CN111813669B - Adaptive random test case generation method based on multi-target group intelligence - Google Patents
Adaptive random test case generation method based on multi-target group intelligence Download PDFInfo
- Publication number
- CN111813669B CN111813669B CN202010633401.6A CN202010633401A CN111813669B CN 111813669 B CN111813669 B CN 111813669B CN 202010633401 A CN202010633401 A CN 202010633401A CN 111813669 B CN111813669 B CN 111813669B
- Authority
- CN
- China
- Prior art keywords
- test case
- population
- test
- algorithm
- fitness
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3684—Test management for test design, e.g. generating new test cases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Complex Calculations (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The method provides an adaptive random test case generation method based on multi-target group intelligence, and belongs to an adaptive random test technology in the field of software testing. The basic flow of the method comprises six steps, including calculating the fitness based on the maximization of the minimum distance, calculating the fitness based on the information entropy of the distance, calculating the fitness based on the sample difference, evolving iteration, generating and executing test cases, and finally iterating to generate the test cases. Simulation comparison experiments were also designed for eAR and FSCS-ART. Aiming at the existing adaptive random test method based on evolution and heuristic adaptive random test method, the test case generated by the invention has higher diversity and stronger failure detection capability.
Description
Technical Field
The invention belongs to an adaptive random test technology in the field of software testing, and relates to an adaptive random test case generation method based on multi-target group intelligence.
Background
With the continuous development of the computer industry, the software testing industry is also becoming increasingly important. Particularly, as the software is applied to the daily life of people in a large scale, the quality of the software is directly caused, and the quality of life of people is influenced in a certain sense. The most common means of ensuring the quality of software is through software testing, so in the software development cycle, software testing plays an indispensable role therein.
Among the numerous methods of test data generation, the Random Test (RT) algorithm is the simplest and easily implemented algorithm. However, the RT algorithm has a disadvantage that the algorithm cannot ensure that the test data can be distributed in the input domain "uniformly", so that the algorithm cannot obtain a good effect of detecting failure. Therefore, on the basis of RT, many optimization methods are created, for example, adaptive random test (Adaptive Random Test, ART) is proposed by t.y. Chen of the university of the technology of the schwannoma, and the ART makes a certain limit on randomness while preserving the randomness of RT (random test method), so that test cases can be distributed uniformly in the whole input domain, the test cases are more diversified, and the effect of detecting the failure domain is improved.
On the basis of ART, many improved algorithms based on heuristic strategies have also been generated. In 2009, tappenden first added an evolution algorithm to ART, and an adaptive random test method (eAR) algorithm based on evolution was proposed. In conventional ART, test data is often generated by heuristic strategies. However, eAR adopts a meta-heuristic search algorithm to search out better test data in the whole input domain through continuous iterative evolution. In actual life, single-objective optimization often cannot solve the problem well, so that meta-heuristic search algorithms such as NSGA, SPEA and the like based on multi-objective optimization are generated in the field of group intelligence.
Disclosure of Invention
Aiming at the existing adaptive random test method based on evolution, an adaptive random test case generation method based on multi-target group intelligence is provided. In order to solve the defect of single-objective optimization, a genetic algorithm is used as a representative of a group intelligent algorithm on the basis of a eAR algorithm, an algorithm NSGA-II which is popular in the field of multi-objective optimization at present is adopted, and an original single-objective optimization function is converted into a multi-objective optimization function, so that feasible solutions are more uniformly distributed in an input space, and finally generated test cases have stronger diversity, and the failure detection capability of the test cases is improved.
The algorithm comprises the following three optimization objective functions: an optimization function based on minimum distance maximization, an optimization function based on distance information entropy, and an optimization function based on sample difference values. The optimization function based on the maximization of the minimum distance is used for enabling the next test case to be far away from the previously generated test case as far as possible so as to enhance the searching capability of a feasible solution and enable the distribution range of the test case to be wider. The distance-based information entropy optimization function is used for ensuring that the distance between the next test case and the neighbor is more uniform, so that each test case can be locally uniform as much as possible. The optimization function based on the sample difference value is generated in a region with smaller test cases in a local range as far as possible for the next test case, so that the test case is more uniform in a global angle. In summary, through the above three optimization objective functions, the test cases generated by the test case generation method provided by the invention are more uniformly distributed in the input space.
The technical scheme of the invention is as follows:
step 1, determining the range of an input domain through information given by a program, and randomly generating a population with the size of NP；
Step 2, evaluating the population according to a multi-objective function one, an optimized objective function based on minimum distance maximizationP；
Step 3, evaluating the population according to a multi-objective function II, namely an optimized objective function of the distance-based information entropyP；
Step 4, evaluating the population according to a multi-objective function III, namely an optimized objective function based on sample differencesP；
Step 5, selecting, crossing and mutating the population according to NSGA-II algorithm;
and 6, judging whether an evolution search suspension condition is reached, if not, jumping to the step 2, otherwise, outputting the optimal solution as the next test case and executing the test case.
The specific implementation of the step 1 comprises the following steps:
step 1.1, determining the range and the dimension of an input domain according to information given by a tested program;
step 1.2, coding, namely taking a test case as a chromosome (individual), wherein the gene correspondence on the chromosome is a numerical value in the dimension corresponding to the test case, the whole test case set is a population, and the population size N is 20;
and 1.3, initializing the whole population by a random generation method.
The specific implementation of the step 2 comprises the following steps:
step 2.1, judging whether the test case set T is empty, setting the fitness 1 of all individuals to 0 if the test case set T is empty, ending the function, otherwise, entering the step 2.2;
step 2.2, if the set T is not empty, calculating a minimum distance between each individual in the population and all test cases in the set T, wherein a calculation formula of the distance is as follows:
step 2.3, frontkThe large minimum distances are all saved and the maximum value is set to be the fitness of the individual 1, whereinkDefault to 5, when the number of test cases in the test case set T is less than 5,kthe number of test cases in the actual test case set T.
The specific implementation of the step 3 comprises the following steps:
step 3.1, judging whether the number of the test case sets T is less than or equal to 1, if so, setting the fitness 2 of all individuals to 0 and ending the function, otherwise, entering step 3.2;
step 3.2, if the number of test cases in the test case set is greater than 1 and less than the constantkThenkEqual to the number of test cases in the current test case set, wherekThe initial value is 5;
step 3.3, according to the pre-calculated in step 2.3kThe maximum minimum distance is calculated, and the information entropy value is set as the fitness 2 of the individual, wherein the information entropyEntropyThe calculation formula is as follows:
d is the frontkThe sum of the distances is used to determine,dist i is the i-th minimum distance of the individual.
The specific implementation of the step 4 comprises the following steps:
step 4.1, judging whether the number of the test case sets T is smaller than a constantθIf yes, setting the fitness 3 of all individuals to 0 and ending the function, otherwise, turning to the next step, wherein the constant θmerThe recognition value is 5;
step 4.2, calculating the number of divided input spacespWhereinpThe calculation formula of (2) is as follows:
，
dis the dimension of the input field;
step 4.3, dividing each dimension of the input domain space intopDetermining the subareas of each individual and each test case in the set T, and calculating the sample difference value of the subareas as fitness 3, wherein the calculation formula of the sample difference value is as follows:
，
where D is the size of the sub-area, S is the size of the entire input field area,tthe number of test cases in the collection T is contained in the subarea, and I T is the number of test cases of the whole test case set.
The specific implementation of the step 5 comprises the following steps:
step 5.1, selecting the population based on NSGA-II algorithm realized by Jenetics framework, wherein the selection algorithm adoptsn-a tournament algorithm,nset to 2;
step 5.2, performing cross operation on the population based on NSGA-II algorithm realized by Jenetics framework, wherein the cross algorithm adopts single-point cross, and the cross probability is set to be 0.6;
step 5.3, carrying out mutation operation on the population based on NSGA-II algorithm realized by Jenetics framework, wherein the mutation algorithm adopts single-point mutation, and the mutation probability is 0.1;
and 5.4, evaluating the population based on an NSGA-II algorithm realized by the Jenetics framework to obtain an optimal individual.
The specific implementation of the step 6 comprises the following steps:
step 6.1, setting iteration times for the step 5, if the iteration times reach 100, namely when the population evolves to 100 generations, outputting an optimal solution as a next test case, otherwise, continuing to enter a step two to continue iteration evolution;
step 6.2, executing the test case, and putting the test case into a test case set T;
and 6.3, judging whether a suspension condition for generating the test cases is reached, wherein the suspension condition is generally that the program fails or the size of the test case set T reaches a certain number, if the suspension condition is reached, ending the method and returning the test case set T, otherwise, turning to the step 1 to reinitialize the population, and continuing to iteratively evolve to generate the next test case.
Compared with the prior art, the invention has the beneficial effects that:
1. compared with the adaptive random test method of the eAR single-target optimization function, the multi-target group intelligent algorithm adopted by the invention can optimize the test cases on a plurality of target components, thereby enhancing the diversity of the test cases and achieving the purpose of uniform distribution of the test cases on a plurality of dimensions. The NSGA-II algorithm adopted by the invention is a popular multi-objective optimization algorithm at present, and can well select a multi-objective optimal solution from a plurality of feasible solutions.
2. The target optimization function adopted by the invention adopts an optimization function based on the information entropy of the distance besides the most classical minimum distance-based maximization function. The function can ensure that the minimum distance between the test case and the neighbor is as uniform as possible, thereby compensating the boundary effect brought by the maximum objective function with the minimum distance, reducing the trend that the test case approaches to the boundary as much as possible, and enabling the test case to have higher probability to be generated in the central area of the input domain.
3. The target optimization function based on the sample difference is used for generating the test cases in the subareas with lower density as far as possible, so that the density uniformity of each part of the area can be better achieved. The function of maximizing the minimum distance and the function of the distance-based information entropy may cause the distribution of the test cases to be distributed insufficiently uniformly in the global, so that the targets of the sample differences can better compensate the defects of the optimized functions before.
Drawings
FIG. 1 is an algorithm flow diagram of an adaptive random test case generation method based on multi-objective population intelligence.
FIG. 2 is a schematic diagram of computing a target optimization function based on minimum distance maximization.
FIG. 3 is a schematic diagram of a target optimization function that calculates distance-based information entropy.
FIG. 4 is a schematic diagram of calculating a sample difference based objective optimization function.
FIG. 5 is a F-ratio line graph of the results of a two-dimensional block simulation experiment under 5000 replicates.
FIG. 6 is a F-ratio line graph of the results of a three-dimensional block simulation experiment under 5000 replicates.
Detailed Description
In order to more clearly understand the content of the adaptive random test case generation method based on multi-objective group intelligence, the invention is further described below with reference to the accompanying drawings and specific embodiments, and it should be noted that the embodiments described and given are intended to facilitate understanding of the invention, and are not limited in any way.
The overall algorithm flow chart of the adaptive random test case generation method based on multi-objective group intelligence is shown in figure 1, and the group is initialized by a random method in the first step; secondly, evaluating the population according to three optimized objective functions; the third step is to adopt NSGA-II to select, cross and mutate the population, obtain new population and cover the original population; the fourth step is to evaluate the generated new population through three objective optimization functions; step five, judging whether an evolution search stopping condition is reached, if the evolution search stopping condition is not reached, switching to the step three to continue evolution, otherwise, jumping to the next step; the sixth step is to select the optimal individual from the new population after evaluation as the next test case t; the seventh step is to execute the test case t; the eighth step is to add the test case T into the test case set T; step nine, judging whether the suspension condition of test case generation is met, if not, turning to a step two, reinitializing the population, continuing to generate the next test case, otherwise, jumping to the next step; the tenth step is to output the test case set T. The abort condition for evolution search is set here to 100 evolution iterations, and the abort condition for test case generation is that the discovery program fails or a certain number of test cases is reached. Three optimization objective functions refer to the sub-flow: the first step is to evaluate the population according to an optimized objective function based on minimum distance maximization; the second step is to evaluate the population according to an optimized objective function based on the information entropy of distance; the third step is to evaluate the population based on an optimized objective function of sample differences.
Referring to fig. 2, the objective optimization function based on the minimum distance maximization is calculated as follows:
step 201, traversing each individual in the population, calculating fitness 1 for each individual, and setting the current individual as c;
step 202, traversing the tested case set T, if the number of the tested case sets T is 0, setting all individual fitness to 0 and ending the function, otherwise, calculating the minimum distance between the individual c and all other individuals in the tested case set T;
step 203, setting the maximum minimum distance to fitness 1, and setting the frontkThe largest minimum distance is stored to facilitate the calculation of the distance entropy.
Referring to fig. 3, the objective optimization function for calculating the distance-based information entropy is as follows:
step 301, traversing each individual in the population, calculating fitness 2 for each individual, and setting the current individual as c;
step 303, judging whether the number of test cases in the test case set T is greater than 1, if not, setting the fitness 2 of the individual to 0 and ending the function, otherwise, turning to the next step;
step 303, based on the stored previous step 203kThe minimum distance calculates the information entropy of the distance according to the formula, and sets the value as fitness 2.
Referring to fig. 4, a sample difference-based objective optimization function is calculated as follows:
step 401, traversing each individual in the population, calculating fitness 1 for each individual, and setting the current individual as c;
step 402, judging whether the number of test cases of the test case set T is 0, if so, setting the fitness 3 of all individuals to 0, otherwise, calculating the number of space division according to a formulapPositioning all test cases in the test case set T to corresponding subareas;
step 403, calculating the area of the sub-area where the individual c is located and the number of test cases including the test case set T in the sub-area, calculating a sample difference value for the individual c according to the formula, and setting the difference value as fitness 3.
Based on studies of ART algorithms, failure zone models of simulation experiments were analyzed herein. Finally, two-dimensional and three-dimensional block failure models are constructed, and repeated experiments are carried out on FSCS, eAR and the invention (MOART) under the condition of four failure rates of 0.01, 0.005, 0.002 and 0.001 respectively. Also employed herein is an indicator (F-measure) Metrics (F-ratio) The experimental results of the three algorithms are compared. In addition, F-measureThe values were rank-sum checked to check for significant superiority of the MOART algorithm. The experimental parameter settings of the genetic algorithm are shown in table 1, the crossover operator adopts single-point crossover, the mutation operator adopts single-point mutation, and the selection operator adopts n-tournament. The two-dimensional experimental results are shown in table 2 and fig. 5, the three-dimensional experimental result data are shown in table 3 and fig. 6, the two-dimensional and three-dimensional rank sum test results are shown in table 4 and fig. 5, and the experimental results are 5000 times of repeated experiments.
Experimental results show that the MOART algorithm has better failure detection capability than the other two algorithms aiming at a model with higher failure rate when in low dimensionality. With reduced failure rates, the effect of this algorithm is no longer advantageous over eAR, but is still better than the FSCS algorithm. When the high latitude is input into the domain, the failure detection effect of the MOART algorithm in the failure rate range of 0.01-0.001 is better than that of the eAR algorithm. The MOART effect is significantly better than FSCS with reduced failure rate compared to FSCS.
Table 1 experimental parameters
| Parameter name | Description of the invention | Parameter value |
| N | Population size | 20 |
| P c | Crossover probability | 0.6 |
| P t | Probability of variation | 0.1 |
| n | Championship match capacity | 2 |
| i | Number of evolution iterations | 100 |
| k | Number of partial maximum minimum distances | 5 |
| θ | Average number of expected test cases in partitioned sub-regions | 5 |
TABLE 2F-ratio values for each algorithm in two-dimensional input field
| FSCS | eAR | MOART | |
| 0.01 | 67.0789 | 67.0457 | 63.5478 |
| 0.005 | 65.6424 | 63.4633 | 61.9725 |
| 0.002 | 64.9998 | 61.4403 | 60.8600 |
| 0.001 | 64.0730 | 59.7913 | 59.1222 |
TABLE 3F-ratio values for the various algorithms in the three-dimensional input field
| FSCS | eAR | MOART | |
| 0.01 | 84.8218 | 92.6200 | 82.7448 |
| 0.005 | 80.2536 | 84.7559 | 79.0998 |
| 0.002 | 77.9776 | 77.4597 | 73.4101 |
| 0.001 | 74.8610 | 74.4151 | 70.9618 |
TABLE 4F-measurement rank sum test results under two-dimensional input field
TABLE 5F-measure rank sum test results under three-dimensional input field
The foregoing description is only for the purpose of clearly showing the embodiments of the present invention and is not intended to limit the scope of the present invention, but any modification, finish or the like will fall within the scope of the present invention without departing from the spirit and scope of the present invention.
Claims (1)
1. The adaptive random test case generation method based on the multi-target group intelligence is characterized by comprising the following steps:
step 1, determining the range of an input domain through information given by a program, and randomly generating a population with the size of NP；
Step 2, evaluating the population according to a multi-objective function one, a function based on minimum distance maximizationP；
Step 3, evaluating the population according to a multi-objective function II, which is a function of the information entropy based on the distanceP；
Step 4, evaluating the population according to a multi-objective function III, which is a function based on sample differencesP；
Step 5, selecting, crossing and mutating the population according to NSGA-II algorithm;
step 6, judging whether an evolution search suspension condition is reached, if not, jumping to the step 2, otherwise, outputting an optimal solution as a next test case and executing the test case;
the specific implementation of the step 1 comprises the following steps:
step 1.1, determining the range and the dimension of an input domain according to information given by a tested program;
step 1.2, coding, namely taking a test case as a chromosome (individual), wherein the gene correspondence on the chromosome is a numerical value in the dimension corresponding to the test case, the whole test case set is a population, and the population size N is 20;
step 1.3, initializing the whole population entirely by a random generation method;
the specific implementation of the step 2 comprises the following steps:
step 2.1, judging whether the test case set T is empty, setting the fitness 1 of all individuals to 0 if the test case set T is empty, ending the function, otherwise, entering the step 2.2;
step 2.2, if the set T is not empty, calculating a minimum distance between each individual in the population and all test cases in the set T, wherein a calculation formula of the distance is as follows:
；
step 2.3, frontkThe large minimum distances are all saved and the maximum value is set to be the fitness of the individual 1, whereinkDefault to 5, when the number of test cases in the test case set T is less than 5,kthe number of test cases in the actual test case set T is determined;
the specific implementation of the step 3 comprises the following steps:
step 3.1, judging whether the number of the test case sets T is less than or equal to 1, if so, setting the fitness 2 of all individuals to 0 and ending the function, otherwise, entering step 3.2;
step 3.2, if the number of test cases in the test case set is greater than 1 and less than the constantkThenkEqual to the number of test cases in the current test case set, wherekThe initial value is 5;
step 3.3, according to the pre-calculated in step 2.3kThe maximum minimum distance is calculated, and the information entropy value is set as the fitness 2 of the individual, wherein the information entropyEntropyThe calculation formula is as follows:
，/>，
d is the frontkThe sum of the distances is used to determine,dist i an ith minimum distance for an individual;
the specific implementation of the step 4 comprises the following steps:
step 4.1, judging whether the number of the test case sets T is smaller than a constantθIf so, setting the fitness 3 of all individuals to 0 and ending the function, otherwise, proceeding to the next step, wherein the constantsθDefault value is 5;
step 4.2, calculating the number of divided input spacespWhereinpThe calculation formula of (2) is as follows:
，
dfor the dimension of the input domain, |T| is the size of the test case set T;
step 4.3, dividing each dimension of the input domain space intopThe identity and the sub-region of each individual and each test case in the set T are determinedAnd calculating a sample difference value of the sub-region as fitness 3, wherein a calculation formula of the sample difference value is as follows:
，
where D is the size of the sub-area, S is the size of the entire input field area,tfor the number of test cases in the collection T contained in the subarea, |T| is the number of test cases of the whole test case set;
the specific implementation of the step 5 comprises the following steps:
step 5.1, selecting the population based on NSGA-II algorithm realized by Jenetics framework, wherein the selection algorithm adoptsn-a tournament algorithm,nset to 2;
step 5.2, performing cross operation on the population based on NSGA-II algorithm realized by Jenetics framework, wherein the cross algorithm adopts single-point cross, and the cross probability is set to be 0.6;
step 5.3, carrying out mutation operation on the population based on NSGA-II algorithm realized by Jenetics framework, wherein the mutation algorithm adopts single-point mutation, and the mutation probability is 0.1;
step 5.4, evaluating the population based on NSGA-II algorithm realized by Jenetics framework to obtain the optimal individual;
the specific implementation of the step 6 comprises the following steps:
step 6.1, setting iteration times for the step 5, if the iteration times reach 100, namely when the population evolves to 100 generations, outputting an optimal solution as a next test case, otherwise, continuing to enter a step two to continue iteration evolution;
step 6.2, executing the test case, and putting the test case into a test case set T;
and 6.3, judging whether a suspension condition for generating the test cases is reached, wherein the suspension condition is generally that the program fails or the size of the test case set T reaches a certain number, if the suspension condition is reached, ending the method and returning the test case set T, otherwise, turning to the step 1 to reinitialize the population, and continuing to iteratively evolve to generate the next test case.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010633401.6A CN111813669B (en) | 2020-07-04 | 2020-07-04 | Adaptive random test case generation method based on multi-target group intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010633401.6A CN111813669B (en) | 2020-07-04 | 2020-07-04 | Adaptive random test case generation method based on multi-target group intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111813669A CN111813669A (en) | 2020-10-23 |
| CN111813669B true CN111813669B (en) | 2023-10-13 |
Family
ID=72856030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010633401.6A Active CN111813669B (en) | 2020-07-04 | 2020-07-04 | Adaptive random test case generation method based on multi-target group intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111813669B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114610600A (en) * | 2022-01-21 | 2022-06-10 | 南京航空航天大学 | RESTful API test case generation and optimization method based on information physical system |
| CN116578611B (en) * | 2023-05-16 | 2023-11-03 | 广州盛成妈妈网络科技股份有限公司 | Knowledge management method and system for inoculated knowledge |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102508770A (en) * | 2011-10-10 | 2012-06-20 | 南京大学 | Test case suite amplification method based on self-adaption random testing of predicate |
| CN106445821A (en) * | 2016-09-23 | 2017-02-22 | 郑州云海信息技术有限公司 | Method for automatically generating test case based on genetic algorithm |
| CN107844835A (en) * | 2017-11-03 | 2018-03-27 | 南京理工大学 | Multiple-objection optimization improved adaptive GA-IAGA based on changeable weight M TOPSIS multiple attribute decision making (MADM)s |
| CN110059015A (en) * | 2019-04-28 | 2019-07-26 | 西安邮电大学 | Evolution of Population multiple target priorities of test cases sort method |
| CN111143195A (en) * | 2019-12-03 | 2020-05-12 | 江苏大学 | Self-adaptive random test method based on iteration of candidate test case set |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AT510328A2 (en) * | 2011-12-12 | 2012-03-15 | Avl List Gmbh | METHOD FOR EVALUATING THE SOLUTION OF A MULTICRITERIAL OPTIMIZATION PROBLEM |
| US10733332B2 (en) * | 2017-06-08 | 2020-08-04 | Bigwood Technology, Inc. | Systems for solving general and user preference-based constrained multi-objective optimization problems |
-
2020
- 2020-07-04 CN CN202010633401.6A patent/CN111813669B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102508770A (en) * | 2011-10-10 | 2012-06-20 | 南京大学 | Test case suite amplification method based on self-adaption random testing of predicate |
| CN106445821A (en) * | 2016-09-23 | 2017-02-22 | 郑州云海信息技术有限公司 | Method for automatically generating test case based on genetic algorithm |
| CN107844835A (en) * | 2017-11-03 | 2018-03-27 | 南京理工大学 | Multiple-objection optimization improved adaptive GA-IAGA based on changeable weight M TOPSIS multiple attribute decision making (MADM)s |
| CN110059015A (en) * | 2019-04-28 | 2019-07-26 | 西安邮电大学 | Evolution of Population multiple target priorities of test cases sort method |
| CN111143195A (en) * | 2019-12-03 | 2020-05-12 | 江苏大学 | Self-adaptive random test method based on iteration of candidate test case set |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111813669A (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Got et al. | Hybrid filter-wrapper feature selection using whale optimization algorithm: A multi-objective approach | |
| US20060230018A1 (en) | Mahalanobis distance genetic algorithm (MDGA) method and system | |
| CN111813669B (en) | Adaptive random test case generation method based on multi-target group intelligence | |
| CN109284766A (en) | A kind of feature selection approach of Multivariate Discrete, device, equipment and storage medium | |
| Falahiazar et al. | Determining the Parameters of DBSCAN Automatically Using the Multi-Objective Genetic Algorithm. | |
| Łapa | Meta-optimization of multi-objective population-based algorithms using multi-objective performance metrics | |
| CN117236278A (en) | Chip production simulation method and system based on digital twin technology | |
| CN115690476A (en) | Automatic data clustering method based on improved harmony search algorithm | |
| CN111126560A (en) | Method for optimizing BP neural network based on cloud genetic algorithm | |
| Chen et al. | A network community-based differential evolution for multimodal optimization problems | |
| Phan et al. | Efficiency enhancement of evolutionary neural architecture search via training-free initialization | |
| CN107798429B (en) | Method for positioning water supply pipe network pollution source based on collaborative expensive optimization algorithm | |
| CN117892209A (en) | Oversampling method based on support vector machine and evolutionary computation | |
| CN113688565A (en) | Supercritical CO based on optimal trade-off2Bayesian optimization method for centrifugal compressor | |
| CN117973009A (en) | Parallel sequence sampling method applied to ship type optimization | |
| KR101462349B1 (en) | Method for multi-objective optimizing based on Pareto-optimality | |
| CN107766887A (en) | A kind of local weighted deficiency of data mixes clustering method | |
| CN112183598A (en) | Feature selection method based on genetic algorithm | |
| CN116010291A (en) | Multipath coverage test method based on equalization optimization theory and gray prediction model | |
| CN111488903A (en) | Decision tree feature selection method based on feature weight | |
| CN114528094A (en) | Distributed system resource optimization allocation method based on LSTM and genetic algorithm | |
| CN113141272A (en) | Network security situation analysis method based on iteration optimization RBF neural network | |
| Liu et al. | A reference points-based evolutionary algorithm for many-objective optimization | |
| Dhivya et al. | Weighted particle swarm optimization algorithm for randomized unit testing | |
| CN117932444B (en) | Soil heavy metal laboratory detection sample point screening method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |