An Efficient Method for Reliability Based Design Optimization 2004-01-1125
In search of improved design, Reliability Based Design Optimization using probabilistic constraints has evolved as a powerful tool, however extensive computational efforts necessary for RBDO being one of the biggest challenges in this field. In order to evaluate probabilistic constraints, single loop and double loop strategies have been successfully used, the number of function evaluations being a decisive comparison criteria. However, when applied to complicated functions, above techniques lead to much higher number of function evaluations.
Efforts to make RBDO technique computationally feasible have generated superior single loop strategy presented in this paper, dealing with the approximated form of the original functions. At the same time, improved Most Probable Point search algorithm has been used providing modified gradient information for fast convergence of the constraints. Above method is successfully tested on some problems, and results are compared with those obtained from other standard methods of RBDO dealing with original functions.