Genetic algorithms-aided reliability analysis

A hybrid procedure of Genetic Algorithms (GAs) and reliability analysis is described, discussed, and summarized. The procedure is specifically referred to as a Genetic Algorithms-aided (GAs-aided) reliability analysis. Two classes of GAs, namely simple GAs and multimodal GAs, are introduced to solve a number of important problems in reliability analysis. The problems cover the determination of Point of Maximum Likelihood in failure domain (PML), the computation of failure probability using the GAs-determined PML, and the determination of multiple design points. The MCS-based method using the GAs-determined PML is specifically implemented in the so-called an Importance Sampling around PML (ISPML). The application of GAs to each respective problem is then demonstrated via numerical examples in order to clarify the procedures. With an aid from GAs, reliability analysis is possible even if there is no information about the geometry or landscape of limit state surfaces and the total number of crucial likelihood points. In addition, GAs significantly improve the computational efficiency and realize the analysis of rare events under constrained computational resources. The implementation of GAs to reliability analysis for building up the hybrid procedure is readily because of their algorithmic simplicity.