Attachment 799 Details for Bug 1480 – Demonstrates problem using eigens AutoDiffScalar type.

This bugzilla service is closed. All entries have been migrated to https://gitlab.com/libeigen/eigen

Demonstrates problem using eigens AutoDiffScalar type.

eigen_ad.cpp (text/x-c++src), 1.37 KB, created by Brad Bell on 2017-10-28 02:47:09 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Brad Bell

Created: 2017-10-28 02:47:09 UTC

Size: 1.37 KB

patch

obsolete

># include <iostream>
># include <Eigen/Sparse>
># include <eigen3/unsupported/Eigen/AutoDiff>
>int main(void)
>{
>	typedef Eigen::AutoDiffScalar<Eigen::VectorXd>     a_scalar;
>	typedef Eigen::Matrix<a_scalar, Eigen::Dynamic, 1> a_vector;
>	//
>	typedef Eigen::SparseMatrix<a_scalar, Eigen::ColMajor>  sparse_matrix;
>    typedef Eigen::SparseLU<sparse_matrix>                  sparse_solver;
>    typedef sparse_matrix::InnerIterator                    column_itr;
>	//
>	a_scalar a;
>	//
>	// declare independent variable vector
>	int n = 2;
>	sparse_matrix a_x(n, 1);
>	//
>	// first independent variable
>	a = 0.0;
>	a.derivatives() = Eigen::VectorXd::Unit(n,0);
>	a_x.insert(0, 0) = a;
>	//
>	// second independent variable
>	a = 4.0;
>	a.derivatives() = Eigen::VectorXd::Unit(n, 1);
>	a_x.insert(1, 0) = a;
>	//
>	// Identity matrix
>	sparse_matrix a_eye(n, n);
>	for(int i = 0; i < n; i++)
>		a_eye.insert(i, i) = 1.0;
>	//
>	sparse_solver solver;
>	solver.analyzePattern(a_eye);
>	solver.factorize(a_eye);
>	//
>	// Instantiate a dependent variable 
>	sparse_matrix a_y = solver.solve( a_x );
>	//
>	for(column_itr itr(a_y, 0); itr; ++itr)
>	{	int r = itr.row();
>		int c = itr.col();
>		a     = itr.value();
>		assert( c == 0 );
>		for(int j = 0; j < n; j++)
>			std::cout << "dy" << r << "_dx" << j << " = " << 
>				a.derivatives()[j] << "\n"; 
>	}
>	std::cout << "Note that y = x so dy0_dx0 should be 1 but is missing.\n";
>	//
>	return 0;
>}

Actions: View

Attachments on bug 1480: 799