PDE Methods in Computational Physics

The field equations of physics, from diffusion and waves to electrostatics, are partial differential equations, and solving them numerically means discretizing space and time into a grid and propagating or relaxing the field on it.

Definition

PDE methods in computational physics are numerical schemes that approximate the solution of partial differential equations on a discrete grid, replacing spatial and temporal derivatives by finite differences or related operators.

Scope

This topic covers finite-difference discretization of the canonical PDE classes, elliptic, parabolic and hyperbolic, together with explicit and implicit time stepping, relaxation and multigrid methods for boundary-value problems, and the stability criteria that govern them. Finite-element and spectral approaches are treated as neighboring methods.

Core questions

• How are elliptic, parabolic and hyperbolic PDEs discretized and solved differently?
• What is the Courant-Friedrichs-Lewy condition and why does it limit explicit time steps?
• How do relaxation and multigrid methods solve large boundary-value problems efficiently?
• When is an implicit scheme worth its extra cost compared to an explicit one?

Key theories

Finite-difference discretization

Spatial and temporal derivatives are replaced by difference quotients on a grid, converting a PDE into a large system of algebraic equations whose accuracy is set by the grid spacing and stencil order.

CFL stability condition

For explicit schemes solving hyperbolic and parabolic equations, the Courant-Friedrichs-Lewy condition bounds the time step relative to the grid spacing and propagation speed, beyond which the numerical solution blows up.

Relaxation and multigrid

Elliptic boundary-value problems such as the Poisson equation are solved by iterative relaxation, with multigrid methods accelerating convergence by correcting errors across a hierarchy of grid resolutions.

Clinical relevance

PDE solvers compute electrostatic and magnetostatic fields, heat conduction and diffusion, wave propagation and the Schrodinger equation, forming the backbone of computational electromagnetism, fluid dynamics and continuum physics simulation.

History

The systematic theory of finite-difference solutions of PDEs began with the 1928 Courant-Friedrichs-Lewy paper on stability, expanded greatly with computers in the mid-twentieth century, and was made efficient for large problems by the development of multigrid methods in the 1970s.

Key figures

• Richard Courant
• Kurt Friedrichs
• Randall J. LeVeque

Seminal works

• leveque2007
• press2007

Frequently asked questions

What is the difference between explicit and implicit time stepping?

Explicit schemes compute the next time level directly from the current one and are cheap per step but limited by a stability condition on the step size. Implicit schemes solve a coupled system at each step, costing more per step but remaining stable for much larger steps, which pays off for stiff or diffusive problems.

Why classify PDEs as elliptic, parabolic or hyperbolic?

The classification reflects how information propagates: elliptic equations describe equilibrium fields with global coupling, parabolic equations describe smoothing diffusion in time, and hyperbolic equations describe waves traveling at finite speed. Each class calls for different discretization and stability strategies.