RELIABILITY-BASED OPTIMIZATION AS AN INFORMATION TOOL

Reliability-based optimization is presented as an information tool in the process of achieving knowledge about the nature of a problem. Sensitivity analysis in pre- and post-evaluation of reliability-based optimization models is introduced for the purpose of identification of important problem parameters. A reliability-based optimization problem is formulated and a solution procedure is suggested. Various techniques for achievement of the sensitivities both with respect to model parameters and optimization variables are described. Finally, the way in which the sensitivities used in pre- and post-evaluations and the optimal design give additional insight into the nature of the problem is discussed and illustrated.