Maximum entropy snapshot sampling for reduced basis modelling

Purpose The maximum entropy snapshot sampling (MESS) method aims to reduce the computational cost required for obtaining reduced basis purpose of model reduction. Hence, it can significantly original system dimension whilst maintaining an adequate level accuracy. this paper is show how these beneficial results are obtained. Design/methodology/approach so-called MESS used reducing two nonlinear circuit models. directly reduces number snapshots by recursively identifying and selecting that strictly increase estimate correlation considered systems. Reduced bases then obtained with orthogonal-triangular decomposition. Findings Two case studies have been validating reduction performance MESS. These numerical experiments verify advocated approach, in terms costs accuracy, relative gappy proper orthogonal Originality/value novel has successfully circuits: particular, a diode chain thermal-electric coupled system. In both cases, removed unnecessary data, hence, matrix, before calling QR generation routine. As result, QR-decomposition called on offline stage scaled down, central processing unit time.