Surrogate assisted active subspace and active subspace assisted surrogate—A new paradigm for high dimensional structural reliability analysis

Performing reliability analysis on complex systems is often computationally expensive. In particular, when dealing with having high input dimensionality, estimation becomes a daunting task. A popular approach to overcome the problem associated time-consuming and expensive evaluations building surrogate model. However, these efficient models suffer from curse of dimensionality. Hence, training model for high-dimensional problems not straightforward. Henceforth, this paper presents framework solving problems. The basic premise train low-dimensional manifold, discovered using active subspace algorithm. learning manifold non-trivial as it requires information gradient response variable. To address issue, we propose sparse algorithms in conjunction algorithm; resulting algorithm referred (SAS) We project inputs onto identified SAS. high-fidelity used map manifolds output response. illustrate efficacy proposed by three benchmark literature. results obtained indicate accuracy efficiency compared already established methods