DistributionGaussian Class Reference

A class for representing a Gaussian distribution. More...

#include <DistributionGaussian.hpp>

Public Member Functions
	DistributionGaussian (const Eigen::VectorXd &mean, const Eigen::MatrixXd &covar)
	Construct the Gaussian distribution with a mean and covariance matrix. More...
void	generateSamples (int n_samples, Eigen::MatrixXd &samples) const
	Generate samples from the distribution. More...
double	maxEigenValue (void) const
	Get the largest eigenvalue of the covariance matrix of this distribution. More...
DistributionGaussian *	clone (void) const
	Make a deep copy of the object. More...
const Eigen::VectorXd &	mean (void) const
	Accessor get function for the mean. More...
const Eigen::MatrixXd &	covar (void) const
	Accessor get function for the covariance matrix. More...
void	set_mean (const Eigen::VectorXd &mean)
	Accessor set function for the mean. More...
void	set_covar (const Eigen::MatrixXd &covar)
	Accessor set function for the covar. More...

Friends
class	boost::serialization::access
	Give boost serialization access to private members. More...
std::ostream &	operator<< (std::ostream &output, const DistributionGaussian &distribution)
	Print to output stream. More...

Detailed Description

A class for representing a Gaussian distribution.

This is mainly a wrapper around boost functionality The reason to make the wrapper is to provide functionality for serialization/deserialization.

Definition at line 44 of file DistributionGaussian.hpp.

Constructor & Destructor Documentation

DistributionGaussian	(	const Eigen::VectorXd &	mean,
		const Eigen::MatrixXd &	covar
	)

Construct the Gaussian distribution with a mean and covariance matrix.

Parameters

[in]	mean	Mean of the distribution
[in]	covar	Covariance matrix of the distribution

Definition at line 55 of file DistributionGaussian.cpp.

{
  mean_ = mean;
  set_covar(covar);
}

Member Function Documentation

void generateSamples	(	int	n_samples,
		Eigen::MatrixXd &	samples
	)		const

Generate samples from the distribution.

Parameters

[in]	n_samples	Number of samples to sample
[in]	samples	the samples themselves (size n_samples X dim(mean)

Definition at line 97 of file DistributionGaussian.cpp.

{
  if (covar_decomposed_.size()==0)
  {
    // Now perform the Cholesky decomposition, which makes it easier to generate samples. 
    MatrixXd A(covar_.llt().matrixL());
    covar_decomposed_ = A;
    // Remark: it would have been better to do this in the constructor and set_covar, 
    // but I couldn't get it to work with boost::serialization (I tried hard)
  }
  
  
  int n_dims = mean_.size();
  samples.resize(n_samples,n_dims);
  
  // http://en.wikipedia.org/wiki/Multivariate_normal_distribution#Drawing_values_from_the_distribution
  VectorXd z(n_dims);
  for (int i_sample=0; i_sample<n_samples; i_sample++)
  {
    // Generate vector with samples from standard normal distribution N(0,1) 
    for (int i_dim=0; i_dim<n_dims; i_dim++)
      z(i_dim) = (*normal_distribution_unit)();

    // Compute x = mu + Az
    samples.row(i_sample) = mean_ + covar_decomposed_*z;
  }  
}

double maxEigenValue

(

void

)

const

Get the largest eigenvalue of the covariance matrix of this distribution.

Returns

largest eigenvalue of the covariance matrix of this distribution.

Definition at line 72 of file DistributionGaussian.cpp.

{
  if (max_eigen_value_<0.0)
  {
    SelfAdjointEigenSolver<MatrixXd> eigensolver(covar_);
    if (eigensolver.info() == Success)
    {
      // Get the eigenvalues
      VectorXd eigen_values = eigensolver.eigenvalues();
      max_eigen_value_ = eigen_values.maxCoeff();
    }
  }
  return max_eigen_value_;
}

DistributionGaussian * clone

(

void

)

const

Make a deep copy of the object.

Returns

A deep copy of the object.

Definition at line 61 of file DistributionGaussian.cpp.

{
  return new DistributionGaussian(mean(),covar()); 
}

const Eigen::VectorXd& mean ( void  ) const

inline

Accessor get function for the mean.

Returns

The mean of the distribution

Definition at line 73 of file DistributionGaussian.hpp.

{ return mean_;   }

const Eigen::MatrixXd& covar ( void  ) const

inline

Accessor get function for the covariance matrix.

Returns

The covariance matrix of the distribution

Definition at line 79 of file DistributionGaussian.hpp.

{ return covar_; }

Here is the call graph for this function:

void set_mean

(

const Eigen::VectorXd &

mean

)

Accessor set function for the mean.

Parameters

[in]

mean

The new mean of the distribution

Definition at line 66 of file DistributionGaussian.cpp.

{ 
  assert(mean.size()==mean_.size());  
  mean_ = mean;
}

void set_covar

(

const Eigen::MatrixXd &

covar

)

Accessor set function for the covar.

Parameters

[in]

covar

The new covariance matrix of the distribution

Definition at line 89 of file DistributionGaussian.cpp.

                                                          { 
  assert(covar.cols()==covar.rows());
  assert(covar.rows()==mean_.size());
  covar_ = covar;
  covar_decomposed_ = MatrixXd::Zero(0,0);
  max_eigen_value_ = -1.0;
}

Friends And Related Function Documentation

friend class boost::serialization::access

friend

Give boost serialization access to private members.

Definition at line 119 of file DistributionGaussian.hpp.

std::ostream& operator<< ( std::ostream &  output, const DistributionGaussian &  distribution  )

friend

Print to output stream.

Parameters

[in]	output	Output stream to which to write to
[in]	distribution	Distribution to write

Returns

Output stream

Definition at line 126 of file DistributionGaussian.cpp.

{
  output << "N([" << toString(distribution.mean_) << "], ["<< toString(distribution.covar_) << "])";
  return output;
}

---

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

• DistributionGaussian.hpp
• DistributionGaussian.cpp