Automated Test Data Generation Based On Individual Constraints and Boundary Value
Testing is an important activity in software development.
Unfortunately till today testing is done manually by most of
the industry due to high cost and complexity of automation.
Automated testing can reduce the cost of software
significantly. Automated Software Test Data Generation is
an activity that in the course of software testing
automatically generates test data for the software under
test. Most of the automated test data generation uses
constraint solver to generate test data. But it cannot
generate test data when the constraints are not solvable.
Although method can be found to generate test data even if
the constraints are unsolvable, but it is poor in terms of
code coverage.
In this paper, we propose a test data generation method to
improve test coverage and to avoid the unsolvable
constraints problem. Our method is based on the individual
constraints and same or dependent variable to create the
path table which holds the information about the path
traversed by various input test data. For generating unique
test data for all the linearly independent feasible path we
created equivalence class from the path table on the basis
of path traversed by the various input test data. The input
data is taken based on individual constraints or boundary
value. Our results are compared with cyclomatic
complexity and number of possible infeasible paths. The
comparison shows the effectiveness of our method.
Keywords: Independent feasible path, scalability,
equivalence class
Download Full-Text
