irls.3.man

COOOL：CWP面向对象最优化库（CWP Object Oriented Optimization Library） COOOL是C++类的一个集合

IRLS(derived)     OPTIMIZATION ALGORITHMS      IRLS(derived)

                                                     Jun  1 15:18
NAME
    Irls - Iterative reweighted least squares

SYNOPSIS
    #include <Irls.hh>

    class IterativeReweightedLS : public QuadraticOptima

        \fIPublic members\fP
            IterativeReweightedLS(int, LinearForward*, Vector<double>*,
            int, int, int, double, double);
            IterativeReweightedLS(int, LinearForward*, Vector<double>*,
            int, int, int, double, double, int);
            ~IterativeReweightedLS();
            Model<double> optimizer(Model<double>&);
            Model<long> optimizer(Model<long>&);

        \fIPrivate members\fP
            PreconditionedCGLS*		pcgls;
            DiagMatrix<double>*		weight;
            double			taper;
            int				niterWeight;
            int				order;
            void			updateWeight(Vector<double>&);

DESCRIPTION
    IterativeReweightedLS()
    This algorithm gives you the flexibility to choose 
    different norms (other than the norm 2) to solve 
    linear systems. A comprehensible description of
    this method can be found in "Robust methods in inverse
    theory", 1988, Inverse Problems 4, by J. Scales and A.
    Gersztenkorn.

DESCRIPTION
    Constructors:
     IterativeReweightedLS(int ???, LinearForward* ???, 
    Vector<double>* ???, int ???, int ???, int ???, 
    double ???, double ???);
    ???...
    
    Methods:
     optimizer(Model<double>&)
      model0:  Initial model for the Irls procedure.
      The optimum model is returned by the function.

CAVEATS
    This procedure is still under development. 

DEFINED MACROS
    IRLS_HH

INCLUDED FILES
    "PreCGLS.hh"

SOURCE FILES
    Irls.cc