Abstract:In this talk, I would like to discuss the quantum size effects and the scale invariance of some physical quantities such as elastic constant , thermal conductivity, dielectric constant and so forth. It can be proven that the elastic constants and the dielectric constants are approximate scale invariants for a nanocrystal in the framework of the local density approximation of density functional theory(DFT-LDA). From this, we try to present the unified electromagnetic models from microto macro-scale. Finally, we shall give some typical applications of the above theoretical results in CMOS transistor at nano-scale, photonic crystals and the protein structures in solution. What is the smallest possible constant in Cea's lemma Wei Chen School of Economics, Shandong University Abstract We consider finite element approximations of a second order elliptic problem on a bounded polytopic domain in R with d∈{1; 2; 3...}. The constant C≥1 appearing in Cea's lemma and coming from its standard proof can be very large when coefficient of an an ellipic operator attain considerably different values. We restrict ourselves to regular families of uniform partitions and linear simplicial elements. Using a lower bound of the interpolation error and the supercloseness between the finite element solution and Lagrange interpolant of the exact solution, we prove that the constant in Cea's lemma can be reduced to 1+Ο (h) as the discretization parameter h tends to zero. Numerical results in on and two-dimensional case illustrating this reduction are presented.We consider finite element approximations of a second order elliptic problem on a bounded polytopic domain in R with d∈{1; 2; 3...}. The constant C≥1 appearing in Cea's lemma and coming from its standard proof can be very large when coefficient of an an ellipic operator attain considerably different values. We restrict ourselves to regular families of uniform partitions and linear simplicial elements. Using a lower bound of the interpolation error and the supercloseness between the finite element solution and Lagrange interpolant of the exact solution, we prove that the constant in Cea's lemma can be reduced to 1+Ο (h) as the discretization parameter h tends to zero. Numerical results in on and two-dimensional case illustrating this reduction are presented. International Conference on Multilevel Iterative Methods Analysis of mortar-type Q /Q rot 1 0 element and multigrid method for the incompressible Stokes problem* Jinru Chen and Peiqi Huang 1. School of Mathematics and Computer Science, Nanjing Normal University 2. Department of Applied Mathematics, Nanjing Forestry University Abstract In this paper, we discuss mortar element method for Q /Q rot 1 0 element for the incompressible Stokes problem and prove the optimal error estimate. Meanwhile a W-cycle multigrid method is presented for solving the discrete equation and the optimal convergence of the multigrid is proven. Finally numerical experiments are given to confirm our theoretical results.In this paper, we discuss mortar element method for Q /Q rot 1 0 element for the incompressible Stokes problem and prove the optimal error estimate. Meanwhile a W-cycle multigrid method is presented for solving the discrete equation and the optimal convergence of the multigrid is proven. Finally numerical experiments are given to confirm our theoretical results. *This work was supported by the National Natural Science Foundation of China under the Grant 10471067 and the National Basic Research Program under the Grants 2005CB321704. 1. E-mail address: jrchen@njnu.edu.cn. 2. E-mail address: pqhuang1979@163.com. A Self-Adapting Block Reduction Method for Solving Multi-scale Linear Systems in QMC Simulation Wenbin Chen, Ichitaro Yamazaki, Zhaojun Bai, Richard Scalettar 1. Department of Mathematics, Fudan University, China 2. Departments of Computer science, University of California, Davis,USA 3. Departments of Computer science, University of California, Davis,USA 4. Departments of Physics, University of California, Davis, USA Abstract In this talk, we discuss the multi-scale linear systems arising in Quantum Monte Carlo simulation for Hubbard model, and challenges of multi-scale phenomena to existing linear system solvers.In this talk, we discuss the multi-scale linear systems arising in Quantum Monte Carlo simulation for Hubbard model, and challenges of multi-scale phenomena to existing linear system solvers. We present a robust and efficient solver for solving multi-scale linear systems, the Features of this new solver: 1) delivers solutions with a sufficient accuracy (just need to make the computational/roundoff error in finite precision is smaller International Conference on Multilevel Iterative Methods than the "model'' approximation error, such as using the checkerboard method to approximate matrix exponential), 2) robustly handles a wide range of conditioning of systems causing by multi-scale parameters, 3) is efficient in terms of memory requirement and flops. The numerical results shows that the new algorithm are efficient and stable. The preconditioning for the linear systems are also introduced. An adaptive Galerkin-characteristic finite element method with reliable and efficient error control for convection-dominated diffusion problem Ming Cui, Hongsen Chen, Richard E.Ewing, AND Guan Qin School of Mathematics and System Science, Shandong University Institute for Scientific Computation, Texas A M University, USA Abstract An efficient and reliable a posteriori error estimate is derived for the Characteristic-Galerkin method for convection-dominated problems. An adaptive algorithm with variable time-step sizes and space meshes is proposed and analyzed. It is proved that at each time step, the adaptive algorithm reduces the error indicators below any given tolerance within a finite number of iterations.An efficient and reliable a posteriori error estimate is derived for the Characteristic-Galerkin method for convection-dominated problems. An adaptive algorithm with variable time-step sizes and space meshes is proposed and analyzed. It is proved that at each time step, the adaptive algorithm reduces the error indicators below any given tolerance within a finite number of iterations. Affine conjugate inexact Newton methods for convex optimization problems and their adaptive finite element realization

Domain Decomposition Survey 143 Methods for Partial Diierential Equationsddm Preconditioner for 4th Order Problems by Using B- Spline Nite Element Method', in Fourth Int. Symp. on Domain Decomposition Methods for Partial Diierential Equationsiterative Substructuring Methods: Algorithms and Theory fo

Joint Australia-China meeting on Non-linear Partial Differential Equations

Current Trends in Scientific Computing : ICM 2002 Beijing Satellite Conference on Scientific Computing, August 15-18, 2002, Xi'an Jiaotong University, Xi'an, China

A Stability Extensible Symplectic FDTD(4,4) Method and Its Application in Subgrid Technique for Arbitrary Grid Ratios

A Class of Iterative Solvers for the Helmholtz Equation: Factorizations, Sweeping Preconditioners, Source Transfer, Single Layer Potentials, Polarized Traces, and Optimized Schwarz Methods

Optimized Operator-Splitting Methods in Numerical Integration of Maxwell's Equations

Scalable parallel multilevel algorithms for solving partial differential equations

Shi Shu An algebraic multigrid method for higher-order finite element equations in two dimensional linear elasticity 15:30-16:00 Qingsong Zou Convergence analysis of adaptive FVM Ying Yang A finite element correction scheme for Green's functions

Stability estimates of Nyström discretizations of Helmholtz decomposition boundary integral equation formulations for the solution of Navier scattering problems in two dimensions with Dirichlet boundary conditions

Domain decomposition: 3D full-anisotropy coupling solver

Innovative Methods and Theories in Numerical Algebra

Solving Frontier Problems of Physics: The Decomposition Method

A Novel Domain Decomposition Method for Highly Oscillating Partial Differential Equations

Domain Decomposition Methods for the Helmholtz Equation: A Numerical Investigation.

Investigation of 3D Internal Flow Using New Flux-Reconstruction High Order Method

Search for CP Violation in t ! Kpnt Decays

A Second-Order Decoupled Implicit/Explicit Method of the 3d Primitive Equations of Ocean Ii: Finite Element Spatial Discretization

Schwarz Methods by Domain Truncation

3-D Domain Decomposition Method With Nonconformal Meshes for Thermoelastic Modeling

Parallel Domain Decomposition Methods with Mixed Order Discretization for Fully Implicit Solution of Tracer Transport Problems on the Cubed-Sphere.

Accurate and efficient multiscale simulation of a heterogeneous elastic beam via computation on small sparse patches