Package PyDSTool :: Package Toolbox :: Package optimizers :: Package step :: Module conjugate_gradient_step
[hide private]
[frames] | no frames]

Module conjugate_gradient_step

source code

Computes the conjugate gradient steps for a specific function at a specific point

Classes [hide private]
  ConjugateGradientStep
The basic conjugate gradient step
Functions [hide private]
 
CWConjugateGradientStep()
The Crowder-Wolfe or Hestenes-Stiefel conjugate gradient step
source code
 
DConjugateGradientStep()
The Dixon conjugate gradient step
source code
 
DYConjugateGradientStep()
The Dai Yan conjugate gradient step Has good convergence capabilities (same as the FR-PRP gradient)
source code
 
FRConjugateGradientStep()
The Fletcher Reeves conjugate gradient step Needs an exact line search for convergence or the strong Wolfe-Powell rules for an inexact line search
source code
 
PRPConjugateGradientStep()
The Polak-Ribiere-Polyak conjugate gradient step Can restart automatically, but needs an exact line search with a uniformely convex function to globally converge
source code
 
FRPRPConjugateGradientStep()
The Fletcher-Reeves modified Polak-Ribiere-Polyak conjugate gradient step Can restart automatically and has the advantages of the PRP gradient and of the FR gradient
source code
 
HZConjugateGradientStep()
The Hager-Zhang conjugate gradient step Has good convergence capabilities (same as the FR-PRP gradient)
source code