Papers › XCS as a reinforcement learning approach to automatic test case prioritization
XCS as a reinforcement learning approach to automatic test case prioritization
Lukas Rosenbauer, Anthony Stein, Roland Maier, David Pätzel, Jörg Hähner
Testing is a crucial part in the development of new products. With the rise of test automation methods, companies start relying on an even higher number of tests. Sometimes it is not feasible to run all tests and the goal is to determine which tests are crucial and which are less important. This prioritization problem has just recently gotten into the focus of reinforcement learning. A neural network combined with prioritized experience replay (ER) was used to identify critical tests. We are the first to apply XCS classifier systems (XCS) for this use case and reveal that XCS is not only suitable for this problem, but can also be superior to the aforementioned neural network and leads to more stable results. In this work, we adapt XCS's learning mechanism to the task by introducing a batch update which is based on Monte Carlo control. Further, we investigate if prioritized ER has the same positive effects on XCS as on the neural network for this test prioritization problem. Our experiments show that in general this is not the case for XCS.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
No leaderboard rows for this paper in the archive.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections