Industrial, systems, and cybernetic engineers explore methods for estimating safety levels of complex systems with incomplete data on equipment failures and risk factors. Traditionally, risk assessment has relied on analyzing cumulative date from previous failures, and on simulating the dynamics of the structure, they explain, but these approaches are not feasible in many-dimensional models, and in systems that are so reliable that failure data is not available. As alternatives, they present numerical methods for analyzing super large failure trees with a large number of multiple vertices, which allow finding minimal sections and reducing computational time. Among specific topics are the light-traffic analysis of some queuing models with losses, developing new mathematical models and methods for determining critical states of nuclear plant reactors (many of the contributors are from Ukraine), and investigating the sensitivity of Bayesian estimates for exponential failure models. Annotation ©2009 Book News, Inc., Portland, OR (booknews.com)
