A T-Way test suite generation strategy for sequence based interaction testing

Abstract

Complementing existing test design techniques (e.g. boundary value analysis, equivalent partitioning and cause and effect graphing), t-way testing is a test design technique that specifically used to cater bugs due to interaction. Many t-way strategies have been proposed in literature including General T-way (GTWay), In Parameter Order General (IPOG), Automatic Efficient Test Generator (AETG), and Jenny for the past 20 years. Although proposed t-way strategies have been proven to detect bugs (in many published case studies demonstrate that the effectiveness of t-way test suite comparable to exhaustive test suite), these strategies only focus on sequence-less interaction. For control and reactive system (i.e. input signals arrived at different time), the implementation of sequence-less t-way strategy is not possible. As a result, researchers nowadays start to focus on sequence based t-way strategy. However, as generating t-way test suite is an NP-Hard problem, no single strategy can claims it producing the optimal test suite for every system configuration. Motivated by the aforementioned challenges, this thesis presented a new t-way strategy, named Sequence Covering Array Test Suite Data Generation (SCATS), which support sequence based tway test suite generation. SCATS implements three main components which is sequence tree, tuple generator and test case generator in order to produe the optimum test suite size. Evaluations have been done by comparing SCATS with existing strategies with various strength( 3 ≤ t ≤ 5) and events ( 3 ≤ s ≤ 30) in term of test suite size generated. Experimental result demonstrates that in most cases SCATS produces competitive test suite size compare to other competing strategies