I am now want to obtain the eigenvectors as:

(D-W)y=tDy
where D and W are sparse matrices.

The problem can be transformed into a standard eigenvalue problem of

D^(-1/2)*(D-W)*D^(-1/2)*x=t*x

I use EigenSolver to solve this problem but the time is too long.

As one paper suggested, "Lanczos algorithm" should be adopted. So in Eigen, is there any module to solve this kind of eigenvectors of sparse matrix problem?


P.S. I noticed that there is SparseMatrix module in Eigen/Sparse. But I am not sure if this can be used.

so far we don't have any sparse Eigen decomposition algorithm. You will have to use another library, like slepc.

Thanks for your reply. I will find another library.

ggael wrote:so far we don't have any sparse Eigen decomposition algorithm. You will have to use another library, like slepc.