Reliability-based design optimization (RBDO) ensures design requirements are satisfied at required reliability levels in the presence of uncertainty. the recent RBDO methodologies can accommodate two types of uncertainty - aleatory type, represented by random variables, and epistemic type, represented by interval variables. However, only independent interval variables can be handled. the objective of this research is to introduce dependent interval variables into RBDO. the dependency of interval variables comes from equality and inequality constraints. with the intervals, the reliability is also an interval, and its lower bound is considered to preserve conservativeness. an efficient optimization method is developed to solve RBDO problems. the method is an extension of the existing framework of sequential optimization and reliability analysis with the inclusion of inequality constraints of interval variables. the new method is efficient because it decouples the optimization process from the process of reliability and interval analyses. the new method is applied to the RBDO of a slider mechanism with joint clearances, which are known as dependent interval variables. © 2012 John Wiley & Sons, Ltd.
Available at: http://works.bepress.com/xiaoping-du/62/