CN114036069A - Multi-target test case sequencing method based on decomposition weight vector self-adaption - Google Patents

Multi-target test case sequencing method based on decomposition weight vector self-adaption Download PDF

Info

Publication number
CN114036069A
CN114036069A CN202111522112.XA CN202111522112A CN114036069A CN 114036069 A CN114036069 A CN 114036069A CN 202111522112 A CN202111522112 A CN 202111522112A CN 114036069 A CN114036069 A CN 114036069A
Authority
CN
China
Prior art keywords
test case
weight vector
target
sequence
test
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.)
Pending
Application number
CN202111522112.XA
Other languages
Chinese (zh)
Inventor
陈信
殷嘉铖
俞东进
范旭麟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Dianzi University
Original Assignee
Hangzhou Dianzi University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hangzhou Dianzi University filed Critical Hangzhou Dianzi University
Priority to CN202111522112.XA priority Critical patent/CN114036069A/en
Publication of CN114036069A publication Critical patent/CN114036069A/en
Pending legal-status Critical Current

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
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Biophysics (AREA)
  • Computer Hardware Design (AREA)
  • Quality & Reliability (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computational Linguistics (AREA)
  • Genetics & Genomics (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Physiology (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Complex Calculations (AREA)

Abstract

The invention discloses a decomposition-based weight vector self-adaptive multi-target test case sequencing method, which creatively introduces an MOEA/D-VW algorithm aiming at the problems of single target, high time cost, non-ideal effect and the like in the conventional test case sequencing problem and takes average branch coverage rate and effective execution time as optimization targets. The method can effectively realize multi-target test case sequencing in a shorter time. The invention improves the early detection rate of the defects and effectively reduces the regression testing cost. In the method level, the traditional multi-objective optimization algorithm is improved, and a self-adaptive weight vector change strategy is added, so that the result has better distribution, and the obtained test case sequencing sequence has more selectivity.

Description

Multi-target test case sequencing method based on decomposition weight vector self-adaption
Technical Field
The invention relates to the field of software maintenance, in particular to a multi-target test case sequencing method based on self-adaption of a weight vector of decomposition.
Background
During software updates and evolution, testers are often required to perform more diverse software regression tests. The software regression test means that after the software introduces a new function, a tester must test the original function again to ensure that the introduction of the new function does not affect the original function. The test case priority ranking is to rank the test cases which need to be subjected to regression testing so as to improve the early detection rate of the defects and reduce the testing cost. With the continuous improvement of the industrial test requirements, in the actual test process, the influence of various factors on the software quality, such as test cost, time, code coverage, repair difficulty and the like, needs to be considered in the test case sequencing. The Multi-Objective Test Case priority ordering problem (MOTCP) is an important problem to be solved urgently in the current regression Test.
Currently, researchers mostly adopt greedy strategies or heuristic algorithms to solve the MoTCP. Yoo et al, based on the Pareto idea in the process of selecting test cases, set the priority of the test cases by using code coverage, test case execution cost and the like as optimization targets and respectively adopting multi-objective optimization methods such as NSGA-II and greedy strategy, although NSGA-II has high running speed and good convergence of solution set, the congestion degree calculation process in each iteration process is complex and the algorithm performance is unsatisfactory. Persevere and the like introduce a CPU and a GPU into the MOTCP problem and solve the MOTCP problem by using a NSGA-II pre-optimization technology, so that the algorithm efficiency is obviously improved, but optimization is still performed by using a dominance-based principle, the algorithm complexity is higher, and the performance is not ideal when a test case set is larger. The method is characterized in that a program is used for performing the operation of MOTCP by using an ant colony algorithm, a gene segment principle in biology is introduced, the mode of updating pheromone is expanded, and a good effect is achieved in solving the MOTCP problem. Additionally, Megala Tulasiraman et al used a peer-to-peer optimal clone selection algorithm, Tyagi Manika et al used a discrete particle swarm algorithm, and Deb Kalyanmoy used the NSGA-II algorithm with elite strategy to rank the test cases. The algorithms are realized according to a multi-objective optimization algorithm based on dominance in principle, the multi-objective optimization algorithm based on dominance is higher in complexity than a decomposition-based method, and meanwhile, preference information of users cannot be introduced. The invention creates a MOEA/D-VW introducing self-adaptive multi-objective optimization algorithm. In the process of solving MOTCP, the method can effectively improve regression testing efficiency and is beneficial to finding software defects as early as possible.
Disclosure of Invention
When software regression testing is performed, due to the influence of multiple factors such as testing cost and time, multiple testing targets generally need to be met, and therefore the quality of the sequencing result of one test case set needs to be evaluated from multiple indexes. In the multi-target test case sequencing, a plurality of targets are difficult to achieve the optimum simultaneously, a non-dominated solution set of the problem needs to be solved, but the non-dominated solution set obtained by the traditional method is not wide enough in distribution and poor in adaptive value, so that a more efficient multi-target optimization algorithm needs to be found for solving the problem.
The technical scheme of the invention is as follows: a multi-target test case sequencing method based on decomposition weight vector self-adaption comprises the following steps:
s1, supposing that a certain program to be tested has m branches, performing regression testing on the program by using n test cases, and expressing the test case set as psi ═ { T1，T2，...，TnWhere T isi(i is more than or equal to 1 and less than or equal to n) is the ith test case in the test case set, and a branch coverage matrix A is constructedn×mAnd the effective execution time vector V ═ t1，t2，...，tn) Wherein t isi(i ═ 1, 2.. times, n) is the effective execution time corresponding to the ith test case, and if the jth branch is covered in the execution of the ith test case, a is the effective execution time corresponding to the jth branch ij1, otherwise Aij＝0；
S2, encoding: giving a test program and a test case set, numbering each test case in the set by 1 to n, wherein a test case priority sequence is the full arrangement of test case numbers;
s3, setting a fitness function:
s4, weight vector initialization:
s5, initialization:
s6, an iteration process: let gen be 0; according to each weight vector wuAnd fitness functionSet F (x) get Q subproblems ζu(x)，u＝1，2，...，Q；
S7, if gen is less than or equal to maxGen, returning to S6, otherwise, turning to S8;
s8, outputting the EP, wherein the sequence corresponding to each individual in the EP is the sequence of the optimized test cases; the user selects the appropriate solution based on the preference for the target.
Further, the specific method of S3 is as follows:
selecting Average Branch Coverage of Branch Coverage (APBC) and Effective Execution Time (EET) as optimization targets to measure the quality of a test case priority sequence; assuming that the first n '(n' ≦ n) test cases on the sequence are executed, all branches in the test program can be covered, thereby constructing a fitness function set f (x), f (x) ═ APBC, EET, where
Figure BDA0003408023110000031
Figure BDA0003408023110000032
Figure BDA0003408023110000033
TBjRepresenting the position of the test case of the jth branch in the first covering program in the test case priority sequence;
further, the specific method of S4 is as follows:
creating a set E ═ E containing l two-dimensional unit vectors1，e2，...，elAs candidate set of weight vectors, and an identity matrix
Figure BDA0003408023110000041
Calculating each unit vector
Figure BDA0003408023110000042
The sum of Euclidean distances to each sub-vector in W is selected, and the unit vector e with the largest sum of the distances is selectedsAdding to W and removing it from ERemoving; this operation is repeated; when the number of unit vectors in W reaches Q (Q is the number of solutions in the final solution set expected to be obtained, and Q is less than or equal to L +2), stopping the operation to obtain a final weight vector set L;
further, the specific method of S5 is as follows:
setting the maximum iteration times maxGen; randomly generating a scale N (N ═ m)*n) of the initial population P0＝{x1，x2，...，xN}; calculating each individual
Figure BDA0003408023110000043
Figure BDA0003408023110000044
Is a full permutation of n test cases) and each weight vector wu(wuE.g. L), find each weight vector w fromuNearest npIndividual and marked as wuNeighborhood of (2)
Figure BDA0003408023110000045
Setting an initial ideal point z*When the storage container EP is initialized to (0, 0):
further, the specific method of S6 is as follows:
s61, a genetic process: let iter equal to 0;
s611, selecting: for each sub-question ζuCalculate its neighborhood BuEach of the individuals xuv(v＝1，2，...，np) Fitness f ofuvFurther calculating the probability of its selection
Figure BDA0003408023110000046
Figure BDA0003408023110000047
Two individuals x are selected using a roulette algorithmuaAnd xub；
S612, crossing: by location-based crossover Algorithm (PBX) pairs xuaAnd xubExecuteThe cross operation comprises the following specific processes: random generation
Figure BDA0003408023110000051
Each different integer
Figure BDA0003408023110000052
And is
Figure BDA0003408023110000053
Figure BDA0003408023110000054
Generating sequences
Figure BDA0003408023110000055
In
Figure BDA0003408023110000056
Value of position and xuaIn
Figure BDA0003408023110000057
The values of the positions are the same; find xuaIn
Figure BDA0003408023110000058
Position value in xubIn (c) and then xubThe values corresponding to the rest positions are put in sequence
Figure BDA0003408023110000059
In the free position of (2), a new sequence is obtained
Figure BDA00034080231100000510
By the same principle
Figure BDA00034080231100000511
S613, variation: to pair
Figure BDA00034080231100000512
Performing mutation operations separately, i.e. on
Figure BDA00034080231100000513
And
Figure BDA00034080231100000514
respectively and randomly generating two positions, and exchanging values at the two positions to obtain
Figure BDA00034080231100000515
Figure BDA00034080231100000516
S614, updating: comparison
Figure BDA00034080231100000517
And
Figure BDA00034080231100000518
size of (1), if
Figure BDA00034080231100000519
Figure BDA00034080231100000520
Then remember
Figure BDA00034080231100000521
Otherwise, record
Figure BDA00034080231100000522
Will be provided with
Figure BDA00034080231100000523
And xuaPerforming Chebyshev comparison if
Figure BDA00034080231100000524
Then use
Figure BDA00034080231100000525
Replacement of xua(ii) a Comparison of the same theory
Figure BDA00034080231100000526
And
Figure BDA00034080231100000527
size of (1), if
Figure BDA00034080231100000528
Then remember
Figure BDA00034080231100000529
Figure BDA00034080231100000530
Otherwise, record
Figure BDA00034080231100000531
Will be provided with
Figure BDA00034080231100000532
And xubPerforming Chebyshev comparison if
Figure BDA00034080231100000533
Then use
Figure BDA00034080231100000534
Replacement of xub；iter＝iter+1；
S615, if
Figure BDA00034080231100000535
This returns to S611;
s62, selecting optimal x from each neighborhooduvPut into EP and delete all quilt x from EPuvPareto dominant individuals;
s63, updating the ideal point z*: comparing each individual
Figure BDA00034080231100000539
The fitness value of each target in the F (x) target space, and taking the minimum value as z*A value in that dimension;
s64, updating the weight vector:
Figure BDA00034080231100000536
Figure BDA00034080231100000537
and unitize it
Figure BDA00034080231100000538
S65, updating each subproblem zetau(x) Neighborhood B ofu(ii) a Setting gen as gen + 1;
compared with the traditional classification method, the method has the following creativity:
1. the invention creatively uses the multi-target optimization algorithm based on decomposition to process the multi-target test case priority ordering problem, and the performance is better.
2. In the method level, the traditional multi-objective optimization algorithm is improved, and a self-adaptive weight vector change strategy is added, so that the result has better distribution, and the obtained test case sequencing sequence has more selectivity.
Drawings
FIG. 1 is a flow chart of the present invention;
figure 2 is a depiction of a PBX crossover process;
Detailed Description
Taking the test case priority ranking of the JavaScript unit test framework Jasmine as an example, the specific implementation of the multi-target test case priority ranking method of the regression test provided by the invention is explained by combining the attached figure 1.
S1, aiming at a program to be tested Jasmine, using n-24 test cases as a regression test suite. The Jasmine source program has 95 branches, and represents a test case set as Ψ ═ T1，T2，...，Ti，...，T24Where T isi(i is more than or equal to 1 and less than or equal to 24) is the ith test case in the test case set, and a branch coverage matrix A is constructed24×95And the effective execution time vector V ═ t1，t2，...，ti，...，t24) Wherein t isi(i 1, 2.., 24) isCorresponding to the effective execution time of the ith test case, if the jth branch is covered in the execution of the ith test case, A ij1, otherwise Aij＝0；
S2, encoding: given a test program and a test case set, each test case in the set is numbered with 1 to 24, and the test case priority sequence is the full arrangement of the test case numbers.
S3, setting a fitness function: selecting Average Branch Coverage of Branch Coverage (APBC) and Effective Execution Time (EET) as optimization targets to measure the quality of a test case priority sequence; assuming that the first n '(n' ≦ 24) test cases on the sequence are executed, all branches in the test program can be covered, thereby constructing a fitness function set f (x), f (x) ═ APBC, EET, where
Figure BDA0003408023110000071
Figure BDA0003408023110000072
TBjRepresenting the position of the test case of the jth branch in the first covering program in the test case priority sequence;
s4, weight vector initialization: creating a set E-E comprising l-2000 two-dimensional unit vectors1，e2，...，e2000As candidate set of weight vectors, and an identity matrix
Figure BDA0003408023110000073
Calculate each unit vector ek(k ═ 1, 2.., 2000) and the sum of euclidean distances for each subvector in W, and the unit vector e with the largest sum of distances is selectedsAdded to W and removed from E; this operation is repeated; when the number of unit vectors in W reaches Q ═ 10 (Q is the number of solutions in the final solution set desired to be obtained), the operation is stopped, and a final weight vector set L is obtained;
s5, initialization: setting the maximum iteration number maxGen as 300; randomly generating oneInitial population P of size N1500＝{x1，x2，...，x150}; calculating each individual
Figure BDA0003408023110000074
Figure BDA0003408023110000075
Is a full permutation of n test cases) and each weight vector wu(wuE.g. L), find each weight vector w fromuThe last 10 individuals and labeled as wuNeighborhood of (2)
Figure BDA0003408023110000076
Setting an initial ideal point z*Initializing the storage container EP Φ (0, 0);
s6, an iteration process: let gen be 0; according to each weight vector wuAnd the fitness function set F (x) obtain Q subproblems zetau(x)，u＝1，2，...，10；
S61, a genetic process: let iter equal to 0;
s611, selecting: for each sub-question ζuCalculate its neighborhood BuEach of the individuals xuv(v＝1，2，...，np) Fitness f ofuvFurther calculating the probability of its selection
Figure BDA0003408023110000081
Figure BDA0003408023110000082
Two individuals x are selected using a roulette algorithmuaAnd xub；
S612, crossing: by location-based crossover Algorithm (PBX) pairs xuaAnd xubPerforming a crossover operation as follows (fig. 2 is a specific process of the PBX crossover algorithm, and for clarity of description, the size of simulation n is 10); randomly generating 6 mutually different integers
Figure BDA0003408023110000083
And is
Figure BDA0003408023110000084
Generating a new sequence
Figure BDA0003408023110000085
Figure BDA0003408023110000086
In
Figure BDA0003408023110000087
Value of position and xuaIn
Figure BDA0003408023110000088
The position values are the same, so that the position of the selected gene in the filial generation is ensured to be the same as that of the parent generation; find xuaIn
Figure BDA0003408023110000089
Position value in xubIn (c) and then xubThe values corresponding to the rest positions are put in sequence
Figure BDA00034080231100000810
In the free position of (2), a new sequence is obtained
Figure BDA00034080231100000811
Generating sequences
Figure BDA00034080231100000812
In
Figure BDA00034080231100000813
Value of position and xubIn
Figure BDA00034080231100000814
The values of the positions are the same; find xubIn
Figure BDA00034080231100000815
Position value in xuaIn the position (a) of (b),then x is putuaThe values corresponding to the rest positions are put in sequence
Figure BDA00034080231100000816
In the free position of (2), a new sequence is obtained
Figure BDA00034080231100000817
S613, variation: to pair
Figure BDA00034080231100000818
Performing mutation operations separately, i.e. on
Figure BDA00034080231100000819
Or
Figure BDA00034080231100000820
Two positions are randomly generated, and values at the two positions are exchanged to obtain
Figure BDA00034080231100000821
S614, updating: comparison
Figure BDA00034080231100000822
And
Figure BDA00034080231100000823
size of (1), if
Figure BDA00034080231100000824
Figure BDA00034080231100000825
Then remember
Figure BDA00034080231100000826
Otherwise, record
Figure BDA00034080231100000827
Will be provided with
Figure BDA00034080231100000828
And xuaPerforming Chebyshev comparison if
Figure BDA00034080231100000829
Then use
Figure BDA00034080231100000830
Replacement of xua(ii) a Comparison of the same theory
Figure BDA00034080231100000831
And
Figure BDA00034080231100000832
size of (1), if
Figure BDA00034080231100000833
Then remember
Figure BDA00034080231100000834
Otherwise, record
Figure BDA00034080231100000835
Will be provided with
Figure BDA00034080231100000836
And xubPerforming Chebyshev comparison if
Figure BDA0003408023110000091
Then use
Figure BDA0003408023110000092
Replacement of xub；iter＝iter+1；
S615, if iter is less than 12, returning to S611;
s62, selecting optimal x from each neighborhooduvPut into EP and delete all quilt x from EPuvPareto dominant individuals;
s63, updating the ideal point z*: comparing each individual xkpThe fitness value of each target in the F (x) target space, and taking the minimum value as z*In this dimensionA value of (d);
s64, updating the weight vector:
Figure BDA0003408023110000093
Figure BDA0003408023110000094
and unitize it
Figure BDA0003408023110000095
S65, updating each subproblem zetau(x) Neighborhood B ofu，gen＝gen+1；
S7, if gen is less than or equal to maxGen, returning to S6, otherwise, turning to S8;
s8, outputting the EP, wherein the sequence corresponding to each individual in the EP is the sequence of the optimized test cases; the user may select an appropriate solution based on the preference for the target. A decomposition-based multi-target test case sequencing algorithm combined with a self-adaptive weight vector aims at branch coverage information of an Iasmine program, 300 iterations are adopted in each experiment, and independent execution is repeated for 30 times. Example results show that the test case sequencing scheme obtained by the multi-target test case sequencing algorithm based on decomposition and combined with the self-adaptive weight vector has the advantages of good quality, low time cost and wide solution set distribution range, and is an effective regression test case priority sequencing method.

Claims (5)

1. The multi-target test case sequencing method based on the self-adaption of the decomposed weight vector is characterized by comprising the following steps of:
s1, supposing that a certain program to be tested has m branches, performing regression testing on the program by using n test cases, and expressing the test case set as psi ═ { T1，T2，...，TnWhere T isi(i is more than or equal to 1 and less than or equal to n) is the ith test case in the test case set, and a branch coverage matrix A is constructedn×mAnd the effective execution time vector V ═ t1，t2，...，tn) Wherein t isi(i ═ 1, 2.. times, n) is the effective execution time corresponding to the ith test case, and if the jth branch is covered in the execution of the ith test case, a is the effective execution time corresponding to the jth branchij1, otherwise Aij＝0；
S2, encoding: giving a test program and a test case set, numbering each test case in the set by 1 to n, wherein a test case priority sequence is the full arrangement of test case numbers;
s3, setting a fitness function:
s4, weight vector initialization:
s5, initialization:
s6, an iteration process: let gen be 0; according to each weight vector wuAnd the fitness function set F (x) obtain Q subproblems zetau(x)，u＝1，2，...，Q；
S7, if gen is less than or equal to maxGen, returning to S6, otherwise, turning to S8;
s8, outputting the EP, wherein the sequence corresponding to each individual in the EP is the sequence of the optimized test cases; the user selects the appropriate solution based on the preference for the target.
2. The decomposition-based weight vector adaptive multi-target test case sequencing method according to claim 1, wherein the S3 is specifically as follows:
selecting Average Branch Coverage of Branch Coverage (APBC) and Effective Execution Time (EET) as optimization targets to measure the quality of a test case priority sequence; assuming that the first n '(n' ≦ n) test cases on the sequence are executed, all branches in the test program can be covered, thereby constructing a fitness function set f (x), f (x) ═ APBC, EET, where
Figure FDA0003408023100000021
Figure FDA0003408023100000022
Figure FDA0003408023100000023
TBjAnd the position of the test case of the jth branch in the first covering program in the test case priority sequence is shown.
3. The decomposition-based weight vector adaptive multi-target test case sequencing method according to claim 1, wherein the S4 is specifically as follows:
creating a set E ═ E containing l two-dimensional unit vectors1，e2，...，elAs candidate set of weight vectors, and an identity matrix
Figure FDA0003408023100000024
Calculating each unit vector
Figure FDA0003408023100000025
The sum of Euclidean distances to each sub-vector in W is selected, and the unit vector e with the largest sum of the distances is selectedsAdded to W and removed from E; this operation is repeated; when the number of the unit vectors in W reaches Q, Q is the number of the solutions in the final solution set expected to be obtained, and Q is less than or equal to L +2, the operation is stopped, and a final weight vector set L is obtained.
4. The decomposition-based weight vector adaptive multi-target test case sequencing method according to claim 1, wherein the S5 is specifically as follows:
setting the maximum iteration times maxGen; randomly generating an initial population P of size N (N-m N)0＝{x1，x2，...，xN}; calculating each individual
Figure FDA0003408023100000026
(
Figure FDA0003408023100000027
Figure FDA0003408023100000028
Is a full permutation of n test cases) and each weight vector wu(wuE.g. L), find each weight vector w fromuNearest npIndividual and marked as wuNeighborhood of (2)
Figure FDA0003408023100000029
Setting an initial ideal point z*When the storage container EP is initialized to (0, 0), the storage container EP is initialized to Φ.
5. The decomposition-based weight vector adaptive multi-target test case sequencing method according to claim 4, wherein the S6 is specifically as follows:
s61, a genetic process: let iter equal to 0;
s611, selecting: for each sub-question ζuCalculate its neighborhood BuEach of the individuals xuv(v＝1，2，...，np) Fitness f ofuvFurther calculating the probability of its selection
Figure FDA0003408023100000031
Figure FDA0003408023100000032
Two individuals x are selected using a roulette algorithmuaAnd xub；
S612, crossing: by location-based crossover Algorithm (PBX) pairs xuaAnd xubAnd performing cross operation, wherein the specific process is as follows: random generation
Figure FDA0003408023100000033
Each different integer
Figure FDA0003408023100000034
And is
Figure FDA0003408023100000035
Figure FDA0003408023100000036
Generating sequences
Figure FDA0003408023100000037
Figure FDA0003408023100000038
In
Figure FDA0003408023100000039
Value of position and xuaIn
Figure FDA00034080231000000310
The values of the positions are the same; find xuaIn
Figure FDA00034080231000000311
Position value in xubIn (c) and then xubThe values corresponding to the rest positions are put in sequence
Figure FDA00034080231000000312
In the free position of (2), a new sequence is obtained
Figure FDA00034080231000000313
By the same principle
Figure FDA00034080231000000314
S613, variation: to pair
Figure FDA00034080231000000315
Performing mutation operations separately, i.e. on
Figure FDA00034080231000000316
And
Figure FDA00034080231000000317
respectively and randomly generating two positions, and exchanging values at the two positions to obtain
Figure FDA00034080231000000318
Figure FDA00034080231000000319
S614, updating: comparison
Figure FDA00034080231000000320
And
Figure FDA00034080231000000321
size of (1), if
Figure FDA00034080231000000322
Figure FDA00034080231000000323
Then remember
Figure FDA00034080231000000324
Otherwise, record
Figure FDA00034080231000000325
Will be provided with
Figure FDA00034080231000000326
And xuaPerforming Chebyshev comparison if
Figure FDA00034080231000000327
Then use
Figure FDA00034080231000000328
Replacement of xua(ii) a Comparison of the same theory
Figure FDA00034080231000000329
And
Figure FDA00034080231000000330
size of (1), if
Figure FDA00034080231000000331
Then remember
Figure FDA00034080231000000332
Figure FDA00034080231000000333
Otherwise, record
Figure FDA00034080231000000334
Will be provided with
Figure FDA00034080231000000335
And xubPerforming Chebyshev comparison if
Figure FDA00034080231000000336
Then use
Figure FDA00034080231000000337
Replacement of xub；iter＝iter+1；
S615, if
Figure FDA00034080231000000338
This returns to S611;
s62, selecting optimal x from each neighborhooduvPut into EP and delete all quilt x from EPuvPareto dominant individuals;
s63, updating the ideal point z*: comparing each individual
Figure FDA0003408023100000041
The fitness value of each target in the F (x) target space, and taking the minimum value as z*In the dimensionA value of (d) above;
s64, updating the weight vector:
Figure FDA0003408023100000042
Figure FDA0003408023100000043
and unitize it
Figure FDA0003408023100000044
S65, updating each subproblem zetau(x) Neighborhood B ofu(ii) a Gen + 1.
CN202111522112.XA 2021-12-13 2021-12-13 Multi-target test case sequencing method based on decomposition weight vector self-adaption Pending CN114036069A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111522112.XA CN114036069A (en) 2021-12-13 2021-12-13 Multi-target test case sequencing method based on decomposition weight vector self-adaption

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111522112.XA CN114036069A (en) 2021-12-13 2021-12-13 Multi-target test case sequencing method based on decomposition weight vector self-adaption

Publications (1)

Publication Number Publication Date
CN114036069A true CN114036069A (en) 2022-02-11

Family

ID=80140632

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111522112.XA Pending CN114036069A (en) 2021-12-13 2021-12-13 Multi-target test case sequencing method based on decomposition weight vector self-adaption

Country Status (1)

Country Link
CN (1) CN114036069A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115809202A (en) * 2023-01-04 2023-03-17 南京邮电大学 Test case priority ordering method facing parameter value switching

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115809202A (en) * 2023-01-04 2023-03-17 南京邮电大学 Test case priority ordering method facing parameter value switching

Similar Documents

Publication Publication Date Title
Jiang et al. Efficient network architecture search via multiobjective particle swarm optimization based on decomposition
Higashi et al. Particle swarm optimization with Gaussian mutation
Menhas et al. Comparative performance analysis of various binary coded PSO algorithms in multivariable PID controller design
CN108038538A (en) Multi-objective Evolutionary Algorithm based on intensified learning
CN112700060A (en) Station terminal load prediction method and prediction device
Rattray The dynamics of a genetic algorithm under stabilizing selection
Krömer et al. Differential evolution for the optimization of low-discrepancy generalized Halton sequences
CN114036069A (en) Multi-target test case sequencing method based on decomposition weight vector self-adaption
Kommadath et al. Parallel computing strategies for sanitized teaching learning based optimization
Broni-Bediako et al. Evolutionary NAS with gene expression programming of cellular encoding
Minerva et al. Building ARMA models with genetic algorithms
Lin et al. An efficient hybrid Taguchi-genetic algorithm for protein folding simulation
Olariu et al. Biology-derived algorithms in engineering optimization
CN113141272B (en) Network security situation analysis method based on iteration optimization RBF neural network
CN115293430A (en) Unmanned node cooperation method and system based on cooperative coevolution algorithm
Chang et al. Simpler and enhanced multifactorial evolutionary algorithms for continuous optimization tasks
Panwar et al. Genetic algorithm for solving simple mathematical equality problem
Ali et al. Balancing search direction in cultural algorithm for enhanced global numerical optimization
Gál et al. Modified bacterial memetic algorithm used for fuzzy rule base extraction
Sivaraj et al. An efficient grouping genetic algorithm
Hedar et al. Genetic algorithm with population partitioning and space reduction for high dimensional problems
Wang et al. A hybrid genetic algorithm for multiobjective structural optimization
Pericleous et al. Generic hybridization of MOEA/D with learning for permutation flow shop scheduling problem
de Lacerda et al. A study of cross-validation and bootstrap as objective functions for genetic algorithms
Zhang et al. Local Search Enhanced Multi-Objective Evolutionary Algorithm for Fuzzy Flexible Job Shop Scheduling

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