util::GaussianEliminationAlg Class Reference

#include <GaussianEliminationAlg.h>

Public Member Functions
	GaussianEliminationAlg (float, float)
double	GetDistance (float) const
const std::vector< float > &	SolveEquations (const std::vector< float > &meanVector, const std::vector< float > &sigmaVector, const std::vector< float > &heightVector)
void	FillAugmentedMatrix (const std::vector< float > &meanVector, const std::vector< float > &sigmaVector, const std::vector< float > &heightVector)
void	GaussianElimination ()
const std::vector< float > &	GetSolutions ()
void	Print ()

Private Member Functions
void	FillDistanceLookupTable ()

Private Attributes
float	fDistanceStepSize
float	fDistanceMax
std::vector< double >	fDistanceLookupTable
std::vector< std::vector < double > >	fMatrix
std::vector< float >	fSolutions

Detailed Description

Definition at line 20 of file GaussianEliminationAlg.h.

Constructor & Destructor Documentation

util::GaussianEliminationAlg::GaussianEliminationAlg	(	float	step,
		float	max
	)

Title: GaussianEliminationAlg Class Author: Wes Ketchum (wketc.nosp@m.hum@.nosp@m.lanl..nosp@m.gov)

Description: Class that solves system of linear equations via Gaussian Elimination. Intended for use with RFFHitFitter

Definition at line 17 of file GaussianEliminationAlg.cxx.

{
  fDistanceStepSize = step;
  fDistanceMax = max;

  if(fDistanceStepSize<0 || fDistanceMax<0)
    throw std::runtime_error("Error in GaussianEliminationAlg: Cannot construct with negative step or max.");

  FillDistanceLookupTable();
}

Member Function Documentation

void util::GaussianEliminationAlg::FillAugmentedMatrix	(	const std::vector< float > &	meanVector,
		const std::vector< float > &	sigmaVector,
		const std::vector< float > &	heightVector
	)

Definition at line 72 of file GaussianEliminationAlg.cxx.

{

  fMatrix.resize(meanVector.size());
  for(size_t i=0; i<meanVector.size(); i++){

    fMatrix[i].resize(meanVector.size()+1);
    for(size_t j=0; j<meanVector.size(); j++){
      if(sigmaVector[j] < std::numeric_limits<float>::epsilon()){
        if(i==j)
          fMatrix[i][j] = 1.0;
        else
          fMatrix[i][j] = 0.0;
      }
      else
        fMatrix[i][j] = GetDistance( (meanVector[i]-meanVector[j])/sigmaVector[j] );
    }
    fMatrix[i][meanVector.size()] = heightVector[i];

  }

}

void util::GaussianEliminationAlg::FillDistanceLookupTable ( )

private

Definition at line 28 of file GaussianEliminationAlg.cxx.

{

  fDistanceLookupTable.clear();
  fDistanceLookupTable.reserve(std::ceil(fDistanceMax/fDistanceStepSize)+2);

  //let's be extra safe, and make sure we have more sampling points than we will need
  double x_val=0.0;
  while(x_val <= fDistanceMax+fDistanceStepSize){
    fDistanceLookupTable.push_back( std::exp(x_val*x_val*0.5*-1) );
    x_val += fDistanceStepSize;
  }
  //do one more to be sure to push beyond...
  fDistanceLookupTable.push_back( std::exp(x_val*x_val*0.5*-1) );
}

void util::GaussianEliminationAlg::GaussianElimination

(

)

Definition at line 97 of file GaussianEliminationAlg.cxx.

{

  fSolutions.resize(fMatrix.size(),0.0);

  for(size_t i=0; i<fMatrix.size(); i++){

    for(size_t j=i+1; j<(fMatrix[i].size()-1); j++){
      float scale_value = fMatrix[j][i] / fMatrix[i][i];
      for(size_t k=i; k<fMatrix[i].size(); k++)
        fMatrix[j][k] -= fMatrix[i][k]*scale_value;
    }//end column loop

  }//end row loop

  for(int i=fMatrix.size()-1; i>=0; i--){
    fSolutions[i] = fMatrix[i].back();

    for(size_t j=i+1; j<fMatrix.size(); j++)
      fSolutions[i] -= fMatrix[i][j]*fSolutions[j];

    fSolutions[i] /= fMatrix[i][i];
  }

}

double util::GaussianEliminationAlg::GetDistance

(

float

d

)

const

Definition at line 44 of file GaussianEliminationAlg.cxx.

{
  double d_abs = std::abs(d);
  if(d_abs > fDistanceMax)
    return 0.0;

  size_t low_bin = std::floor(d_abs/fDistanceStepSize);
  return fDistanceLookupTable[low_bin] -
    (d_abs/fDistanceStepSize-(double)low_bin)*(fDistanceLookupTable[low_bin]-fDistanceLookupTable[low_bin+1]);

}

const std::vector<float>& util::GaussianEliminationAlg::GetSolutions ( )

inline

Definition at line 34 of file GaussianEliminationAlg.h.

{ return fSolutions; }

void util::GaussianEliminationAlg::Print

(

)

Definition at line 123 of file GaussianEliminationAlg.cxx.

{
  std::cout << "GaussianEliminationAlg." << std::endl;

  std::cout << "\tLookup table (step=" << fDistanceStepSize << ", max=" << fDistanceMax << ")" << std::endl;
  for(size_t i=0; i<fDistanceLookupTable.size(); i++)
    std::cout << "\t\tGaussian(" << fDistanceStepSize*i << ") = " << fDistanceLookupTable[i] << std::endl;


  std::cout << "\tAugmented matrix " << std::endl;
  for(size_t i=0; i<fMatrix.size(); i++){
    std::cout << "\t\t | ";
    for(size_t j=0; j<fMatrix[i].size()-1; j++)
      std::cout << fMatrix[i][j] << " ";
    std::cout << " | " << fMatrix[i][fMatrix[i].size()-1] << " |" << std::endl;
  }

  std::cout << "\tSolutions" << std::endl;
  for(size_t i=0; i<fSolutions.size(); i++)
    std::cout << "\t\t" << i << " " << fSolutions[i] << std::endl;
}

const std::vector< float > & util::GaussianEliminationAlg::SolveEquations	(	const std::vector< float > &	meanVector,
		const std::vector< float > &	sigmaVector,
		const std::vector< float > &	heightVector
	)

Definition at line 56 of file GaussianEliminationAlg.cxx.

{

  if(meanVector.size()!=sigmaVector.size() ||
     meanVector.size()!=heightVector.size())
    throw std::runtime_error("Error in GaussianEliminationAlg: Vector sizes not equal.");

  FillAugmentedMatrix(meanVector,sigmaVector,heightVector);
  GaussianElimination();

  return fSolutions;

}

Member Data Documentation

std::vector<double> util::GaussianEliminationAlg::fDistanceLookupTable

private

Definition at line 41 of file GaussianEliminationAlg.h.

float util::GaussianEliminationAlg::fDistanceMax

private

Definition at line 40 of file GaussianEliminationAlg.h.

float util::GaussianEliminationAlg::fDistanceStepSize

private

Definition at line 39 of file GaussianEliminationAlg.h.

std::vector< std::vector<double> > util::GaussianEliminationAlg::fMatrix

private

Definition at line 45 of file GaussianEliminationAlg.h.

std::vector<float> util::GaussianEliminationAlg::fSolutions

private

Definition at line 46 of file GaussianEliminationAlg.h.

---

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

• GaussianEliminationAlg.h
• GaussianEliminationAlg.cxx