Falsification of Hybrid Systems Using Adaptive Probabilistic Search

• Published in: ACM transactions on modeling and computer simulation 2021-07, Vol.31 (3), p.1-22
• Main Authors: Ernst, Gidon, Sedwards, Sean, Zhang, Zhenya, Hasuo, Ichiro
• Format: Article
• Language: English
• ISSN: 1049-3301; 1558-1195
• DOI: 10.1145/3459605

We present and analyse an algorithm that quickly finds falsifying inputs for hybrid systems. Our method is based on a probabilistically directed tree search, whose distribution adapts to consider an increasingly fine-grained discretization of the input space. In experiments with standard benchmarks, our algorithm shows comparable or better performance to existing techniques, yet it does not build an explicit model of a system. Instead, at each decision point within a single trial, it makes an uninformed probabilistic choice between simple strategies to extend the input signal by means of exploration or exploitation. Key to our approach is the way input signal space is decomposed into levels, such that coarse segments are more probable than fine segments. We perform experiments to demonstrate how and why our approach works, finding that a fully randomized exploration strategy performs as well as our original algorithm that exploits robustness. We propose this strategy as a new baseline for falsification and conclude that more discriminative benchmarks are required.