Fast Implicit Methods For Elliptic Moving Interface Problems

Abstract

Two notable advances in numerical methods were supported by this grant. First, a fast algorithm was derived, analyzed, and tested for the Fourier transform of piecewise polynomials given on d-dimensional simplices in D-dimensional Euclidean space. These transforms play a key role in computational problems ranging from medical imaging to partial differential equations, and existing algorithms are inaccurate and/or prohibitively slow for d > 0. The algorithm employs low-rank approximation by Taylor series organized in a butterfly scheme, with moments evaluated by a new dimensional recurrence and simplex quadrature rules. For moderate accuracy and problem size it runs orders of magnitude faster than direct evaluation, and one to three orders of magnitude slower than the classical uniform Fast Fourier Transform. Second, bilinear quadratures ---which numerically evaluate continuous bilinear maps, such as the L2 inner product, on continuous f and g belonging to known finite-dimensional function spaces---were analyzed and developed.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 11, 2015
Accession Number
AD1001342

Entities

People

  • John Strain

Organizations

  • University of California Regents

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Accuracy
  • Air Force
  • Air Force Research Laboratories
  • Applied Mathematics
  • Computational Fluid Dynamics
  • Computational Science
  • Computer Programs
  • Computer Science
  • Differential Equations
  • Eigenvalues
  • Gaussian Quadrature
  • Integral Equations
  • Mathematics
  • Polynomials
  • Precision
  • Standards
  • Two Dimensional

Readers

  • Finite Element Method (FEM) for solving Partial Differential Equations (PDEs)
  • Graph Algorithms and Convex Optimization.

Technology Areas

  • Space