Randomized test case generation for hybrid systems: metric selection

Abstract

We are developing a randomized approach to test generation for hybrid systems, and control systems in general, using techniques from robotic path planning which have proved successful in solving high dimensional nonlinear problems. A critical component of the proposed algorithm is the choice of "metric" - how one decides the closeness of two states - which is nontrivial in the hybrid state space. In this paper we introduce four metrics for hybrid systems; and benchmark the algorithm using each of these metrics on a popular example problem from the literature and compare the impact of metric choice on computational efficiency.