Two-Layered Falsification of Hybrid Systems Guided by Monte Carlo Tree Search

Abstract

Few real-world hybrid systems are amenable to formal verification, due to their complexity and black box components. Optimization-based falsification-a methodology of search-based testing that employs stochastic optimization-is thus attracting attention as an alternative quality assurance method. Inspired by the recent work that advocates coverage and exploration in falsification, we introduce a two-layered optimization framework that uses Monte Carlo tree search (MCTS), a popular machine learning technique with solid mathematical and empirical foundations (e.g. in computer Go). MCTS is used in the upper layer of our framework; it guides the lower layer of local hill-climbing optimization, thus balancing exploration and exploitation in a disciplined manner. We demonstrate the proposed framework through experiments with benchmarks from the automotive domain.

Fast Falsification of Hybrid Systems Using Probabilistically Adaptive Input

“…Monte Carlo tree search (MCTS) has been applied to a model of aircraft collisions in [27]; and in a falsification context more recently [32] to guide global optimization, building on the previous idea of time-staging [31]. That work noted the strong similarities between falsification using MCTS and statistical model checking (SMC) using importance splitting [20].…”

Hybrid System Falsification Under (In)equality Constraints via Search Space Transformation

“…Input Constraints. Falsification typically considers a hyperrectangle as the search space for input signals: this is the problem setting adopted in many works [6]- [12], and also in tools Breach [4] and S-TaLiRo [5]. However, in real hybrid systems input signals are not free to assume any value within the hyperrectangle, because there exist some constraints ψ among them.…”

Multi-armed Bandits for Boolean Connectives in Hybrid System Falsification

“…The trade-off has been already discussed in some works on falsification. A recent example is [35], where they use Monte Carlo tree search to force systematic exploration of the space of input signals. Besides MCTS, Gaussian process learning (GP learning) has also attracted attention in machine learning as a clean way of balancing exploitation and exploration.…”