Output details

Probabilistic reachability and safety for controlled discrete time stochastic hybrid systems

<11>

This work provides the first theoretical and computable characterization of probabilistic reachability (safety) for stochastic processes evolving over a continuous or hybrid state space. Unlike related approaches in the literature, it shows practical schemes for the precise computation of the quantities of interest. It formalizes, competes, and extends earlier versions of this work appeared in conference proceedings.

Technically, it investigates probabilistic reachability over a finite horizon for a class of discrete-time stochastic hybrid systems with control inputs (controlled nondeterminism), and characterises the set of initial conditions related to any given probabilistic safety guarantee by solving a dynamic programming problem.

[PART D]: This journal paper builds significantly on a conference paper presented at HSCC 2006. In particular it provides a complete theoretical characterization of probabilistic reachability and safety and elaborates on their duality. It includes proofs of all statements in general measure-theoretical terms, unlike the conference publication. None of this work was submitted to RAE2008.