Safety of stochastic systems: An analytic and computational approach

We refine the concept of stochastic reach avoidance for a general class Markov processes introducing threshold p reaching probability. This new problem is called -safety, and it aims to ensure that given process reaches forbidden set before leaving its ‘working’ state space with probability less than . In situation when an initial measure characterizes states, variant -safety put forward. call this form safety weak -safety. work, we characterize both show how compute them. employ semi-definite programming linear To get point, use certificates positivity polynomials translated into sum squares Bernstein forms.