Abstract:

In this paper we propose a novel approach to deriving uncertainty model in reliability applications. It is based on a state space representation of the system life cycle which is similar to techniques widely used in the control theory (CT). Such a model in turn allows applying CT methods (e.g., optimal control algorithms) to assess the reliability in case of high level of uncertainty. In particular, a CT-based model is effective even if the system performances cannot be described in terms of single-valued statistical characteristics. We demonstrate that in such a situation a CT-based model combined with the so-called ‘interval-valued probabilities’ becomes a very promising tool for reliability computations.