Multi objective genetic algorithm for regression

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 03 Issue: 01 | Jan-2014, Available @ http://www.ijret.org 368
MULTI-OBJECTIVE GENETIC ALGORITHM FOR REGRESSION
TESTING REDUCTION
Ravneet Kaur
M.Tech scholars Department of Computer Science and Engineering, GIMET, Amritsar, Punjab, India
Abstract
Regression Testing is type of testing which is used to cut off the directly price associated with testing of different modules.
Generally regression testing performs on test cases so that the resource utilization should be very low. Test cases in it are
challenging task to achieve and suitable test cases for regression testing are difficult to judge and process. Optimization of
various testing processes is done by optimization algorithms such as Genetic and Ant colony which usually provide solution to the
good regression testing. Still these algorithms lack of some features which are required for better optimization of test cases in
regression testing. The regression testing is the most expensive phase of the software testing, regression testing reduction
eliminates the redundant test cases in the regression testing suite and saves the cost of the regression testing. In our proposed
work, we will focus on optimization of regression testing with multi-objective genetic algorithm which will cover parameters like
simplicity and complexity for test cases for regression testing. The complexity and simplicity for test cases will be judged and
according to a common fitness function threshold we will proceed with optimization of the regression testing phases. Finally the
paper evaluates the basic genetic algorithm for optimizing the test cases based on execution time; implement the multi-objective
genetic algorithm with simplicity and complexity of the test cases along with execution time for test case prioritization for
regression testing.
Index Terms: Genetic Algorithm, Regression Testing Reduction, Test Cases, and Fitness Function etc…
--------------------------------------------------------------------***----------------------------------------------------------------------
1. INTRODUCTION
Regression testing is any type of software testing that seeks
to uncover software errors by retesting a modified program.
The intent of regression testing is to provide a general
assurance that no additional errors were introduced in the
process of fixing other problems or modifications of
software. Regression test suites are often simply test cases
that software engineers have previously developed, and that
have been saved so that they can be used later to perform
regression testing. Re executing all the test cases require
enormous amount of time thus make the testing process
inefficient. For example, one industrial collaborator reports
that for one of its products of about 20,000 lines of code, the
entire test suite requires seven weeks to run. Prioritizing test
cases provide the opportunity to maximize some
performance goals or effectiveness. One of the performance
goals may be rate of dependency detection among faults.
During software testing, pragmatic experiences show that
independent faults can be directly detected and removed, but
mutually dependent faults can be removed if and only if the
leading faults have been removed. That is, dependent faults
may not be immediately removed, and the +fault removal
process lags behind the fault detection process. For
example, if software takes limited number of inputs and
after functioning generates several types of outputs then a
single fault in input module may generate a large number of
faults in output module if they are not mutually independent.
So, in regression testing if the test cases that reveal the faults
of output module execute first and test cases reveals faults
of input module executes later then it will be delayed and in
many cases will take long time to detect the original cause
of output faults.
1.1 Unit Regression
Unit Regression will be done during Unit testing. Unit
Regression is done only when a defect is fixed in a module,
if that module has relationship with other sections within
that Module, then we perform Unit Regression. We will be
testing only the fixed part of the module.
1.2 Partial Regression
Partial regression is when regression will be done after
Impact Analysis and we would go for Partial regression
means complete Regression suite will not be executed.
2. GENETIC ALGORITHM
Genetic algorithm is a random searching method that has a
better optimization ability and internal implicit parallelism.
It can obtain, and instruct the optimized searching space and
adjust the searching direction automatically through the
optimization method of probability. Genetic algorithm is a
random searching method. After the first population is
generated, it evolves better and better approximate solutions
using the law of survival of the fittest from the generations.
An individual is chosen in every generation based on the
fitness of different individuals in certain problem domains.
A new population representing a new solution set is

_______________________________________________________________________________________
produced when different individuals combine, cross, and
vary by genetic operators in natural genetics.
Chart -1: Genetic Algorithm
2.1 TEST CASES
Test cases are the combination of testing units for software
or any coding. Normally the complete regression testing of
the codes take many resources. To cut off the resources for
the development phases, code is divided into various groups
so that partial groups can be tested on the bees of the
essential testing for partial groups. Now partial groups used
to test which save resources. Further for more resource
saving and accuracy, priority is provided to the groups on
bases of various attributes to reduce further cost of testing.
The genetic and multi genetic are similar in their approach
but difference lies in the performance of the multi objective.
The comparison can be done on the bases of fitness function
of the genetic process.
1. The simplicity of the process can be obtained with
accordance to accessing process of the test case. In our
research, we have processed the test cases for testing
reduction. Number of test cases will be used for the breaking
of testing affords. Test cases will be set as input and
according to the fitness functions we will calculate the
simplicity by find the chances of occurrence of the particular
test case. As occurrence possibility is more than simplicity
considered to be good for test case.
2. In case of complexity, we will calculate the process time
for the test case processing, if in a single iteration, the time
taken is more and in other iteration time taken is less for
same test case then we will consider it complex in
processing
2.2 Related Work
Our research will start with study of test case prioritization
process for various types of codes and find the existing
process for regression testing available. Our research will
start with developing a language code (probably in java) and
then to test it for errors. Total regression testing process will
be done on it. Once done, we will find the test cases for
complete code which will be used as test case in further
testing. Fitness function of the multi-objective genetic
algorithm will be decided on the individual fit of the test
case. We will try to reduce the regression testing resources
by reducing the test cases to have reduced test suites for
testing. Less test suite will be used for regression testing so
that resources can be saved. We will start with simple
regression strategies such as testing of whole code without
use of any test case. After that we will find simple test case
scenario which will save resources and improve efficiency
of bug finding also. Finally for providing better test case
selection process, we will process the set of test cases
created so far with multi objective genetic algorithm which
will consider the optimized filtering of the test cases.
Chart-2: Fault Finding Algorithm

_______________________________________________________________________________________
CONCLUSIONS
This paper proposed a multi-objective genetic algorithm for
regression testing reduction. The multi objective genetic
algorithm overcomes the short comes of genetic algorithm.
This paper will focus on optimization of regression testing
with multi-objective genetic algorithm which will cover
parameters like simplicity and complexity for test cases for
regression testing. The complexity and simplicity for test
cases will be judged and according to a common fitness
function threshold we will proceed with optimization of the
regression testing phases.
REFERENCES
[1] Liang You, Yansheng Lu, ‘‘A Genetic Algorithm For
The Time-Aware Regression Testing Reduction Problem’’,
8th International Conference on Natural Computation, ICNC
2012.
[2] R. Savenkov,”How to become a software tester”, Roman
Savenkov Consulting, 2004.
[3] Huang,Chin-Yu, Lin,Chu-Ti,”Software reliability
analysis by considering fault dependency and debugging
time lag”, IEEE Transactions, vol. 55(3), 2006, pp. 436-
450.
[4] Christof Budnik, Siemens,” Software Testing, Software
Quality and Trust in Software-Based Systems”,
Siemens.com.
[5] Md. Imrul Kayes, “Test Case Prioritization for
Regression Testing Based on Fault Dependency”, IEEE
International Conference of Software Engineering, February
2012.
[6] Alexey G. Malishevsky, Joseph R. Ruthruff, Gregg
Rothermel, Sebastian Elbaum, “Cost-cognizant Test Case
Prioritization,” Technical Report TR-UNL-CSE-2006-004,
Department of Computer Science and Engineering,
University of Nebraska–Lincoln, Lincoln, Nebraska, U.S.A.,
March 2006.
[7] S. Elbaum, A. Malishevsky, and G. Rothermel,“Test
case prioritization: A family of empirical studies,” IEEE
Transactions on Software Engineering, vol. 28(2), 2002, pp.
159–182.

Multi objective genetic algorithm for regression

More Related Content

What's hot (17)

Viewers also liked (20)

Similar to Multi objective genetic algorithm for regression (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Multi objective genetic algorithm for regression