High-speed train top-level design index definition method based on comprehensive evaluation optimization

Abstract

The invention discloses a high-speed train top-level design index definition method based on comprehensive evaluation optimization. The high-speed train top-level design index definition method comprises the following steps from S1 to S4. The method comprises the following steps: S1, determining the weight of a top design index; comprising the following steps: determining the weight of each evaluation criterion, and calculating the relative weight of a top design index; s2, initially defining a top design index; comprising the steps of establishing a competitive product reference, and determining key top design indexes of the novel high-speed train; s3, carrying out nonlinear standardization processing on the top-layer design indexes; comprising the steps that a nonlinear S-type derivable function is constructed and solved; s4, constructing and solving a comprehensive competitiveness optimization model; comprising the steps of constructing and solving a comprehensive competitiveness calculation model of the whole high-speed train, constructing a comprehensive competitiveness optimization model of the high-speed train, and solving the optimization model by utilizing a genetic algorithm. By adopting the method, the reasonable value of the high-speed train top layer design index can be quantitatively determined.

Description

Technical Field

[0001] The present invention relates to a method for defining design indicators, and specifically, to a method for defining top-level design indicators of high-speed trains based on comprehensive evaluation and optimization. Background Art

[0002] At present, the methods for defining product design indicators mainly focus on two aspects: QFD-based modeling and data-driven. The QFD-based method mainly realizes the transformation from customer requirements to product indicators. For example, the mapping relationship between product requirements and engineering indicators is established through fuzzy optimization models, customer satisfaction models, etc.; the data-driven method extracts design information from a large amount of real data to determine the relationship between the design space and constraints, and customer preferences and design indicators. The data sources include product data, sales data, operation data, etc. However, in the above methods, the former needs to obtain satisfaction information such as customer preferences and value perception through surveys and joint analysis. Due to the individual differences in the understanding of different products among the respondents, it is difficult to evaluate the perceived performance and the reliability is low. The latter has high requirements for the scale and quality of data, and the design information extracted from a single or small amount of data sources is incomplete.

[0003] Product design indicators, as the input of forward design, are technical measures that describe product design conditions, usage efficiency, and engineering quality, and are characterized by a series of engineering characteristics and their corresponding values. They will directly affect product performance and quality, and thus affect its core competitiveness. Top-level design indicators of high-speed trains such as speed, axle load, and noise can reflect the overall state, core technology, and system interfaces of high-speed trains, and are the basic basis for their forward design. Therefore, the definition of top-level design indicators is the primary link in the forward design of high-speed trains. However, the current process of defining top-level design indicators of high-speed trains mainly relies on the empirical judgment of R & D personnel or the research on incomplete information and data of competing products and product performance tests, resulting in a long design cycle, strong subjectivity of the definition results, and poor interpretability. Therefore, how to quantitatively determine the reasonable values of top-level design indicators of high-speed trains is an urgent problem to be solved. Summary of the Invention

[0004] A method for defining top-level design indicators of high-speed trains based on comprehensive evaluation and optimization, including the following steps S1 - S4;

[0005] Step S1: Determining the weights of top-level design indicators; including determining the weights of each evaluation criterion and calculating the relative weights of top-level design indicators;

[0006] Step S2: Initial definition of top-level design indicators; including establishing a benchmark for competing products and determining the key top-level design indicators of new high-speed trains;

[0007] Step S3: Nonlinear normalization processing of top-level design indicators; including constructing a nonlinear S-shaped differentiable function and solving it.

[0008] Step S4: Construction and solution of the comprehensive competitiveness optimization model; including constructing and solving the comprehensive competitiveness calculation model of the entire high-speed train, constructing the comprehensive competitiveness optimization model of the high-speed train, and using the genetic algorithm to solve the optimization model.

[0009] Preferably, step S1 specifically includes the following steps S11 - S12:

[0010] Step S11: Use the group comprehensive analytic hierarchy process to determine the weights of each evaluation criterion.

[0011] Step S12: Calculate the relative weights of the top-level design indicators.

[0012] Preferably, S111: Define the goal and criteria; the goal is to determine the weights of high-speed train performance; the criteria are transportation capacity, safety, comfort, and environmental friendliness.

[0013] S112: Form a decision-making group; select experts or decision-makers in multiple relevant fields to form a decision-making group.

[0014] S113: Design a judgment matrix; the judgment matrix A is an n*n square matrix, and each element a in the matrix ij represents the relative importance of criterion i relative to criterion j; each expert makes pairwise comparisons of the criteria according to the 1 - 9 scale method to obtain the elements in the matrix and constructs the judgment matrix.

[0015]

[0016] S114: Calculate the individual ranking vector; in the case of s experts and n evaluation objects, the individual ranking vector of the n evaluation objects by the i-th expert is:

[0017]

[0018] S115: Calculate the similarity relationship matrix; according to the individual ranking vectors, calculate the similarity relationship matrix R between experts, R = (r ij ) n×n ; denote the decision similarity degree between expert i and expert j as r ij , 0 ≤ r ij ≤ 1, i, j ∈ {1, 2,..., s}, the closer r ij is to 0, the less similar the decisions of expert i and expert j are, and the closer r ij is to 1, the more similar the decisions of expert i and expert j are; the calculation formula for the similarity degree r ij is:

[0019]

[0020] Among them, w ik and w jk are the weights of expert i and expert j for the k-th criterion respectively; the expression of the similarity relationship matrix R is as follows:

[0021]

[0022] S116: Fuzzy clustering analysis; including the following step1-step3:

[0023] step1: According to each expert judgment matrix, find the individual ranking vector of the expert and conduct a consistency test. For the judgment matrix that fails the consistency test, feedback its result to each expert and readjust it;

[0024] step2: Calculate the transitive closure of the similarity matrix R R = (r ij ) n×n is the similarity relationship matrix, then there exists a smallest natural number k, k ≤ n, such that the transitive closure For all natural numbers 1 greater than m, there is R 1 = R m ; at this time is the fuzzy equivalence matrix; reflects the comprehensive similarity degree relationship among experts. Use the successive squaring method to find the transitive closure of the similarity matrix R

[0025] step3: Select a threshold λ, where 0 ≤ λ ≤ 1, and conduct clustering analysis on experts through ; the larger λ is, the higher the requirement for the similarity between experts. Only when the similarity degree between two experts reaches or exceeds this threshold will they be classified into the same category; conversely, the smaller λ is, the lower the requirement for the similarity between experts classified into the same category;

[0026] Use the transitive closure Conduct clustering on experts according to the selected threshold λ; check the values of the elements in ; if the element values corresponding to the i-th expert and the j-th expert in are greater than or equal to λ, then classify these two experts into the same category;

[0027] S117: Determination of expert weights;

[0028] Through fuzzy clustering analysis, divide experts into different categories; experts in the same category have the same weight, and the weights of experts in different categories are determined according to the number of experts in the category;

[0029] Let the number of experts in the \(i\)-th category be \(\beta\). i Then the weight calculation formula for the \(i\)-th category of experts is as shown in Equation (5):

[0030]

[0031] where \(b\) represents the number of expert categories, represents the sum of the squares of the number of experts in all categories, and \(\varepsilon\) i represents the weight of the experts in the \(i\)-th category; the expert weights can be calculated;

[0032] S118: Calculate the final weights; according to the expert weights and the individual ranking vectors, calculate the final weights of each criterion; the individual ranking vectors reflect the judgment of each expert on the weights of each criterion, and the expert weights reflect the relative importance of expert opinions in the group; calculating the final weights includes the following step1-step2:

[0033] step1: For each criterion, multiply the weight value corresponding to this criterion in the individual ranking vector of each expert by the weight of this expert; assume there are \(n\) criteria and \(m\) experts. For the \(k\)-th criterion, calculate the product corresponding to the \(i\)-th expert is the weight value of the \(k\)-th criterion in the individual ranking vector of the \(i\)-th expert, and \(\varepsilon\) i represents the weight of the experts in the \(i\)-th category;

[0034] step2: Sum up the products calculated by all experts for the same criterion to obtain the final weight of this criterion; as shown in Equation (6):

[0035]

[0036] where: \(W\) k is the final weight of the \(k\)-th criterion; calculate for each criterion to obtain the weights of all evaluation criteria.

[0037] Preferably, step S12 includes steps S121 - S122;

[0038] S121: Establish a relationship matrix;

[0039] Use the relationship degree matrix in quality function deployment to establish the relationship between the top-level design indicators of high-speed trains and each evaluation criterion. The rows of the relationship degree matrix represent different evaluation criteria, the columns represent specific top-level design indicators, and each element \(R\) in the matrix i,j represents the relationship degree between the \(i\)-th evaluation criterion and the \(j\)-th top-level design indicator, and \(W\) j is the relative weight of the \(j\)-th top-level design indicator, which is calculated through the relationship degree matrix and the evaluation criterion weights;

[0040] S122: Establish the relative weight formula for top-level design indicators;

[0041] W j is the relative weight of the j-th top-level design indicator; W j The calculation formula is as shown in Equation (7):

[0042]

[0043] In the formula, m represents the number of evaluation criteria items, n represents the number of top-level design indicator items, and R i,j is the relationship degree between the i-th evaluation criterion and the j-th top-level design indicator, with values of 0, 1, 3, 5, 7, 9, and the corresponding correlation degrees are: irrelevant, weakly correlated, moderately weakly correlated, moderately correlated, strongly moderately correlated, strongly correlated, and ω i is the weight of the i-th evaluation criterion.

[0044] Preferably, step S2 specifically includes the following steps S21 - S22:

[0045] Step S21: Establish a benchmark for competing products;

[0046] Step S22: Determine the key top-level design indicators of the new high-speed train based on the competing products.

[0047] Preferably, step S3 is specifically to construct and solve a non-linear S-shaped differentiable function:

[0048] Use a non-linear S-shaped differentiable function to map the original value data of the high-speed train top-level design indicators to the interval (0, 1). The non-linear S-shaped differentiable function is as shown in Equation (8):

[0049]

[0050] In the formula, V j and are the top-level design indicator values before and after standardization respectively. V jmin and V jmax represent the minimum and maximum values of the j-th top-level design indicator respectively. S and T are the S-shaped curve shape adjustment parameters, and k is the S-shaped curve direction adjustment parameter.

[0051] Preferably, step S4 specifically includes the following steps S41 - S43:

[0052] Step S41: Construct and solve the comprehensive competitiveness calculation model for the entire high-speed train;

[0053] Step S42: Construct the comprehensive competitiveness optimization model for the high-speed train;

[0054] Step S43: Use the genetic algorithm to solve the optimization model.

[0055] Preferably, step S41 is specifically as follows: Construct a calculation model for comprehensive competitiveness as shown in Equation (9):

[0056]

[0057] In the formula, is the value of the j-th top-level design index after standardization, and W j is the relative weight of the j-th top-level design index, and n is the number of top-level design indices.

[0058] Preferably, step S42 is specifically as follows:

[0059] Taking the comprehensive competitiveness of high-speed trains as the optimization goal, transform the definition problem of different top-level design indices of high-speed trains into an optimization problem of the comprehensive competitiveness of high-speed trains, and construct the following optimization model:

[0060]

[0061] Subject to: V jmin ≤V j ≤V jmax (11)

[0062]

[0063] Among them, Equation (11) is the definition range of each top-level design index; Equation (12) is the relevant mathematical model of the j-th top-level design index and other top-level design indices; Equation (13) is the relevant constraint model of the j-th top-level design index and other top-level design indices.

[0064] Preferably, step S43 is specifically as follows:

[0065] Select a genetic algorithm to solve the top-level design index scheme after continuous iterative optimization, and compare and evaluate it with the initial definition scheme. If it meets the design requirements, the reasonable values of the top-level design indices of high-speed trains can be obtained. Otherwise, continue to perform iterative optimization to solve a new top-level design index scheme until the design requirements are met, that is, the optimization definition of the indices is completed.

[0066] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0067] (1) The inventor found in practice that the reasonable definition of product design indices aims to improve product competitiveness, and product competitiveness may be affected by the index combination rather than a single index. The present invention proposes to use "transport capacity, safety, comfort, and environmental friendliness" as the evaluation criteria for the definition of top-level design indices, which can more comprehensively and accurately evaluate the overall level of high-speed trains.

[0068] (2) In practice, the inventors found that the process of defining top-level design indicators based on traditional QFD faces problems such as strong subjectivity and low reliability in the results of customer participation in performance perception evaluation. The present invention integrates expert experience evaluation into this process, makes full use of the comprehensive prior design information in expert experience knowledge, and can effectively improve the reliability of evaluation results.

[0069] (3) In practice, the inventors found that when determining the weights of top-level design indicators, common methods such as subjective weighting, objective weighting, and subjective-objective combined weighting cannot fully reflect expert preferences. In response to this, the present invention proposes to use the group analytic hierarchy process (GAHP) method to determine weights, which can more comprehensively reflect expert preferences, combines subjective judgment and objective data at the same time, and improves the accuracy and scientificity of weight determination.

[0070] (4) In practice, the inventors found that due to the different dimensions and orders of magnitude of the values of different top-level design indicators, the requirements for data consistency are not met, and it is difficult to comprehensively evaluate each top-level design indicator with the same standard to obtain the comprehensive competitiveness expression of high-speed train products. In response to this, the present invention proposes to use a non-linear S-shaped differentiable function to map the original value data of high-speed train top-level design indicators to the interval (0,1), realizing the standardization of the numerical values of high-speed train top-level design indicators.

[0071] (5) In practice, the inventors found that for a high-speed train vehicle, it is composed of multiple complex systems, with intricate structures and functions. Due to physical and technical limitations in its design, while meeting the requirements of a certain top-level design indicator, other aspect indicator requirements will be sacrificed correspondingly to ensure the overall vehicle comprehensive performance requirements. Therefore, the present invention constructs a comprehensive competitiveness calculation model for high-speed train vehicles, comprehensively considers each top-level design indicator, and realizes the reasonable calculation of the overall vehicle comprehensive performance.

[0072] (6) In practice, the inventors found that there are differences in the influence of top-level design indicators under different evaluation criteria on the comprehensive competitiveness of high-speed train products. In response to this, the present invention establishes a hierarchical structure as shown in Figure 3 between the evaluation criteria and the comprehensive competitiveness, providing a basis for more reasonable calculation of the comprehensive competitiveness.

[0073] (7) In practice, the inventors found that there are problems such as long design cycles, strong subjectivity in definition results, and poor interpretability in the process of traditional high-speed train top-level design indicator definition. The present invention constructs an optimization model for the comprehensive competitiveness of high-speed train top-level design indicators based on historical product and competitive product data and solves it through a genetic algorithm to determine the optimal top-level design indicator combination scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a schematic diagram of the optimization definition process of high-speed train top-level design indicators;

[0075] Figure 2 Schematic diagram of the composition of the top-level design indicators for high-speed trains;

[0076] Figure 3 Schematic diagram of the hierarchical structure of the evaluation criteria for the top-level design indicators of high-speed trains;

[0077] Figure 4 Schematic diagram of the basic structure of the relational degree matrix. Detailed implementation manner

[0078] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention.

[0079] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely represents some embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0080] A method for defining the top-level design indicators of a high-speed train based on comprehensive evaluation optimization, comprising the following steps:

[0081] Step S1: Determining the weights of the top-level design indicators. This includes determining the weights of each evaluation criterion and calculating the relative weights of the top-level design indicators.

[0082] Step S2: Initial definition of the top-level design indicators. This includes establishing a benchmark for competing products and determining the key top-level design indicators of the new high-speed train.

[0083] Step S3: Nonlinear normalization processing of the top-level design indicators. This includes constructing a nonlinear S-shaped differentiable function and solving it.

[0084] Step S4: Construction and solution of the comprehensive competitiveness optimization model. This includes constructing a comprehensive competitiveness calculation model for the entire high-speed train and solving it, constructing a comprehensive competitiveness optimization model for the high-speed train, and using a genetic algorithm to solve the optimization model.

[0085] Preferably, step S1 specifically includes the following steps S11 - S12:

[0086] Step S11: Using the group analytic hierarchy process (GAHP) method to determine the weights of each evaluation criterion.

[0087] Step S11 specifically includes steps S111 - S118.

[0088] S111: Define the goals and criteria. Preferably, the goal is to determine the weights of the high-speed train performance. The criteria are: transport capacity, safety, comfort, and environmental friendliness.

[0089] S112: Form a decision-making group. Preferably, select multiple experts or decision-makers in relevant fields to form a decision-making group.

[0090] S113: Design a judgment matrix. The judgment matrix A is an n*n square matrix, and each element a ij in the matrix represents the relative importance of criterion i relative to criterion j. Each expert makes pairwise comparisons of the criteria according to the 1-9 scale method (AHP scale) to obtain the elements in the matrix and construct the judgment matrix.

[0091]

[0092] S114: Calculate the individual ranking vector. Preferably, in the case of s experts and n evaluation objects, the individual ranking vector of the n evaluation objects by the i-th expert is:

[0093]

[0094] S115: Calculate the similarity relationship matrix. Preferably, according to the individual ranking vectors, calculate the similarity relationship matrix R among experts, R=(r ij ). Denote the decision similarity degree between expert i and expert j as r n×n . For 0≤r ij ≤1, i, j∈{1, 2, …, s}, the closer r ij is to 0, the less similar the decisions of expert i and expert j are, and the closer r ij is to 1, the more similar the decisions of expert i and expert j are. The calculation formula for the similarity degree r ij is: ij

[0095]

[0096] where w ik and w jk are the weights of the k-th criterion by expert i and expert j respectively. The expression of the similarity relationship matrix R is as follows:

[0097]

[0098] S116: Fuzzy clustering analysis. Preferably, it includes step1-step3:

[0099] ​Step 1: Based on each expert's judgment matrix, calculate the individual ranking vector of the expert and conduct a consistency test. For the judgment matrix that fails the consistency test, feedback its results to each expert and readjust it.

[0100] Step 2: Calculate the transitive closure of the similarity matrix R R=(r ij ) n×n is the similarity relation matrix. Then there exists a smallest natural number k (k ≤ n) such that the transitive closure For all natural numbers l greater than m, there is R 1 =R m . At this time is the fuzzy equivalence matrix. reflects the comprehensive similarity degree relationship among experts. Use the successive squaring method to find the transitive closure of the similarity matrix R

[0101] Step 3: Select a threshold λ (0 ≤ λ ≤ 1), and conduct cluster analysis on experts through The selection of λ is determined according to actual needs. The larger λ is, the higher the requirement for the similarity between experts. Only when the similarity degree between two experts reaches or exceeds this threshold will they be grouped into the same category; conversely, the smaller λ is, the lower the requirement for the similarity between experts grouped into the same category.

[0102] Use the transitive closure Conduct cluster analysis on experts according to the selected threshold λ. Check the values of the elements in . If the element value corresponding to the i-th expert and the j-th expert in is greater than or equal to λ, then group these two experts into the same category.

[0103] S117: Determination of expert weights.

[0104] Through fuzzy cluster analysis, divide experts into different categories. Experts in the same category have the same weight, and the weights of experts in different categories are determined according to the number of experts in the category.

[0105] Let the number of experts clustered into the i-th category be β i , then the weight calculation formula for the i-th category of experts is as shown in Equation (5):

[0106]

[0107] where b represents the number of expert categories, represents the sum of the squares of the number of experts in all categories, and ε i represents the weight of the i-th category of experts. The expert weights can be calculated.

[0108] S118: Calculate the final weights. Based on the weights of the experts and the individual ranking vectors, calculate the final weights of each criterion. The individual ranking vectors reflect each expert's judgment on the weights of the criteria, while the expert weights reflect the relative importance of the experts' opinions in the group. Calculating the final weights includes the following steps 1 - 2:

[0109] Step 1: For each criterion, multiply the weight value corresponding to this criterion in each expert's individual ranking vector by the weight of that expert. Suppose there are n criteria and m experts. For the kth criterion, calculate the product corresponding to the ith expert is the weight value of the kth criterion in the individual ranking vector of the ith expert, and ε i represents the weight of the experts in the ith category.

[0110] Step 2: Sum up the products calculated by all experts for the same criterion to obtain the final weight of this criterion. As shown in Equation (6):

[0111]

[0112] where: W k is the final weight of the kth criterion. Calculate for each criterion to obtain the weights of all evaluation criteria.

[0113] Step S12: Calculate the relative weights of the top - level design indicators.

[0114] Preferably, Step S12 includes Steps S121 - S122.

[0115] S121: Establish the relationship matrix.

[0116] Use the relationship degree matrix in Quality Function Deployment (QFD) to establish the relationship between the top - level design indicators of high - speed trains and each evaluation criterion.

[0117] Preferably, the basic structure of the relationship degree matrix is as Figure 4 shown. The rows of the relationship degree matrix represent different evaluation criteria, the columns represent specific top - level design indicators, and each element R i,j in the matrix represents the relationship degree between the ith evaluation criterion and the jth top - level design indicator, and W j is the relative weight of the jth top - level design indicator, which is calculated through the relationship degree matrix and the weights of the evaluation criteria.

[0118] S122: Establish the formula for the relative weights of the top - level design indicators.

[0119] W j is the relative weight of the jth top - level design indicator. W jThe calculation formula is as shown in Equation (7):

[0120]

[0121] In the formula, m represents the number of items of evaluation criteria, n represents the number of items of top-level design indicators, and R i,j is the correlation degree between the i-th evaluation criterion and the j-th top-level design indicator, and the value range is "0, 1, 3, 5, 7, 9", and the corresponding correlation degrees are: irrelevant, weakly relevant, relatively weakly relevant, moderately relevant, relatively strongly relevant, strongly relevant. ω i is the weight of the i-th evaluation criterion.

[0122] Preferably, step S2 specifically includes the following steps S21 - S22:

[0123] Step S21: Establish a benchmark for competing products.

[0124] Preferably, based on the determined relative weights of the top-level design indicators, select typical domestic and foreign competing products of the same type as the benchmark for competing products according to the design objectives and design features of the newly designed high-speed train.

[0125] Step S22: Determine the key top-level design indicators of the new high-speed train according to the competing products. The differences between the key top-level design indicators of the new high-speed train and the indicators of the competing products are adaptively set by those skilled in the art according to the performance improvement requirements of the new high-speed train.

[0126] Preferably, step S3 is specifically: construct a non-linear S-shaped differentiable function and solve it.

[0127] Since the dimensions and orders of magnitude of the values of different top-level design indicators are different, which do not meet the data consistency requirements, it is difficult to comprehensively evaluate each top-level design indicator with the same standard to obtain the comprehensive competitiveness expression of the high-speed train product. Preferably, use a non-linear S-shaped differentiable function to map the original value data of the high-speed train top-level design indicators to the interval (0, 1). The non-linear S-shaped differentiable function is as shown in Equation (8):

[0128]

[0129] In the formula, V j and are the top-level design indicator values before and after standardization respectively. V jmin and V jmax represent the minimum and maximum values of the j-th top-level design indicator respectively. S and T are the shape adjustment parameters of the S-shaped curve, and k is the direction adjustment parameter of the S-shaped curve. For the values of the above parameters, the following are: obtain V jmin and V jmax by statistically analyzing the data of previous-generation products and competing products; S is usually taken as 1 - 2, T is usually taken as 1; k takes the value of 1 or - 1. For the top - level design indicators where the larger the value is, the better, k takes 1, otherwise, k takes - 1.

[0130] Preferably, step S4 specifically includes the following steps S41 - S43:

[0131] Step S41: Construct and solve the comprehensive competitiveness calculation model of the high - speed train vehicle.

[0132] For the high - speed train vehicle, it consists of multiple complex systems, with intricate structures and functions. Due to physical and technical limitations in its design, it is difficult to fully meet the customer's preferences for the product. Therefore, while meeting the requirements of a certain top - level design indicator, other indicator requirements will be sacrificed accordingly to ensure the overall performance requirements of the vehicle. Thus, the comprehensive competitiveness of the high - speed train vehicle is the comprehensive result after considering various top - level design indicators. The calculation model of the comprehensive competitiveness is as shown in Equation (9):

[0133]

[0134] In the formula, is the value of the j - th top - level design indicator after standardization, W j is the relative weight of the j - th top - level design indicator, and n is the number of top - level design indicators.

[0135] Step S42: Construct the comprehensive competitiveness optimization model of the high - speed train.

[0136] Taking the comprehensive competitiveness of the high - speed train as the optimization goal, transform the definition problem of different top - level design indicators of the high - speed train into the optimization problem of the comprehensive competitiveness of the high - speed train, and construct the optimization model as follows:

[0137]

[0138] Subject to: V jmin ≤V j ≤V jmax (11)

[0139]

[0140] Among them, Equation (11) is the definition range of each top - level design indicator; Equation (12) is the relevant mathematical model of the j - th top - level design indicator and other top - level design indicators (the relevant mathematical model refers to the mathematical relationship model between indicators); Equation (13) is the relevant constraint model of the j - th top - level design indicator and other top - level design indicators (the relevant constraint model refers to the range constraint of one indicator on another indicator. For example: when indicator x1 is greater than a certain value, indicator x2 should be less than a certain value).

[0141] Step S43: Solve the optimization model using a genetic algorithm.

[0142] Select a genetic algorithm to solve the top-level design index scheme after continuous iterative optimization, and compare and evaluate it with the initially defined scheme. If it meets the design requirements, the reasonable values of the top-level design indexes of the high-speed train are obtained. Otherwise, continue the iterative optimization to solve a new top-level design index scheme until the design requirements are met, and the optimization definition of the indexes is completed.

[0143] The above embodiments are only used to illustrate the present invention and do not limit the technical solutions described in the present invention. Although this specification has described the present invention in detail with reference to the above respective embodiments, the present invention is not limited to the above specific embodiments. Therefore, any modification or equivalent replacement to the present invention; and all technical solutions and their improvements that do not depart from the spirit and scope of the invention are covered by the scope of the claims of the present invention.

Claims

1. A method for defining top-level design indicators of high-speed trains based on comprehensive evaluation optimization, characterized in that: It includes the following steps S1 - S4; Step S1: Determination of the weights of top - level design indicators; It includes determining the weights of each evaluation criterion and calculating the relative weights of top - level design indicators; Step S2: Initial definition of top - level design indicators; It includes establishing a benchmark of competing products and determining the key top - level design indicators of the new high - speed train; Step S3: Non - linear normalization processing of top - level design indicators; It includes constructing and solving a non - linear S - type differentiable function; Step S4: Construction and solution of the comprehensive competitiveness optimization model; It includes constructing and solving a comprehensive competitiveness calculation model for the whole high - speed train, constructing a comprehensive competitiveness optimization model for the high - speed train, and using a genetic algorithm to solve the optimization model.

2. The method for defining the top-level design indexes of a high-speed train based on comprehensive evaluation optimization as described in claim 1, wherein: Step S1 specifically includes the following steps S11 - S12: Step S11: Use the group comprehensive analytic hierarchy process to determine the weights of each evaluation criterion; Step S12: Calculate the relative weights of top - level design indicators.

3. The method for defining top-level design indicators of a high-speed train based on comprehensive evaluation optimization according to claim 2, wherein: Step S11 specifically includes steps S111 - S118; S111: Define the goal and criteria; The goal is: to determine the weights of high - speed train performance; The criteria are: transport capacity, safety, comfort, environmental friendliness; S112: Form a decision - making group; Select experts or decision - makers in multiple related fields to form a decision - making group; S113: Design judgment matrix; the judgment matrix A is an n*n square matrix, and each element a in the matrix ij represents the relative importance of criterion i relative to criterion j; each expert makes pairwise comparisons of the criteria according to the 1-9 scale method to obtain the elements in the matrix and constructs the judgment matrix; S114: Calculate the individual ranking vector; In the case of s experts and n evaluation objects, the individual ranking vector of the i - th expert for these n evaluation objects is: S115: Calculate the similarity relation matrix; According to the individual sorting vector, calculate the similarity relationship matrix R between experts, R = (r ij ) n×n ; Denote the decision similarity degree between expert i and expert j as r ij , 0 ≤ r ij ≤ 1, i, j ∈ {1, 2,..., s}, the closer r ij is to 0, it indicates that the decisions of expert i and expert j are less similar, and the closer r ij is to 1, it indicates that the decisions of expert i and expert j are more similar; The calculation formula of the similarity degree r ij is: where, w ik and w jk are the weights of the k-th criterion given by experts i and j respectively; the expression of the similarity relationship matrix R is as follows: S116: Fuzzy clustering analysis; It includes the following step1 - step3: step1: According to the judgment matrices of each expert, find the individual ranking vectors of the experts and conduct a consistency test. For the judgment matrices that do not pass the consistency test, feedback their results to each expert and readjust; Step 2: Calculate the transitive closure of the similarity matrix R R = (r ij ) n×n is the similarity relation matrix, then there exists a smallest natural number k, k ≤ n, such that the transitive closure For all natural numbers 1 greater than m, there is R 1 = R m ; At this time is the fuzzy equivalence matrix; reflects the comprehensive similarity degree relationship among experts. The transitive closure of the similarity matrix R is obtained by the successive squaring method Step 3: Select a threshold value λ, where 0 ≤ λ ≤ 1, and perform clustering analysis on the experts through When the value of λ is larger, it indicates that a higher similarity requirement is imposed on the experts. Only when the similarity degree between two experts reaches or exceeds this threshold will they be classified into the same category. On the contrary, when the value of λ is smaller, the similarity requirement for the experts classified into the same category is lower. Using the transitive closure Cluster the experts according to the selected threshold λ; Check the value of the element in If the element values corresponding to the i-th expert and the j-th expert in are greater than or equal to λ, then classify these two experts into the same category; S117: Determination of expert weights; Through fuzzy clustering analysis, divide the experts into different categories; Experts in the same category have the same weight, and the weights of experts in different categories are determined according to the number of experts in the category; Let the number of experts in the $i$-th category be $\beta$. i , then the weight calculation formula for the $i$-th category of experts is as shown in Equation (5): where b represents the number of categories of experts, represents the sum of the squares of the number of experts in all categories, and ε i represents the weight of the experts in the i-th category; the expert weights can be calculated; S118: Calculate the final weights; According to the expert weights and individual ranking vectors, calculate the final weights of each criterion; The individual ranking vector reflects the judgment of each expert on the weights of each criterion, and the expert weight reflects the relative importance of expert opinions in the group; Calculating the final weights includes the following step1 - step2: Step 1: For each criterion, multiply the weight value corresponding to this criterion in the individual ranking vector of each expert by the weight of this expert. Let there be n criteria and m experts. For the k-th criterion, calculate the product corresponding to the i-th expert is the weight value of the k-th criterion in the individual ranking vector of the i-th expert, ε i represents the weight of the expert in the i-th category; step2: Sum the products calculated by all experts for the same criterion to obtain the final weight of this criterion; As shown in Equation (6): where: W k is the final weight of the k-th criterion; each criterion is calculated to obtain the weights of all evaluation criteria.

4. The method for defining top-level design indicators of a high-speed train based on comprehensive evaluation optimization according to claim 3, characterized in that: Step S12 includes steps S121 - S122; S121: Establish a relation matrix; Establish the relationship between the top-level design indicators of high-speed trains and various evaluation criteria using the relationship degree matrix in quality function deployment. The rows of the relationship degree matrix represent different evaluation criteria, the columns represent specific top-level design indicators, and each element R in the matrix i,j represents the relationship degree between the i-th evaluation criterion and the j-th top-level design indicator. W j is the relative weight of the j-th top-level design indicator, which is calculated through the relationship degree matrix and the weights of the evaluation criteria; S122: Establish a formula for the relative weights of top - level design indicators; W j is the relative weight of the j-th top-level design indicator; W j The calculation formula of is as shown in Equation (7): Wherein, m represents the number of items of the evaluation criteria, n represents the number of items of the top-level design indicators, and R i,j is the relationship degree between the i-th evaluation criterion and the j-th top-level design indicator, and the values are 0, 1, 3, 5, 7, 9, and the corresponding correlation degrees are: irrelevant, weakly relevant, relatively weakly relevant, moderately relevant, relatively strongly relevant, and strongly relevant. ω i is the weight of the i-th evaluation criterion.

5. A method for defining top - level design indicators of high - speed trains optimized based on comprehensive evaluation, as claimed in claim 3, wherein: Step S2 specifically includes the following steps S21 - S22: Step S21: Establish a benchmark of competing products; Step S22: According to the competing products, determine the key top - level design indicators of the new high - speed train.

6. The method for defining top - level design indicators of high - speed trains optimized based on comprehensive evaluation according to claim 5, wherein: Step S3 specifically is to construct and solve a non - linear S - type differentiable function: Adopt a non - linear S - type differentiable function to map the original value data of high - speed train top - level design indicators to the interval (0,1). The non - linear S - type differentiable function is as shown in Equation (8): where, V j and are the top-level design index values before and after standardization respectively, V jmin and V jmax represent the minimum and maximum values of the j-th top-level design index respectively, S and T are the S-curve shape adjustment parameters, and k is the S-curve direction adjustment parameter.

7. The method for defining top-level design indicators of a high-speed train based on comprehensive evaluation optimization according to claim 6, characterized in that: Step S4 specifically includes the following steps S41 - S43: Step S41: Construct and solve a comprehensive competitiveness calculation model for the entire high - speed train; Step S42: Construct an optimization model for the comprehensive competitiveness of the high - speed train; Step S43: Use the genetic algorithm to solve the optimization model.

8. The method for defining top-level design indicators of a high-speed train based on comprehensive evaluation optimization as claimed in claim 7, wherein: Specifically, step S41 is: Construct a calculation model for comprehensive competitiveness as shown in Equation (9): In the formula, is the value of the j-th top-level design indicator after standardization, and W j is the relative weight of the j-th top-level design indicator, and n is the number of top-level design indicators.

9. A method for defining the top-level design indicators of a high-speed train based on comprehensive evaluation optimization, characterized in that: Specifically, step S42 is: Taking the comprehensive competitiveness of the high - speed train as the optimization goal, transform the definition problem of different top - level design indicators of the high - speed train into an optimization problem of the comprehensive competitiveness of the high - speed train, and construct the following optimization model: Subject to: V jmin ≤ V j ≤ V jmax (11) Among them, Equation (11) is the definition range of each top - level design indicator; Equation (12) is the related mathematical model of the j - th top - level design indicator and other top - level design indicators; Equation (13) is the related constraint model of the j - th top - level design indicator and other top - level design indicators.

10. A method for defining top-level design indicators of a high-speed train based on comprehensive evaluation optimization, characterized in that: Specifically, step S43 is: Select the genetic algorithm to solve the top - level design indicator scheme after continuous iterative optimization, and compare and evaluate it with the initial definition scheme. If it meets the design requirements, the reasonable values of the top - level design indicators of the high - speed train are obtained; otherwise, continue the iterative optimization to solve a new top - level design indicator scheme until the design requirements are met, and the optimization definition of the indicators is completed.

Images