What is the difference between an explicit and an implicit scheme?

An explicit scheme computes each new grid value directly from known values but is stable only for small time steps, whereas an implicit scheme solves a coupled system for all new values at once, allowing much larger stable time steps at the cost of a linear solve per step.

Why might finite differences be preferred over finite elements?

On simple, regular geometries finite differences are easy to implement, cheap, and straightforward to make high-order. Finite elements become advantageous mainly when the domain has complex shape or the problem has a natural variational formulation.

Finite Difference Methods

Finite difference methods approximate derivatives by difference quotients on a grid, turning a differential equation into a system of algebraic equations for the values of the solution at the grid points.

Hitta ämne med PaperMindSnartFind papers & topics

Tools & resources

Ladda ner bildspel

Learn & explore

VideoSnart

Definition

A finite difference method is a discretization of a differential equation in which derivatives are replaced by difference quotients of the unknown evaluated on a structured grid, producing algebraic equations whose solution approximates the differential equation's solution at the grid points.

Scope

This topic covers the construction of difference approximations from Taylor expansions, discretization of elliptic, parabolic, and hyperbolic PDEs, explicit and implicit time-stepping schemes (such as forward Euler, backward Euler, and Crank-Nicolson), von Neumann stability analysis, and the consistency-stability-convergence framework specialized to difference schemes.

Core questions

How are accurate difference approximations to derivatives derived and their truncation error quantified?
How do explicit and implicit time-stepping schemes differ in stability and cost?
How does von Neumann analysis determine the stability of a difference scheme?
How does the equation type set the appropriate scheme and any step-size restrictions?

Key theories

Consistency, stability, and convergence: A difference scheme is consistent if its truncation error vanishes as the grid is refined and stable if errors do not grow unboundedly; by the Lax equivalence theorem these together guarantee convergence to the true solution for well-posed linear problems.
Von Neumann stability analysis: Decomposing the error into Fourier modes on a uniform grid reduces stability to bounding an amplification factor for each mode; the scheme is stable when no mode is amplified, yielding explicit step-size conditions such as the diffusion and CFL limits.

Mechanisms

Difference formulas are built by combining Taylor expansions at neighbouring grid points to cancel low-order terms and isolate a derivative, with the leading remaining term giving the truncation error and the method's order. For time-dependent problems, explicit schemes update each new value directly from old ones but must respect a stability limit (a diffusion-number bound for parabolic equations, the CFL condition for hyperbolic ones), while implicit schemes like Crank-Nicolson couple the new values into a linear system that is unconditionally stable but requires a solve each step. Von Neumann analysis substitutes Fourier modes to test stability and pin down these limits.

Clinical relevance

Finite difference methods are widely used for problems on regular domains and structured grids: heat conduction and diffusion, wave propagation and seismic modelling, computational electromagnetics (the finite-difference time-domain method), and option pricing via the Black-Scholes equation; their simplicity and ease of high-order extension make them a first choice when geometry is simple.

History

The mathematical foundation was laid by the 1928 Courant-Friedrichs-Lewy paper on difference equations for PDEs; von Neumann's wartime stability analysis and the Lax equivalence theorem of the 1950s established the modern theory, and difference methods remain a staple of computational physics and engineering.

Key figures

Richard Courant
Kurt Friedrichs
Hans Lewy
John von Neumann
Randall J. LeVeque

Seminal works

leveque2007
morton2005

Frequently asked questions

What is the difference between an explicit and an implicit scheme?: An explicit scheme computes each new grid value directly from known values but is stable only for small time steps, whereas an implicit scheme solves a coupled system for all new values at once, allowing much larger stable time steps at the cost of a linear solve per step.
Why might finite differences be preferred over finite elements?: On simple, regular geometries finite differences are easy to implement, cheap, and straightforward to make high-order. Finite elements become advantageous mainly when the domain has complex shape or the problem has a natural variational formulation.