Abstract interface for BundleSolver providing routines that determine the weight of the quadratic term in the augmented model. It also allows the user to specify bounds on the weights by setting minweight and maxweight or to choose the weight to be used in the next iteration. More...
#include <BundleWeight.hxx>
Public Member Functions | |
| BundleWeight (const CBout *cb=0, int incr=-1) | |
| default constructor with the possibility to set the output | |
| virtual void | set_defaults ()=0 |
| Sets default values for 'constant' parameters, e.g. minweight and maxweight. | |
| virtual void | clear ()=0 |
| Resets all adaptive variables and parameters. | |
| virtual int | init (CH_Matrix_Classes::Real aggr_dnormsqr, Groundset *groundset, BundleModel *model)=0 |
| Compute the first weight u and set some parameters,. More... | |
| virtual void | set_next_weight (CH_Matrix_Classes::Real u)=0 |
| Allows the user to set the weight of the very next iteration of the bundle method. More... | |
| virtual void | set_minweight (CH_Matrix_Classes::Real umin)=0 |
| Sets a lower bound for the weight. Nonpositive values may be used to indicate that the weight is allowed to get arbitrarily close to zero. | |
| virtual CH_Matrix_Classes::Real | get_minweight () const =0 |
| Returns the current value of the minweight. | |
| virtual void | set_maxweight (CH_Matrix_Classes::Real umax)=0 |
| Sets an upper bound for the weight. Nonpositive values may be used to indicate that the weight is allowed to get arbitrarily large. | |
| virtual CH_Matrix_Classes::Real | get_maxweight () const =0 |
| Returns the current value of the maxweight. | |
| virtual CH_Matrix_Classes::Real | get_weight () const =0 |
| Returns the current value of the weight. | |
| virtual bool | get_next_weight_set () const =0 |
| returns true if the current weight has been set externally by the user | |
| virtual bool | weight_changed () const =0 |
| Returns false only if the last call to descent_update() or nullstep_update() did not modify the weight and no other routine influenced the weight since then, otherwise it returns true. | |
| virtual int | descent_update (CH_Matrix_Classes::Real newval, CH_Matrix_Classes::Real oldval, CH_Matrix_Classes::Real modelval, const CH_Matrix_Classes::Matrix &y, const CH_Matrix_Classes::Matrix &newy, CH_Matrix_Classes::Real aggr_dnormsqr, BundleProxObject *Hp)=0 |
| The BundleSolver calls this for computing the next weight if the candidate will result in a descent step. More... | |
| virtual int | nullstep_update (CH_Matrix_Classes::Real newval, CH_Matrix_Classes::Real oldval, CH_Matrix_Classes::Real modelval, const CH_Matrix_Classes::Matrix &y, const CH_Matrix_Classes::Matrix &newy, MinorantPointer &new_minorant, MinorantPointer &aggregate, CH_Matrix_Classes::Real nullstep_bound, CH_Matrix_Classes::Real aggr_dnormsqr, BundleProxObject *Hp)=0 |
| The BundleSolver calls this for computing the next weight if if the candidate will result in a null step. More... | |
| virtual int | apply_modification (const GroundsetModification &gsmdf)=0 |
| if the BundleSolver is called to modify the groundset it also calls this More... | |
 Public Member Functions inherited from ConicBundle::CBout | |
| virtual void | set_out (std::ostream *out=0, int print_level=1) |
| Specifies the output level (out==NULL: no output at all, out!=NULL and level=0: errors and warnings, level>0 increasingly detailed information) More... | |
| virtual void | set_cbout (const CBout *cb, int incr=-1) |
| Specifies the output level relative to the given CBout class. More... | |
| void | clear_cbout () |
| reset to default settings (out=0,print_level=1) | |
| CBout (const CBout *cb=0, int incr=-1) | |
| calls set_cbout | |
| CBout (std::ostream *outp, int pl=1) | |
| initialize correspondingly | |
| CBout (const CBout &cb, int incr=0) | |
| copy constructor | |
| virtual bool | cb_out (int pl=-1) const |
| Returns true if out!=0 and (pl<print_level), pl<0 should be used for WARNINGS and ERRORS only, pl==0 for usual output. | |
| std::ostream & | get_out () const |
| If cb_out() returned true, this returns the output stream, but it will abort if called with out==0. | |
| std::ostream * | get_out_ptr () const |
| returns the pointer to the output stream | |
| int | get_print_level () const |
| returns the print_level | |
| virtual int | mfile_data (std::ostream &out) const |
| writes problem data to the given outstream | |
Detailed Description
Abstract interface for BundleSolver providing routines that determine the weight of the quadratic term in the augmented model. It also allows the user to specify bounds on the weights by setting minweight and maxweight or to choose the weight to be used in the next iteration.
Member Function Documentation
◆ apply_modification()
|
pure virtual |
if the BundleSolver is called to modify the groundset it also calls this
The task is then to update the stored information appropriately so that dimensions match and the proximal term makes sense for the modified problem
Implemented in ConicBundle::BundleRQBWeight, and ConicBundle::BundleHKWeight.
Referenced by BundleWeight().
◆ descent_update()
|
pure virtual |
The BundleSolver calls this for computing the next weight if the candidate will result in a descent step.
- Parameters
-
[in] newval objective value in the candidate (next center) [in] oldval objective value in the current (old) center [in] modelval value of the current model in the candidate [in] y current (old) center [in] newy candidate (next center) [in] aggr_dnormsqr squared dual norm of new_aggregate [in] Hp pointer to scaling information
- Returns
- 0 on success
- 1 on failure
Implemented in ConicBundle::BundleRQBWeight, and ConicBundle::BundleHKWeight.
Referenced by BundleWeight().
◆ init()
|
pure virtual |
Compute the first weight u and set some parameters,.
- Parameters
-
[in] aggr_dnormsqr the squared dual norm of the current aggregate [in] groundset pointer to the current groundset [in] model pointer to the current model
- Returns
- 0 on success
- 1 on failure
Implemented in ConicBundle::BundleRQBWeight, and ConicBundle::BundleHKWeight.
Referenced by BundleWeight().
◆ nullstep_update()
|
pure virtual |
The BundleSolver calls this for computing the next weight if if the candidate will result in a null step.
- Parameters
-
[in] newval objective value in the candidate [in] oldval objective value in the current center [in] modelval value of the current model in the candidate [in] y current center [in] newy candidate [in] new_minorant the new subgradient obtained in the candidate (only some rules need this) [in] aggregate the aggregate generating the candidate [in] nullstep_bound evaluationts giving a higer value than this result in null steps [in] aggr_dnormsqr squared dual norm of the aggregate [in] Hp pointer to scaling/variable metric information
- Returns
- 0 on success
- 1 on failure
Implemented in ConicBundle::BundleRQBWeight, and ConicBundle::BundleHKWeight.
Referenced by BundleWeight().
◆ set_next_weight()
|
pure virtual |
Allows the user to set the weight of the very next iteration of the bundle method.
As it is set by the user, the weight will be accepted as long as it is strictly positive, even if it is outside the interval [minweight,maxweight]. Nonpositive values, however, will leave everything unchangend.
Implemented in ConicBundle::BundleRQBWeight, and ConicBundle::BundleHKWeight.
Referenced by BundleWeight().
The documentation for this class was generated from the following file:
