Multi-factor four-factor near-orthogonal near-uniform experimental scheme design method and system

Abstract

The invention provides a multi-factor four-factor near-orthogonal near-uniform experimental scheme design method and system, and relates to the technical field of experimental design. According to the method, the advantages of an orthogonal experimental scheme design method and a near-orthogonal Latin hypercube design method are combined, the design process meeting near-orthogonality and near-uniformity experimental data is simplified, and through the pre-constructed basic table of the multi-factor four-factor discrete variable near-orthogonality near-uniformity design meeting the near-orthogonality near-uniformity requirement, the multi-factor four-factor discrete variable near-orthogonality near-uniformity requirement is obtained. When the enumeration times of the experiment exceed 512 times, a typical general experiment scheme design scheme which only needs 512 times and accords with near orthogonality and near uniformity is given, and all characteristics of multi-factor 4-factor enumeration data are covered. The experiment scheme design work is greatly optimized, and the experiment scheme design efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of experimental design, and in particular to a method and system for designing a nearly orthogonal and nearly uniform experimental scheme of multiple factors and four factors. Background Art

[0002] In experiments in fields such as aerospace, medicine, and biology, there are often a large number of influencing factors and intricate interactions between the factors. It is not possible to conduct experimental verification for each situation in terms of time, manpower, and material resources, which requires sampling and assumptions of the data.

[0003] However, the current common experimental designs, such as completely randomized design, factorial experimental design, orthogonal experimental design, and Latin hypercube experimental design, all require a lot of calculations and verifications during the application process, which brings inconvenience to the application of experimental design and low efficiency of experimental design. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for designing a multi-factor 4-factor nearly orthogonal and nearly uniform experimental scheme, which solves the technical problem of low efficiency of the existing experimental scheme design method.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] In a first aspect, the present invention provides a method for designing a near-orthogonal and near-uniform experimental scheme with multiple factors and four factors, comprising the following steps:

[0009] S1. Obtain experimental variable factor data and variable change factors of the product to be tested;

[0010] S2, determine whether the value range of each variable change factor can be converted into 4 discrete variable values. If so, enter S3, otherwise, end;

[0011] S3, determine whether the number of experimental variable factors is less than the number of columns M of the basic table of the 4-factor discrete variable nearly orthogonal and nearly uniform design. If so, directly transfer to S4; otherwise, classify and merge the experimental variable factors until the number of experimental variable factors is less than or equal to M, and return to S2; M is a positive integer, M≤20;

[0012] S4, determine whether the number m of experimental variable factors is less than 5. If so, arrange and combine the m groups of 4-factor variables to obtain 512 groups of all data combinations that meet the nearly orthogonal and nearly uniform calculation target experimental data sets, and end; otherwise, select the same number of columns from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design in sequence, and execute the next step;

[0013] S5, determine the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design;

[0014] S6, determining the mapping relationship between each variable value and the numerical value in each column;

[0015] S7. According to the mapping relationship, all experimental variable values ​​are used to replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design, and the experimental schemes of 512 groups of experimental data sets that meet the target experimental data set calculation of near-orthogonality and near-uniformity for the experimental products are output;

[0016] Among them, the basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design is a table with M columns and 512 rows. The basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design has near-orthogonality and near-uniformity.

[0017] Preferably, the process of constructing the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design includes:

[0018] a. Given a 512*1 column vector;

[0019] b. Add a 512*1 column vector to form a 512*2 matrix, calculate the orthogonality of the two vectors, and if the result is less than 0.05, add it;

[0020] c. Add a 512*1 column vector to form a 512*3 matrix, and calculate the orthogonality of the two vectors. If the results of the orthogonality of the three two vectors are all less than 0.05, keep the matrix. Otherwise, start again to find another 512*1 vector and calculate it; and so on, find and establish a 512*M matrix;

[0021] d. Use the two-dimensional plane distribution diagram to judge the uniformity of the design content. If it meets the uniformity, the set of 512*M matrices is retained, otherwise return to step 1 and start the construction again;

[0022] Each column vector includes 4 different values, and each different value has the same frequency of occurrence.

[0023] Preferably, the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable near-orthogonal near-uniform design includes:

[0024] Suppose there is a set A with m variable factors, where A = {1, ..., m}, and m columns are selected from the basic table of multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design to form a set B, where set B = {1, ..., m}, and the relationship between sets A and B is f: A → B, f(a1)≠f(a2), and b∈B∧a∈A→f(a)=b; establish variable mapping relationship { i ,b j >}, where a i ∈A,b j ∈B,i={1,…,m},j={1,…,m}.

[0025] Preferably, determining the mapping relationship between each variable value and the numerical value in each column includes:

[0026] a i ={a i1 ,a i2 ,a i3 ,a i4}, b j = {b j1 ,b j2 ,b j3 ,b j4}, i = {1, ..., m}, j = {1, ..., m},

[0027] Determine the mapping relationship between each variable value and the corresponding column of the 4-factor basic table, { i1 ,b j1 >} or { i1 ,b j2 >} or { i1 ,b j3 >} or { i1 ,b j4 >}.

[0028] In a second aspect, the present invention provides a multi-factor 4-factor nearly orthogonal nearly uniform experimental design system, comprising:

[0029] The product data acquisition module to be tested is used to execute S1 and obtain the experimental variable factor data and variable change factor of the product to be tested;

[0030] The first judgment module is used to execute S2 and judge whether the value range of each variable change factor can be converted into four discrete variable value situations. If so, enter S3, otherwise, end;

[0031] ​​​​​The second judgment module is used to execute S3 and judge whether the number of experimental variable factors is less than the number of columns M of the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design. If so, directly transfer to S4; otherwise, classify and merge the experimental variable factors until the number of experimental variable factors is less than or equal to M, and return to S2; M is a positive integer, M≤20;

[0032] The third judgment module is used to execute S4 and judge whether the number m of experimental variable factors is less than 5. If so, the m groups of 4-factor variables are arranged and combined to obtain 512 groups of all data combinations that meet the nearly orthogonal and nearly uniform calculation target experimental data sets, and then end; otherwise, the same number of columns are sequentially selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design, and the next step is executed;

[0033] A first mapping relationship determination module is used to execute S5, determine the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design;

[0034] A second mapping relationship determination module, used to execute S7, determine the mapping relationship between each variable value and the numerical value in each column;

[0035] The experimental scheme output module is used to execute S7, replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design with all experimental variable values ​​according to the mapping relationship, and output 512 groups of experimental schemes of the experimental products that meet the target experimental data sets of near-orthogonality and near-uniformity calculation;

[0036] The basic table of the near-orthogonal and near-uniform design of multi-factor 4-factor discrete variables is a table with M columns and 512 rows. The basic table of the near-orthogonal and near-uniform design of multi-factor 4-factor discrete variables has near-orthogonality and near-uniformity.

[0037] Preferably, the process of constructing the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design includes:

[0038] a. Given a 512*1 column vector;

[0039] b. Add a 512*1 column vector to form a 512*2 matrix, calculate the orthogonality of the two vectors, and if the result is less than 0.05, add it;

[0040] c. Add a 512*1 column vector to form a 512*3 matrix, and calculate the orthogonality of the two vectors. If the results of the orthogonality of the three two vectors are all less than 0.05, keep the matrix. Otherwise, start again to find another 512*1 vector and calculate it; and so on, find and establish a 512*M matrix;

[0041] d. Use the two-dimensional plane distribution diagram to judge the uniformity of the design content. If it meets the uniformity, the set of 512*M matrices is retained, otherwise return to step 1 and start the construction again;

[0042] Each column vector includes 4 different values, and each different value has the same frequency of occurrence.

[0043] Preferably, the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable near-orthogonal near-uniform design includes:

[0044] Suppose there is a set A with m variable factors, where A = {1, ..., m}, and m columns are selected from the basic table of multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design to form a set B, where set B = {1, ..., m}, and the relationship between sets A and B is f: A → B, f(a1)≠f(a2), and b∈B∧a∈A→f(a)=b; establish variable mapping relationship { i ,b j >}, where a i ∈A,b j ∈B,i={1,…,m},j={1,…,m}.

[0045] Preferably, determining the mapping relationship between each variable value and the numerical value in each column includes:

[0046] a i ={a i1 ,a i2 ,a i3 ,a i4}, b j = {b j1 ,b j2 ,b j3 ,b j4}, i = {1, ..., m}, j = {1, ..., m},

[0047] Determine the mapping relationship between each variable value and the corresponding column of the 4-factor basic table, { i1 ,b j1 >} or { i1 ,b j2 >} or { i1 ,b j3 >} or { i1 ,b j4 >}.

[0048] ​​​​​In the third aspect, the present invention provides a computer-readable storage medium, which stores a computer program for the design of nearly orthogonal and nearly uniform experimental schemes for multiple factors and four factors, wherein the computer program enables the computer to execute the nearly orthogonal and nearly uniform experimental scheme design method for multiple factors and four factors as described above.

[0049] In a fourth aspect, the present invention provides an electronic device, comprising:

[0050] One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include a method for executing the multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method as described above.

[0051] (III) Beneficial effects

[0052] The present invention provides a method and system for designing a near-orthogonal and near-uniform experimental scheme with multiple factors and four factors. Compared with the prior art, it has the following beneficial effects:

[0053] The present invention provides a multi-factor 4-factor near-orthogonal near-uniform experimental scheme design method and system, which combines the advantages of the orthogonal experimental scheme design method and the near-orthogonal Latin hypercube design method, and simplifies the design process of experimental data that meet the near-orthogonality and near-uniformity requirements, through the pre-constructed near-orthogonal near-uniform design basic table of multi-factor 4-factor discrete variable near-orthogonal near-uniformity, when the number of enumerations of the experiment exceeds 512 times, a typical general experimental scheme design scheme that only needs 512 times and meets the near-orthogonal near-uniformity is given, covering all the characteristics of the multi-factor 4-factor enumeration data. The experimental scheme design work is greatly optimized and the efficiency of the experimental scheme design is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0055] Figure 1 A block diagram of a multi-factor 4-factor nearly orthogonal and nearly uniform experimental scheme design method in an embodiment of the present invention;

[0056] Figure 2 A two-dimensional plane distribution effect diagram is designed for the uniformity of 20 groups of discrete variables with a variation range of 4 ({1,2,3,4}). DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] The embodiment of the present application solves the technical problem of low efficiency of existing experimental design methods by providing a multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method and system, optimizes the experimental design work, and improves the application efficiency of the experimental design.

[0059] The technical solution in the embodiment of the present application is to solve the above technical problems, and the overall idea is as follows:

[0060] 1. Completely randomized design

[0061] Completely randomized design is also called single-factor design, which means randomly assigning subjects to treatment groups for experimental observation or randomly sampling from different populations for comparative observation. This method is suitable for comparing two or more samples. The sample sizes between groups can be equal or unequal. When the samples are equal, the statistical analysis efficiency is higher.

[0062] In general, completely randomized design has obvious limitations: first, it requires that non-experimental factors have little effect on the effect index; second, it balances the influence of non-research factors through random grouping. The advantages are: experimental design and statistical analysis are relatively simple. However, the disadvantages are obvious: only one factor can be studied at a time, and the required sample size is relatively large, so the experimental design efficiency is not high, and the experimental design results are not necessarily completely affected by random factors, the parameter estimation of the model may be biased, and the experimental error is large.

[0063] 2. Factorial Experimental Design

[0064] Factorial design is a method of conducting experiments by cross-grouping the levels of two or more factors. It can not only test whether there are differences between different levels of each factor, but also check whether there are interactions between two or more factors. This design is combined by cross-grouping different levels of each factor, so the total number of experimental groups is equal to the product of the levels of each factor. Therefore, when conducting factorial design, the number of factors and the number of levels of the factors should not be too many. This method can accurately estimate the size of the main effect of each experimental factor, and can also estimate the size of the interaction between factors. It is an efficient experimental design method. However, this experimental design method requires too many experiments, consumes more manpower, material resources and time, and is not efficient when there are many factors and levels to be studied.

[0065] 3. Orthogonal experimental design

[0066] Orthogonal experimental design is an efficient and fast design method for studying multiple factors and multiple levels. Through a set of standardized orthogonal tables and interaction tables, the research factors are reasonably arranged, and the results are statistically analyzed to obtain the corresponding conclusions. Although this experimental design method eliminates the correlation and mutual influence between factors, when the overall sample is large, the experiment designed using this method cannot well represent the entire experimental space, that is, the experiment is not very representative; and the designed scheme can only be limited to a determined level, rather than the optimal scheme within a certain experimental range.

[0067] 4. Latin Hypercube Experimental Design

[0068] On the basis of random design, the idea of ​​"quota sampling" is added, and the Latin hypercube sampling method is proposed. In Latin hypercube sampling, random variables are considered to be generated by known distribution functions, and it is also considered that "the sampled input variable sample can represent the entire sample space". The variable sample generated by Latin hypercube sampling is called a Latin hypercube matrix, and the matrix dimension is n*k, where n represents the number of simulation runs and k represents the number of influencing factors. In order to improve the usability of Latin hypercube experimental design for regression analysis, a construction method of orthogonal Latin hypercube matrix is ​​proposed, and the orthogonality is that the correlation coefficient between any two different columns of the matrix is ​​0. Both Latin hypercube experimental design and orthogonal Latin hypercube experimental design consider reducing or even eliminating the correlation between variables, but they do not consider whether the sample space of the experiment is representative, that is, whether it can be "space filled" into the entire sample population.

[0069] 5. Uniform design

[0070] Uniform design is a statistical experimental design method that complements many other experimental methods. The degree of uniformity can be judged by the two-dimensional plane distribution effect of the experimental data. For uniform design, when the sample size and experimental operation plan are small, its efficiency and accuracy are very high. The disadvantages are: first, it does not reduce or eliminate the correlation between variables, and experiments may be repeated; second, when the variables are large, the experimental plan obtained by using uniform design is large, and the experimental efficiency is not high.

[0071] 6. Nearly orthogonal Latin hypercube experimental design

[0072] Whether it is orthogonal Latin hypercube experimental design or uniform design, there are still some deficiencies in orthogonality and uniformity. The nearly orthogonal Latin hypercube experimental design method provides a sampling method for a large number of input variables or influencing factors. Not only does it meet the orthogonality to a certain extent, but it also has good uniformity, that is, the "filling space" property described by it. Compared with other experimental design schemes, this method requires relatively fewer experimental schemes.

[0073] From the above, it can be seen that although the orthogonal experimental design method eliminates the correlation and mutual influence between factors, when the overall sample is large, the experiment designed using this method cannot represent the entire experimental space well, that is, the experiment is not very representative; and the designed scheme can only be limited to a certain level, rather than the optimal scheme within a certain experimental range. The orthogonal experimental design needs to select an established orthogonal experimental design table based on the experimental scheme, the number of levels and the number of factors to be verified. Since the orthogonal table is known, it saves the workload of data table design to achieve orthogonal characteristics. Although the orthogonal design reduces or even eliminates the correlation between variables, the experimental sample space is not representative.

[0074] The specific construction method of Latin hypercube experimental design is relatively complex, and its method only reduces or even eliminates the correlation between variables, but the experimental sample space is not representative.

[0075] Nearly orthogonal Latin hypercube experimental design solves the shortcomings of orthogonal design, Latin hypercube design and uniform design in orthogonality and uniformity, but the calculation of its design method is relatively complicated.

[0076] From the above description, it can be seen that the existing experimental scheme designs all require a lot of calculations and verifications during the application process, which brings inconvenience to the application of experimental scheme design.

[0077] To solve the above problems, the embodiment of the present invention proposes a multi-factor 4-factor near-orthogonal and near-uniform experimental scheme design method and system. The method is mainly aimed at the experimental scheme design that meets the requirements of near-orthogonality and near-uniformity when the number of variables is less than or equal to 20, the variation range of each variable is 4 results, the number of experiments does not exceed 512 times, and the number of experiments is less than orthogonal.

[0078] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0079] The embodiment of the present invention provides a multi-factor 4-factor nearly orthogonal nearly uniform experimental scheme design method, such as Figure 1 As shown, the following steps are included:

[0080] S1. Obtain experimental variable factor data and variable change factors of the product to be tested;

[0081] S2, determine whether the value range of each variable change factor can be converted into 4 discrete variable values. If so, enter S3, otherwise, end;

[0082] S3, determine whether the number of experimental variable factors is less than the number of columns M of the basic table of the 4-factor discrete variable nearly orthogonal and nearly uniform design. If so, directly transfer to S4; otherwise, classify and merge the experimental variable factors until the number of experimental variable factors is less than or equal to M, and return to S2; M is a positive integer, M≤20;

[0083] S4, determine whether the number m of experimental variable factors is less than 5. If so, arrange and combine the m groups of 4-factor variables to obtain 512 groups of all data combinations that meet the nearly orthogonal and nearly uniform calculation target experimental data sets, and end; otherwise, select the same number of columns from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design in sequence, and execute the next step;

[0084] S5, determine the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design;

[0085] S6, determining the mapping relationship between each variable value and the numerical value in each column;

[0086] S7. According to the mapping relationship, all experimental variable values ​​are used to replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design, and the experimental schemes of 512 groups of experimental data sets that meet the target experimental data set calculation of near-orthogonality and near-uniformity for the experimental products are output;

[0087] Among them, the basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design is a table with M columns and 512 rows. The basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design has near-orthogonality and near-uniformity.

[0088] Among them, the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design is a table with M columns and 512 rows. The basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design has good nearly orthogonality, that is, it ensures the irrelevance of each design scheme, and good nearly uniformity, that is, the representativeness of each scheme in the experimental design space. Therefore, the experimental data generated by this table has good coverage of the experimental input variables.

[0089] The embodiment of the present invention combines the advantages of the orthogonal experimental scheme design method and the nearly orthogonal Latin hypercube design method, and simplifies the design process of experimental data that meets the nearly orthogonal and nearly uniform requirements. Through the pre-constructed basic table of the nearly orthogonal and nearly uniform design of multi-factor 4-factor discrete variables that meets the nearly orthogonal and nearly uniform requirements, when the number of enumerations of the experiment exceeds 512 times (512<4^5, that is, when the number of variables exceeds 5, the number of experimental design schemes enumerated will exceed 512. This method can control the experimental design schemes to 512, reduce the design schemes, and ensure the experimental coverage), a typical general experimental scheme design scheme that only needs 512 times and meets the nearly orthogonal and nearly uniform requirements is given, covering all the characteristics of the multi-factor 4-factor enumeration data. The experimental scheme design work is greatly optimized and the efficiency of the experimental scheme design is improved.

[0090] It should be noted that the experimental scheme obtained can be used for simulation experiments or engineering experiments.

[0091] The construction method of the basic table of multi-factor 4-factor discrete variable near-orthogonal and near-uniform design is as follows:

[0092] Step 1: Given a 512*1 column vector, the column vector includes 4 different values, and the frequencies of occurrence of the 4 different values ​​are the same;

[0093] Step 2: Add a 512*1 column vector to form a 512*2 matrix. Calculate the orthogonality of the two column vectors. If the result is less than 0.05, add it.

[0094] Step 3: Add a 512*1 column vector to form a 512*3 matrix, and calculate the orthogonality of two column vectors. Here, we need to calculate the orthogonality of three two column vectors. If the results are all less than 0.05, keep the matrix. Otherwise, start again to find another 512*1 column vector and calculate it. And so on, find and build a 512*20 matrix;

[0095] Step 4: Use the two-dimensional plane distribution graph to determine the uniformity of the design content. If it is uniform, keep the set of 512*20 matrices, otherwise start again from step 1. The uniformity judgment criteria are as follows: if the two-dimensional plane distribution graph between any two variables in the two-dimensional plane distribution graph is not approximately an "X" shape, that is, there is no area without sampling points, then it is uniform.

[0096] It should be noted that the basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design in the embodiment of the present invention is shown in Table 1:

[0097] Table 1 Basic table of multi-factor 4-factor discrete variable near-orthogonal near-uniform design

[0098]

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] The methods for proving the near-orthogonality of Table 1 include:

[0112] The near-orthogonality of this near-orthogonal and near-uniform base table is verified by calculating the maximum absolute value of the correlation coefficient. The maximum absolute value of the correlation coefficient is denoted as ρ 相关系数绝对值的最大值 , the absolute value of the correlation coefficient is recorded as |ρ ij |, for any two different column vectors X i and X j , and i≠j, the maximum absolute value of the correlation coefficient is:

[0113]

[0114] Among them, ρ 相关系数绝对值的最大值 The smaller the value of indicates the better the orthogonality, and when the value is less than or equal to 0.05, it is considered to be a nearly orthogonal design. i and X j ρ 相关系数绝对值的最大值 = 0, it means orthogonal. l represents the column vector column vector X i and X j The lth row of .

[0115] According to the maximum value calculation method of the absolute value of the correlation coefficient, the maximum value set of the absolute value of the correlation coefficient of all columns in the basic table is obtained, that is, ρ = {0.031504612, 0.029947452, 0.033128724, 0.035631082, 0.030559936, 0.033856035, 0.034997662, 0.033425133, 0.03404319, 0.028886638, 0.029330456, 0.03259797, 0.034782293, 0.028804324, 0.030021852, 0.031375466, 0.033704908, 0.034519632, 0.03513909, 0.029795316}. Where i represents the maximum absolute value of the correlation coefficient between the i-th column and other columns in the base table, where 1≤i≤20. Therefore, it can be obtained that the columns in this matrix table have good near-orthogonality.

[0116] The uniformity proof includes: through the uniformity design of 20 discrete variables with a range of 4 ({1,2,3,4}), the dispersion and uniformity of the experimental design data can be intuitively demonstrated. Figure 2 shown.

[0117] Embodiment 1:

[0118] This embodiment designs an experimental scheme for a test plan data with 20 test points, and each test point has four states of "valid left boundary value", "valid right boundary value", "valid value" and "invalid value". The experimental scheme is designed to ensure the comprehensive coverage of the combined test of each test point. The specific design process is as follows:

[0119] S101, obtaining test plan data of 20 test points of the product to be tested as experimental variables, and using four states of "valid left boundary value", "valid right boundary value", "valid value" and "invalid value" of each test point as variable change factors;

[0120] S102, determining whether the value range of each variable change factor can be converted into four discrete variable value situations. In this embodiment, it is obviously convertible.

[0121] S103, determine the relationship between the number of experimental variable factors and the number of columns M of the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design. In the embodiment of the present invention, the value of M is 20, that is, the entire basic table is selected.

[0122] S104, determine the mapping relationship between the variables (set A) and the columns in the selected basic table (set B) { i ,b j >}, where a i ∈A,b j ∈B,i={1,…,20},j={1,…,20};

[0123] S105, determine the mapping relationship between each variable value and the value in each column { i1 ,b j1 >} or { i1 ,b j2 >} or { i1 ,b j3 >} or { i1 ,b j4 >}, where a i ={a i1 ,a i2 ,a i3 ,a i4} = {"valid left boundary value", "valid right boundary value", "valid value", "invalid value"}, b j = {b j1 ,b j2 ,b j3 ,b j4}={1, 2, 3, 4}, i={1,...,20}, j={1,...,20};

[0124] S106, replace 1, 2, 3, 4 in the basic table with "valid left boundary value", "valid right boundary value", "valid value" and "invalid value" respectively, and obtain 512 test schemes that meet the requirements of near-orthogonality and near-uniformity and have comprehensive coverage of 20 test points. The two-dimensional distribution effect of the design data in the experimental scheme is as follows: Figure 2 Shown

[0125] Embodiment 2:

[0126] In this embodiment, an experimental scheme is designed for a product to be tested with 20 experimental variables, each of which has a variable change factor of 4 discrete variable values. The design process is as follows:

[0127] S201, obtaining 20 experimental variable factor data and value ranges of variable change factors of the product to be tested;

[0128] S202, determining whether the value range of each variable change factor can be converted into four discrete variable value situations;

[0129] ​​​​​S203, determine the relationship between the number of experimental variable factors and the number of columns M of the basic table of the multi-factor 4-factor discrete variable near-orthogonal near-uniform design. In the embodiment of the present invention, the value of M is 20; because the number of experimental variable factors of the product to be tested is 20, 20 columns are sequentially selected from the basic table of the multi-factor 4-factor discrete variable near-orthogonal near-uniform design;

[0130] S204, determine the mapping relationship between the experimental variable factors (set A) and the columns in the selected basic table (set B) { i ,b j >}, where a i ∈A,b j ∈B,i={1,…,20},j={1,…,20};

[0131] S205, determine the mapping relationship between each variable value and the value in each column { i1 ,b j1 >} or { i1 ,b j2 >} or { i1 ,b j3 >} or { i1 ,b j4 >}, where a i ={a i1 ,a i2 ,a i3 ,a i4}, b j = {b j1 ,b j2 ,b j3 ,b j4}, i={1,…,20}, j={1,…,20};

[0132] S206. According to the mapping relationship, all experimental variable values ​​are used to replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design, and the experimental scheme of 512 groups of experimental products that meet the target experimental data set of near-orthogonality and near-uniformity calculation is output. The two-dimensional distribution effect of the design data in the experimental scheme is as follows: Figure 2 shown.

[0133] The embodiment of the present invention further provides a multi-factor 4-factor nearly orthogonal nearly uniform experimental scheme design system, the system comprising:

[0134] The product data acquisition module to be tested is used to execute S1 and obtain the experimental variable factor data and variable change factor of the product to be tested;

[0135] ​​​​​The first judgment module is used to execute S2 and judge whether the value range of each variable change factor can be converted into four discrete variable value situations. If so, enter S3, otherwise, end;

[0136] The second judgment module is used to execute S3 and judge whether the number of experimental variable factors is less than the number of columns M of the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design. If so, directly transfer to S4; otherwise, classify and merge the experimental variable factors until the number of experimental variable factors is less than or equal to M, and return to S2; M is a positive integer, M≤20;

[0137] The third judgment module is used to execute S4 and judge whether the number m of experimental variable factors is less than 5. If so, the m groups of 4-factor variables are arranged and combined to obtain 512 groups of all data combinations that meet the nearly orthogonal and nearly uniform calculation target experimental data sets, and then end; otherwise, the same number of columns are sequentially selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design, and the next step is executed;

[0138] A first mapping relationship determination module is used to execute S5, determine the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design;

[0139] A second mapping relationship determination module, used to execute S7, determine the mapping relationship between each variable value and the numerical value in each column;

[0140] The experimental scheme output module is used to execute S7, replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design with all experimental variable values ​​according to the mapping relationship, and output 512 groups of experimental schemes for the experimental products that meet the near-orthogonality and near-uniformity calculation target experimental data sets.

[0141] The basic table of the near-orthogonal and near-uniform design of multi-factor 4-factor discrete variables is a table with M columns and 512 rows. The basic table of the near-orthogonal and near-uniform design of multi-factor 4-factor discrete variables has near-orthogonality and near-uniformity.

[0142] It can be understood that the near-orthogonal and near-uniform experimental scheme design system for multiple factors and four factors provided in an embodiment of the present invention corresponds to the above-mentioned near-orthogonal and near-uniform experimental scheme design method for multiple factors and four factors. The explanations, examples, beneficial effects and other parts of the relevant contents can refer to the corresponding contents in the near-orthogonal and near-uniform experimental scheme design method for multiple factors and four factors, and will not be repeated here.

[0143] An embodiment of the present invention also provides a computer-readable storage medium, which stores a computer program for designing a nearly orthogonal and nearly uniform experimental scheme for multiple factors and four factors, wherein the computer program enables a computer to execute the nearly orthogonal and nearly uniform experimental scheme design method for multiple factors and four factors as described above.

[0144] An embodiment of the present invention further provides an electronic device, including:

[0145] one or more processors;

[0146] Memory; and

[0147] One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including a method for executing a multi-factor 4-factor nearly orthogonal and nearly uniform experimental design as described above.

[0148] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0149] The embodiment of the present invention combines the advantages of the orthogonal experimental scheme design method and the nearly orthogonal Latin hypercube design method, and simplifies the design process of experimental data that meets the nearly orthogonal and nearly uniform requirements. Through the pre-constructed basic table of the nearly orthogonal and nearly uniform design of multi-factor 4-factor discrete variables that meets the nearly orthogonal and nearly uniform requirements, when the number of enumerations of the experiment exceeds 512 times, a typical general experimental scheme design scheme that only requires 512 times and meets the nearly orthogonal and nearly uniform requirements is given, covering all the characteristics of the multi-factor 4-factor enumeration data. The experimental scheme design work is greatly optimized and the efficiency of the experimental scheme design is improved.

[0150] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0151] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method, characterized in that: The following steps are involved: S1. Obtain experimental variable factor data and variable change factors of the product to be tested; S2, determine whether the value range of each variable change factor can be converted into 4 discrete variable values. If so, enter S3, otherwise, end; S3, determine whether the number of experimental variable factors is less than the number of columns M of the basic table of the 4-factor discrete variable nearly orthogonal and nearly uniform design. If so, directly transfer to S4; otherwise, classify and merge the experimental variable factors until the number of experimental variable factors is less than or equal to M, and return to S2; M is a positive integer, M≤20; S4, determine whether the number m of experimental variable factors is less than 5. If so, arrange and combine the m groups of 4-factor variables to obtain 512 groups of all data combinations that meet the nearly orthogonal and nearly uniform calculation target experimental data sets, and end; otherwise, select the same number of columns from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design in sequence, and execute the next step; S5, determine the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design; S6, determining the mapping relationship between each variable value and the numerical value in each column; S7. According to the mapping relationship, all experimental variable values ​​are used to replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design, and the experimental schemes of 512 groups of experimental data sets that meet the target experimental data set calculation of near-orthogonality and near-uniformity for the experimental products are output; Among them, the basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design is a table with M columns and 512 rows. The basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design has near-orthogonality and near-uniformity.

2. The multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method according to claim 1, characterized in that: The construction process of the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design includes: a. Given a 512*1 column vector; b. Add a 512*1 column vector to form a 512*2 matrix, calculate the orthogonality of the two vectors, and if the result is less than 0.05, add it; c. Add a 512*1 column vector to form a 512*3 matrix, and calculate the orthogonality of the two vectors. If the results of the orthogonality of the three two vectors are all less than 0.05, keep the matrix. Otherwise, start again to find another 512*1 vector and calculate it; and so on, find and establish a 512*M matrix; d. Use the two-dimensional plane distribution diagram to judge the uniformity of the design content. If it meets the uniformity, the set of 512*M matrices is retained, otherwise return to step 1 and start the construction again; Each column vector includes 4 different values, and each different value has the same frequency of occurrence.

3. The multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method according to claim 1 or 2, characterized in that: The mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design includes: Suppose there is a set A with m variable factors, where A = {1, ..., m}, and m columns are selected from the basic table of multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design to form a set B, where set B = {1, ..., m}, and the relationship between sets A and B is f: A → B, f(a1)≠f(a2), and b∈B∧a∈A→f(a)=b; establish variable mapping relationship { i ,b j >}, where a i ∈A,b j ∈B,i={1,…,m},j={1,…,m}.​ 4. The multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method as claimed in claim 3, characterized in that: Determining the mapping relationship between each variable value and the numerical value in each column includes: a i = {a i1 , a i2 , a i3 , a i4}, b j = {b j1 , b j2 , b j3 , b j4}, i = {1, …, m}, j = {1, …, m}, sure Determine the mapping relationship between each variable value and the corresponding column of the 4-factor basic table, { i1 ,b j1 >} or { i1 ,b j2 >} or { i1 ,b j3 >} or { i1 ,b j4 >}.​​​​ 5. A multi-factor 4-factor nearly orthogonal and nearly uniform experimental design system, characterized by: include: The product data acquisition module to be tested is used to execute S1 and obtain the experimental variable factor data and variable change factor of the product to be tested; The first judgment module is used to execute S2 and judge whether the value range of each variable change factor can be converted into four discrete variable value situations. If so, enter S3, otherwise, end; The second judgment module is used to execute S3 and judge whether the number of experimental variable factors is less than the number of columns M of the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design. If so, directly transfer to S4; otherwise, classify and merge the experimental variable factors until the number of experimental variable factors is less than or equal to M, and return to S2; M is a positive integer, M≤20; The third judgment module is used to execute S4 and judge whether the number m of experimental variable factors is less than 5. If so, the m groups of 4-factor variables are arranged and combined to obtain 512 groups of all data combinations that meet the nearly orthogonal and nearly uniform calculation target experimental data sets, and then end; otherwise, the same number of columns are sequentially selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design, and the next step is executed; A first mapping relationship determination module is used to execute S5, determine the mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable nearly orthogonal nearly uniform design; A second mapping relationship determination module, used to execute S7, determine the mapping relationship between each variable value and the numerical value in each column; The experimental scheme output module is used to execute S7, replace the values ​​in the basic table of the selected multi-factor 4-factor discrete variable near-orthogonal and near-uniform design with all experimental variable values ​​according to the mapping relationship, and output 512 groups of experimental schemes of the experimental products that meet the target experimental data sets of near-orthogonality and near-uniformity calculation; The basic table of the near-orthogonal and near-uniform design of multi-factor 4-factor discrete variables is a table with M columns and 512 rows. The basic table of the near-orthogonal and near-uniform design of multi-factor 4-factor discrete variables has near-orthogonality and near-uniformity.

6. The multi-factor 4-factor nearly orthogonal and nearly uniform experimental design system as claimed in claim 5, characterized in that: The construction process of the basic table of the multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design includes: a. Given a 512*1 column vector; b. Add a 512*1 column vector to form a 512*2 matrix, calculate the orthogonality of the two vectors, and if the result is less than 0.05, add it; c. Add a 512*1 column vector to form a 512*3 matrix, and calculate the orthogonality of the two vectors. If the results of the orthogonality of the three two vectors are all less than 0.05, keep the matrix. Otherwise, start again to find another 512*1 vector and calculate it; and so on, find and establish a 512*M matrix; d. Use the two-dimensional plane distribution diagram to judge the uniformity of the design content. If it meets the uniformity, the set of 512*M matrices is retained, otherwise return to step 1 and start the construction again; Each column vector includes 4 different values, and each different value has the same frequency of occurrence.

7. The multi-factor 4-factor nearly orthogonal and nearly uniform experimental design system according to claim 5 or 6, characterized in that: The mapping relationship between the experimental variable factors and the columns selected from the basic table of the multi-factor 4-factor discrete variable near-orthogonal and near-uniform design includes: Suppose there is a set A with m variable factors, where A = {1, ..., m}, and m columns are selected from the basic table of multi-factor 4-factor discrete variable nearly orthogonal and nearly uniform design to form a set B, where set B = {1, ..., m}, and the relationship between sets A and B is f: A → B, f(a1)≠f(a2), and b∈B∧a∈A→f(a)=b; establish variable mapping relationship { i ,b j >}, where a i ∈A,b j ∈B,i={1,…,m},j={1,…,m}.​ 8. The multi-factor 4-factor nearly orthogonal and nearly uniform experimental design system as claimed in claim 7, characterized in that: Determining the mapping relationship between each variable value and the numerical value in each column includes: a i = {a i1 , a i2 , a i3 , a i4}, b j = {b j1 , b j2 , b j3 , b j4}, i = {1, …, m}, j = {1, …, m}, sure Determine the mapping relationship between each variable value and the corresponding column of the 4-factor basic table, { i1 ,b j1 >} or { i1 ,b j2 >} or { i1 ,b j3 >} or { i1 ,b j4 >}.​​​​ 9. A computer-readable storage medium, characterized in that: It stores a computer program for the nearly orthogonal and nearly uniform experimental scheme design of multiple factors and four factors, wherein the computer program enables the computer to execute the nearly orthogonal and nearly uniform experimental scheme design method of multiple factors and four factors as described in any one of claims 1 to 4.

10. An electronic device, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include a method for executing a multi-factor 4-factor nearly orthogonal and nearly uniform experimental design method as described in any one of claims 1 to 4.

Images