CGAL 6.0.1  Shape Regularization

A concept that describes the set of methods used by the class QP_regularization
to access various data required for setting up the the global regularization problem.
Public Member Functions  
FieldNumberType  bound (const std::size_t query_index) const 
returns the maximum bound on a regularization characteristic (angleorientation/ distanceoffset/etc.) with respect to which a geometric object with the index query_index is being regularized.  
FieldNumberType  target (const std::size_t i, const std::size_t j) 
returns an objective function value between two geometric objects, which are direct neighbors, that is they form a neighbor pair i <> j .  
void  update (const std::vector< FieldNumberType > &solution) 
updates regularization characteristics (angleorientation/distanceoffset/etc.) of the geometric objects being regularized using values from solution , one value per one regularized object.  
FieldNumberType CGAL::Shape_regularization::RegularizationType::bound  (  const std::size_t  query_index  )  const 
returns the maximum bound on a regularization characteristic (angleorientation/ distanceoffset/etc.) with respect to which a geometric object with the index query_index
is being regularized.
QP_regularization
calls this method once for each object from the input range.
FieldNumberType CGAL::Shape_regularization::RegularizationType::target  (  const std::size_t  i, 
const std::size_t  j  
) 
returns an objective function value between two geometric objects, which are direct neighbors, that is they form a neighbor pair i <> j
.
Neighbors are provided by the concept NeighborQuery
.
QP_regularization
calls this method once for each neighbor pair being regularized.
void CGAL::Shape_regularization::RegularizationType::update  (  const std::vector< FieldNumberType > &  solution  ) 
updates regularization characteristics (angleorientation/distanceoffset/etc.) of the geometric objects being regularized using values from solution
, one value per one regularized object.
These values depend on what is being regularized, they could be angle or offset differences for example. The solution vector is computed by the QuadraticProgramTraits
.
Number of values in solution
equals to the number n of geometric objects being regularized + the number m of neighbor pairs between these objects. The first n values are the values that should be used.
QP_regularization
calls this method once after the global regularization QP problem has been solved.