Linearly constrained positive definite completions (Q2277549): Difference between revisions

A partial matrix is a matrix, where some of the elements are fixed and the rest are unspecified complex numbers. It is shown that if \(A(z)\) is a partial matrix (with \(z\) the vector of unspecified elements) such that \(A(z)\) is positive semidefinite and \(z\) satisfies some linear constraints then if the problem is feasible then there is a unique such matrix \(A(z)\) maximizing the determinant of \(A(z)\). This is applied to the case when \(A(z)\) is positive semidefinite and Toeplitz.