TaoBNCGType#

Determine the conjugate gradient update formula used in the TAOBNCG algorithm.

Synopsis#

typedef enum {
  TAO_BNCG_GD,
  TAO_BNCG_PCGD,
  TAO_BNCG_HS,
  TAO_BNCG_FR,
  TAO_BNCG_PRP,
  TAO_BNCG_PRP_PLUS,
  TAO_BNCG_DY,
  TAO_BNCG_HZ,
  TAO_BNCG_DK,
  TAO_BNCG_KD,
  TAO_BNCG_SSML_BFGS,
  TAO_BNCG_SSML_DFP,
  TAO_BNCG_SSML_BRDN
} TaoBNCGType;

Values#

  • TAO_BNCG_GD - basic gradient descent, no CG update

  • TAO_BNCG_PCGD - preconditioned/scaled gradient descent

  • TAO_BNCG_HS - Hestenes-Stiefel

  • TAO_BNCG_FR - Fletcher-Reeves

  • TAO_BNCG_PRP - Polak-Ribiere-Polyak (PRP)

  • TAO_BNCG_PRP_PLUS - Polak-Ribiere-Polyak “plus” (PRP+)

  • TAO_BNCG_DY - Dai-Yuan

  • TAO_BNCG_HZ - Hager-Zhang (CG_DESCENT 5.3)

  • TAO_BNCG_DK - Dai-Kou (2013)

  • TAO_BNCG_KD - Kou-Dai (2015)

  • TAO_BNCG_SSML_BFGS - Self-Scaling Memoryless BFGS (Perry-Shanno)

  • TAO_BNCG_SSML_DFP - Self-Scaling Memoryless DFP

  • TAO_BNCG_SSML_BRDN - Self-Scaling Memoryless (Symmetric) Broyden

See Also#

Tao, TAOBNCG, TaoBNCGSetType(), TaoBNCGGetType()

Level#

advanced

Location#

include/petsctao.h

Index of all Tao routines
Table of Contents for all manual pages
Index of all manual pages