num::inverse_iteration

Defined in header <linear/eigen/power.hpp>

power_result inverse_iteration(const mat<real> &A, real sigma, real tol = 1e-10,
                               idx max_iter = 1000);

Inverse iteration -- finds the eigenvalue closest to a shift sigma.

Factorizes (A - sigmaI) once then solves repeatedly.

Parameters

• A: Square matrix (symmetric recommended)
• sigma: Shift -- should be near the target eigenvalue
• tol: Tolerance on eigenvalue change between iterations
• max_iter: Maximum iterations