Reinforcement Learning (RL) is increasingly adopted to train agents that can deal with complex sequential tasks, such as driving an autonomous vehicle or controlling a complex environment. Correspondingly, novel approaches are needed to ensure that RL agents have been tested adequately before going to production. Among them, mutation testing is quite promising, especially under the assumption that the injected faults (mutations) mimic the real ones.

In this paper, we first describe a taxonomy of real RL faults obtained by repository mining. Then, we present the mutation operators derived from such real faults and implemented in the tool µPRL. Finally, we discuss the experimental results, which show that µPRL is extremely effective at discriminating strong from weak test generators, hence providing useful feedback to developers about the adequacy of the test scenarios generated and executed so far.