Image Component Library (ICL)
Public Types | Public Member Functions | Protected Member Functions | Private Attributes | List of all members
icl::math::StochasticOptimizer< T > Class Template Referenceabstract

The StochasticOptimizer is a tiny frame-work for simple stochastic optimization processes. More...

#include <StochasticOptimizer.h>

Public Types

typedef StochasticOptimizerResult< T > Result
 Result structure. More...
 

Public Member Functions

 StochasticOptimizer (int dataDim)
 create a stochastic optimizer with given data dimension More...
 
Result optimize (int maxTimeSteps)
 start optimization process with given step count More...
 
Result optimize (T minError, int maxSteps)
 start optimization process with given step count and minimal error stop-criterion More...
 

Protected Member Functions

virtual T * getData ()=0
 must return the current data vector More...
 
virtual T getError (const T *data)=0
 must return current error value (>=0) More...
 
virtual const T * getNoise (int currentTime, int endTime)=0
 returns a noise vector (of size dataDim, which was passed to the constructor) More...
 
virtual void reinitialize ()=0
 this function is called before the optimization is started More...
 
virtual void notifyProgress (int t, int numSteps, int startError, int currBestError, int currError, const T *data, int dataDim)
 a pure utility function, which can be implemented in derived classes to notify optization progress (somehow) More...
 

Private Attributes

int m_dataDim
 internal data-dimension variable More...
 

Detailed Description

template<class T = float>
class icl::math::StochasticOptimizer< T >

The StochasticOptimizer is a tiny frame-work for simple stochastic optimization processes.

Stochastic optimization is minimization of an error function, which depends on a set of latent variables (here: called data). The naive approach origins from starting configuration, which must be created by the pure virtual reinitialize function. Now in each optimization step, the current data vector is changed slightly using an additive noise vector which must be provided by the also pure virtual getNoise-function. If the error arising from the changed data is less then the current minimal error, then the change data vector use hold, otherwise the last change of data is reverted. This procedure is iterated until either a maximum number of iterations is reached or until a given minimal error is reached.

NEW** now this class is a template (defined for float and double)

Member Typedef Documentation

template<class T = float>
typedef StochasticOptimizerResult<T> icl::math::StochasticOptimizer< T >::Result

Result structure.

Constructor & Destructor Documentation

template<class T = float>
icl::math::StochasticOptimizer< T >::StochasticOptimizer ( int  dataDim)

create a stochastic optimizer with given data dimension

Member Function Documentation

template<class T = float>
virtual T* icl::math::StochasticOptimizer< T >::getData ( )
protectedpure virtual

must return the current data vector

data is un-const, as it is changed in each step if the reached error is less then the current best error. Returned pointer must have at least the length of dataDim passed to the constructor

template<class T = float>
virtual T icl::math::StochasticOptimizer< T >::getError ( const T *  data)
protectedpure virtual

must return current error value (>=0)

returns the error measurement dependent on given data vector

template<class T = float>
virtual const T* icl::math::StochasticOptimizer< T >::getNoise ( int  currentTime,
int  endTime 
)
protectedpure virtual

returns a noise vector (of size dataDim, which was passed to the constructor)

optionally, the noise-strength might depend on the current time-progress. Therefore, currentTime and endTime is also passed to this functions

template<class T = float>
virtual void icl::math::StochasticOptimizer< T >::notifyProgress ( int  t,
int  numSteps,
int  startError,
int  currBestError,
int  currError,
const T *  data,
int  dataDim 
)
protectedvirtual

a pure utility function, which can be implemented in derived classes to notify optization progress (somehow)

template<class T = float>
Result icl::math::StochasticOptimizer< T >::optimize ( int  maxTimeSteps)
inline

start optimization process with given step count

template<class T = float>
Result icl::math::StochasticOptimizer< T >::optimize ( minError,
int  maxSteps 
)

start optimization process with given step count and minimal error stop-criterion

template<class T = float>
virtual void icl::math::StochasticOptimizer< T >::reinitialize ( )
protectedpure virtual

this function is called before the optimization is started

Internally data must be initialized

Member Data Documentation

template<class T = float>
int icl::math::StochasticOptimizer< T >::m_dataDim
private

internal data-dimension variable


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