248 views
Let \(A\) be a real matrix, not necessarily square.

a) Show that both \(A^{*} A\) and \(A A^{*}\) are self-adjoint.
b) Show that both \(A^{*} A\) and \(A A^{*}\) are positive semi-definite.
c) Show that \(\operatorname{ker} A=\operatorname{ker} A^{*} A\). [HINT: Show separately that \(\operatorname{ker} A \subset \operatorname{ker} A^{*} A\) and ker \(A \supset \operatorname{ker} A^{*} A\). The identity \(\left\langle\vec{x}, A^{*} A \vec{x}\right\rangle=\langle A \vec{x}, A \vec{x}\rangle\) is useful.]
d) If \(A\) is one-to-one, show that \(A^{*} A\) is invertible
e) If \(A\) is onto, show that \(A A^{*}\) is invertible.
f) Show that the non-zero eigenvalues of \(A^{*} A\) and \(A A^{*}\) agree. Generalize.
g) Show that image \(\left(A A^{*}\right)=\left(\operatorname{ker} A A^{*}\right)^{\perp}=\left(\operatorname{ker} A^{*}\right)^{\perp}=\operatorname{image} A\).
55% Accept Rate Accepted 8191 answers out of 14789 questions

Please log in or register to answer this question.

This site uses cookies to provide quality services and to analyze traffic. For more information, see the Privacy Policy