On the numerical solution of \(AX-XB=C\) (Q2565276)

The paper is concerned with the Galerkin method for computing an approximate solution to \(AX-XB=C\), when \(C\) is a tall rectangular matrix. It is shown that the Galerkin solution can be written as the truncation of the finite power series of the exact solution. Experimental results show that the method can be effective if \(B\) is small so that restart can be applied at low computational cost. Analysis of the pseudo-spectrum of \(A\) and \(B\) also provides some insight into the study of the sensitivity of the method with respect to standard linear solver applied on the Kronecker formulation.