Spectral Upscaling for Graph Laplacian Problems with Application to Reservoir Simulation
Here, we consider coarsening procedures for graph Laplacian problems written in a mixed saddlepoint form. In that form, in addition to the original (vertex) degrees of freedom (dofs), we also have edge degrees of freedom. We extend previously developed aggregationbased coarsening procedures applied to both sets of dofs to now allow more than one coarse vertex dof per aggregate. Those dofs are selected as certain eigenvectors of local graph Laplacians associated with each aggregate. Additionally, we coarsen the edge dofs by using traces of the discrete gradients of the already constructed coarse vertex dofs. These traces are defined on the interface edges that connect any two adjacent aggregates. The overall procedure is a modification of the spectral upscaling procedure developed in for the mixed finite element discretization of diffusion type PDEs which has the important property of maintaining infsup stability on coarse levels and having provable approximation properties. We consider applications to partitioning a general graph and to a finite volume discretization interpreted as a graph Laplacian, developing consistent and accurate coarsescale models of a finescale problem.
 Authors:

^{[1]};
^{[1]};
^{[2]}
 Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
 Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States); Portland State Univ., Portland, OR (United States)
 Publication Date:
 Report Number(s):
 LLNLJRNL693123
Journal ID: ISSN 10648275
 Grant/Contract Number:
 AC5207NA27344
 Type:
 Accepted Manuscript
 Journal Name:
 SIAM Journal on Scientific Computing
 Additional Journal Information:
 Journal Volume: 39; Journal Issue: 5; Journal ID: ISSN 10648275
 Publisher:
 SIAM
 Research Org:
 Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
 Sponsoring Org:
 USDOE
 Country of Publication:
 United States
 Language:
 English
 Subject:
 97 MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE; 58 GEOSCIENCES; graph Laplacian; finite volume methods; numerical upscaling; algebraic multigrid; reservoir simulation
 OSTI Identifier:
 1438732
Barker, Andrew T., Lee, Chak S., and Vassilevski, Panayot S.. Spectral Upscaling for Graph Laplacian Problems with Application to Reservoir Simulation. United States: N. p.,
Web. doi:10.1137/16M1077581.
Barker, Andrew T., Lee, Chak S., & Vassilevski, Panayot S.. Spectral Upscaling for Graph Laplacian Problems with Application to Reservoir Simulation. United States. doi:10.1137/16M1077581.
Barker, Andrew T., Lee, Chak S., and Vassilevski, Panayot S.. 2017.
"Spectral Upscaling for Graph Laplacian Problems with Application to Reservoir Simulation". United States.
doi:10.1137/16M1077581. https://www.osti.gov/servlets/purl/1438732.
@article{osti_1438732,
title = {Spectral Upscaling for Graph Laplacian Problems with Application to Reservoir Simulation},
author = {Barker, Andrew T. and Lee, Chak S. and Vassilevski, Panayot S.},
abstractNote = {Here, we consider coarsening procedures for graph Laplacian problems written in a mixed saddlepoint form. In that form, in addition to the original (vertex) degrees of freedom (dofs), we also have edge degrees of freedom. We extend previously developed aggregationbased coarsening procedures applied to both sets of dofs to now allow more than one coarse vertex dof per aggregate. Those dofs are selected as certain eigenvectors of local graph Laplacians associated with each aggregate. Additionally, we coarsen the edge dofs by using traces of the discrete gradients of the already constructed coarse vertex dofs. These traces are defined on the interface edges that connect any two adjacent aggregates. The overall procedure is a modification of the spectral upscaling procedure developed in for the mixed finite element discretization of diffusion type PDEs which has the important property of maintaining infsup stability on coarse levels and having provable approximation properties. We consider applications to partitioning a general graph and to a finite volume discretization interpreted as a graph Laplacian, developing consistent and accurate coarsescale models of a finescale problem.},
doi = {10.1137/16M1077581},
journal = {SIAM Journal on Scientific Computing},
number = 5,
volume = 39,
place = {United States},
year = {2017},
month = {10}
}