Probabilistic bounded reachability for hybrid systems with continuous nondeterministic and probabilistic parameters

Abstract: We develop an algorithm for computing bounded reachability probability for hybrid systems, i.e., the probability that the system reaches an unsafe region within a finite number of discrete transitions. In particular, we focus on hybrid systems with continuous dynamics given by solutions of nonlinear ordinary differential equations (with possibly nondeterministic initial conditions and parameters), and probabilistic behaviour given by initial parameters distributed as continuous (with possibly infinite support) and discrete random variables. Our approach is to define an appropriate relaxation of the (undecidable) reachability problem, so that it can be solved by $\delta$-complete decision procedures. In particular, for systems with continuous random parameters only, we develop a validated integration procedure which computes an arbitrarily small interval that is guaranteed to contain the reachability probability. In the more general case of systems with both nondeterministic and probabilistic parameters, our procedure computes a guaranteed enclosure for the range of reachability probabilities. We have applied our approach to a number of nonlinear hybrid models and validated the results by comparison with Monte Carlo simulation.