The paper introduces AAg, a nonlinear-residual variant of Anderson acceleration, proves sharp linear convergence rates for arbitrary depth on Picard iterations for Navier-Stokes, and proposes an adaptive depth strategy validated by numerical tests.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.NA 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A modified Anderson acceleration with sharp linear convergence rate predictions and application to incompressible flows
The paper introduces AAg, a nonlinear-residual variant of Anderson acceleration, proves sharp linear convergence rates for arbitrary depth on Picard iterations for Navier-Stokes, and proposes an adaptive depth strategy validated by numerical tests.