Resumen de A deterministic robust optimisation method under interval uncertainty based on the reverse model
Qi Zhou, Xinyu Shao, Ping Jiang, Hui Zhou, Longchao Cao, Lin Zhang
Uncertainty is inevitable for real-world engineering optimisation. Most existing robust optimisation (RO) approaches consider many drastically different alternatives in an effort to design an engineering system under uncertainty. In this paper, a deterministic RO approach named variable adjustment robust optimisation (VARO) is proposed to improve the robustness of a preferred preexisting design. Firstly, a formulation to obtain design alternatives with acceptable difference from the preferred preexisting design is constructed. Secondly, the robustness indices, estimated sensitivity regions using worst-case scenario analysis, are generated by the reverse model that maps the given acceptable objective variations and feasibility variations into the space of variable variations. The obtained robustness indices are incorporated into the constructed formulation to evaluate the robustness of the design alternatives. Thirdly, metamodels are built for robustness indices to transform the nested optimisation structure of the proposed VARO approach into a single loop optimisation structure for the purpose of easing its computational burden. Seven examples with differing complexity are used to demonstrate the applicability and efficiency of the proposed approach. Verifications of robustness for the optimum obtained are also performed via design of experiment.
