In modern simulation-driven engineering activities such as multidisciplinary design optimization, approaches that reduce the required computational cost of optimization are required. One powerful approach to efficient optimization is indirect optimization. With indirect optimization first a surrogate model is trained to predict the simulation results. The optimization is then performed directly on the surrogate, rather than the original more computationally expensive simulation.
However, the success of an indirect optimization approach utilizing a surrogate model hinges on the prediction accuracy of the surrogate model. Inaccurate prediction can lead to solutions that are suboptimal, infeasible, or have catastrophic consequences.
Join us for this webinar in which SmartUQ principal application engineer, Gavin Jones, will demonstrate why surrogate model accuracy is so important for indirect optimization success. How SmartUQ addresses the need for accuracy with its best-in-class machine learning models will also be discussed.
