#include <MoochoPack_UpdateReducedSigma_Step.hpp>
Inheritance diagram for MoochoPack::UpdateReducedSigma_Step:
Overridden from AlgorithmStep | |
| bool | do_step (Algorithm &algo, poss_type step_poss, IterationPack::EDoStepType type, poss_type assoc_step_poss) |
| | |
Public Member Functions | |
| void | update_method (const e_update_methods &update_method) |
| update method for the reduced sigma term update_method = always_explicit; update_method = BFGS_primal; update_method = BFGS_dual_no_correction; update_method = BFGS_dual_explicit_correction; *** (default) update_method = BFGS_dual_scaling_correction; These options determine exactly how the reduced sigma term will be updated. | |
| UpdateReducedSigma_Step (const e_update_methods update_method=ALWAYS_EXPLICIT) | |
Definition at line 50 of file MoochoPack_UpdateReducedSigma_Step.hpp.
|
|
Constructor. Definition at line 54 of file MoochoPack_UpdateReducedSigma_Step.cpp. |
|
|
update method for the reduced sigma term update_method = always_explicit; update_method = BFGS_primal; update_method = BFGS_dual_no_correction; update_method = BFGS_dual_explicit_correction; *** (default) update_method = BFGS_dual_scaling_correction; These options determine exactly how the reduced sigma term will be updated. always_explicit : the full Z_kT*Sigma*Zk at each step (expensive) BFGS_primal : a BFGS update of mu*X^-2 (exact at solution) BFGS_dual_no_correction : update with Z_kT*Sigma*Z_k*pz (no correction when mu changes) BFGS_dual_explicit_correction : same as above (do an explicit calculation when mu changes) BFGS_dual_scaling_correction : same as above (scale by mu_kp1/mu_k when mu changes) Definition at line 82 of file MoochoPack_UpdateReducedSigma_Step.hpp. |
|
||||||||||||||||||||
|
Definition at line 61 of file MoochoPack_UpdateReducedSigma_Step.cpp. |
1.3.9.1