mets::abstract_search< move_manager_type > Class Template Reference

An abstract search. More...

#include <abstract-search.hh>

Inheritance diagram for mets::abstract_search< move_manager_type >:

Collaboration diagram for mets::abstract_search< move_manager_type >:

Public Types
enum	{   MOVE_MADE = 0 , IMPROVEMENT_MADE , ITERATION_BEGIN , ITERATION_END ,   LAST }

Public Member Functions
	abstract_search (feasible_solution &working, solution_recorder &recorder, move_manager_type &moveman)
	Set some common values needed for neighborhood based metaheuristics.
	abstract_search (const abstract_search< move_manager_type > &)
	purposely not implemented (see Effective C++)
abstract_search &	operator== (const abstract_search< move_manager_type > &)
	purposely not implemented (see Effective C++)
virtual	~abstract_search ()
	Virtual destructor.
virtual void	search () throw (no_moves_error) =0
	This method starts the search.
const solution_recorder &	recorder () const
	The solution recorder instance.
const feasible_solution &	working () const
	The current working solution.
feasible_solution &	working ()
const move &	current_move () const
	The last move made.
move &	current_move ()
	The last move made.
const move_manager_type &	move_manager () const
	The move manager used by this search.
move_manager_type &	move_manager ()
	The move manager used by this search.
int	step () const
	The current step of the algorithm (to be used by the observers).
Public Member Functions inherited from mets::subject< abstract_search< move_manager_type > >
virtual void	attach (observer< abstract_search< move_manager_type > > &o)
	Attach a new observer to this subject.
virtual void	detach (observer< abstract_search< move_manager_type > > &o)
	Detach a new observer to this subject.
virtual void	notify ()
	Notify all attached observers.

Protected Attributes
solution_recorder &	solution_recorder_m
feasible_solution &	working_solution_m
move_manager_type &	moves_m
move_manager_type::iterator	current_move_m
int	step_m
Protected Attributes inherited from mets::subject< abstract_search< move_manager_type > >
std::set< observer< abstract_search< move_manager_type > > * >	observers_m

Detailed Description

template<typename move_manager_type>
class mets::abstract_search< move_manager_type >

An abstract search.

See also

mets::tabu_search, mets::simulated_annealing, mets::local_search

Member Enumeration Documentation

anonymous enum

template<typename move_manager_type>

anonymous enum

Enumerator
MOVE_MADE	We just made a move.
IMPROVEMENT_MADE	Our solution_recorder_chain object reported an improvement.
ITERATION_BEGIN	We are about to start a new iteration.
ITERATION_END	We have done the iteration.
LAST	Placeholer for next values.

Constructor & Destructor Documentation

abstract_search()

template<typename move_manager_type>

mets::abstract_search< move_manager_type >::abstract_search ( feasible_solution & working, solution_recorder & recorder, move_manager_type & moveman )

inline

Set some common values needed for neighborhood based metaheuristics.

Parameters

working	The starting point solution (this will be modified during search as the working solution)
recorder	A solution recorder instance used to record the best solution found
moveman	A problem specific implementation of the move_manager_type used to generate the neighborhood.

Member Function Documentation

search()

template<typename move_manager_type>

virtual void mets::abstract_search< move_manager_type >::search ( ) throw ( no_moves_error )

pure virtual

This method starts the search.

Remember that this is a minimization.

An exception mets::no_moves_error can be risen when no move is possible.

Implemented in mets::local_search< move_manager_type >, mets::simulated_annealing< move_manager_type >, and mets::tabu_search< move_manager_type >.

step()

template<typename move_manager_type>

int mets::abstract_search< move_manager_type >::step ( ) const

inline

The current step of the algorithm (to be used by the observers).

When you implement a new type of search you should set step_m protected variable to the status of the algorithm (0 = "MOVE_MADE", 1 = "IMPROVEMENT_MADE", etc.).

---

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

• metslib/abstract-search.hh