minlca::utils::SimulatedAnnealing< Node > Class Template Reference

Generic simulated annealing algorithm. More...

#include <simulated-annealing.hh>

Public Member Functions
	SimulatedAnnealing (int steps, int stIter, int k, double lambda, int seed=0)
	Constructor with seed. More...
	SimulatedAnnealing (int maxSteps, int stIter, int k, double lambda, Random &r)
	Constructor using pre-built random number generator. More...
	~SimulatedAnnealing ()
	Default destructor. More...
void	init (Node *start)
	Initialiser for the class. More...
virtual double	temperature () const
	Returns the current temperature. More...
Node *	run ()
	Run the algorithm. More...
Node *	best ()
	Get the best node. More...

Protected Types
typedef lemon::Random	Random

Protected Member Functions
void	runStep ()
	Run a single step. More...

Protected Attributes
Node *	_current
	Current node.
Node *	_best
int	_steps
	Total steps done.
int	_max_steps
	Maximum number of steps.
int	_stiter
	Steps per iteration.
int	_k
	Coefficient for the temperature.
double	_lambda
	Coefficient in the exponent of the temperature.
bool	_delete_random
	Delete the random number generator in destructor.
Random *	_r
	Pointer to random number generator.

Detailed Description

template<typename Node>
class minlca::utils::SimulatedAnnealing< Node >

Generic simulated annealing algorithm.

Definition at line 17 of file simulated-annealing.hh.

Constructor & Destructor Documentation

template<typename Node>

minlca::utils::SimulatedAnnealing< Node >::SimulatedAnnealing ( int  steps, int  stIter, int  k, double  lambda, int  seed = 0  )

inline

Constructor with seed.

Constructor for the class requesting an initial seed for the random number generator.

Parameters

[in]	steps	Maximum number of steps
[in]	stIter	Number of steps per iteration group
[in]	k	Coefficient for temperature function
[in]	lambda	Factor in the exponent of the temperature
[in]	seed	Seed for the random number generator

Definition at line 89 of file simulated-annealing.hh.

template<typename Node>

minlca::utils::SimulatedAnnealing< Node >::SimulatedAnnealing ( int  maxSteps, int  stIter, int  k, double  lambda, Random &  r  )

inline

Constructor using pre-built random number generator.

Constructor for the class requiring an existing random number generator.

Parameters

[in]	maxSteps	Maximum number of steps
[in]	stIter	Number of steps per iteration group
[in]	k	Coefficient for temperature function
[in]	lambda	Factor in the exponent of the temperature
[in]	r	Already initialised random number generator

Definition at line 106 of file simulated-annealing.hh.

template<typename Node>

minlca::utils::SimulatedAnnealing< Node >::~SimulatedAnnealing ( )

inline

Default destructor.

This destructor takes care of deleting the random number generator if it was not given by the user.

Warning

The destructor does not delete the returned search node from run(), you need to delete it on your own.

See also

run()

Definition at line 127 of file simulated-annealing.hh.

Member Function Documentation

template<typename Node>

Node* minlca::utils::SimulatedAnnealing< Node >::best ( )

inline

Get the best node.

Returns the best node found, the same obtained after executing run().

See also

run()

Definition at line 203 of file simulated-annealing.hh.

template<typename Node>

void minlca::utils::SimulatedAnnealing< Node >::init ( Node *  start )

inline

Initialiser for the class.

Resets the status of the search and sets the initial node to start.

Warning

This function does not remove the returned node from run(), you need to delete it for yourself.

See also

run()

Parameters

start

Initial node for the search

Definition at line 146 of file simulated-annealing.hh.

template<typename Node>

Node* minlca::utils::SimulatedAnnealing< Node >::run ( )

inline

Run the algorithm.

Runs the whole search. Needs to have the initial search node set by using function SimulatedAnnealing::init(Node*).

Warning

Make sure to delete the contents of the returned pointer when it is not needed anymore. Neither the destructor nor init(Node*) take care of this.

See also

init(Node*)

~SimulatedAnnealing()

Definition at line 178 of file simulated-annealing.hh.

template<typename Node>

void minlca::utils::SimulatedAnnealing< Node >::runStep ( )

inlineprotected

Run a single step.

This function runs a single simulated annealing step. If SA_VERBOSE is defined during compilation, this function writes to standard output the information about the current step.

Definition at line 36 of file simulated-annealing.hh.

template<typename Node>

virtual double minlca::utils::SimulatedAnnealing< Node >::temperature ( ) const

inlinevirtual

Returns the current temperature.

Returns the current temperature using the formula

\[ k\,\exp\left( -\lambda\,\left\lfloor\frac{steps}{stiter}\right\rfloor\,stiter \right) \]

Definition at line 162 of file simulated-annealing.hh.

---

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

• utils/simulated-annealing.hh