Pairwise Test Data Generation Based On Flower Pollination Algorithm
Main Article Content
Abstract
Owing to an exponential increase in computational time associated with increasing number of system components, exhaustive testing is increasingly become impractical. Here, many researchers opt to adopt pairwise testing to minimize the overall number of tests. Recently, many existing works are focusing on the use of Search-Based algorithms as the basis of the implementation algorithm for pairwise test suite generation; however, there is no single strategy that can be the best for all cases. Currently, researches on Flower Pollination Algorithm (FPA) are very active and its applications have been proven successful to solve many optimization problems. This paper proposes a new search-based strategy for generating the pairwise test suite, called Pairwise Flower Strategy (PairFS). The main feature of PairFS is that it is the first pairwise strategy that adopts FPA as its core implementation. To evaluate and benchmark our proposed strategy, several existing comparative experiments are adopted from the literature. The results of the experiment show that PairFS in many cases are more efficient than the existing strategies in terms of the generated pairwise test suite size.