Fast Wavelet Transforms and the Numerical Solution of Initial Value Problems.

Abstract

Fast algorithms to evaluate the approximate solution of time dependent problems were developed by taking advantage of the sparse wavelet representation of finite difference operators and using only part of the representation to compute the local solution. For example, we can evaluate the solutions at a point to parabolic equations with variable coefficients in O(log4N) operations when the equation has time independent coefficients. For time dependent coefficients; the complexity is O(N log3N). Additionally, high resolution numerical methods for the high frequency asymptotic expansion to electromagnetic propagation and scattering codes were developed. This replaces ray tracing by a direct solution to the eikonal equation. Moreover, we developed and solved generalized eikonal equations for diffraction phenomena.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1997
Accession Number
ADA323399

Entities

People

  • Stanley Osher

Organizations

  • University of California, Los Angeles

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Asymptotic Series
  • Boltzmann Equation
  • Differential Equations
  • Equations
  • Formulas (Mathematics)
  • High Resolution
  • Mathematical Analysis
  • Mathematics
  • Military Research
  • Partial Differential Equations
  • Ray Tracing
  • Time Domain
  • Wave Equations
  • Wavelet Transforms

Fields of Study

  • Mathematics

Readers

  • Finite Element Method (FEM) for solving Partial Differential Equations (PDEs)
  • Graph Algorithms and Convex Optimization.
  • Wave Propagation and Nonlinear Chaotic Dynamics.