A Random search based effective algorithm for pairwise test data generation

Khatun, Sabira; Rabbi, Khandakar; Yaakub, Che Yahaya; Klaib, Mohammad F. J.

doi:10.1109/inecce.2011.5953894

Cited by 15 publications

(10 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Automatic Efficient Test Generator or AETG [9,14] and its variant mAETG [31] employ the computational approach. This approach uses 'Greedy technique' to construct test cases based on the criteria that every test case covers as many uncovered combinations as possible.…”

Section: Related Workmentioning

confidence: 99%

“…As the number of classes of elements increases, the number of interactions between the elements also increases exponentially [9][10][11] which leads to the problem of combinatorial explosion. Thus, combinatorial explosion [21,22] occurs when a huge number of possible combinations are produced by increasing the number of entities or elements, which have to interact with one another for successful functioning of a product.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Anatomy of the Tree Based Strategy for High Strength Interaction Testing

Klaib¹

2015

The 7th International Conference on Information Technology

View full text Add to dashboard Cite

Abstract-The amount of resources consumed for a complete and exhaustive testing becomes unreasonable and unaffordable. While it is vital to assure the quality and the reliability of any system, it is impossible to do an exhaustive testing due to the huge number of possible combinations. To bring a balance between exhaustive testing and lack of testing combinatorial interactions testing has been adopted. Although it is stated in literature that a complete pairwise interaction testing ensures the detection of 50-97 percent of faults, it is not sufficient to stop with pairwise testing alone for highly interactive systems. Therefore, there is a need to extend the level of testing for a general multi way combinatorial interactions testing. This paper enhanced the previous strategies "A tree based strategy for test data generation and cost calculation" and "3-way interaction testing using the tree strategy" to support a general multi-way combinatorial interaction testing involving uniform and non uniform parametric values. In this strategy, two algorithms have been adopted; a tree construction algorithm which constructs the possible test cases and an iterative cost calculation algorithm that constructs efficient multiway test suites which cover all parameter interactions between input components. Both algorithms are presented in details.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Anatomy of the Tree Based Strategy for High Strength Interaction Testing

Klaib¹

2015

The 7th International Conference on Information Technology

View full text Add to dashboard Cite

show abstract

“…The set of ad-hoc techniques contains those which are not using any of the other techniques. An adhoc approach typically selects the test cases randomly or on the basis of some input distribution (e.g., [53] [54]).…”

Section: Bmentioning

confidence: 99%

An Orchestrated Survey of Available Algorithms and Tools for Combinatorial Testing

Khalsa

Labiche

2014

2014 IEEE 25th International Symposium on Software Reliability Engineering

View full text Add to dashboard Cite

For functional testing based on the input domain of a functionality, parameters and their values are identified and a test suite is generated using a criterion exercising combinations of those parameters and values. Since software systems are large, resulting in large numbers of parameters and values, a technique based on combinatorics called Combinatorial Testing (CT) is used to automate the process of creating those combinations. CT is typically performed with the help of combinatorial objects called Covering Arrays. The goal of the present work is to determine available algorithms/tools for generating a combinatorial test suite. We tried to be as complete as possible by using a precise protocol for selecting papers describing those algorithms/tools. The 75 algorithms/tools we identified are then categorized on the basis of different comparison criteria, including: the test suite generation technique, the support for selection (combination) criteria, mixed covering array, the strength of coverage, and the support for constraints between parameters. Results can be of interest to researchers or software companies who are looking for a CT algorithm/tool suitable for their needs.

show abstract

“…The set of ad-hoc techniques contains those tools and algorithms that are not using any of the other techniques. An ad-hoc approach typically selects the test cases randomly or on the basis of some input distribution (e.g., [94,95]).…”

Section: Rq2: What Covering Array Generation Techniques Are Used By Tmentioning

confidence: 99%

“…Yes CATS [140] Yes Test Cover [168] Yes Algebraic method [169] yes Deterministic density algorithm [141] Yes Density Based Greedy [142] Yes DA-RO [143] Yes Yes Yes DA-FO [143] Yes Yes Yes Test Case Generator [132], Yes R2Way [94] Yes ART-CT [93] Yes Yes MIPOG [161] Yes ITTDG [104] Yes Yes Yes AURA [144] Yes Yes Yes Harmony search strategy [65] Yes Yes * Particle Swarm Test Generator VS-PSTG [85] yes Yes* HSTCG [158] Yes Yes* EXACT [164,165] Yes Branch and Bound [166] Yes* Yes Tabu Search [159] Yes* Yes Distance Based Technique [79] Yes Yes Yes T-Gen SYS/3 -a Software Development Tool [92] Yes Sequence Covering Array Generator [145] Yes S.No Algorithm/tool Maximum Strength support (t) Configurations Ant Colony System(ACS) [157] 3 VSCA(N,2,3 20 ,10 2 , {MCA(3,3 20 ,10 2 )}) Greedy Algorithm with Hill Climbing [86] MCA(N, t, 2 10 ,3 3 ,4 2 ,5 ) Greedy Algorithm with Simulated annealing [86] 4…”

Section: Future Workmentioning

confidence: 99%

An Analysis and Extension of Category Partition Testing in the Presence of Constraints

Khalsa¹

View full text Add to dashboard Cite

Category Partition (CP) is a Black Box testing technique that formalizes the specification of the input domain of the system under test. A CP specification is driven by the tester's expertise and comprises of parameters, categories (characteristics of parameters) and choices (acceptable values for categories) required for extensively testing the system. To ensure completeness, choices not only correspond to permitted input values but also correspond to values that account for boundaries or robustness. These choices are combined on the basis of various selection criterion (e.g., Base Choice, Pairwise) to form test frames which given input values form test cases. To ensure that the combinations of choices are feasible and account for valid sets of user requirements, constraints are introduced. A constraint can be a specification for permitted choice combination or a choice annotation as Error or Single. In a typical development environment where testing is driven by stringent deadlines, a tester might have to decide how many constraints (constraints among choices, Error or Single annotation) are enough to attain the maximum level of test completeness. The present work will assist a test engineer in making this decision. This thesis contributes by concluding, based on experimental evaluation of academic and industrial case studies, that in case of limited resources an equally effective test suite can be attained by meticulously defining Error and Single annotations in a CP specification. The present work also contributes by challenging the notion that introducing constraints reduces the cost of the test suite by restricting the combination of choices. This thesis asserts, based on experimental evaluation, that introducing constraints does not always reduce cost and that the cost of the test suite depends on various other factors. iv Combining choices while accounting for constraints and forming each choice adequate test set, for instance, is achievable, thanks to Constrained Covering Arrays from combinatorial testing. But, Base Choice (BC) criterion besides being an integral and domain specific criterion has not been defined to specifically account for constraints on choices. This thesis contributes by introducing two extensions to the BC criterion: Constrained Base Choice and Extended Constrained Base Choice, to account for constraints resulting from complex input domain. A number of academic and industrial case studies have been used to compare different selection criteria, including the new ones, in terms of data flow coverage, control flow coverage, cost and effectiveness at finding faults. Results show the performance of the new criteria equivalent to a 3-way selection criterion with a much smaller cost. The use of CP on large systems can result in a large number of parameters (categories) and values (choices). A technique based on combinatorics called Combinatorial Testing (CT) is then used to automate the process of creating those combinations of choices to form test frames. CT is typically performed with the...

show abstract

A Random search based effective algorithm for pairwise test data generation

Cited by 15 publications

References 13 publications

Anatomy of the Tree Based Strategy for High Strength Interaction Testing

Anatomy of the Tree Based Strategy for High Strength Interaction Testing

An Orchestrated Survey of Available Algorithms and Tools for Combinatorial Testing

An Analysis and Extension of Category Partition Testing in the Presence of Constraints

Contact Info

Product

Resources

About