Developing multithreaded techniques and improved constraint handling for the tool CAgen

CAgen is a state-of-the-art combinatorial test generation tool that is known for its execution speed. In addition, it supports an extensive list of features such as constraint handling, higher-index arrays, and import and export of models/test sets in various different formats. It is based on the FIPO algorithm, which can be considered an improved version of the widely used In-Parameter-Order strategy. In order to further speed up CAgen, this work first discusses how multithreading can be effectively used to optimally utilize available resources, particularly for large instances. We evaluate three different multithreaded variations of the horizontal extension and use the obtained insights to design the mFIPOG algorithm. In addition, we adopt methods that have previously been utilized to speed up constraint handling of CSP solvers in IPO algorithms into a forbidden tuple approach. In order to evaluate the performance of the improved tool, we provide results of benchmarks on the instances offered by the CT competition of IWCT 2022.