An Algorithm for Generating Very Large Covering Arrays

This note describes a covering array algorithm that can be parallelized, making it possible to handle a much larger number of variables than other know algorithms. The algorithm trades test case optimization for speed ? it produces roughly 3% to 15% more tests than other known for 10 or more... See full abstract

This note describes a covering array algorithm that can be parallelized, making it possible to handle a much larger number of variables than other know algorithms. The algorithm trades test case optimization for speed ? it produces roughly 3% to 15% more tests than other known for 10 or more variables, but this ratio improves as the number of variables increases. It thus complements but does not replace other algorithms, which produce somewhat fewer tests. A significant advantage of the algorithm is that it can be distributed across any number of processors.
Hide full abstract