Int. J. Advance Soft Compu. Appl, Vol. 5, No. 3, December 2013
ISSN 2074-8523; Copyright © SCRG Publication, 2013
Variable-Strength Interaction for T-Way Test
Generation Strategy
Syahrul A. C. Abdullah, Zainal H. C. Soh, and Kamal Z. Zamli
Faculty of Electrical Engineering,
Universiti Teknologi MARA (UiTM), Shah Alam, Malaysia.
e-mail: bekabox181343@salam.uitm.edu.my
Faculty of Electrical Engineering,
Universiti Teknologi MARA (UiTM), Pulau Pinang, Malaysia.
e-mail: zainal872@ppinang.uitm.edu.my
Faculty of Computer Systems & Software Engineering,
Universiti Malaysia Pahang (UMP), Pahang, Malaysia.
e-mail: kamalz@ump.edu.my
Abstract
Often, t-way testing is usually adopted to trigger faults due to interactions. As a
result, a myriad of useful t-way test generation strategies have been developed in
order to generate t-way interaction test suite that is small in size while maintaining
adequate t-way interaction coverage. Even though finding an efficient strategy to
construct an optimal test suite is very valuable, another aspect to consider is the
cost benefits of running the tests, i.e. at high level of interaction strength; the test
suite size can become enormous. To balance the need for stronger interaction with
the cost of running the tests, variable-strength interaction has been recommended.
Hence, to enable support for construction of test suite with variable-strength
interaction, a step by step procedure that extends our t-way test generation strategy,
called Test Suite Generator (TSG) is highlighted. Benchmarking results against
most existing strategies that support variable-strength interaction demonstrate that
TSG is able to give competitive results.
Keywords: combinatorial testing, software testing, t-way testing, t-way minimization
strategy, variable-strength interaction.
1 Introduction
One of the key challenges in software testing is to identify selection of input values to test as
most software systems have a very high range of valid and invalid inputs. Moreover, it is not
always feasible to test for all the input values because of constraint in cost, resources and time.
To balance between these desirable but conflicting features, software tester wishes that the
testing activities will eventually demonstrate that the software systems are fit for purpose and to