Abstract

Over the past decades, it is more and more acknowledged that including uncertainty in numerical analysis embodies a very useful and valuable application of the increasing capabilities of modern (high-performance) computer facilities, both in science and in industry. Non-deterministic methodologies enable the numerical simulation of uncertain or variable processes, providing information on ranges of possible behavior rather than unique predictions. This information can only increase the trustworthiness of the underlying analysis, and proves to be of great value in the virtual prototyping process that is becoming a standard in current design engineering practice.