An algorithm for symmetric generalized inverse eigenvalue problems (Q1124955)

The paper is devoted to the following generalized inverse eigenvalue problem: given \(n\) real numbers \(\lambda_1 \leq \lambda_2 \leq \dots \leq \lambda_n\) find a \(c \in \mathbb{R}^n\) such that \(A(c)x=\lambda B(c)x\) has the prescribed eigenvalues \(\lambda_1, \dots, \lambda_n\) where \(A(c)\) and \(B(c)\) are \(n \times n\) twice continuously differentiable symmetric matrix-valued functions. A new iterative algorithm based on QR-like decomposition with column pivoting and least squares techniques is described. The algorithm is suitable for distinct and multiple eigenvalues as well. Locally quadratic convergence of the iteration process is also proved. Numerical results are reported to illustrate the behavior of the algorithm.