作者: P.S. Koutsourelakis
DOI: 10.1016/J.CMA.2008.04.012
关键词:
摘要: The present paper proposes an algorithmic framework for designing complex systems in the presence of large uncertainties. It is highly applicable to realistic engineering problems as it directly parallelizable and can interact a non-intrusive manner with any deterministic solver (e.g. finite element codes) order quantify response statistics their dependence on design variables. efficiency proposed procedure, which equivalent number calls solver, high due optimized sampling process employed conjunction Bayesian, statistical learning component. Several numerical examples, dealing static dynamic, linear nonlinear demonstrate accuracy effectiveness methodology. In addition robust sensitivity measures uncertainties are provided.