Positive definite constrained least-squares estimation of matrices (Q1902117): Difference between revisions

The problem of finding a symmetric matrix \(X\) minimizing \(B-XA\) in the least squares sense, under the condition that all eigenvalues of \(X\) exceed a given positive value, is solved by means of a sequence of quadratic programming problems. A suboptimal variant, which is guaranteed to converge in a finite number of \(QP\) steps, is given as an alternative to the optimal algorithm, which may approach a solution asymptotically.