A method and system for intelligently setting parameters of a test system

Abstract

The present invention discloses a method and system for intelligently setting parameters for a test system, and belongs to the field of intelligent test design. The method comprises: using a theoretical mapping model of test system parameters to construct an initial test record data set; using the initial test record data set to construct an actual mapping model of test system parameters, expanding the data set during the test, and using the expanded data set to iteratively correct the above-mentioned regression model to obtain the optimal actual mapping model of system parameters; substituting the expected output parameters of the test system into the optimal actual mapping model of system parameters to obtain the input parameters required by the test system. The present invention records the input and output data of the test system multiple times, and then uses these recorded data as a data set to construct a mapping model. The obtained model is closer to the actual mapping model of the test system. When applied, it can make the output of the test system closer to the expected value, and effectively improves the adjustment efficiency of the input parameters to be set during the test.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent test design, and more specifically, relates to a method and system for intelligently setting parameters of a test system. Background Art

[0002] When testing an application hardware or software system, the input parameters need to be set in order to make the system output reach the preset expected value. However, due to the limited understanding of the characteristics of the system itself or the influence of environmental factors, the output parameters may be different from the expected value, or even have a large deviation. This requires the operator to manually try different input parameters to obtain the expected output within a certain error range. For a test system with a high operating cost, how to efficiently and quickly find the required input parameter setting value to make the output as close to the expected value as possible has important application value.

[0003] For parameter setting problems in the test system, operators are often required to use theoretical parameter mapping models to guide parameter setting. However, there may be deviations between the initial theoretical parameter mapping model and the actual parameter mapping model, which results in a gap between the actual output parameter value of the test and the expected output result when the theoretical parameter mapping model is used to guide parameter setting. Operators are required to make multiple adjustments to the test input parameters to obtain output results close to the expected values ​​of the test output parameters. After a long period of accumulated test work experience, operators can often gain a certain degree of perceptual understanding of the test system model, but they are still unable to quantify and correct the model, and the parameter setting efficiency cannot be improved. When operators are replaced, previous operating experience and perceptual understanding of the system model will be reset to zero, and new operators can only start from scratch, explore the system parameter setting rules, and repeat previous mistakes or search tests.

[0004] Therefore, it is necessary to propose an intelligent parameter setting method for experimental design, taking the operator's experimental records as learning samples, and constructing a more accurate parameter mapping model to guide the parameter setting work of subsequent experiments and reduce the difficulty and workload of subsequent experiments. Summary of the invention

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a method and system for intelligent parameter setting of a test system, which aims to construct a more accurate parameter mapping model for the test system to guide the parameter setting work of subsequent tests and reduce the difficulty and workload of subsequent tests.

[0006] To achieve the above object, according to one aspect of the present invention, a method for intelligently setting parameters of a test system is provided, comprising:

[0007] S1. Use the test system parameter theoretical mapping model to construct an initial test record data set; the initial data set includes input parameters and actual output parameters of the test system;

[0008] S2. constructing an actual mapping model of test system parameters using the initial test record data set, expanding the data set during the test, and iteratively correcting the above regression model using the expanded data set to obtain the optimal actual mapping model of system parameters;

[0009] S3. Substitute the expected output parameters of the test system into the optimal system parameter actual mapping model to obtain the input parameters required by the test system.

[0010] Furthermore, step S1 specifically includes:

[0011] S1.1. Determine the fixed input parameters x1, x2, x3, ..., x1 of the test system according to the test requirements. P-1 、Input parameter x to be set P And the expected output parameters y; x1, x2, x3, ..., x P Indicates different parameters;

[0012] S1.2. Initialize fixed input parameters x1, x2, x3, ..., x P-1 The value of the expected output parameter y is obtained by the experimental system parameter theoretical mapping model y = f(x1, x2, x3, ..., x P ), and obtain the theoretical value x of the input parameter to be set P ;

[0013] S1.3. Combine prior knowledge and try to find the answer in x P The input parameters to be set are adjusted until the error between the actual output parameters of the test system and the expected output parameters is less than the set threshold; the output parameters and the corresponding input parameters that meet the above conditions are added to the initial test record data set;

[0014] S1.4. Repeat steps S1.2-S1.3 until the initial test record data set reaches the set size.

[0015] Furthermore, step S2 specifically includes:

[0016] S2.1. Calculate the mean absolute error of the initial test record data set Among them, k0 is the number of data in the initial test record data set, It means that through y=f1 j (x1,x2,x3,…x P ) predicts the experimental input parameter x to be set P The estimated value of Represents xP The actual setting value of

[0017] S2.2. Initialize j = 1, the current optimal parameter mapping model M is the theoretical parameter mapping model y = f(x1, x2, x3, ... x P ) and the optimal parameter mapping model mean absolute error e M =e0; where j is the iteration number in the model training process;

[0018] S2.3. Using the existing test record data set, a regression algorithm is used to construct the optimal parameter mapping model y=f1 for the jth iteration j (x1,x2,x3,…x P ), using the optimal parameter mapping model to generate k j New test record data, and calculate the mean absolute error of the model

[0019] S2.4. Judgement j <e T Is it true? If so, the current optimal parameter mapping model is output, and the optimal parameter mapping model search process ends; otherwise, go to step S2.5; where e T is the preset test error threshold;

[0020] S2.5. Judgement j <e M Is it true? If so, update the current optimal parameter mapping model to y = f1 j (x1,x2,x3,…x P ); otherwise, keep the current optimal parameter mapping model unchanged;

[0021] S2.6. Determine whether j < n holds; if so, proceed to step S2.7; otherwise, output the current optimal parameter mapping model, and end the search process for the optimal parameter mapping model; wherein n is the maximum number of iterations in the parameter mapping model search process;

[0022] S2.7. The iteration round is increased by 1, j = j + 1, and the current optimal parameter mapping model M is used to generate k j New test record data, and calculate the mean absolute error of the new test record data set

[0023] S2.8. Take the initial test record data set and the test record data generated during the first j-1 iterations as the existing test record data set, and return to execute step S2.3.

[0024] Furthermore, the regression algorithm includes a linear regression algorithm, a decision tree regression algorithm and a random forest regression algorithm.

[0025] Furthermore, in step S2.3 and step S2.7, the optimal parameter mapping model M is used to generate k j New test record data, including:

[0026] 01. Initialize x1, x2, x3, …x according to the test requirements p-1 , the value of y;

[0027] 02. Through y=f1 j (x1,x2,x3,…x P ) predicts the experimental input parameter x to be set P The estimated value of the predicted i-th data is

[0028] 03. Search for x near P The actual setting value Input it into the test system and record the current test output parameters

[0029] 04.Judgment Is it true? If so, record As the i-th test record data of this round, and go to step 05; otherwise, return to step 03; T2(j) is the preset error threshold, y i is the expected value of the test output parameter of the i-th test record data in this round of iteration;

[0030] 05. Judge i<k j Is it true? If so, set i=i+1 and return to step 01 to collect the next test record data; otherwise, j The new test record data is added to the original test record data set, ending the new data generation process; where k j The size of the experimental record dataset in the current iteration.

[0031] Furthermore, the maximum number of iterations of the parameter mapping model n and the size of the test record data set in each iteration k j And the method for determining the size of the initial test record data set k0 is as follows:

[0032] Set the maximum number of iterations n of the parameter mapping model and the size of the data set k for each iteration test record j and the range of values ​​of the initial test record data set size k0;

[0033] Set a number of different theoretical parameter mapping models and assumed actual parameter mapping models and perform one-to-one combinations, with the number of combinations being greater than or equal to 3;

[0034] List n, k for each combination of conditions j and all value combinations of k0;

[0035] For each parameter combination, perform the following steps in sequence; set different inputs (x1, x2, x3, ... x P ), and get the actual output y of the test system, then (x1,x2,x3,…x P , y) is a piece of data, and k0 pieces of data are obtained as the initial test record data set; the optimal parameter mapping model search process of the test system is carried out by simulation, and when the condition e is met j <e T Under the premise that x p The parameters n and k with the minimum total search times j The optimal values ​​of n and k0 are obtained under various combinations of theoretical parameter mapping models and assumed actual parameter mapping models. j and k0 are averaged respectively, and the average value of each parameter is taken as the final determined value.

[0036] According to another aspect of the present invention, a system for intelligent parameter setting of a test system is provided, comprising: a computer-readable storage medium storing a computer program; and a device for controlling the storage medium to execute the above-mentioned method for intelligent parameter setting of a test system when the computer program is executed by a processor.

[0037] In general, compared with the prior art, the above technical solution conceived by the present invention can achieve the following beneficial effects.

[0038] (1) The intelligent parameter setting method based on the test system proposed in the present invention records the input and output data of the test system for multiple times, and then uses these recorded data as a data set to construct a mapping model. The obtained model is closer to the actual mapping model of the test system, and it is iteratively applied to the guidance of the test parameter setting, which effectively improves the adjustment efficiency of the input parameters to be set during the test process.

[0039] (2) The method of the present invention can obtain a specific actual mapping model of the test system and use it to replace the operator's intuitive understanding of the test system. Therefore, even when the operator is replaced, the new operator can set the input parameters under the guidance of the model without having to explore the system parameter setting rules and repeat previous erroneous or search test work. The time cost of operating the test system can be greatly reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flowchart of iterative training of the optimal parameter mapping model and optimal model search process;

[0041] Figure 2 It is to use the optimal parameter mapping model M to generate k j A new test record data flow chart. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0043] The embodiment of the present invention provides a method for intelligently setting parameters of a test system, comprising:

[0044] S1. Use the test system parameter theoretical mapping model to construct an initial test record data set; the initial data set includes input parameters and actual output parameters of the test system;

[0045] Step S1 specifically includes:

[0046] S1.1. Determine the fixed input parameters x1, x2, x3, ..., x1 of the test system according to the test requirements. P-1 、Input parameter x to be set P And the expected output parameters y; x1, x2, x3, ..., x P Represents different parameters; S1.2. Initialize fixed input parameters x1, x2, x3, ..., x P-1 The value of the expected output parameter y is obtained by the experimental system parameter theoretical mapping model y = f(x1, x2, x3, ..., x P ), and obtain the theoretical value x of the input parameter to be set P ; S1.3. Combine prior knowledge and try it out in x P Adjust the input parameters to be set nearby until the error between the actual output parameters of the test system and the expected output parameters is less than the set threshold; add the output parameters that meet the above conditions and the corresponding input parameters to the initial test record data set; S1.4. Repeat steps S1.2-S1.3 until the initial test record data set reaches the set scale.

[0047] Taking the laser attenuation system as an example, the laser attenuation system is used to attenuate laser energy and obtain the specified laser-to-target power. The system input parameters are current, repetition rate, number of pulses, pulse width and attenuation coefficient, and the output parameter is power. Because there are 5 input parameters, p=5 in step S1.1; x1, x2, x3, x4 are current, repetition rate, number of pulses and pulse width, which are fixed; x5 is the attenuation coefficient, which is to be set, and x1, x2, x3, x4, x5, y represent a data record.

[0048] S2. constructing an actual mapping model of test system parameters using the initial test record data set, expanding the data set during the test, and iteratively correcting the above regression model using the expanded data set to obtain the optimal actual mapping model of system parameters;

[0049] like Figure 1 As shown, step S2 specifically includes:

[0050] S2.1. Calculate the mean absolute error of the initial test record data set Among them, k0 is the number of data in the initial test record data set, It means that through y=f1 j (x1,x2,x3,…x P ) predicts the experimental input parameter x to be set P The estimated value of Represents x P The actual setting value; S2.2. Initialize j = 1, the current optimal parameter mapping model M is the theoretical parameter mapping model y = f (x1, x2, x3, ... x P ) and the optimal parameter mapping model mean absolute error e M =e0; where j is the iteration round in the model training process; S2.3. Using the existing test record data set, a regression algorithm (such as linear regression algorithm, decision tree regression algorithm and random forest regression algorithm, etc.) is used to construct the optimal parameter mapping model y=f1 for the jth iteration j (x1,x2,x3,…x P ), using the optimal parameter mapping model to generate k j New test record data, and calculate the mean absolute error of the model S2.4. Judgement j <e T Is it true? If so, the current optimal parameter mapping model is output, and the optimal parameter mapping model search process ends; otherwise, go to step S2.5; where e T is the preset test error threshold; S2.5. j <e MIs it true? If so, update the current optimal parameter mapping model to y = f1 j (x1,x2,x3,…x P ); otherwise, keep the current optimal parameter mapping model unchanged; S2.6. Determine whether j<n is true; if so, go to step S2.7; otherwise, output the current optimal parameter mapping model and end the search process for the optimal parameter mapping model; where n is the maximum number of iterations in the parameter mapping model search process; S2.7. The iteration number is increased by 1, j=j+1, and the current optimal parameter mapping model M is used to generate k j New test record data, and calculate the mean absolute error of the new test record data set S2.8. Take the initial test record data set and the test record data generated during the first j-1 iterations as the existing test record data set, and return to execute step S2.3.

[0051] like Figure 2 As shown, in steps S2.3 and S2.7, the optimal parameter mapping model M is used to generate k j New test record data, including:

[0052] 01. Initialize x1, x2, x3, …x according to the test requirements p-1 , the value of y; 02. through y = f1 j (x1,x2,x3,…x P ) predicts the experimental input parameter x to be set P The estimated value of the predicted i-th data is 03. Search for x near P The actual setting value Input it into the test system and record the current test output parameters 04.Judgment Is it true? If so, record As the i-th test record data of this round, and go to step 05; otherwise, return to step 03; T2(j) is the preset error threshold, y i is the expected value of the test output parameter of the i-th test record data in this round of iteration; 05. Judge i<k j Is it true? If so, set i=i+1 and return to step 01 to collect the next test record data; otherwise, j The new test record data is added to the original test record data set, ending the new data generation process; where k j The size of the experimental record dataset in the current iteration.

[0053] Step S2 needs to first determine the maximum number of iterations n in the optimal model search process, the size of the data set k for each iteration test record j (j={1,2,3...,n}) and the empirical value of the initial test record data set size k0. This process does not require operation on the laser attenuation system. It only needs to assume several sets of actual mapping models of the laser attenuation system and use simulation to determine the empirical value. It should be noted that here, because the simulation method is used to obtain n and k j (j={1,2,3...,n}) and k0 are empirical values ​​of parameters, so an assumed actual parameter mapping model can be set in advance. However, in the actual operation process of step S2, the actual parameter mapping model is unknown, and it is also expected to be approximated in the optimal model search process of this method. The specific method is:

[0054] Set the maximum number of iterations n of the parameter mapping model and the size of the data set k for each iteration test record j and the range of values ​​of the initial test record data set size k0; for example, n ranges from 1 to 4, k j The range of is 0~200, and the range of k0 is 0~200; set several different theoretical parameter mapping models and assumed actual parameter mapping models and make one-to-one combinations, and the number of combinations is greater than or equal to 3; list n, k under each combination condition j and all value combinations of k0; for each parameter combination, perform the following steps in sequence; set different inputs (x1, x2, x3, x4, x5) to obtain the actual output y of the laser attenuation system, then (x1, x2, x3, x4, x5, y) is a piece of data, and k0 pieces of data are obtained as the initial test record data set; the optimal parameter mapping model search process of the laser attenuation system is carried out by simulation. In the optimal parameter mapping model search simulation process, when the condition e is met j <e T Under the premise of , the parameters n and k that minimize the total number of searches for x5 (i.e., the number of executions of step 03 above) j The values ​​of n and k0 are the optimal values; the optimal n and k are obtained under various combinations of the theoretical parameter mapping model and the assumed actual parameter mapping model. j and k0 are averaged respectively, and the average value of each parameter is taken as its empirical value.

[0055] Because the simulation method for searching the laser attenuation system parameter mapping model is much more convenient than the actual operation of the laser attenuation system, and does not require manual system operation, thus avoiding the loss of the hardware system, the embodiment of the present invention determines n, k jWhen the empirical values ​​of parameters such as (j={1,2,3...,n}) and k0 are obtained, the simulation method is used. The difference from the optimal parameter mapping model search process in step S2 is that the test record data set of the simulation method is constructed by simulation, that is, the data set is obtained by the actual parameter mapping model assumed in advance.

[0056] S3. Substitute the expected output parameters of the test system into the optimal system parameter actual mapping model to obtain the input parameters required by the test system.

[0057] When verifying the effectiveness of the method proposed in the present invention, the regression algorithm used in the embodiment of the present invention is the XGBoost algorithm, the fixed test input parameters are 4 x1, x2, x3, x4, the test input parameter to be set is 1 x5, and the test parameter of the expected output is 1 y. The verification process is as follows:

[0058] (a) Calculate n and k through the simulation process j and k0 experience value;

[0059] (b) Using the empirical values ​​calculated in step (a), a simulation process is performed to find the optimal parameter mapping model and verify the validity of the empirical values.

[0060] The detailed process of step (a) is as follows: first, the theoretical parameter mapping model expressions and actual parameter mapping model expressions required in the test process are selected from equations (1) to (6), and then four groups of experiments are carried out using the selected expressions. The expression combination method is as follows:

[0061] As shown in Table 1.

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068] Table 1 Combination of theoretical parameter mapping model and actual parameter mapping model in simulation test

[0069]

[0070] Then, some parameters involved in the parameter mapping model training simulation process are set, and the setting method is shown in Figure 2:

[0071] Table 2 Parameter settings during model training simulation

[0072]

[0073]

[0074] Finally, the empirical values ​​of parameters such as n, k0, k1, k2, k3, and k4 are calculated. The simulation test results of the optimal parameter mapping model search process for experiment ad are shown in Table 3.

[0075] Table 3 Simulation test results of the optimal parameter mapping model search process for AD

[0076]

[0077] At this time, the empirical values ​​of n, k0, k1, k2, k3, and k4 are calculated as follows:

[0078]

[0079] After calculating the empirical values ​​of parameters such as n, k0, k1, k2, k3, k4, etc., proceed to step (b) to verify the validity of the empirical values, as detailed below:

[0080] In order to verify the validity of the empirical value, the present invention adds four groups of tests for test eh. The combination of the theoretical parameter mapping model and the actual parameter mapping model for test eh is shown in Table 4, and the test results are shown in Table 5. are the data sets used in the simulation process of searching for the optimal parameter mapping model from test e to test h, They are the test sets for experiments e to h respectively (each group of experiments uses 200 sets of new data as the test set).

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087] Table 4 Combination of theoretical parameter mapping model and actual parameter mapping model of test eh

[0088]

[0089] Table 5 Simulation test results of the optimal parameter mapping model search process for the test eh when using parameter experience values

[0090]

[0091]

[0092] In Table 5, the fixed theoretical parameter mapping model is used to simulate manual search, and the efficiency of the present invention is compared with that of manual search. The number of searches for x5 in the data set when the fixed theoretical parameter mapping model is used represents the number of searches for x5 in the data set during manual search. It can be seen from Table 5 that after using the empirical values ​​of n, k0, k1, k2, k3, k4 obtained by experiment ad, the empirical values ​​are applied to the new experiment (experiment eh), and the search efficiency of parameter x5 is significantly improved, indicating that the present invention can indeed greatly improve the efficiency of parameter setting, and the empirical value setting of n, k0, k1, k2, k3, k4 has a certain rationality. When the empirical values ​​are applied to new data testing, the search efficiency of parameter x5 can still be improved.

[0093] The test system of the present invention refers to a hardware or software system in which the actual output has an error with the theoretical output for a specific input, and the operator often needs to make multiple adjustments to the input data and conduct experimental explorations in order to obtain the desired test output results. In addition, the system may be required to output different test results due to application needs. For example, the input parameters of the laser attenuation system are current x1, repetition rate x2, number of pulses x3, pulse width x4, and attenuation coefficient x5, and the output is the target power y. In theory, the input and output have a mapping relationship y=f(x1,x2,x3,x4,x5), but due to the limited understanding of the characteristics of the laser attenuation system itself or the influence of environmental factors, the actual input and output mapping relationship of the laser attenuation system is y′=f′(x1,x2,x3,x4,x5). The method of the present invention is applicable to similar test systems.

[0094] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for intelligently setting parameters of a test system, characterized in that: Including: S1. Using the theoretical mapping model of the test system parameters, construct an initial test record data set; the initial data set includes input parameters and the actual output parameters of the test system. S2. Using the initial test record data set to construct the actual mapping model of the test system parameters, expand the data set during the test process, and use the expanded data set to iteratively correct the above-mentioned actual mapping model of the parameters to obtain the optimal actual mapping model of the system parameters. S3. Substitute the expected output parameters of the test system into the optimal actual mapping model of the system parameters to obtain the input parameters required by the test system. Step S1 specifically includes: S1.

1. Determine the fixed input parameters x1, x2, x3, ..., x1 of the test system according to the test requirements. P-1 、Input parameter x to be set P And the expected output parameters y; x1, x2, x3, ..., x P Indicates different parameters; S1.

2. Initialize fixed input parameters x1, x2, x3, ..., x P-1 The value of the expected output parameter y is obtained by the experimental system parameter theoretical mapping model y = f(x1, x2, x3, ..., x P ), and obtain the theoretical value x of the input parameter to be set P ; S1.

3. Combine prior knowledge and try to find the answer in x P The input parameters to be set are adjusted until the error between the actual output parameters of the test system and the expected output parameters is less than the set threshold; the output parameters and the corresponding input parameters that meet the above conditions are added to the initial test record data set; S1.

4. Repeat steps S1.2 - S1.3 until the initial test record data set reaches the set scale. Step S2 specifically includes: S2.

1. Calculate the mean absolute error of the initial test record data set Among them, k0 is the number of data in the initial test record data set, It means that through y=f1 j (x1,x2,x3,…x P ) predicts the experimental input parameter x to be set P The estimated value of Represents x P The actual setting value of S2.

2. Initialize j = 1, the current optimal parameter mapping model M is the theoretical parameter mapping model y = f(x1, x2, x3, ... x P ) and the optimal parameter mapping model mean absolute error e M =e0; where j is the iteration number in the model training process; S2.

3. Using the existing test record data set, a regression algorithm is used to construct the optimal parameter mapping model y=f1 for the jth iteration j (x1,x2,x3,…x P ), using the optimal parameter mapping model to generate k j New test record data, and calculate the mean absolute error of the model S2.

4. Judgement j <e T Is it true? If so, the current optimal parameter mapping model is output, and the optimal parameter mapping model search process ends; otherwise, go to step S2.5; where e T is the preset test error threshold; S2.

5. Judgement j <e M Is it true? If so, update the current optimal parameter mapping model to y = f1 j (x1,x2,x3,…x P ); otherwise, keep the current optimal parameter mapping model unchanged; S2.

6. Determine whether j < n holds; if so, proceed to step S2.7; otherwise, output the current optimal parameter mapping model and end the search process for the optimal parameter mapping model; where n is the maximum number of iteration rounds in the search process of the parameter mapping model. S2.

7. The iteration round is increased by 1, j = j + 1, and the current optimal parameter mapping model M is used to generate k j New test record data, and calculate the mean absolute error of the new test record data set S2.

8. Use the initial test record data set and the test record data generated in the previous j - 1 iteration processes as the existing test record data set, and return to execute step S2.

3.

2. A method for intelligently setting parameters of a test system according to claim 1, characterized in that: The regression algorithm includes linear regression algorithm, decision tree regression algorithm, and random forest regression algorithm.

3. The method for intelligently setting parameters of a test system according to claim 1, characterized in that: In step S2.3 and step S2.7, the optimal parameter mapping model M is used to generate k j New test record data, including:

01. Initialize x1, x2, x3, …x according to the test requirements p-1 , the value of y; 02. Through y=f1 j (x1,x2,x3,…x P ) predicts the experimental input parameter x to be set P The estimated value of the predicted i-th data is 03. Search for x near P The actual setting value Input it into the test system and record the current test output parameters 04.Judgment Is it true? If so, record As the i-th test record data of this round, and go to step 05; otherwise, return to step 03; T2(j) is the preset error threshold, y i is the expected value of the test output parameter of the i-th test record data in this round of iteration; 05. Judge i<k j Is it true? If so, set i=i+1 and return to step 01 to collect the next test record data; otherwise, j The new test record data is added to the original test record data set, ending the new data generation process; where k j Record the size of the dataset for the experiment in the current iteration.

4. A method for intelligently setting parameters of a test system according to claim 1 or 3, characterized in that: The maximum number of iterations of the parameter mapping model n and the size of the test record data set in each iteration k j And the method for determining the size of the initial test record data set k0 is as follows: Set the maximum number of iterations n of the parameter mapping model and the size of the data set k for each iteration test record j and the range of values ​​of the initial test record data set size k0; Set a number of different theoretical parameter mapping models and assumed actual parameter mapping models and perform one-to-one combinations, and the number of combinations is greater than or equal to 3. List n, k for each combination of conditions j and all value combinations of k0; For each value combination, perform the following steps in sequence: Set different inputs (x1, x2, x3, ... x P ), and get the actual output y of the test system, then (x1,x2,x3,…x P , y) is a piece of data, and k0 pieces of data are obtained as the initial test record data set; the optimal parameter mapping model search process of the test system is carried out by simulation, and when the condition e is met j <e T Under the premise that x p n, k with the smallest total number of searches j The optimal values ​​of n and k0 are obtained under various combinations of theoretical parameter mapping models and assumed actual parameter mapping models. j and k0, and use the average values ​​as n and k j and the final determined value of k0.

5. A parameter intelligent setting system for a test system, characterized in that: Including: A computer-readable storage medium storing a computer program; when the computer program is run by a processor, it controls the device where the storage medium is located to execute the method according to any one of claims 1 to 4.

Images