OpenANN::IPOPCMAES Class Reference

Evolution Strategies with Covariance Matrix Adaption and a restart strategy that increases the population size (IPOP-CMA-ES). More...

#include <IPOPCMAES.h>

Inheritance diagram for OpenANN::IPOPCMAES:

Public Member Functions
	IPOPCMAES ()
	Create an instance of IPOPCMAES. More...
virtual	~IPOPCMAES ()
virtual void	setOptimizable (Optimizable &opt)
	Pass the objective function. More...
virtual void	setStopCriteria (const StoppingCriteria &stop)
	Pass the stop criteria. More...
bool	restart ()
	Restart the optimizer. More...
virtual void	optimize ()
	Optimize until the optimization meets the stop criteria. More...
virtual bool	step ()
	Execute one optimization step. More...
Eigen::VectorXd	getNext ()
	Get next parameter vector. More...
void	setError (double fitness)
	Set fitness of last individual. More...
bool	terminated ()
	Did the optimizer finish? More...
virtual Eigen::VectorXd	result ()
	Determine the best result. More...
virtual std::string	name ()
	Get the name of the optimization algorithm. More...
void	setSigma0 (double sigma0)
	Set the initial step size. More...
Public Member Functions inherited from OpenANN::Optimizer
virtual	~Optimizer ()

Detailed Description

Evolution Strategies with Covariance Matrix Adaption and a restart strategy that increases the population size (IPOP-CMA-ES).

IPOP-CMA-ES is an evolutionary optimization algorithm that requires no gradient. The following stopping criteria will be regarged:

• maximalFunctionEvaluations
• maximalIterations
• minimalValue
• minimalValueDifferences
• minimalSearchSpaceStep
• maximalRestarts

IPOPCMAES does not support step-wise execution with step(). Use the functions getNext() and setError() instead to get the next parameter vector and set fitness values respectively.

[1] Hansen and Ostermeier: Completely Derandomized Self-Adaptation in Evolution Strategies. Evolutionary Computation, 9 (2), pp. 159-195, 2001.

[2] Auger and Hansen: A Restart CMA Evolution Strategy With Increasing Population Size. IEEE Congress on Evolutionary Computation, pp. 1769-1776, 2005.

Constructor & Destructor Documentation

OpenANN::IPOPCMAES::IPOPCMAES

(

)

Create an instance of IPOPCMAES.

virtual OpenANN::IPOPCMAES::~IPOPCMAES ( )

virtual

Member Function Documentation

Eigen::VectorXd OpenANN::IPOPCMAES::getNext

(

)

Get next parameter vector.

Returns

parameter vector

virtual std::string OpenANN::IPOPCMAES::name ( )

virtual

Get the name of the optimization algorithm.

Returns

name of the optimization algorithm

Implements OpenANN::Optimizer.

virtual void OpenANN::IPOPCMAES::optimize ( )

virtual

Optimize until the optimization meets the stop criteria.

Implements OpenANN::Optimizer.

bool OpenANN::IPOPCMAES::restart

(

)

Restart the optimizer.

Returns

optimizer is still running

virtual Eigen::VectorXd OpenANN::IPOPCMAES::result ( )

virtual

Determine the best result.

Returns

the best parameter the algorithm found

Implements OpenANN::Optimizer.

void OpenANN::IPOPCMAES::setError

(

double

fitness)

Set fitness of last individual.

Parameters

fitness

fitness

virtual void OpenANN::IPOPCMAES::setOptimizable ( Optimizable &  optimizable)

virtual

Pass the objective function.

Parameters

optimizable

objective function, e. g. error function of an ANN

Implements OpenANN::Optimizer.

void OpenANN::IPOPCMAES::setSigma0

(

double

sigma0)

Set the initial step size.

Parameters

sigma0

initial step size

virtual void OpenANN::IPOPCMAES::setStopCriteria ( const StoppingCriteria &  sc)

virtual

Pass the stop criteria.

Parameters

sc	the parameters used to stop the optimization

Implements OpenANN::Optimizer.

virtual bool OpenANN::IPOPCMAES::step ( )

virtual

Execute one optimization step.

Implements OpenANN::Optimizer.

bool OpenANN::IPOPCMAES::terminated

(

)

Did the optimizer finish?

Returns

terminated

---

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

• OpenANN/optimization/IPOPCMAES.h