CN111666221B - Software test data set amplification method based on longicorn stigma search - Google Patents
Software test data set amplification method based on longicorn stigma search Download PDFInfo
- Publication number
- CN111666221B CN111666221B CN202010631557.0A CN202010631557A CN111666221B CN 111666221 B CN111666221 B CN 111666221B CN 202010631557 A CN202010631557 A CN 202010631557A CN 111666221 B CN111666221 B CN 111666221B
- Authority
- CN
- China
- Prior art keywords
- test data
- target
- test
- longicorn
- branch
- 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
Images
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
- 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
- G06F11/3676—Test management for coverage analysis
-
- 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
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
-
- 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 Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Debugging And Monitoring (AREA)
Abstract
A software test data set amplification method based on longicorn stigma search belongs to the field of software regression testing. Aiming at the problem that the original test case is difficult to meet the test requirement of a new program modified due to software evolution, the test case amplification method based on program execution information and longicorn stigma search is provided. Statically analyzing new and old versions of programs, acquiring method call graphs and program execution information to obtain a target method set to be tested, and sequencing target methods by calculating the influence degree of errors contained in the method; selecting part of test cases as initial population according to the method coverage information of the original test case set, designing a fitness function according to the branch distance and the branch nesting depth, and carrying out test data amplification on the ordered target method set by adopting an improved longicorn searching algorithm. The invention utilizes the method to call the graph, the program execution information and the longicorn whisker search algorithm to amplify the regression test data, improves the utilization rate of the original test case and reduces the cost of the regression test.
Description
Technical Field
The invention belongs to the technical field of software testing, in particular to a software testing data set amplification method which is used for amplifying a testing data set, fully utilizes an original testing set to meet testing requirements in a software evolution process and belongs to the technical field of regression testing.
Background
In order to effectively guarantee the quality of the modified software product, the correctness of a code modification module is generally ensured by executing regression testing and side effects of the code modification module on other modules of a tested program are avoided. Regression testing refers to retesting the evolving software after the existing software has been modified to confirm that the newly introduced modifications do not introduce new errors or interfere with the functionality of the existing unmodified parts.
The software evolution causes the modification and increase of programs, and the existing test cases in the regression test cannot completely test the programs. In order to meet the test requirements of software evolution and ensure that a test data set can cover the modification and new increase parts of a program, the original test case set needs to be amplified. When the regression test data amplification is carried out, firstly, a target needs to be analyzed and extracted, then, the test data amplification is carried out on the regression test data, and the test target mainly refers to a modified part and an added part in a program. The test case set amplification needs to make full use of executed test information and assist the generation of new test cases by combining with software evolution information.
Compared with the traditional test case generation, the target covered by the test case amplification is more targeted, and the new test data is generated by using the executed test data information and the program evolution information in an auxiliary mode, so that the case set generation efficiency of the regression test can be greatly improved. At present, the existing methods for generating test case sets are roughly divided into two categories: class 1 is a behavior-oriented test case augmentation technique, i.e., a new test case is expected to show the execution changes of the evolving software; class 2 is a coverage-oriented test case set augmentation technique, i.e., it is expected that new test cases cover new and modified portions of the evolving software.
The traditional test data set amplification technology has multiple limiting conditions, and an effective method for automatically amplifying test cases is still lacked in the regression test. Therefore, a software test data set amplification method based on longicorn stigma search is provided. And performing static analysis on the new version program and the old version program to obtain the execution information of the method call graph and the original test case in the program to obtain a modified method set to be tested, and sequencing by calculating the influence degree of the method call graph and the original test case containing errors. And selecting part of cases from the original test case set as an initial population according to execution information of the original test cases, and amplifying test data of the ordered target method set by adopting an improved longicorn searching algorithm.
Disclosure of Invention
The invention aims to solve the problems that: the problem that the test case of an original version program in the existing test data generation technology cannot meet the test requirement of a new version software easily is solved, an efficient new regression test data amplification method is needed, an error method possibly introduced by a granularity analysis program is improved under the condition of consuming less resources, and the utilization rate of the original test case and the regression test efficiency are improved.
The technical scheme of the invention is as follows: the method calling graph and program execution information are adopted in the early-stage preprocessing, a modification method part is possibly introduced into a method level analysis program, a target method set is obtained, and amplified initial test data is selected; and performing regression test data amplification on the target method set by adopting an improved longicorn stigma search algorithm in the later stage. The invention is characterized by comprising the following steps:
(1) for old new version program P old And P new Acquiring a method call Graph of the program by adopting a file generation tool Doxygen and a Graph drawing tool Graph viz, and establishing adjacency matrixes G and G' of the method call Graph:
if method f i And f j If there is a call relation, the matrix element g ij 1, otherwise g ij And (2) returning to be 0 and 1, traversing G and G', and counting the number w of the callable methods of each method k And w' k ；
(2) To P old And P new Executing the original test data set T and collecting program execution information, represented by matrices A and A':
test data t 1 ,t 2 ,…,t n The test method is that M ═ f 1 ,…,f m T in matrix A ij (1. ltoreq. i.ltoreq.m, 1. ltoreq. j.ltoreq.n) represents t j For f i If the test data t j Cover method f i ，t ij 1, otherwise t ij 0; using the vector e ═ e 1j ,…,e ij ,…e mj ]. -1 Represents the test data t j Is performed as a result of element e ij Representing jth test data for method f i If successful, e ij 1, otherwise e ij ＝0；
For the same in matrices A and APerforming XOR operation on the coverage method information and the execution result information of the test case to obtain a target method set
(3) Assuming that the sets of methods are independent of each other, the initial method f 1 Is performed with probability P (f) 1 ) 1.00, f is obtained by calling the matrix G according to the method obtained in the step (1) 1 The called method is f 2 ,…,f k (2. ltoreq. k. ltoreq.m) in total of w 1 Then P (f) 1 )/w 1 ＝1.00/w 1 ＝P(f 2 )＝…＝P(f k ) (ii) a If f k Calling f a …f b Equal w k A method, then f a Is performed with probability P (f) a ) And f k Method weight N (f) k ) Expressed as:
calculating the execution probability P (c) of the target method in the old and new program k ) And P' (c) k ) Target method c k Probability of containing error P "(c) k ) Comprises the following steps:
P″(c k )＝|P(c k )-P′(c k )|
calculating the weight N (c) of the target method in the new program k ) Multiplying the weight value by the probability of containing errors to obtain the influence degree s (c) of the method containing errors k )：
S(c k )＝P″(c k )·N(c k )
Will S (c) k ) The values are arranged according to a descending order to obtain an ordered target method set C' k ＝{f name1 ，f name2 …, which are subsequently amplified;
(4) obtaining C 'Using Soot code analysis tool and Graph viz' k The control flow graph of each method obtains the theoretical Path number and the Path set Path to be covered according to the branch node number of the Path (Path1, Path2, …), and inserts the branch distance function f in front of each branch node of the target Path i And representing the branch distance of the ith branch, and adjusting the weights of different branch nodes through the branch nesting depth:
wherein d is i The nesting depth of the branch node i. is obtained through static analysis of a process flow diagram, n. is the branch judgment number of the tested program aiming at the given target path, and the fitness function is designed according to the branch distance function and the branch nesting depth as follows:
wherein, epsilon is a very small constant to ensure that the denominator is not 0, and epsilon is set to be 0.01;
(5) according to the execution information matrix of the method of the test case obtained in the step (2) in the old and new programs, aiming at the target method c k All satisfy t ki T 1 (1. ltoreq. i. ltoreq.m) i As method c k Initial test data of test case amplification;
(6) adopting an improved longicorn whisker search algorithm for the initial population in the step (5), and assuming that the position of a longicorn individual in a solution space is x, the left whisker coordinate of the longicorn individual is x l X + l.c, with the coordinate of the right whisker being x r X-l.d, wherein x represents the coordinates of the centroid, l represents the distance between the longicorn centroid and the tentacle,
representing the random orientation of the longicorn in the solution space, r ═ rands (n,5) is a random vector, and the next position of the current longicorn according to the rule perception is:
x′＝x+s·d·sign(f(x l )-f(x r ))
wherein x is the current position, s is the variable step size of the search, sign (·) is a sign function for determining the next moving direction of the longicorn, and f (x) l ) Odor intensity of left palpus, f (x) r ) The intensity of the smell of the right palpus, if x r Is less than the left whisker x l If the fitness of the celestial cow is smaller than the fitness of x, accepting x 'as a next step position, otherwise, accepting x' with a random probability, and the step length follows:
wherein f is the fitness of the current individual, f ave Is the maximum fitness of the population, f min Is the minimum fitness in the population, f ave The average value of the fitness of other individuals after the maximum and minimum fitness is removed is called false average fitness; if the fitness of the individual is less than or equal to the false average fitness of the current group, the performance of the individual is good, the step length is reduced, otherwise, the step length is increased;
if the maximum iteration number is reached or the test data cover the target path, outputting the satisfied test data, and selecting C' k And (5) taking the test case which has successfully tested the target method and the test case selected in the step (5) as initial test data, and continuing to execute the step (6) until all the target methods are tested.
The method comprises the steps of analyzing a program acquisition method call graph through a white box test technology, establishing method execution information according to original test data, and acquiring a target method set through comparison of the execution information of old and new programs; calculating the probability that the method in the old and new program may contain errors, multiplying the probability by the method weight, and calculating the influence degree of the method containing errors so as to sequence the target method set; and comparing the execution information of the old and new program methods, selecting part of test cases associated with the target method as initial test data, and sequentially amplifying the test path sets corresponding to the target method by adopting an improved longicorn stigma search algorithm. The regression test data amplification method comprises the steps of increasing the granularity of a changed part of a new version program to a method level, carrying out static analysis aiming at the interior of the method, finding a method which possibly introduces error modification, obtaining a target path to be tested, and finally optimizing a test case through the orientation and the step length of a longicorn in an improved longicorn beard search algorithm to generate new test data covering the target path. The method calling and the method static analysis of the program enable the generated test case to be more accurate, improve the utilization rate of original test data and reduce the generation cost of new test data.
Drawings
FIG. 1 is a flow chart of the method of the present invention
FIG. 2 is a method call graph resulting from Doxygen analyzing a Java instance program
FIG. 3 is a program control flow graph obtained by a method of Soot analysis
FIG. 4 is a comparison of the coverage of target paths between the Triangle classifier Triangle test case set amplification and the other three methods
Detailed Description
As shown in fig. 1, the method combines evolution information of software and an improved longicorn whisker search algorithm to perform amplification of a regression test case set, a preprocessing part at an early stage acquires an ordered target method set and an initial test case by using a method call graph and method coverage execution information of a program, and an optimization algorithm is used to perform regression test data amplification at a later stage, and the method specifically includes the following steps:
(1) FIG. 2 illustrates a method call graph generated by the Doxygen tool for a Java program;
(2) using the original test set t 1 ＝(-2,1,4)、t 2 ＝(1,2,3)、t 3 ＝(3,4,5)、t 4 (3,3,4) and method call trace setup program execution information, the old new version program execution information is represented by matrices a and a', where f 1 →main(){}，f 2 →getType(triangle){}，f 3 →judgeType(){}，f 4 →isTriangle(triangle){}，f 5 →pType(){}，f 6 →diffOfBorders(){}，f 7 → getbytes () { }, carrying out XOR operation on A and A' to obtain a target method set C k ＝{f 3 ,f 4 ,f 6 }；
(3) Calculating to obtain the probability P (c) of the target method being called in the old program 3 )＝0.25，P(c 4 )＝0.25，P(c 6 ) The probability of the target method being invoked in the new program, P' (c), is calculated according to fig. 2 at 0.25 3 )＝0.33，P′(c 4 )＝0.33，P′(c 6 ) The weight of a node is N (c) 0.16 3 )＝1，N(c 4 )＝2，N(c 6 ) Then the probability that the target method modified may include an error is P "(c) 3 )＝0.08，P″(c 4 )＝0.08，P″(c 6 ) The target method obtains the influence degree of the error contained in the target method in the new program as S (c) 0.09 3 )＝0.08，S(c 4 )＝0.16，S(c 6 ) 0.09, and obtaining an ordered target method set C 'by descending order according to the influence degree' k ＝{f 4 ,f 6 ,f 3 }；
(4) By test method f 6 For example, the instrumentation of the branch distance function is first performed, FIG. 3 is a graph of analysis by Soot tool method f 6 And analyzing the obtained control flow graph to obtain 2 theoretical paths according to the number of the branch points, wherein the target path set is a path { p ═ p 1 ,p 2 }; by the target path p 1 For example, in branch predicate p ri Front insert branch distance function f [ p ] ri ]When the branch predicate is true, f [ p ] ri ]When the branch predicate is false, f [ p ] is equal to 0 ri ]The construction method of (1) is shown in Table 1, wherein K represents a constant;
TABLE 1 distance function for different branch conditions
(5) Overriding method f by a matrix of new program execution information 6 Test case t of 2 ＝(1,2,3)，t 3 (3,4,5) and t 4 Iterating as an initial population (3,3, 4);
(6) calculating the fitness of each individual in the population, if a case with the fitness value of 0 exists, meeting the condition, outputting a test case, otherwise, calculating the next step position through an improved longicorn stigma search algorithm, updating the next step position of the population by using a Metropolis criterion, updating the step length, continuing local search or global search operation, and outputting the test case until the maximum iteration times is reached or the fitness value is 0; continuing to select a next target path in the test path set to generate test data covering the target path; and selecting the next method of the ordered target method set until all target methods are covered.
The method set which is possibly introduced with errors after modification can be identified through the process, regression test data set amplification is carried out on the method set, in order to explain the effectiveness of the method, one or more modification points are defined for a test program under the same condition, the Method (MBAS), the traditional Genetic Algorithm (GA) and the traditional particle swarm algorithm (PSO) are respectively used for running for 30 times to generate test data covering a target path, evolution generations of each method are recorded, and the average value and the standard deviation of the methods are respectively calculated, so that the efficiency and the stability of the test data amplification method are measured. The test program selects schedule and Tcas in Siemens industry program set, Triangle judgment program Triangle and benchmark program NextDay, each method is operated under the same condition, the iteration times of the test data is recorded when the test data is amplified, and the experimental result is shown in Table 2:
TABLE 2 number of iterations for different algorithms to amplify test data
Taking the program NextDay as an example, the genetic algorithm needs to be iterated for 39.0 times, the particle swarm algorithm needs to be iterated for 27.1 times, and the method needs to be iterated for about 17.8 times to generate test data covering the target path. Compared with the test data amplification method based on genetic algorithm and particle swarm algorithm, the test data amplification efficiency of the method is respectively improved by about 54.35% and 34.31%.
Taking Triangle classification program Triangle as an example, the coverage of test data paths generated by different methods is shown in fig. 4, and different regression test data amplification methods are evaluated from the coverage of the test data.
The standard deviation of the iteration times of the method in the table 2 is lower than that of the other two methods, which shows that the method has higher stability in the aspect of software test data amplification.
According to the invention, the method call graph, the program execution information, the control flow graph of the program and the improved longicorn whisker search algorithm are combined under the same condition by combining the comprehensive experimental data, the test data amplification efficiency is superior to that of a Genetic Algorithm (GA) and a particle swarm algorithm (PSO), the test data amplification efficiency can be effectively improved, the original test case is fully utilized, and the regression test efficiency is improved.
Claims (1)
1. The software test data set amplification method based on longicorn stigma search is characterized by comprising the following steps of:
(1) for old new version program P old And P new Establishing adjacent matrixes G and G' of method call graphs and traversing, and counting the number w of the callable methods of each method k And w' k ；
(2) To P old And P new Executing an original test data set T and collecting program execution information A and A';
in matrix A, t ij For testing cases t j For f i Coverage of the method, e ij In order to execute the result, the matrix A and the matrix A' are subjected to exclusive OR operation to obtain a target method set
(3) If f is k Calling f a …f b Equal w k The method then executes the probability P (f) a ) And method weight N (f) k ) Expressed as:
target method c k Influence degree S (c) including error k ) Comprises the following steps:
S(c k )＝|P(c k )-P′(c k )|·N(c k )
wherein | P (c) k )-P′(c k ) I represents the target method c k Including the probability of error, will S (c) k ) The values are arranged according to the descending order to obtain an ordered target method set C k ′＝{f name1 ，f name2 … } in turn;
(4) using Soot code analysis tool and Graph viz, C 'was obtained' k Scanning and extracting a path set to be covered, and inserting a branch distance function f in front of each branch node of a target path i According toThe branch distance function and the branch nesting depth design fitness function are as follows:
wherein d is i Obtaining the nesting depth of the branch node i for the static analysis of the flow chart of the general process program, wherein n is the branch judgment number of the tested program aiming at the given target path, g i Is the weight of the branch node i, d i Obtaining the nesting depth of a branch node i for static analysis of a general process flow chart, wherein n is the branch judgment number of a tested program aiming at a given target path, and epsilon is a very small constant to ensure that a denominator is not 0;
(5) according to the execution information matrix obtained in the step (2), aiming at the target method c k All of which conform to t ki T 1 (1. ltoreq. i. ltoreq.m) i As initial test data for test case amplification;
(6) adopting an improved longicorn beard search algorithm for the initial population in the step (5), and sensing the next position of the current longicorn according to a rule:
x′＝x+s·d·sign(f(x l )-f(x r ))
r＝rands(n,5)
the next step position of the longicorn is updated using Metropolis criteria, step s follows:
if the fitness of the individual is less than or equal to f ave In this case, the individual can be well,decreasing the step size, and conversely, increasing the step size;
calculating the adaptability of the individual by using the fitness function in the step (4), outputting satisfied test data if the maximum iteration number is reached or the test data covers the target path, and continuously selecting C' k And (5) taking the test case which has successfully tested the target method and the test case selected in the step (5) as initial test data, and executing the step (6) until all the target methods are tested.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010631557.0A CN111666221B (en) | 2020-07-03 | 2020-07-03 | Software test data set amplification method based on longicorn stigma search |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010631557.0A CN111666221B (en) | 2020-07-03 | 2020-07-03 | Software test data set amplification method based on longicorn stigma search |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111666221A CN111666221A (en) | 2020-09-15 |
| CN111666221B true CN111666221B (en) | 2022-09-27 |
Family
ID=72390947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010631557.0A Active CN111666221B (en) | 2020-07-03 | 2020-07-03 | Software test data set amplification method based on longicorn stigma search |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111666221B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113900942B (en) * | 2021-10-09 | 2023-04-07 | 电子科技大学 | Method for generating simplified test case set of flight control machine-mounted model |
| CN114978974B (en) * | 2022-05-20 | 2023-09-19 | 国网重庆市电力公司电力科学研究院 | Wireless communication module testing method, device, equipment and storage medium |
| CN115809203B (en) * | 2023-02-07 | 2023-04-25 | 杭州罗莱迪思科技股份有限公司 | Dynamic nesting method and device for software test cases and application of dynamic nesting method and device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718368A (en) * | 2016-01-15 | 2016-06-29 | 西安邮电大学 | Software testing data amplification method |
| CN110109822A (en) * | 2019-03-30 | 2019-08-09 | 华南理工大学 | The regression testing method of priorities of test cases sequence is carried out based on ant group algorithm |
| CN110134590A (en) * | 2019-04-18 | 2019-08-16 | 上海大学 | A kind of longicorn palpus fuzz testing case generation method for Modbus/TCP |
| CN110288074A (en) * | 2018-12-04 | 2019-09-27 | 东华大学 | A kind of improvement longicorn palpus searching algorithm based on the mechanism of encirclement |
| CN110554952A (en) * | 2018-06-04 | 2019-12-10 | 西安邮电大学 | search-based hierarchical regression test data generation method |
| CN110674752A (en) * | 2019-09-25 | 2020-01-10 | 广东省智能机器人研究院 | Hidden Markov model-based tool wear state identification and prediction method |
| CN111209192A (en) * | 2019-12-30 | 2020-05-29 | 山东科技大学 | Test case automatic generation method based on double-chaos whale optimization algorithm |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10671517B2 (en) * | 2017-12-15 | 2020-06-02 | International Business Machines Corporation | Generating mobile test sequences |
| US10725899B2 (en) * | 2018-06-29 | 2020-07-28 | Wipro Limited | Method and system of performing automated exploratory testing of software applications |
-
2020
- 2020-07-03 CN CN202010631557.0A patent/CN111666221B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718368A (en) * | 2016-01-15 | 2016-06-29 | 西安邮电大学 | Software testing data amplification method |
| CN110554952A (en) * | 2018-06-04 | 2019-12-10 | 西安邮电大学 | search-based hierarchical regression test data generation method |
| CN110288074A (en) * | 2018-12-04 | 2019-09-27 | 东华大学 | A kind of improvement longicorn palpus searching algorithm based on the mechanism of encirclement |
| CN110109822A (en) * | 2019-03-30 | 2019-08-09 | 华南理工大学 | The regression testing method of priorities of test cases sequence is carried out based on ant group algorithm |
| CN110134590A (en) * | 2019-04-18 | 2019-08-16 | 上海大学 | A kind of longicorn palpus fuzz testing case generation method for Modbus/TCP |
| CN110674752A (en) * | 2019-09-25 | 2020-01-10 | 广东省智能机器人研究院 | Hidden Markov model-based tool wear state identification and prediction method |
| CN111209192A (en) * | 2019-12-30 | 2020-05-29 | 山东科技大学 | Test case automatic generation method based on double-chaos whale optimization algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 基于天牛须搜索粒子群优化的改进克里金模型;刘岩等;《国外电子测量技术》;20200415(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111666221A (en) | 2020-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111666221B (en) | Software test data set amplification method based on longicorn stigma search | |
| CN110175120B (en) | Fuzzy test case self-adaptive mutation method and device based on reinforcement learning | |
| Mauša et al. | Co-evolutionary multi-population genetic programming for classification in software defect prediction: An empirical case study | |
| CN108304316B (en) | Software defect prediction method based on collaborative migration | |
| Alshraideh et al. | A multiple-population genetic algorithm for branch coverage test data generation | |
| CN106250461A (en) | A kind of algorithm utilizing gradient lifting decision tree to carry out data mining based on Spark framework | |
| CN111240995B (en) | Multi-path coverage method and system combining key point probability and path similarity | |
| CN103559129B (en) | Statistical regression test data generating method based on genetic algorithm | |
| CN112732577A (en) | Evolution generation method for multi-task software test case | |
| CN113704771B (en) | Service vulnerability mining method based on artificial intelligence analysis and big data mining system | |
| CN114697128B (en) | Big data denoising method and big data acquisition system through artificial intelligence decision | |
| CN111782460A (en) | Large-scale log data anomaly detection method and device and storage medium | |
| CN111209192A (en) | Test case automatic generation method based on double-chaos whale optimization algorithm | |
| CN105718368A (en) | Software testing data amplification method | |
| CN104915680B (en) | Multi-tag transformation Relationship Prediction method based on Ameliorative RBF Neural Networks | |
| CN112836735A (en) | Optimized random forest processing unbalanced data set method | |
| CN115510981A (en) | Decision tree model feature importance calculation method and device and storage medium | |
| CN114238078A (en) | Method for extracting dependency relationship between programs based on high-order function | |
| Salinas et al. | Optimizing hyperparameters with conformal quantile regression | |
| CN110377525B (en) | Parallel program performance prediction system based on runtime characteristics and machine learning | |
| CN110554952B (en) | Search-based hierarchical regression test data generation method | |
| CN115062227B (en) | User behavior activity analysis method adopting artificial intelligence analysis and big data system | |
| CN114780967A (en) | Mining evaluation method based on big data vulnerability mining and AI vulnerability mining system | |
| CN112599194A (en) | Method and device for processing methylation sequencing data | |
| Zhao et al. | Plant breeding evaluation with rank entropy-based decision tree |
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 |