Definition
An operator that maps coarse-level variables or solutions to a finer-level grid or representation, typically by interpolation or reconstruction so that information defined on a coarse discretization is represented on a fine discretization.
Principle
Principle
Construct a mapping that preserves consistency with the coarse representation, controls numerical stability and error amplification, and suits the discrete structure (e.g., nested grids, basis functions, or finite-element spaces).
Demonstration
Demonstration
In geometric multigrid for a 1D Poisson problem, a prolongation operator inserts linearly interpolated values on fine-grid nodes between coarse-grid nodes to lift a coarse-grid correction into the fine space for smoothing.
Misapplication
Misapplication
Using a prolongation operator derived for nested uniform grids on a non-nested or strongly anisotropic mesh without modification, producing spurious oscillations and loss of convergence.
Consequence
Consequence
When designed appropriately, prolongation enables effective transfer of coarse corrections to the fine level, accelerates iterative solvers, and preserves important solution structure such as conservation or continuity.
Reversal
Reversal
The natural inverse concept is the restriction operator, which maps fine-level residuals or fields to a coarse-level representation, usually by averaging or projection rather than interpolation.
Boundary
Boundary
Applies to mappings between discretizations or model resolutions; it does not by itself solve equations or invent fine-scale physics and may be invalid when no clear geometrical or functional embedding exists.
Semantic Tension
Semantic Tension
Prolongation is often conflated with generic interpolation or upsampling; the tension is between mathematically consistent prolongation operators tied to discretization spaces and ad hoc interpolation that ignores operator compatibility.
Synthesis
Synthesis
A prolongation operator is the discretization-aware interpolation/reconstruction used to lift coarse solutions into a finer representational space while maintaining numerical compatibility and controlled error behavior.