LagHeu2 Class Reference

Another lagrange heuristic. More...

#include <lagheu.h>

Inheritance diagram for LagHeu2:

[legend]Collaboration diagram for LagHeu2:

[legend]List of all members.

Public Member Functions
	LagHeu2 (Pointer< MinlpProblem > orig_prob_, Pointer< LinearRelax > linrelax_, bool is_gams_prob=false, double closeval_tol_=0., Pointer< dvector > diam_=NULL, Pointer< Param > param_=NULL, Pointer< ostream > out_solver_p_=out_out_p, Pointer< ostream > out_solver_log_p_=out_log_p)
int	solve (Pointer< MinlpNode > node_)
void	set_reform (Pointer< Reformulation > reform_)
Private Member Functions
void	search (map< pair< double, int >, multimap< double, list< ExtremePoint >::iterator > >::reverse_iterator it_k, double ns)
bool	project (dvector &res)
	Tries to make x feasible.
double	key (const dvector &x)
	Gives key for x.
Private Attributes
dvector	x
map< pair< double, int >, multimap< double, list< ExtremePoint >::iterator > >	W
	{ ((key, k), { (z*_kj, w_kj) }_j }_k
Pointer< MinlpPenaltyFunc >	penalty
map< double, dvector >	candidates
int	max_candidates
int	max_ns
double	ns_done
double	ns_all
int	lastprint

---

Detailed Description

Another lagrange heuristic.

Parameters:

	LagHeu2	max locopt options integer $ 0$ default 100 level 0 Number of best candidates to keep for local optimization.
	LagHeu2	max combinations options integer $ 0$ default 10000 level 0 Upper bound on the number of combinations of extreme points to explore.

Definition at line 198 of file lagheu.h.

---

Constructor & Destructor Documentation

LagHeu2::LagHeu2	(	Pointer< MinlpProblem >	orig_prob_,
		Pointer< LinearRelax >	linrelax_,
		bool	is_gams_prob = false,
		double	closeval_tol_ = 0.,
		Pointer< dvector >	diam_ = NULL,
		Pointer< Param >	param_ = NULL,
		Pointer< ostream >	out_solver_p_ = out_out_p,
		Pointer< ostream >	out_solver_log_p_ = out_log_p
	)

---

Member Function Documentation

void LagHeu2::search	(	map< pair< double, int >, multimap< double, list< ExtremePoint >::iterator > >::reverse_iterator	it_k,
		double	ns
	)			[private]

bool LagHeu2::project

(

dvector &

res

)

[private]

Tries to make x feasible.

Parameters:

res

The vector to store the result in.

Returns:

True, if project_LP was feasible, false else.

double LagHeu2::key

(

const dvector &

x

)

[private]

Gives key for x.

int LagHeu2::solve

(

Pointer< MinlpNode >

node_

)

[virtual]

Implements LagHeu.

void LagHeu2::set_reform

(

Pointer< Reformulation >

reform_

)

[virtual]

Reimplemented from RelaxationSolver.

---

Member Data Documentation

dvector LagHeu2::x [private]

Definition at line 200 of file lagheu.h.

map<pair<double, int>, multimap<double, list<ExtremePoint>::iterator> > LagHeu2::W [private]

{ ((key, k), { (z*_kj, w_kj) }_j }_k

Definition at line 204 of file lagheu.h.

Pointer<MinlpPenaltyFunc> LagHeu2::penalty [private]

Definition at line 206 of file lagheu.h.

map<double, dvector> LagHeu2::candidates [private]

Definition at line 208 of file lagheu.h.

int LagHeu2::max_candidates [private]

Definition at line 209 of file lagheu.h.

int LagHeu2::max_ns [private]

Definition at line 211 of file lagheu.h.

double LagHeu2::ns_done [private]

Definition at line 213 of file lagheu.h.

double LagHeu2::ns_all [private]

Definition at line 214 of file lagheu.h.

int LagHeu2::lastprint [private]

Definition at line 215 of file lagheu.h.

---

The documentation for this class was generated from the following file:

• lagheu.h

---