a probability table to generate a second population, thereby selecting the most appropriate offspring generation strategy for individuals according to different evolutionary states, and based on the dynamic feedback dominated by the grid state. The reward strategy is to give the strategy a dynamic feedback reward according to the change of the grid state transition dominance relationship between the parent generation, i.e., the first population individual, and the offspring generation, i.e., the second population individual, so as to affect the preference for the generation strategy, and to grid the second population to obtain the second grid data of the second population; based on the first grid data and the second grid data, the probability table is updated, and the updated probability table is used for the selection operation in the next iteration step, the first population and the second population are merged to obtain the third population, and the third population is subjected to environmental selection to obtain the fourth population, and the fourth population is used as the initial population for iteration, and the optimal parameters of the S-shaped front longitudinal beam are determined based on the iterated population. The present invention provides a dynamic feedback reward strategy based on the grid state dominance in generating the offspring sub-population, and according to the change of the grid state transition dominance relationship between the parent generation, i.e., the first population individual, and the offspring generation, i.e., the second population individual, so as to affect the preference for the generation strategy, thereby improving the convergence accuracy of the algorithm and further improving the accuracy of the parameters of the S-shaped front longitudinal beam.

[0025] It should be noted that the present invention is applicable to the parameter design of the S-shaped front longitudinal beam. It can be understood that in other multi-objective optimization problems, adopting the technical solution of the present invention only requires changing the target model.

[0026] In a specific embodiment of the present invention, Figure 2 As shown in the figure, this crash box is a thin-walled structure, and its characteristics are greatly affected by its wall thickness. The maximum crushing force and mass of this thin-walled material are limited to below a certain value. The front end of the S-shaped front longitudinal beam is connected to the crash box. When a frontal collision occurs, the front longitudinal beam deforms to absorb the energy brought by the collision to protect the safety of the driver and passengers in the car. Its crashproof performance is closely related to the design parameters. The cross-sectional size of the thin-walled structure is the main factor affecting its energy absorption performance and maximum impact resistance, such as Figure 3 As shown, its width w ,high h and thickness 𝑡 are three important parameters considered in the scheme design. Therefore, in order to improve the anti-collision performance of the S-shaped front longitudinal beam, its scheme design can be abstracted as a dual-objective optimization problem, that is, to find a set of ( w , h ,t) , in order to maximize energy absorption While minimizing the impact force .

[0027] The S-shaped front longitudinal beam design optimization problem can be modeled as the following dual-objective optimization problem, that is, the expression of the objective model: (1) ,

[0028] (2) ; In the formula, ; (3) ; represents the objective function of the target model; It means that the S-shaped front longitudinal beam maximizes energy absorption; It means that the S-shaped front longitudinal beam minimizes the impact force; Indicates the height of the S-shaped front longitudinal beam; Indicates the width of the S-shaped front longitudinal beam; Indicates the thickness of the S-shaped front longitudinal member.

[0029] In some embodiments of the present invention, constructing a target model of an S-shaped front longitudinal beam, generating a direction vector in a target space, and generating a random initial population in a decision space includes: Construct a target model of the S-shaped front longitudinal beam to maximize energy absorption, minimize impact force, and the width, height, and thickness of the S-shaped front longitudinal beam; According to the range of decision variables of the target model, random initialization is performed in the decision space to obtain the initial population; Generates uniformly distributed direction vectors in target space.

[0030] In a specific embodiment of the present invention, the parameters are first initialized. Specifically, an initial population is first randomly generated in the decision space, and then their target values ​​are calculated. : (4) is the population size, is the number of subproblems (i.e., subpopulations), is the size of the subpopulation.

[0031] Produced in the target space Evenly distributed direction vectors divide the target space into multiple subspaces. For an optimization problem with m goals, assuming that each goal is evenly divided into H parts, each subproblem will correspond to an m-dimensional vector , where each element of the vector is taken from , and the sum of each dimension is 1, that is, = 1. The number of vectors is ,like Figure 4 As shown in Figure 1, for a two-target problem, when the number of divisions of each target is 3, the four direction vectors in the figure can be obtained. Each direction vector corresponds to a subspace, and the vector coordinates are shown in Table 1.

[0032] Table 1: Direction vector examples

[0033] Then, according to the angle between the individual and the direction vector, the individual is associated with the direction vector with the smallest angle and assigned to the corresponding subspace. The individuals in the same subspace form the corresponding subpopulation. Because the population and direction vector are uniformly distributed during initialization, the size of the obtained sub-population is also relatively average. Figure 4 As shown, according to the direction vector v 1. v 2. v 3. v 4. Divide the space into 4 areas Ω 1. Ω 2. Ω 3. Ω 4. Calculate the individual x i With direction vector v j Angle θ ij ,like , attach the individual to the vector v k On. Figure 4 middle x 1 attached to the direction vector v 1， Divided into regions Ω 1 in; x 2 Attached to the direction vector v 2， Divided into regions Ω 2 in… (5) The above describes the process of population initialization. After the population initialization is completed, the population will enter the iterative evolution stage.

[0034] In some embodiments of the present invention, in step S102, as Figure 5 As shown, the first population and the temporary population are obtained based on the initial population and the direction vector, including: S501, determining the angle between the individuals in the initial population and the direction vector; S502, associating the individuals in the initial population with the direction vector with the smallest angle to obtain a plurality of first associated subpopulations of the initial population; S503: Determine a first population and a temporary population based on the number of the first associated sub-populations and the preset number of sub-populations of the initial population.

[0035] In some embodiments of the present invention, determining the first population and the temporary population based on the number of the first associated sub-population and the preset number of sub-populations of the initial population includes: When it is determined that the number of the first associated subpopulations is less than the preset number of subpopulations of the initial population, a preset number of subpopulations is selected from the initial population to be added to the first associated subpopulation to obtain a first population and a temporary population, wherein the preset number of subpopulations is the preset number of subpopulations of the initial population minus the number of the first associated subpopulations; After determining that the number of the associated sub-populations is not less than the preset number of sub-populations of the initial population, a first population and a temporary population are obtained based on a non-dominated sorting algorithm.

[0036] In a specific embodiment of the present invention, the determining of the first population and the temporary population based on the number of the first associated sub-population and the preset number of sub-populations of the initial population includes the following specific steps: Step 1: Initialize the angle matrix θ =zeros( N , K ), calculate each individual in the first population P according to formula (5) i With each direction vector k Angle θ ik ; Step 2: Find out θ Minimum subscript of the i-th row in the matrix , then the individual i With direction vector j Related, specific examples are Figure 4 As shown. At the same time, individual i is assigned to the subpopulation p j .

[0037] Step 3: If , then randomly select from the first population P Individuals are added to the subpopulation middle; like , then the subpopulation p j Perform non-dominated sorting. Non-dominated sorting method: First, individuals that are not dominated by any individual are placed in the first layer of the non-dominated sorting and stored in ; Then remove the individuals in the first layer, and among the remaining individuals, the non-dominated individuals are placed in the second layer and stored ; and so on, until the number of individuals that have undergone non-dominated sorting exceeds In the last layer of non-dominated sorting, the nearest neighbor distances between individuals in the last layer are calculated and sorted in descending order, and the ones at the back are truncated. individual; Step 4: Traverse each individual in the first population P and find the direction vector associated with the individual j and the corresponding subpopulation ,exist Another individual is randomly selected as the parent and saved in the temporary population P'.

[0038] In some embodiments of the present invention, in step S103, Figure 6 As shown, the first grid data includes a grid state and a grid dominance ranking, and the first grid data of the first population is obtained by gridding the first population, including: S601, determining an upper boundary and a lower boundary of a first target space based on a preset dimension and a target model; S602, dividing the target model into a preset number of parts to obtain a number of equally divided first target models; S603, determining the coordinates of each individual in the first population based on the upper boundary and the lower boundary; S604, determining a grid state of the first population based on the coordinates of each individual in the first population; S605: Determine a grid dominance ranking of the first population based on the grid status and the preset number of shares of the first population.

[0039] In a specific embodiment of the present invention, before describing in detail how to divide the first population into grids, the grid division strategy of the present technical solution is first introduced.

[0040] The grid division strategy is as follows: First, determine the j The lower and upper bounds of the dimensional boundary. Assume They represent the individuals in the population in j The minimum and maximum function values ​​on the targets are divided into div parts. j The upper and lower bounds of the target are set as: (6) (7) The entire target space will be divided into div m Hyperboxes, the first j The side length of a dimension can be expressed as: (8) It can be seen from formulas (6) and (7) that lb j and ub j Respectively The purpose of this design is to ensure that the boundary points do not fall on the edge of the grid. Figure 7 As shown in j Schematic diagram of the determination and division of the grid boundary on the target, the individual x in the group j The grid state coordinates are obtained by lb j and ub j Determined, it can be expressed as: (9) In the formula Represents an individual In the The function value on the target, For individuals In the The grid coordinates on the target.

[0041] Determine the grid state of the individual: After the grid division is completed, the grid state of the individual is calculated according to the coordinates of the individual in the grid distribution. Figure 8 As shown, taking a two-target problem as an example, div is set to 10, so the target space is divided into 10 2 grids, and the grid coordinates of individual x are between (0,0) and (9,9). Then, individual x Grid status s x It can be defined as: (10) in, For individuals x grid coordinates in the first target dimension, For individuals x The grid coordinates in the second target dimension have the following specific correspondence: Figure 8 As shown, the number in each small square in the figure represents the grid status of the individual divided into the corresponding grid.

[0042] Grid state dominance ranking: In order to better evaluate the convergence of the solution set, the grid dominance ranking of the individual is determined according to the grid state. The grid state dominance scheme designed in this paper is as follows: (11) in, s x Is an individual x The grid state of div is the number of copies of each target dimension. rank x Is an individual x The grid is dominant in the ranking. For solving the minimization problem, the smaller the rank value, the more dominant the individual is. For example, in Figure 8 In the example, for individuals with grid states 2 and 11, their dominant ranking values ​​calculated by formula (11) are both 2, indicating that individuals with grid states 2 and 11 do not dominate each other under grid division. Through the coarse-grained grid division method proposed in the present invention, it can be seen that individuals in the grids passed by the same dotted line have the same rank, and this strategy can quickly calculate the dominant relationship between individuals.

[0043] After introducing the grid division strategy, the following describes in detail the grid division of the first population in this embodiment to obtain the first grid data of the first population. The specific steps are as follows: Step 1: Divide the target space into grids according to formulas (6) and (7). The size is set to 10, so if Figure 8 As shown, the target space is divided into 100 state blocks.

[0044] Step 2: Traverse each individual in the parent generation, i.e. the first population P, and calculate the coordinates of each individual in the grid according to formulas (8) and (9). Determine the grid state of each parent generation according to formula (10) and record it as , according to formula (11), determine the grid state dominance ranking p of each parent individual rank .

[0045] In a specific embodiment of the present invention, in step S104, a selection operation is performed for the first population based on the temporary population according to the probability table to generate a second population, that is, an operator selection operation is performed for the first population based on the temporary population according to the probability table in the genetic operation operator pool to generate the second population. First, the strategy for generating offspring in this embodiment is introduced. The genetic operation operator strategy pool constructed in this embodiment includes a simulated binary crossover operator (SBX), a differential evolution operator (DE), and an LLX operator. A genetic operator operation strategy pool is constructed.

[0046] In each state, the parent individual selects a mutation strategy from the mutation strategy pool according to the mutation probability table to generate a child individual. The child individual is compared with the parent individual in terms of state dominance, and forward feedback is performed to give rewards and update the mutation probability table, thereby adaptively adjusting the frequency of use of each strategy according to the current evolutionary state.

[0047] Construct mutation probability table: There are three mutation operators in the strategy pool, such as Figure 8 As shown in Table 2, the target space is divided into 100 grid states. The individuals in each state may select one of the genetic operators to produce offspring with a certain probability. Therefore, the probability of selecting a certain operator in a certain state is recorded in the table, which is a 100×3 matrix, as shown in Table 2. Q ( i , m j ) indicates that in the state i Next Select j Probability of operators. In the initial stage of the evolutionary algorithm, each value in the matrix is ​​first initialized to 1 / 3, indicating that the selection probability of each operator in each state is equal.

[0048] Table 2: Genetic operation selection probability table

[0049] The roulette algorithm is used to select the mutation operator in the strategy pool according to the mutation probability table, and then the parent individual uses the mutation operator to generate the offspring individual.

[0050] The reward feedback value reward is obtained based on the grid status and dominant ranking of the parent generation, i.e., the first population individuals, and the offspring generation, i.e., the second population individuals. Assume that individual a chooses the mutation strategy m i Produced offspring b , calculate the grid state of the two individuals according to formula (10), and convert the individuals a The grid state is recorded as s a ,individual b The grid state is recorded as s b According to formula (11), calculate the parent generation, i.e. the first population individual a and generate offspring, the second population of individuals b The grid-dominant ranking of rank a , rank b ), if the offspring, i.e., the second population, has a better ranking than the parent, i.e., the first population, ( rank b < rank a ), indicating the selected mutation strategy m iexist s a state produces better offspring, then the state s a Using mutation strategy m i The reward value reward is set to 1; on the contrary, if the parent, i.e., the first population, has a higher ranking than the offspring, i.e., the second population, the ranking is better ( rank a < rank b ) or the ranking of the parent, i.e., the first population, is equal to the ranking of the offspring, i.e., the second population. rank a = rank b ), indicating that s a Select strategy m in the state i If no individual with better convergence is generated, the reward value under this change is set to 0. The reward value setting is shown in formula (12): (12) To update the mutation probability table, we need to update the mutation probability table based on the parent, i.e. the individual grid state of the first population. s a and the offspring, i.e. the grid state of the second population s b The change of and the corresponding reward value reward update the probability of using the mutation strategy. According to the following formula:

[0051] in, is the return rate, indicating that the parent state is the first population state s a The second population state s b Level of study; is the discount factor, The larger the number, the stronger the development capability. The smaller it is, the stronger the exploration ability is; m i is the selected strategy. The strategy is executed according to the parent state, that is, the first population state m i Probability , the reward value from the parent state (i.e., the first population state) to the child state (i.e., the second population state) and the maximum probability max(Q(s b ,:)) to update s a Select strategy under status m iThe probability Q(s a ,m i ), which is used in the next roulette selection algorithm. After traversing all individuals in the population, the newly generated offspring set O is obtained, which is the second population.

[0052] The specific steps in the code are as follows: Step 1: Traverse the first population P Each individual i , according to the individual grid status , let s = . Read the evolution operation selection probability table Q s OK, that is s The execution probability of each strategy under the state is the first population individual P i Select the genetic operation to be performed. The other parent participating in the genetic operation is taken from the corresponding subscript in the temporary population P', that is, The specific method is as follows: Generate a random number rand ∈[0,1) CP=0 for j=1 to 3 CP=CP+Q sj if rand <CP Strategy i =j Break; end_if end_for Step 2: Record the selected strategy according to strategy and select the genetic operation operator for individual i.

[0053] switch(strategy i ) Case 1: / / Execute SBX genetic operator operation if rand≤0.5

[0054] else

[0055] end_if break Case 2: / / Execute DE genetic operator operation

[0056] break Default: / / Execute LLX genetic operator operation / / t and T represent the current evolutionary generation and the maximum evolutionary generation respectively

[0057]

[0058] Break end_switch Step 3: Calculate the status and ranking of each individual in the offspring population O according to formulas (8), (9), (10), and (11), and record them in the array , O rank ; Step 4: Update the genetic operation selection probability table Q. Compare the status ranking changes of the offspring, i.e., individuals in the second population, and the parent, i.e., individuals in the first population, in turn, and calculate the reward value according to formula (12). Formula (13) updates the probability that the parent, i.e., individuals in the first population, will use this strategy in this state next time. The specific operations are as follows: for i=1 to NP if <

[0059] j = Strategy i reword=1

[0060]

[0061] end_if end_for In some embodiments of the present invention, updating the probability table based on the first grid data and the second grid data includes: The first individual in the first population generates a second individual as an offspring based on the strategy in the probability table, and the second individual is an individual in the second population; Calculate the mesh states of the first body and the second body; Determining a grid dominance ranking of the first individual based on a grid state of the first individual; determining a grid dominance ranking of the second individual based on the grid state of the second individual; When it is determined that the grid dominance ranking of the second individual is less than the grid dominance ranking of the first individual, determining the reward value to be a first preset value; When it is determined that the grid dominance ranking of the second individual is not less than the grid dominance ranking of the first individual, determining the reward value to be a second preset value; The probability value of the strategy in the probability table is determined based on the reward value and the current probability value of the strategy.

[0062] In some embodiments of the present invention, the step of performing environmental selection on the third population to obtain the fourth population comprises: Calculating the angle between each individual in the third population and the direction vector; Associating the individuals in the third population with the direction vector with the smallest angle to obtain a plurality of second associated subpopulations of the third population; When it is determined that the number of the second associated subpopulations is less than the preset number of subpopulations of the initial population, a preset number of subpopulations is selected from the three populations and added to the second associated subpopulation, where the preset number of subpopulations is the preset number of subpopulations of the initial population minus the number of the second associated subpopulations; After determining that the number of the second associated sub-populations is not less than the preset number of sub-populations of the initial population, a fourth population is obtained based on the angle and grid dominance ranking of each individual of the third population.

[0063] In a specific embodiment of the present invention, a third merged population is obtained, and a dual standard selection mechanism of grid and direction vector angle of the third population is subjected to environmental selection to obtain a fourth population. This dual standard selection not only maintains the convergence of the population but also ensures the distribution of the population. The specific steps are as follows: Step 1: Set each subpopulation , initialize the angle matrix θ =zeros(2 N , K ), calculate each individual in U according to formula (5) i With each direction vector k Angle θ ik .

[0064] Step 2: Find out θ The matrix i Minimum row subscript , then the individual i With direction vector j Related, specific examples are Figure 3 As shown. i Assigned to subpopulation p j .

[0065] Step 3: If , then from outside the subpopulation Set selection and direction vector v j The front with the smallest angle Individuals join the subpopulation , and Join the fourth group; if , then the subpopulation p j The individuals in the grid are ranked by grid dominance and direction vector angle. Set SP = ,Temp= . SP = SP∪Temp, the subpopulation p j All individuals with the smallest grid dominance ranking in the grid are stored in Temp and extracted from p j , and repeat the above steps until |SP|+|Temp|≥ . Use formula (5) to calculate the individual and direction vector in Temp v j The smallest angle Individuals join SP. p j =SP and p j Join the fourth group.

[0066] Step 4: Iteration = Iteration + 1, use the fourth population as the initial population for the next iteration until Iteration>Max_Iteration. Output the population P after the last iteration as the optimal solution obtained by the algorithm.

[0067] The embodiment of the present invention performs environmental selection on the merged population based on a dual selection mechanism of direction vector division and state grid dominance. This dual selection mechanism ensures the convergence of the population and maintains the diversity of the population.

[0068] Example 2: ∅ like Fig. 9 The specific steps of the second embodiment are shown as follows: Step 1: Initialize parameters and the number of problems to be optimized m 2, set the number of direction vectors K =10, population size N =100, the number of copies of each dimension in the target space , the maximum number of iterations Max_Iteration = 20000, the initial probability of each item in the mutation probability table is 1 / 3, the learning rate , discount factor .

[0069] Step 2: Randomly initialize the population in the decision space P ={ x 1, x 2, …, xN},in x i =( x i1 , x i2 , x i3 ),make x i1 ∈[136,146], x i2 ∈[58,66], x i3 ∈[14,22].

[0070] Step 3: Generate 10 uniformly distributed direction vectors in the target space, as shown in Table 3 and Fig.10 shown.

[0071] Table 3: Direction vector

[0072] Step 4: Associate the individual with the direction vector with the smallest angle, and form a subpopulation of individuals in the same sub-region , so that the size of each subpopulation is .

[0073] Step 4.1: Initialize the angle matrix θ =zeros( N , K ), calculate each individual in P according to formula (5) i With each direction vector k Angle θ ik .

[0074] Step 4.2: Find out θ Minimum subscript of the i-th row in the matrix , then the individual i With direction vector j Related, specific examples are Figure 3 As shown. At the same time, individual i is assigned to the subpopulation p j .

[0075] Step 4.3: If , then randomly select from the population P Individuals are added to the subpopulation middle; like , then the subpopulation p jPerform non-dominated sorting. Non-dominated sorting method: First, individuals that are not dominated by any individual are placed in the first layer of the non-dominated sorting and stored in the temporary population. ; Then remove the individuals in the first layer, and among the remaining individuals, the non-dominated individuals are placed in the second layer and stored ; and so on, until the number of individuals that have undergone non-dominated sorting exceeds In the last layer of non-dominated sorting, the nearest neighbor distances between individuals in the last layer are calculated and sorted in descending order, and the ones at the back are truncated. Individual.

[0076] Step 5: Traverse each individual in the first population P and find the direction vector associated with the individual j and the corresponding first associated subpopulation ,exist Another individual is randomly selected as the parent and saved in the temporary population P'.

[0077] Step 6: Divide the grid and determine the grid status of the individuals in the parent generation, i.e. the first population.

[0078] Step 6.1: Divide the target space into grids according to formulas (6) and (7). The size is set to 10, so if Figure 8 As shown, the target space is divided into 100 state blocks.

[0079] Step 6.2: Traverse each individual in the parent generation, i.e., the first population group P, calculate the coordinates of each individual in the grid according to formulas (8) and (9), and determine the grid state of each parent generation, i.e., the first population individual, according to formula (10) and record it as According to formula (11), the grid state dominance ranking p of each parent, i.e., the first population individual, is determined rank .

[0080] Step 7: Fig.11 As shown, it is the parent generation, i.e. the first population P Each individual in i Select genetic operation operators in the operator pool to generate offspring, namely the second population O.

[0081] Step 7.1: Traverse the parent population, i.e. the first population P Each individual i , according to the individual state calculated in step 6 , let s = . Read the first s OK, that is s The execution probability of each strategy under the state is the parent generation, that is, the first population individual P iSelect the genetic operation to be performed. The other parent participating in the genetic operation is taken from the corresponding subscript in the temporary population P', that is, The specific method is as follows: Generate a random number rand ∈[0,1) CP=0 for j=1 to 3 CP=CP+Q sj if rand <CP Strategy i =j Break; end_if end_for Step 7.2: According to the strategy, record the selected strategy and select the genetic operation operator for the first population individual i.

[0082] switch(strategy i ) Case 1: / / Execute SBX genetic operator operation if rand≤0.5

[0083] else

[0084] end_if break Case 2: / / Execute DE genetic operator operation

[0085] break Default: / / Execute LLX genetic operator operation / / t and T represent the current evolutionary generation and the maximum evolutionary generation respectively

[0086]

[0087] Break end_switch Step 8: Calculate the status and ranking of each individual in the offspring population, i.e., the second population O, according to formulas (8), (9), (10), and (11), and record them in the array , O rank .

[0088] Step 9: Update the genetic operation selection probability table Q. Compare the status ranking changes of the offspring individuals, i.e., the individuals of the second population, and the parent individuals, i.e., the individuals of the first population, in turn, and calculate the reward value according to formula (12). Formula (13) updates the probability that the parent individual, i.e., the individuals of the first population, will use this strategy in this state next time. The specific operations are as follows: for i=1 to NP if <

[0089] j = Strategy i reword=1

[0090]

[0091] end_if end_for Step 10: Merge the parent population, i.e. the first population P, and the child population, i.e. the second population O, into a set U, i.e. the third population. Perform environmental selection using the grid and direction vector angle dual-standard selection mechanism designed in this scheme to obtain the fourth population. This dual-standard selection maintains both the convergence of the population and the distribution of the population.

[0092] Step 10.1: Set each subpopulation , initialize the angle matrix θ =zeros(2 N , K ), calculate each individual in U according to formula (5) i With each direction vector k Angle θ ik .

[0093] Step 10.2: Find out θ The matrix i Minimum row subscript , then the individual i With direction vector j Related, specific examples are Figure 3 As shown. i Assigned to subpopulation p j .

[0094] Step 10.3: If , then from outside the subpopulation Set selection and direction vector v j The front with the smallest angle Individuals join the subpopulation P j , and P j Join the fourth group; if , then the subpopulation p j The individuals in the grid are ranked by grid dominance and direction vector angle. Set SP = ,Temp= . SP = SP∪Temp, the subpopulation p j All individuals with the smallest grid dominance ranking in the grid are stored in Temp and extracted from p j , and repeat the above steps until |SP|+|Temp|≥ . Use formula (5) to calculate the individual and direction vector in Temp v j The smallest angle Individuals join SP. p j =SP, and p j Join the fourth population. The specific selection process is as follows Fig.12 shown.

[0095] Step 11: Iteration = Iteration + 1, go to step 5, use the fourth population as the initial population for the next iteration, until Iteration>Max_Iteration. Output the population after the last iteration as the optimal solution obtained by the algorithm.

[0096] Experimental results: According to the above specific implementation, the algorithm is used to solve the S-shaped front longitudinal beam design optimization problem and conduct simulation experiments. In 30 independent experiments, the population size N = 100, the maximum number of iterations Max_Iteration = 20000, the number of direction vectors K = 10, and the learning rate , discount factor .

[0097] Experiments were conducted using the present invention. Fig.13 The visualization data of a set of Pareto solutions in the target space in 30 independent experiments is presented. It can be seen that the optimal solution obtained by the algorithm proposed in the present invention approaches the Pareto frontier and the distribution between solutions is also good.

[0098] A set of experimental results from 30 experiments were screened, and two individuals from the obtained non-dominated solutions were extracted for comparison with the solutions designed in the prior art, as shown in Table 4. The brackets in Table 4 indicate the percentage of improvement of the solution of the present invention over the original solution, and the calculation method is as shown in Formula (14).

[0099] (14) Table 4 Comparison of algorithm solution results

[0100] The experimental results show that compared with the original design, when the vehicle is in a frontal collision, Result 1 improves the energy absorption capacity by 0.47% and reduces the impact force by 1.44%. Result 2 improves the energy absorption capacity by 0.76% and reduces the impact force by 0.62%.

[0101] In order to better implement a method for designing parameters of an S-shaped front longitudinal beam in an embodiment of the present invention, based on a method for designing parameters of an S-shaped front longitudinal beam, correspondingly, as follows Fig.14 As shown, an embodiment of the present invention further provides a device for designing parameters of an S-shaped front longitudinal beam, and a device 1400 for designing parameters of an S-shaped front longitudinal beam includes: Initialization parameter module 1401, used to construct a target model of the S-shaped front longitudinal beam, generate a direction vector in the target space, and generate a random initial population in the decision space; A first population acquisition module 1402, used to obtain a first population and a temporary population based on an initial population and a direction vector; The first grid data acquisition module 1403 is used to perform grid division on the first population to obtain first grid data of the first population; the second population acquisition module 1404 is used to perform a selection operation for the first population based on the temporary population according to the probability table to generate the second population; The second grid data acquisition module 1405 is used to perform grid division on the second population to obtain second grid data of the second population; A probability table updating module 1406, used to update the probability table based on the first grid data and the second grid data, and the updated probability table is used for the selection operation in the next iteration step; A population merging module 1407 is used to merge the first population and the second population to obtain a third population, and perform environmental selection on the third population to obtain a fourth population; The loop iteration module 1408 is used to repeat steps S2-S7 using the fourth population as the initial population, and determine the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

[0102] The device 1400 for designing parameters of an S-shaped front longitudinal beam provided in the above embodiment can implement the technical solution described in the above embodiment of a method for designing parameters of an S-shaped front longitudinal beam. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above embodiment of a method for designing parameters of an S-shaped front longitudinal beam, which will not be repeated here.

[0103] like Fig.15 As shown, the present invention also provides an electronic device 1500. The electronic device 300 includes a processor 1501, a memory 1502 and a display 1503. Fig.15 Only some components of the electronic device 300 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0104] In some embodiments, the processor 1501 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 1502, such as a design method for parameters of an S-shaped front longitudinal beam in the present invention.

[0105] In some embodiments, processor 1501 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, processor 1501 may be local or remote. In some embodiments, processor 1501 may be implemented in a cloud platform. In some embodiments, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0106] In some embodiments, the memory 1502 may be an internal storage unit of the electronic device 1500, such as a hard disk or memory of the electronic device 1500. In other embodiments, the memory 1502 may also be an external storage device of the electronic device 1500, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 1500.

[0107] Furthermore, the memory 1502 may include both an internal storage unit of the electronic device 1500 and an external storage device. The memory 1502 is used to store application software installed in the electronic device 1500 and various data.

[0108] In some embodiments, the display 1503 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 1503 is used to display information on the electronic device 1500 and to display a visual user interface. The components 1501-1503 of the electronic device 1500 communicate with each other via a system bus.

[0109] In one embodiment, when the processor 1501 executes a design program for parameters of an S-shaped front longitudinal beam in the memory 1502, the following steps may be implemented: S1: Construct the target model of the S-shaped front longitudinal beam, generate the direction vector in the target space, and generate a random initial population in the decision space; S2: Get the first population and temporary population based on the initial population and direction vector; S3: gridding the first population to obtain first grid data of the first population; S4: performing a selection operation on the first population based on the temporary population according to the probability table to generate a second population; S5: gridding the second population to obtain second grid data of the second population; S6: updating the probability table based on the first grid data and the second grid data, and the updated probability table is used for the selection operation in S4 in the next iteration step; S7: The first population and the second population are combined to obtain the third population, and the third population is subjected to environmental selection to obtain the fourth population; S8: using the fourth population as the initial population, repeating steps S2-S7, and determining the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

[0110] It should be understood that: when the processor 1501 executes a design program for parameters of an S-shaped front longitudinal beam in the memory 1502, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0111] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 1500 mentioned, and the electronic device 1500 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 1500 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0112] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0113] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for designing parameters of an S-shaped front longitudinal beam, characterized in that: include: S1: Construct the target model of the S-shaped front longitudinal beam, generate the direction vector in the target space, and generate a random initial population in the decision space; S2: Get the first population and temporary population based on the initial population and direction vector; S3: gridding the first population to obtain first grid data of the first population; S4: performing a selection operation on the first population based on the temporary population according to the probability table to generate a second population; S5: gridding the second population to obtain second grid data of the second population; S6: updating the probability table based on the first grid data and the second grid data, and the updated probability table is used for the selection operation in S4 in the next iteration step; S7: The first population and the second population are combined to obtain the third population, and the third population is subjected to environmental selection to obtain the fourth population; S8: using the fourth population as the initial population, repeating steps S2-S7, and determining the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

2. The method for designing parameters of an S-shaped front longitudinal beam according to claim 1, characterized in that: The target model of the S-shaped front longitudinal beam is constructed, a direction vector is generated in the target space, and a random initial population is generated in the decision space, including: Construct a target model of the S-shaped front longitudinal beam to maximize energy absorption, minimize impact force, and the width, height, and thickness of the S-shaped front longitudinal beam; According to the range of decision variables of the target model, random initialization is performed in the decision space to obtain the initial population; Generates uniformly distributed direction vectors in target space.

3. The method for designing parameters of an S-shaped front longitudinal beam according to claim 1, characterized in that: The step of obtaining a first population and a temporary population based on the initial population and the direction vector includes: Determine the angle between individuals in the initial population and the direction vector; Associating the individuals in the initial population with the direction vector with the smallest angle to obtain a plurality of first associated subpopulations of the initial population; The first population and the temporary population are determined based on the number of the first associated sub-populations and the preset number of sub-populations of the initial population.

4. The method for designing parameters of an S-shaped front longitudinal beam according to claim 3, characterized in that: The determining of the first population and the temporary population based on the number of the first associated sub-population and the preset number of sub-populations of the initial population includes: When it is determined that the number of the first associated subpopulations is less than the preset number of subpopulations of the initial population, a preset number of subpopulations is selected from the initial population to be added to the first associated subpopulation to obtain a first population and a temporary population, wherein the preset number of subpopulations is the preset number of subpopulations of the initial population minus the number of the first associated subpopulations; After determining that the number of the associated sub-populations is not less than the preset number of sub-populations of the initial population, a first population and a temporary population are obtained based on a non-dominated sorting algorithm.

5. The method for designing parameters of an S-shaped front longitudinal beam according to claim 1, characterized in that: The first grid data includes a grid state and a grid dominance ranking, and the first grid data of the first population is obtained by gridding the first population, including: Determining an upper boundary and a lower boundary of a first target space based on the preset dimension and the target model; Divide the target model into a preset number of parts to obtain a number of equally divided first target models; determining the coordinates of each individual in the first population based on the upper boundary and the lower boundary; determining a grid state of the first population based on the coordinates of each individual in the first population; A grid dominance ranking of the first population is determined based on the grid status and the preset number of shares of the first population.

6. The method for designing parameters of an S-shaped front longitudinal beam according to claim 1, characterized in that: The updating of the probability table based on the first grid data and the second grid data comprises: The first individual in the first population generates a second individual as an offspring based on the strategy in the probability table, and the second individual is an individual in the second population; Calculate the mesh states of the first body and the second body; Determining a grid dominance ranking of the first individual based on a grid state of the first individual; determining a grid dominance ranking of the second individual based on the grid state of the second individual; When it is determined that the grid dominance ranking of the second individual is less than the grid dominance ranking of the first individual, determining the reward value to be a first preset value; When it is determined that the grid dominance ranking of the second individual is not less than the grid dominance ranking of the first individual, determining the reward value to be a second preset value; The probability value of the strategy in the probability table is determined based on the reward value and the current probability value of the strategy.

7. The method for designing parameters of an S-shaped front longitudinal beam according to claim 2, characterized in that: The step of performing environmental selection on the third population to obtain the fourth population comprises: Calculating the angle between each individual in the third population and the direction vector; Associating the individuals in the third population with the direction vector with the smallest angle to obtain a plurality of second associated subpopulations of the third population; When it is determined that the number of the second associated subpopulations is less than the preset number of subpopulations of the initial population, a preset number of subpopulations is selected from the three populations and added to the second associated subpopulation, where the preset number of subpopulations is the preset number of subpopulations of the initial population minus the number of the second associated subpopulations; After determining that the number of the second associated sub-populations is not less than the preset number of sub-populations of the initial population, a fourth population is obtained based on the angle and grid dominance ranking of each individual of the third population.

8. A design device for parameters of an S-shaped front longitudinal beam, characterized in that: include: Initialization parameter module, used to build the target model of the S-shaped front longitudinal beam, generate direction vectors in the target space, and generate random initial populations in the decision space; A first population acquisition module, used for obtaining a first population and a temporary population based on an initial population and a direction vector; A first grid data acquisition module is used to perform grid division on the first population to obtain first grid data of the first population; a second population acquisition module is used to perform a selection operation for the first population based on the temporary population according to the probability table to generate a second population; A second grid data acquisition module, used for performing grid division on the second population to obtain second grid data of the second population; A probability table updating module, used for updating the probability table based on the first grid data and the second grid data, wherein the updated probability table is used for the selection operation in the next iteration step; A population merging module, used for merging the first population and the second population to obtain a third population, and performing environmental selection on the third population to obtain a fourth population; The loop iteration module is used to repeat steps S2-S7 using the fourth population as the initial population, and determine the optimal parameters of the S-shaped front longitudinal beam based on the iterated population.

9. An electronic device, characterized in that: comprising a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the method for designing parameters of an S-shaped front longitudinal beam as described in any one of claims 1 to 7 above.

10. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the method for designing parameters of an S-shaped front longitudinal beam as described in any one of claims 1 to 7 above.

Images