StopCriteria Class Reference

#include <solver.hpp>

List of all members.

---

Detailed Description

Class of stop criteria used in gradient method minimization of J_{alpha}(v) TODO: rename this class.

Class describes stop criterias used in gradient method minimization of Tikhonov regularization functional J_{alpha}(v) where J_{alpha}(v) = alpha * square_norm_{L2([0;H]x[t1;t2])}(v) + square_norm_{L2[0;H]}(qT) where v is source function, qT(x)=h(x,T)-h_obs(x) where h is result of solving forward linear problem.

Definition at line 135 of file solver.hpp.

Public Types
enum	stop_type { stop_max_iters, stop_max_iters_calc_func, stop_func }
	StopCriteria mode. More...
Public Member Functions
	StopCriteria (unsigned max_iter, bool calc_func=false)
	Initialized stop criteria by max_iters.
	StopCriteria (double tol)
	Initialized stop criteria by relative error of minimized functional.
void	Reset ()
	Reset iterations to start.
stop_type	GetType () const
	Get type of stop criteria.
bool	IsStop () const
	Check is stop criteria satisfied.
void	UpdateCurIter ()
	Update number of current iteration.
unsigned	GetCurIter () const
	Get number of current iteration.
bool	NeedCalcFunc () const
	Check is calculating functional is needed.
void	CalcFunc (unsigned N, double h, const double *qT, double alpha, double square_norm_v)
	Calculate functional J_{alpha}(v).
double	GetCurFunc () const
	Get value of functional on current iteration.
Classes
union	un_param

---

Member Enumeration Documentation

enum StopCriteria::stop_type

StopCriteria mode. Enumeration values: stop_max_iters  stop by maximum number of iterations stop_max_iters_calc_func  as stop_max_iters but also calculate minimized functional stop_func  stop by relative error of minimized functional Definition at line 138 of file solver.hpp.

---

Constructor & Destructor Documentation

StopCriteria::StopCriteria (  unsigned  max_iter, bool  calc_func = false )

Initialized stop criteria by max_iters. Parameters: max_iter  maximum number of iterations calc_func  if true functional will be calculate on each iteration Definition at line 204 of file solver.cpp.

StopCriteria::StopCriteria (  double  tol  )

Initialized stop criteria by relative error of minimized functional. Parameters: tol  tolerance of relative error of functional Definition at line 211 of file solver.cpp.

---

Member Function Documentation

void StopCriteria::Reset (   )

Reset iterations to start. Definition at line 217 of file solver.cpp.

StopCriteria::stop_type StopCriteria::GetType (   )  const

Get type of stop criteria. Definition at line 224 of file solver.cpp. References stop_func, stop_max_iters, and stop_max_iters_calc_func.

bool StopCriteria::IsStop (   )  const

Check is stop criteria satisfied. Definition at line 244 of file solver.cpp. References stop_func.

void StopCriteria::UpdateCurIter (   )

Update number of current iteration. Definition at line 162 of file solver.hpp.

unsigned StopCriteria::GetCurIter (   )  const

Get number of current iteration. Definition at line 164 of file solver.hpp.

bool StopCriteria::NeedCalcFunc (   )  const

Check is calculating functional is needed. Definition at line 239 of file solver.cpp. References stop_max_iters.

void StopCriteria::CalcFunc (  unsigned  N, double  h, const double *  qT, double  alpha, double  square_norm_v )

Calculate functional J_{alpha}(v). Calc J_{alpha}(v) = alpha * square_norm_{L2([0;H]x[t1;t2])}(v) + square_norm_{L2[0;H]}(qT) where qT(x)=h(x,T)-h_obs(x) where h is result of solving forward linear problem. Parameters: N  size of uniform grid on particles h  step of uniform grid on particles qT  double vector of size N+1 alpha  regularization parameter square_norm_v  equal square_norm_{L2([0;H]x[t1;t2])}(v) where v is source function Definition at line 253 of file solver.cpp.

double StopCriteria::GetCurFunc (   )  const

Get value of functional on current iteration. Definition at line 181 of file solver.hpp.

---

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

• solver.hpp
• solver.cpp

---