Prioritizing Test Cases for Regression Testing A Model Based Approach

INTERNATIONAL JOURNAL FOR TRENDS IN ENGINEERING & TECHNOLOGY
VOLUME 4 ISSUE 1 – APRIL 2015 - ISSN: 2349 - 9303
1
Prioritizing Test Cases for Regression Testing
A Model Based Approach
Vinothkumar.N1
1
Sona College of Technology,
Department of CSE.
vinothkumaaraa@gmail.com
Galeebathullah.B2
2
Sona College of Technology,
Department of CSE.
kalifathullah@gmail.com
Abstract— Testing is an important phase of quality control of Software Development Life Cycle (SDLC). There are various
types of testing methodologies involved to test the application. Regression Testing is a type of testing, which is done to ensure whether
the modified features or bug fix had an impact over the existing functionality. Defects are identified by executing the set of test cases.
Regression Test case selection is not at all possible to conclude how much retesting is required to identify the deviation when the test
suites are larger in size. Prioritization of test cases is done to change the order of test case execution based on the severity. In the
proposed a model based approach prioritization of test cases are generated based on UML diagrams (Sequence and State Chart). The
modified features have the reflection in the model generation and the number of states and transitions covered. Prioritized test cases
are then clustered based upon the severities using dendragram approach. It leads to decrease in the time and cost of regression testing.
Index Terms— Model Based Testing, Test cases, Regression Testing, Software Engineering, Test Case Prioritization.
—————————  ——————————
1 INTRODUCTION
OFTWARE Testing is one of the important activities in software
development process. Depending on the testing method
employed, software testing can be implemented at any stage in
the development process. Traditionally, most of the test effort occurs
once the requirements had been collected and the coding process has
completed. Regression testing is used to verify the correctness of the
software under development and in maintenance process. It is very
inefficient and time consuming process to re-execute every test case
in regression testing to find the small changes over the application.
Hence model based test case prioritization is used to schedule the
test cases in a specific order that maximizes the testing efficiency.
In model based prioritization the regression test cases are
prioritized from the models and test cases are clustered based on its
severity factor. The test effort is on-going and test cases are
prioritized from the models before the coding process begins. This
technique involves constructing a graph model for the source code to
represent those controls, data dependence as well as object oriented
relations such as inheritance, polymorphism, data abstraction and
message passing. Model based approach for test case prioritization
involves model augmentation with the information such as message
paths and object states which are available from the UML design
models. Changes in the model will be reflected in UML diagrams
where states can be compared to identify the absence of any
functionality. Forward slice of the constructed model is used to
identify all the possible states in the program that may be affected
during the modification or bug fix. Test information available from
the intersection of the forward slice of the model and the changed
state is then used to prioritize the test cases.
——————————————
 Vinothkumar. N is currently pursuing a masters degree program in software
engineering in Sona College of Technology, India, PH-9865123410.
E-mail: vinothkumaaraa@gmail.com
 Galeebathullah.B is currently working as Assistant Professor in computer
science and engineering in Sona College of Technology, India, PH-
9790722912. E-mail: kalifathullah@gmail.com
Test case prioritization using a model based approach has opened
up new opportunities for software testing, which provides
prioritizing a large number of test cases by clustering factor. It
reduces the execution time of retesting of large applications and lead
to cost-effective solutions. However, it also explores challenges such
as test automation and a lack of testing deadlines to be met.
2 TEST CASE
Test Case is defined as a set of conditions or variables with step by
step procedure in a sequential flow to test the application. It contains
the set of inputs, expected output and actual output. The strategy
which is used to measure whether a software program or system has
passed or failed in a test is stored in the test report and termed as a
test oracle. In some conditions, an oracle could be a requirement or
the use case for defining the system. Various test cases are required
to determine the overall functionality of the system by incorporating
performance, load and stress test cases. Test cases are often referred
as test scripts, particularly when written are stored in a test repository
to share among the test server. A group of test cases is named as Test
suites, where all the possible test cases are stored in the execution.
Each test case should contain minimum of two test cases for
each requirement where it should consist of positive and negative
test cases. In case of any sub-requirements, each sub-requirement
must also have at least two test cases. Requirements can be mapped
to improve the efficiency of testing by mapping with the
Requirement document using Requirement Traceability Matrix. Test
cases should include a description about the performance and
functionality of the application to be tested, and the master test plan
is required to ensure that the test could be conducted formally. A
formal test-case must contain a valid input and the expected output.
Based on the types of testing going to be done, test case format will
differ, where the precondition is the input and the expected output is
the post condition to compare the actual and expected results.
S

2
3 REGRESSION TESTING
Regression testing is the process of re-executing the old test cases in
the new build to ensure that enhancements or defect fixes made to
the software works properly and does not affect the existing
functionality. It is done every time where any changes are made in
the application or any new modules are integrated in the application.
It is very essential in the system testing to test the end functionalities
by selecting the appropriate minimum set of test efforts required to
cover a particular change. Regression testing types are classified as,
i) Unit Regression-Retest
ii) Linear Regression
iii) Full Regression
4 TEST CASE PRIORITIZATION
Test Case Prioritization is the process of scheduling the test cases
which are to be executed in a specified order by assigning the values.
Test cases with the highest priority are executed first in the test
sequence. Test cases are arranged in a logical manner based on the
required conditions and run them according to their logical order. It
offers test cases having highest priority to be run earlier than the
other test cases in the test suite. Based on the criticality of the
function in an application and deadlines the test cases are prioritized.
Test cases having a highest priority need to be run earlier to prevent
blocking state and test cases having low priority can be run later.
Prioritization decreases the time consumption in test planning and
executing where it will lead to early detection of fault during the
integration.
5 RELATED WORK
Santosh Kumar Swain and Subhendu Kumar Pani [1] et al,
demonstrate the various important software models used in Model
Based Testing. A model contains a UML Based Testing, Finite State
Machines, Markov Chains and Grammars [1]. It describes a typical
model based testing process to generate test cases and automatic
execution of test cases. The Behavior of a system is difficult to
extract from the code, but it is easily obtained from design models.
State diagram represents the various states involved in the model.
Gurudiksha and Janpreet Singh [2] stated the approaches used for
regression test case prioritization and its importance in Software
Testing Life Cycle. Defines the prioritization approaches using
Dependency Structure, Model Based Prioritization, Requirement
Based Prioritization and Prioritization based on Clustering. Test case
generation processes are compared and its efficiencies are calculated
with respect to the time and cost. It compares the traditional method
which is entirely code based and used for a post implementation
phase of software development.
A.Hartman and K.Nagin [3] describes about the MBT tool AGEDIS
and its interfaces [3]. The approach includes an environment for
model test case generations, execution and model checking. It is used
for automation and widely accepted tool by industries where it
supports model based testing. AGEDIS tool is suitable for small
projects where it does not support large scale projects.
Hassan Reza et al.[4] discussed a software testing method for
dynamic websites which utilizes the model behavior of the SUT
from State chart models originally devised by Harel (1987,1988).
State chart models can be used in modeling and generating of test
cases to test the applications. However, the major focus was on the
flow of design in the module and testing those interactions. Their
approach was a systematic to test the front-end functionality of web
applications. Usability of links, forms and graphical arrangements
are mostly concerned based on the requirement specification
documents and also they describe the method of modeling the web
applications with the State chart in their work. Generation of test
cases from the State chart diagram depends on five coverage criteria:
(1) all-blobs (2) all- transitions (3) all- transitions -pairs (4) all-
conditions (5) all-paths.
Samaila Musa [5] present an evolutionary prioritized approach that
will select the best test cases from the existing test suite T, which is
used to test the original program P by using Dependence Graph as
an intermediate to identify the changes in P, at statements level.
Identification of the changes using this kind of graph will lead to the
precise detection of changes. The changed statements will be used to
identify the affected statements and test cases that execute the
affected statements are selected for regression testing. The selected
test cases will be prioritized by using genetic algorithm in order to
have a superior rate of fault detection. This approach will reduce the
cost of regression testing by increasing the rate of fault detection and
reducing the number of test cases to be used in testing the modified
program which changes the model. Methodologies available for
regression testing involve retest all, regression test selection, test
suite reduction, test case prioritization.
Sanjukta Mohany, Arup Abhinna Acharya [6], et al defined that in
code-based test case prioritization, source code of the system is used
to prioritize the test cases. Most of the test case prioritization
methods are code based. In model-based test case prioritization a
system‟s model is used to prioritize the test cases. System models are
used to capture some aspects of the system behavior. Fault can be
detected earlier in a model based testing, when compared to code-
based test case prioritization method. Model-based test prioritization
may be an efficient alternative to the existing prioritization methods.
However, model-based test case prioritization may be sensitive to the
correct/incorrect information provided by the testers/developers.
G. Rothermel, R. Untch, M. Harrol [7] described Regression testing
is the process of testing a modified system using the old test suites.
Developers need to make sure that modifications are correct and do
not adversely affect the unchanged portion of the system. During
regression testing the modified parts of the system are first tested.
Then the whole system needs to be retested using the old test suite to
have confidence that the modifications did not introduce new faults
into the system. Because of the large size of a test suite, system
retesting tends to consume a large amount of time and computing
resources; where it may last for hours, or even days. So one of the
issues developer‟s facing during retesting of the system is ordering
the test cases for execution. Test case prioritization tries to address
this issue. Test case prioritization orders tests for execution so that
the test cases with the highest priority, based on some criterion, are
executed before lower priority test cases. Several test prioritization
criteria are compared. For example, tests can be ordered to achieve
selected code coverage at the fastest rate.

3
Siripong and Jirapun [8], made researches on test case prioritization
techniques where testing consumes 40-70% of the time and cost of
software development process. Many techniques are proposed to
reduce the testing time, including test case prioritization techniques.
It introduces and organizes a new „„4C‟ classification of existing
prioritization algorithm such as Customer requirement-based
techniques, Coverage-based techniques, Cost effective-based
techniques and Chronographic history-based techniques.
Salam AL-EMARI and Izzat Mahmoud ALSMADI [9] evaluated the
usage of Spec Explorer from Microsoft Research for model based
test suite generations. It clearly defines the application functionalities
in the form of activity states where the possible states for the
modules can be re-explored to identify the modifications done on the
application. The two necessary states which are used to explore the
functionalities are the initial state and end state. The initial state is to
initiate the flow of the application and based on the GUI and final
state is the end of the FSM. The finite state machine plays an
important role in software testing to plan the flow of the test. State
graphs generated for the selected modules can be compared with the
other states to explore the input and output conditions for testing. It
is also known as finite automation to explore the available finite
number of states.
Shin Yoo, Mark Harman [10], et al performs analysis on criteria for
test case prioritization where single criterion prioritization will not
dominate other methods. They proposed a potential way to enhance a
prioritization criterion to utilize the domain expert judgement by
asking the human tester to compare the importance of different test
cases. Test cases are clustered based on Analytic Hierarchy Process
(AHP) algorithm. Using dendragram approach the test cases are
selected at defined level.
6 PROPOSED SYSTEM
The design model from the requirement defines the state variables
and rules written in the application and the states are identified by
state machine which is written in cord scripts for exploring the
models. The states are compared and prioritized based on severity.
Test case generator generates and clustering the test cases using
Agglomerative Hierarchical Process which is cost effective method
using dendragram. The following architecture describes the
schematic representation of the proposed approach. The architectural
design of the proposed system is shown in Figure 1.
The process starts with exploring the models with an application
using spec explorer. States and transitions are absorbed for the model
and state graph is generated. The model is then modified and a model
program for the model is created by configuration cords where the
test suites are stored. The machine is created for the test suite and the
model programs to explore the possible states are created.
The machine is validated for the available states and transitions
which covers code coverage results. A test code and test suite for the
model is visualized in test suite graph. The model program illustrates
the flow of events in the state based graph. Test case generator
generates the test case using test code and test is executed. Test
results and states executed can be displayed in the test view editor.
Test suites for modified model can be obtained in the test view by
refreshing the tests.
The proposed system has the following modules,
(i) Design Model
(ii) States Identification and Comparison
(iii) Flows Prioritization
(iv) Test Case Generator
(v) Prioritized Test Cases
6.1 Design Model
Fig 1. Architectural Design of the Proposed System
The design model describes the behavior of the specified
functionality in the state based diagram. Models are used to create
the design and possible to find weakness in the design before
implementation. Design models are created using spec explorer with
UML plug-in. The state based graph is generated based on the
requirements and modifications done. Whenever the changes are
made the model program has to be changed for exploring the new
FSM Model. Dependency graph can be generated to understand the
model clearly.
6.2 States Identification and Comparison
The states are explored for a given model using spec explorer. The
machine explores the number of transitions and states available for
the specified model. Test suites are generated from the defined
models. States are compared with previous versions to ensure the
modifications had done. It shows the total number of transitions and
states covered such as (S0, S1, S2….S10) with start and end states in
a graph.
6.3 Flows Prioritization
Flows of a module or event are important to achieve and understand
the functionality of the application. In this module flows are
prioritized to generate the specified test cases in order to achieve
regression test cases .Flows are prioritized to improve testing effort
where functionalities are very important in the application under test.
Based on the start and end states of the model design flow of an
event is achieved.
6.4 Test Case Generation
In this module the test suites are generated and clustered using the
hierarchical method. Test cases are prioritized and group together
based on severity factor such as Low, Medium and High. The
priority value of the test cases depends on the severity factor where
high severity test cases are prioritized first and generated according

4
to the criteria given. Generations of test cases are based on the
prioritization criteria and clustering is done to group similar test
cases having High, Medium and Low priorities.
6.5 Prioritized Test Cases
The test cases that are generated based on the prioritization criteria
for regression testing is stored in test case repositories, with a set of
test data. Test cases are grouped based on the testing to be performed
such as retest or regression test. Regression test cases are executed
on the modified application to ensure the correctness. These test
cases are compared with old test cases in the previous build to verify
the changes.
TABLE 1
SEVERITY TABLE
CONCLUSION
Model Based Testing is an effective approach to understand the
behavior of any applications. Test case generation and prioritization
play an important role in software testing. Test cases are generated
earlier in a model based approach when compared to other
approaches. It offers high reliability where changes in the models
have a direct impact on the application. Existing techniques for
prioritization of test cases are a post implementation testing process.
Fig 2.Testing Efforts
The model based approach proposed here is a pre-
implementation testing process starts at the design phase. Changes in
the requirements have a quicker effect on the models rather than
changes in coding phases. Agglomerative Clustering approach is
used to group similar test cases based on severity factor provide an
efficient method to group similar test cases. Prioritization of test
cases for multiple criteria can be improved in the future work.
REFERENCES
[1] Kumar Swain, Subhendu Kumar Pani, Durga Prasad
Mohapatra,„„Model Based Object-Oriented Software Testing” . In
Proceedings of the Journal of Theoretical and Applied Information
Technology, 2010.
[2] Gurdiksha, Janpreet Singh,„„Approaches Used for Prioritization of
Test Suites ”. In proceedings of the International journal of
Scientific Engineering and Research, 2014.
[3] A.Hartman and K.Nagin,„„The AGEDIS Tools for Model Based
Testing”, IBM Haifa Research Laboratory,ISSTA’04, ACM, 2004.
[4] Hazzan Reza, K. Ogaard and A. Malge," A model based testing
technique to test web applications using statecharts", Proceedings
of 5th International Conference on Information Technology: New
Generations, 2008.
[5] Samaila Musa,„„A Regression Test Case Selection and
Prioritization for object-oriented programs using Dependency graph
and Genetic Algorithm”. In Proceedings of the International
Journal of Engineering and Science, 2014.
[6] Sanjukta Mohanty, Arup Abhinna Acharya, Durga Prasad
Mohapatra,„„A Survey On Model Based Test Case Prioritization”.
In the proceedings of International Journal of Computer Science
and Information Technologies, 2011.
[7] G. Rothermel, R. Untch, M. Harrol, “Prioritizing Test Cases For
Regression Testing,” In the proceedings of IEEE Transactions on
Software Engineering, 2001.
[8] Siripong and Jirapun, „„Test Case Prioritization Techniques”. In the
proceedings of Journal of Theoretical and Applied Information
Technology, 2005-2010.
[9] Salam AL-EMARI, Izzat Mahmoud ALSMADI ,„„Using Spec
Explorer for Automatic Checking of Constraints in Software
controlled Systems”. In the proceedings of Informatica Economica,
2011.
[10] S. Yoo, M. Harman, P. Tonella, and A. Susi,„„Clustering Test Cases
to Achieve Effective and Scalable prioritization Incorporating
expert knowledge”. In the proceedings of International symposium
Software Testing and Analysis, 2009.
[11] R.V.Binder, „„Testing Object-Oriented Systems: Models, Patterns,
and Tools”. In the proceedings of Addison-Wesley, 1999.
[12] Logan Yu, Robert B. France, Indrakshi Ray, „„Scenario-based Static
Analysis of UML Class Models”. Springer-VerlagBerlin
Heideberg, 2008.
[13] Lin Chen, Ziyuan Wang, „„Test Case Prioritization for Web Service
Regression Testing”. In the proceedings of Fifth IEEE International
Symposium on Service Oriented System Engineering, 2010.
[14] Quart-ul-an-Farooq, M.Z, „„A Model Based Regression Testing
Approach for Evolving Software Systems with Flexible Tool
Support”. In the proceedings of IEEE, 2010.
[15] Ryan Carlson, „„A Clustering Approach to Improving Test Case
Prioritization‟‟.An Industrial Case Study. In the proceedings of
IEEE, 2011.
SEVERITY PRIORITY
S1- Low P1- Low
S2- Medium P2- Medium
S3- High P3- High
S4- Severe P4- Very High
0
1
2
3
4
1 2 3 4 5
Regressiontesting
No of Builds
Test Cases

Prioritizing Test Cases for Regression Testing A Model Based Approach

Recommended

More Related Content

What's hot (12)

Similar to Prioritizing Test Cases for Regression Testing A Model Based Approach (20)

More from IJTET Journal (20)

Recently uploaded (20)

Prioritizing Test Cases for Regression Testing A Model Based Approach