Marcus Sarkis

Marcus Sarkis

Auto-generated profile, based on publicly available information

Marcus Sarkis

Instituto Nacional de Matem´atica Pura e Aplicada, Estrada Dona Castorina 110, Rio de Janeiro 22460-320, Brazil

Research Focus & Metrics

Academic Career Length

24Years (2003-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 January 2007 in Computer Methods in Applied Mechanics and Engineering
Verified
Journal Article
Closed access
JKJung-Han Kimn
Jung-Han Kimn
MSMarcus Sarkis
Marcus Sarkis

1 January 2006 in Discrete and Continuous Dynamical Systems-Series B
Verified
Journal Article
Open access
PDF file
LRLuis J. Roman
Luis J. Roman
MSMarcus Sarkis
Marcus Sarkis
An equation that arises in mathematical studies of the transport of pollutants in groundwater and of oil recovery processes is of the form: $-\nabla_{x}\cdot(\kappa(x,\cdot)\nabla_{x}u(x,\omega))=f(x)$, for $x\in D$, where $\kappa(x,\cdot)$, the permeability tensor, is random and models the properties of the rocks, which are not know with certainty. Further, geostatistical models assume $\kappa(x,\cdot)$ to be a log-normal random field. The use of Monte Carlo methods to approximate the expected value of $u(x,\cdot)$, higher moments, or other functionals of $u(x,\cdot)$, require solving similar system of equations many times as trajectories are considered, thus it becomes expensive and impractical. In this paper, we present and explain several advantages of using the White Noise probability space as a natural framework for this problem. Applying properly and timely the Wiener-Itô Chaos decomposition and an eigenspace decomposition, we obtain a symmetric positive definite linear system of equations whose solutions are the coefficients of a Galerkin-type approximation to the solution of the original equation. Moreover, this approach reduces the simulation of the approximation to $u(x,\omega)$ for a fixed $\omega$, to the simulation of a finite number of independent normally distributed random variables.

1 January 2010 in Computational Methods in Applied Mathematics
Verified
Journal Article
Open access
PDF file
MDM. Dryja
M. Dryja
MSM. Sarkis
M. Sarkis
A second order elliptic problem with highly discontinuous coefficients has been considered. The problem is discretized by two methods: 1) continuous finite element method (FEM) and 2) composite discretization given by a continuous FEM inside the substructures and a discontinuous Galerkin method (DG) across the boundaries of these substructures. The main goal of this paper is to design and analyze parallel algorithms for the resulting discretizations. These algorithms are additive Schwarz methods (ASMs) with special coarse spaces spanned by functions that are almost piecewise constant with respect to the substructures for the first discretization and by piecewise constant functions for the second discretization. It has been established that the condition number of the preconditioned systems does not depend on the jumps of the coefficients across the substructure boundaries and outside of a thin layer along the substructure boundaries. The algorithms are very well suited for parallel computations.

16 January 2007 in Numerical Linear Algebra with Applications
Verified
Journal Article
Closed access
PDF file
TMT. P. Mathew
T. P. Mathew
MSM. Sarkis
M. Sarkis
Christian SchaererChristian Schaerer
Christian Schaerer

1 January 2003 in Computational Methods in Applied Mathematics
Verified
Journal Article
Open access
PDF file
MSMarcus Sarkis
Marcus Sarkis
XTXuemin Tu
Xuemin Tu
We consider the Poisson equation with Dirichlet boundary conditions on a polygonal domain with one reentrant corner. We introduce new nonconforming finite element discretizations based on mortar techniques and singular functions. The main idea introduced in this paper is the replacement of cut-off functions by mortar element techniques on the boundary of the domain. As advantages, the new discretizations do not require costly numerical integrations and have smaller a priori error estimates and condition numbers. Based on such an approach, we prove optimal accuracy error bounds for the discrete solution. Based on such techniques, we also derive new extraction formulas for the stress intensive factor. We establish optimal accuracy for the computed stress intensive factor. Numerical examples are presented to support our theory.