Falsification of Cyber-Physical Systems Using Deep Reinforcement Learning
This addresses the challenge of practical falsification for CPS in domains like smart grids and autonomous vehicles, though it appears incremental as it builds on existing robustness-guided methods.
The paper tackles the problem of efficiently detecting defects in Cyber-Physical Systems (CPS) models by using Deep Reinforcement Learning (DRL) to reduce the number of simulation runs needed to find counterexamples, reporting preliminary evaluation results.
With the rapid development of software and distributed computing, Cyber-Physical Systems (CPS) are widely adopted in many application areas, e.g., smart grid, autonomous automobile. It is difficult to detect defects in CPS models due to the complexities involved in the software and physical systems. To find defects in CPS models efficiently, robustness guided falsification of CPS is introduced. Existing methods use several optimization techniques to generate counterexamples, which falsify the given properties of a CPS. However those methods may require a large number of simulation runs to find the counterexample and is far from practical. In this work, we explore state-of-the-art Deep Reinforcement Learning (DRL) techniques to reduce the number of simulation runs required to find such counterexamples. We report our method and the preliminary evaluation results.