1: \begin{abstract}
2:
3: In this paper, we consider local multiscale model reduction for
4: problems with multiple scales in space and time. We developed our approaches within the framework of the Generalized Multiscale Finite Element Method (GMsFEM) using space-time coarse cells. The main idea of GMsFEM is to construct a local snapshot space and a local spectral decomposition in the snapshot space. Previous research in developing multiscale spaces within GMsFEM focused on constructing multiscale spaces and relevant ingredients in space only. In this paper, our main objective is to develop a multiscale
5: model reduction framework within GMsFEM that uses space-time coarse
6: cells. We construct space-time snapshot and offline spaces.
7: We compute these snapshot solutions by solving local problems. A complete snapshot space will use all possible
8: boundary conditions; however, this can be very expensive. We
9: propose using randomized boundary conditions and oversampling
10: (cf. \cite{randomized2014}). We construct the local spectral decomposition
11: based on our analysis, as presented in the paper.
12: We present numerical results to confirm our theoretical findings and
13: to show that using our proposed approaches, we can obtain an accurate
14: solution with low dimensional coarse spaces. We discuss using online basis functions constructed in the online stage
15: and using the residual information. Online basis functions use
16: global information via the residual
17: and provide fast convergence to the exact solution
18: provided a sufficient number of offline basis functions.
19: We present numerical studies for our proposed online procedures.
20: We remark that the proposed method is
21: a significant extension compared to
22: existing methods, which use coarse cells in space only because
23: of (1) the parabolic nature of cell solutions, (2) extra degrees of freedom
24: associated with space-time cells, and (3) local boundary conditions in space-time cells.
25:
26:
27: \end{abstract}
28: