LatNet Builder Manual
2.0.1-11
Software Package for Constructing Highly Uniform Point Sets
|
Class to represent figure of merit based on the t-value of a net. More...
#include <TValue.h>
Inherits NetBuilder::FigureOfMerit::FigureOfMerit.
Public Types | |
typedef std::unique_ptr< LevelCombiner::LevelCombiner > | pCombiner |
Public Member Functions | |
TValue (pCombiner combiner=std::make_unique< LevelCombiner::LevelCombiner >()) | |
Constructor. More... | |
std::unique_ptr< FigureOfMeritEvaluator > | createEvaluator () override |
MeritValue | combine (const RealVector &merits) |
virtual std::string | format () const override |
Output information about the figure of merit. | |
template<> | |
std::string | format () const |
Template specialization of the output information about the figure of merit in the unilevel case. | |
template<> | |
std::string | format () const |
Template specialization of the output information about the figure of merit in the multilevel case. | |
Public Member Functions inherited from NetBuilder::FigureOfMerit::FigureOfMerit | |
virtual | ~FigureOfMerit ()=default |
Virtual default destructor. | |
std::unique_ptr< FigureOfMeritEvaluator > | evaluator () |
Returns a std::unique_ptr to an evaluator for the figure of merit. | |
Class to represent figure of merit based on the t-value of a net.
The computation if the t-value uses Algorithm 1 proposed by Dick and Matsumoto in [5]. The multilevel computation of the t-value is based on the second item of Remark 3. Note that this algorithm has strong similiarities with the coordinate-uniform evaluation algorithm used for kernel-based figures. That is why the Storage class from LatBuilder is used here, with the same trick as for CoordUniformFigureOfMerit.
|
inline |
Constructor.
combiner | Combiner multilevel merits. Only used when template parameter ET equals EmbeddingType::MULTILEVEL . |