Computes the cellwise coefficients for the diffusion operator from the given data: $\tilde{A}^M(x) = E_a(x) + \sum_{n=1}^M \sqrt{\lambda_n} \phi_n(x) y_m,$ where $y_m$ realisations of iid random variables on (-.5,.5) are. More...

#include <cellwise.hh>

Inheritance diagram for CellA_M_of_x_MC:

[legend]

Collaboration diagram for CellA_M_of_x_MC:

[legend]

List of all members.

Public Member Functions
	CellA_M_of_x_MC (concepts::DynArray< Real > &A_M_alpha_of_x, const __gnu_cxx::hash_map< uint, uint > &data_level_map, const __gnu_cxx::hash_map< uint, Real > &Ea_of_x, const concepts::Array< Real > &sqrt_lambda, const concepts::Array< concepts::DynArray< Real > > &phi)
	Constructor.
virtual void	operator() (const Element< Real > &elm)=0
	Application operator.
virtual void	operator() (const concepts::Cell &cell)
	Application operator.
virtual void	operator() (const concepts::Element< Real > &elm)
Protected Member Functions
virtual std::ostream &	info (std::ostream &os) const
	Returns information in an output stream.
Private Attributes
concepts::DynArray< Real > &	A_M_
const __gnu_cxx::hash_map < uint, uint > &	data_level_map_
const __gnu_cxx::hash_map < uint, Real > &	Ea_
const uint	M_
	Sum boundary.
const concepts::Array < concepts::DynArray< Real > > &	phi_
concepts::Array< Real >	random_
	Random values.
const concepts::Array< Real > &	sqrt_lambda_

Detailed Description

Computes the cellwise coefficients for the diffusion operator from the given data: $\tilde{A}^M(x) = E_a(x) + \sum_{n=1}^M \sqrt{\lambda_n} \phi_n(x) y_m,$ where $y_m$ realisations of iid random variables on (-.5,.5) are.

This is used for the Monte Carlo method.

Author:: Philipp Frauenfelder, 2004

Definition at line 103 of file cellwise.hh.

Constructor & Destructor Documentation

CellA_M_of_x_MC::CellA_M_of_x_MC	(	concepts::DynArray< Real > &	A_M_alpha_of_x,
		const __gnu_cxx::hash_map< uint, uint > &	data_level_map,
		const __gnu_cxx::hash_map< uint, Real > &	Ea_of_x,
		const concepts::Array< Real > &	sqrt_lambda,
		const concepts::Array< concepts::DynArray< Real > > &	phi
	)

Constructor.

random_ is filled with M_ realisations of an uniformly(-1/2, 1/2) distributed random variable. The same value of this realisation is later used in every physical point x.

Creating a new instance of this class for each Monte Carlo problem instance to be solved then creates a new realisation of the whole diffusion coefficient.

Parameters:

A_M_alpha_of_x	Result, $\tilde{A}^M_\alpha(x)$
data_level_map	Maps indices from data mesh to computational mesh, `phi` is given on the data mesh and the rest on the computational mesh
Ea_of_x
sqrt_lambda	$\{ \sqrt{\lambda_n} \}_{n=1}^M$
phi	$\{ \phi_n(x) \}_{n=1}^M$