Falsification of Cyber-Physical Systems using Bayesian Optimization

2022-09-14

arXiv.org

AI Summary

This work addresses the formal specification falsification problem for safety-critical cyber-physical systems (CPS). Methodologically, it proposes a Bayesian optimization framework that integrates adaptive local surrogate modeling—built upon Gaussian processes—with domain-informed prior knowledge to guide initial sampling, and employs a customized acquisition function (e.g., an enhanced Expected Improvement) tailored for counterexample search. The core contributions are: (i) a local–global collaborative modeling strategy that balances exploration and exploitation, and (ii) a prior-driven, budget-aware optimization policy. Experiments on multiple benchmark CPS demonstrate that the approach significantly reduces simulation cost: local modeling accelerates convergence, while prior knowledge boosts counterexample detection success by up to 3× under tight budgets (<50 simulations). Moreover, the designed acquisition function substantially improves efficiency over standard Bayesian optimization baselines.

Abstract

Cyber-physical systems (CPSs) are often complex and safety-critical, making it both challenging and crucial to ensure that the system's specifications are met. Simulation-based falsification is a practical testing technique for increasing confidence in a CPS's correctness, as it only requires that the system be simulated. Reducing the number of computationally intensive simulations needed for falsification is a key concern. In this study, we investigate Bayesian optimization (BO), a sample-efficient approach that learns a surrogate model to capture the relationship between input signal parameterization and specification evaluation. We propose two enhancements to the basic BO for improving falsification: (1) leveraging local surrogate models, and (2) utilizing the user's prior knowledge. Additionally, we address the formulation of acquisition functions for falsification by proposing and evaluating various alternatives. Our benchmark evaluation demonstrates significant improvements when using local surrogate models in BO for falsifying challenging benchmark examples. Incorporating prior knowledge is found to be especially beneficial when the simulation budget is constrained. For some benchmark problems, the choice of acquisition function noticeably impacts the number of simulations required for successful falsification.

Problem

• Optimize Bayesian for CPS falsification
• Reduce simulation costs efficiently
• Enhance BO with local models

Innovation

• Bayesian Optimization enhances CPS falsification
• Local surrogate models improve efficiency
• Acquisition functions reduce simulation needs