Syntax:
fluid-minimize <key1> <value1> <key2> <value2> ...
Description:
where possible keys and value types are:
- alphaTincreaseFactor: step size increase factor when curvature is wrong in linmin
- alphaTmin : minimum test step size
- alphaTreduceFactor : step size reduction factor when energy increases in linmin
- alphaTstart : initial test step size (constant step-size factor for Relax linmin)
- dirUpdateScheme : FletcherReeves|HestenesStiefel|L-BFGS|PolakRibiere|SteepestDescent (search direction method)
- energyDiffThreshold : convergence threshold for energy difference between successive iterations
- fdTest : no|yes, whether to perform a finite difference test
- history : number of past states and gradients retained for L-BFGS
- knormThreshold : convergence threshold for gradient (preconditioned) norm
- linminMethod : CubicWolfe|DirUpdateRecommended|Quad|Relax (line minimization method)
- nAlphaAdjustMax : maximum step-size adjustments per linmin
- nEnergyDiff : number of iteration pairs that must satisfy energyDiffThreshold
- nIterations : maximum iterations (single point calculation if 0)
- updateTestStepSize : no|yes, whether test step size is updated
- wolfeEnergy : dimensionless energy threshold for Wolfe linmin stopping criterion
- wolfeGradient : dimensionless gradient threshold for Wolfe linmin stopping criterion
Any number of these key-value pairs may be specified in any order.
Properties:
Requires: fluid
Forbids: (None)
Allow multiple: no
Default:
fluid-minimize \
dirUpdateScheme PolakRibiere \
linminMethod DirUpdateRecommended \
nIterations 100 \
history 15 \
knormThreshold 0 \
energyDiffThreshold 0 \
nEnergyDiff 2 \
alphaTstart 1 \
alphaTmin 1e-10 \
updateTestStepSize yes \
alphaTreduceFactor 0.1 \
alphaTincreaseFactor 3 \
nAlphaAdjustMax 3 \
wolfeEnergy 0.0001 \
wolfeGradient 0.9 \
fdTest no
Back to: Input file documentation or Index of commands