Re: [eigen] Cholesky Factorization with AMD ordering

[Gael Guennebaud <gael.guennebaud@xxxxxxxxx>]

You can simply use the SparseLLT class of the SparseCholesky module which use AMD ordering by default. Then, SparseLLT gives you access to the L factor and symmetric permutation vector.

See the sparse chapter for some examples: http://eigen.tuxfamily.org/dox/group__Sparse__chapter.html

gael

On Tue, Jun 24, 2014 at 10:52 PM, Gokturk <poyrazoglu@xxxxxxxx> wrote:

Hello all,

I need your precious help to complete my task.
I was a MATLAB user for a long time, now I'm trying to write a code in C++.

For those of you who are familiar with MATLAB, I simply want to perform

    [L, ~ , s] = chol ( E, 'lower' , 'vector' );
where E is positive definite matrix,
            L is factorized Lower triangular matrix
            s is permutation matrix associated with AMD Ordering

For those of you who are not familiar with MATLAB,
I simply want to perform cholesky factorization to matrix E with AMD Ordering.
I need AMD Ordering to reduce the fill-in number in lower triangular matrix L.
As you may suspected, matrix L is the lower triangular matrix which gives  L * LT = s * E * sT

Now, my problem is I am not able to perform same process with Eigen.
I've realized Eigen provides AMD Ordering for sparse matrices.
But I couldn't understand how I use it.
Could you please help me to write an equivalent version of MATLAB line given above?

-- 
Gokturk Poyrazoglu
Mitsubishi Electric Research Laboratories (MERL)
201 Broadway, Cambridge, MA