Parallel and Multi-objective Falsification with Scenic and VerifAI
Parallel and Multi-objective Falsification with Scenic and VerifAI
复制标题
使用 Scenic 和 VerifAI 进行并行和多目标证伪
DOI：
10.1007/978-3-030-88494-9_15
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Seshia, S. A.
中科院分区：
文献类型：
--
作者：
Viswanadha, K.;Kim, E.;Indaheng, F.;Fremont, D. J.;Seshia, S. A.
证伪已成为基于仿真的自治系统验证的重要工具。在本文中，我们提出了对scenicscenario规范语言和verifaitoolkit的扩展，通过使用并行性来提高基于采样的证伪方法的可扩展性，并将证伪扩展到多目标规范。我们首先提出了一个并行框架，该框架与scenici的模拟和采样能力以及verifai的伪造能力相结合，减少了基于模拟的测试中固有的执行时间瓶颈。然后，我们提出了verifai证伪算法的扩展，以支持采样期间的多目标优化，使用规则手册的概念来指定可用于指导反例搜索过程的多个指标的偏好顺序。最后，我们用thesiclanguage编写的一组全面的基准测试来评估这些扩展的好处。
Falsification has emerged as an important tool for simulation-based verification of autonomous systems. In this paper, we present extensions to theScenicscenario specification language andVerifAItoolkit that improve the scalability of sampling-based falsification methods by using parallelism and extend falsification to multi-objective specifications. We first present a parallelized framework that is interfaced with both the simulation and sampling capabilities ofScenicand the falsification capabilities ofVerifAI, reducing the execution time bottleneck inherently present in simulation-based testing. We then present an extension ofVerifAI’s falsification algorithms to support multi-objective optimization during sampling, using the concept of rulebooks to specify a preference ordering over multiple metrics that can be used to guide the counterexample search process. Lastly, we evaluate the benefits of these extensions with a comprehensive set of benchmarks written in theSceniclanguage.