UW AMath Numerical Methods for Time-Dependent Differential Equations
 
Applied Math 586
 
Spring Quarter, 2015

Previous topic

Course information

Next topic

Homework

This Page

Outline and TopicsΒΆ

The emphasis of this course is the basic theory of finite difference methods for time-dependent differential equations, including both ODEs and PDEs. The derivation of methods and their analysis in terms of both accuracy and stability are stressed.

  • Numerical methods for time-dependent ODEs
    • Taylor series, Runge-Kutta, and Linear Multistep Methods
    • Consistency, order of accuracy, local and global error
    • Zero-stability, A-stability, L-stability, etc.
    • Stiff equations and implicit methods
  • Stability theory for PDE methods
    • Method of Lines approach
    • Lax-Richtmyer stability
    • von Neumann stability analysis
    • Relation of ODE and PDE stability theories
  • Parabolic PDEs, e.g. diffusion or heat equation
    • Stiffness and the need for implicit solvers
    • Crank-Nicolson method
  • Hyperbolic PDEs, e.g. advection and wave equations
    • Lax-Wendroff, upwind methods, etc.
    • Numerical dissipation / dispersion
    • Modified equation analysis
  • Mixed equations, e.g. reaction-diffusion, advection-diffusion
    • Fractional step methods
    • Unsplit methods
  • Brief introduction to other approaches
    • Finite volume methods
    • Finite element methods
    • Spectral methods