Application of Genetic Algorithm in Software Testing

A Seminar on
Application of Genetic
Algorithm in
Software Testing
Presented By
Gaurav Singh CSE III yr
mmmut gkp

Software Testing
Software testing is an investigation conducted to
provide stakeholders with information about the quality of the
Product or service under test.
Properties evaluated in software testing process indicate the extent
to which the component or system under test:

meets design and development requirement.

Responds correctly to all kinds of inputs.

Performs functions within an acceptable time.

Can be installed and run in its intended environments.

Current area of research in
software testing
To devise methods to generate test data automatically, i.e. different
methods for developing test data/case generators.
Advantages of Automated test case generator

Significantly reduce the cost of developing software.

The test preparation can be done in advance.

The test runs would be considerably fast.

The confidence of the testing result can be increased.

Genetic Algorithm
A genetic algorithm starts with guesses and attempts to improve
the guesses by evolution.
A genetic algorithm will typically have five parts :-

A representation of a guess called a chromosome.

An initial pool of chromosomes.

A fitness function.

A selection function.

A crossover operator and a mutation operator.

A basic algorithm for a GA is as follows:
Initialize (population)
Evaluate (population)
While (stopping condition not satisfied) do
{
Selection (population)
Crossover (population)
Mutate (population)
Evaluate (population)
}

Aim of the algorithm
This Algorithms optimizes and select the software path clusters which are
weighted in accordance with the criticality of the path.
Why this approach to software testing:

Exhaustive software testing intractable for even medium sized software.

Typically only parts of a program can be tested, but these parts are not
necessarily the most error prone.

Therefore, we are developing a more selective approach to testing by focusing on
those parts that are most critical so that these paths can be tested first

By identifying the most critical paths, the testing efficiency can be increased.

Procedure of algorithm
1. Assigning weights to edges of CFG:

Assigning more edges to critical paths.

An initial credit is taken (100 or 10):

Dense

Sparse

At each node of CFG the incoming credit is divided and distributed to all the
outgoing edges of the node.
Distribution of weights is done as follows:

Take ‘n’ to be the number of outgoing edges.

We will considered an 80-20 rule..

If there is only one outgoing edge from a particular node than the incoming
weight is assigned to the outgoing edge.

Application of Genetic Algorithm in Software Testing

2. Selection: The selection of parents for reproduction is done according to a
Probability distribution based on the individual’s fitness values.
 Fitness Function:
 Probability of selection pj
for each path j :
 Cumulative probability ck
is calculated for each path k with equation:
F = Σi = 1 to n wi
pj = Fj / Σi = 1 to n Fi
ck = Σ j = 1 to k pj

3. Reproduction (crossover):

In one-point (or single) crossover, two input data selected as potential
parents by selection process exchange substring information at a random
position in the data to produce two new data.
 Crossover happens according to a crossover probability pc
.

After that, the selected data are formatted randomly.

Each pair of parents generates two new paths, called offspring.

4. Mutation :
 Mutation is performed on a bit-by-bit basis. Every bit of every
chromosome in the offspring has an equal chance to mutate (change from
‘0’ to ‘1’ or from ‘1’ to ‘0’).
 The mutation occurs according to a mutation probability pm
, which is
also an adjustable parameter. To perform mutation, for each
chromosome in the offspring and for each bit within the chromosome,
generate a random real number r in the range [0, 1]; if r < pm
then
mutate the bit.

Initial population: (n, m)
(12, 8), (2, 3), (6, 2), (15, 4)

Conclusion

The Genetic Algorithm is a probabilistic search algorithm that iteratively
transforms a set (population) of mathematical objects (fixed length binary string),
each with an associated fitness value, that maps onto a new population of
offspring objects using crossover and mutation techniques.

genetic algorithm are used for finding the most critical paths based on artefact's
during testing in a software construct and can improve the software testing
efficiency.

Thanks For Your Attention !
QUERY?

Application of Genetic Algorithm in Software Testing

Recommended

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Application of Genetic Algorithm in Software Testing (20)

Recently uploaded (20)

Application of Genetic Algorithm in Software Testing