Reliability-based optimization using bridge importance sampling

This paper introduces an efficient reliability estimation approach to be used in the framework of reliability-based optimization. The key feature of the procedure consists of reusing during the optimization procedure the results of the previous iterations. The reliability analysis is performed subsequently with a reduced number of samples which greatly decreases the computational efforts associated with the reliability-based optimization process. The validity and the advantages of the procedure are demonstrated by means of two reliability-based optimization problems.

• Reliability analysis is performed with a reduced number of samples.
• An efficient reliability estimation approach is integrated in the context of RBO.
• The approach bridges information from different optimization stages.
• Information from previous iterations is used during the design process.
• Advantages are demonstrated in practical problems.