The objective of this study was to determine a methodology for implementing feedback loops into a logical, automated, computer assisted probability assessment tool. A tool exists called the GO program, which allows for systems to be modeled in a block diagram or schematic format and then analyzed in a structured manner to determine the probabilities of outcome events. The challenge was to incorporate a method for analyzing feedback loops. Given the difficulty involved with using computer code to analyze feedback loops, reliability engineers would normally create two separate models. To allow for a single model to be used and achieve consistent and repeatable results, a methodology for creating multiple layers of feedback loops in increasing complexity has been analyzed for use with the GO program. Monte Carlo simulations for each of these representative models have been constructed and analyzed to validate the adequacy of the GO program to effectively create probabilities of event success and failure. With the demonstrated ability of the GO program to correctly model feedback loops, it clears a path for the Department of Defense to investigate the benefits of adopting a standardized approach for the analyses of complex systems.

Subjects: Reliability (Engineering); Statistical methods; System safety; Mathematical models; Feedback control systems; Systems engineering