ABSTRACT
Manufacturers are striving to develop high-performance, low-cost products that meet increasingly demanding reliability requirements and safety standards under stringent constraints for time and resources. In this environment, the traditional deterministic approach involving safety factors has become inadequate. What is needed are alternative methods that can provide greater accuracy and realism in the analysis and design process, thus eliminating unnecessarily high safety margins. This accuracy and realism can only be obtained by adequately accounting for uncertainty, usually by characterizing pertinent quantities as random variables, i.e., as quantities that are not deterministic and that actually exhibit statistical variations. Adequacy of performance is then measured directly in terms of reliability or its converse, probability of failure, rather than in terms of safety factors. The recognition of the inherent
nondeterministic nature of basic physical phenomena led the Boeing Company to start a new project advancing the concept of probabilistic analysis and design system (PADS) in 1989.
