Adaptive Strategies For Efficient Model Reduction In High-Dimensional Inverse Problems

This work explores a novel approach for adaptive, differentiable parametrization of large-scale non-stationary random fields. Coupled with any gradient-based algorithm, the method can be applied to variety of optimization problems, including history matching. The developed technique is based on principal component analysis (PCA), but, in contrast to other PCA-based methods, allows to amend parametrization process regarding objective function behaviour.