I asked these questions to my TA during workshop but none of them figured it out.
Can anyone help me out?
I asked these questions to my TA during workshop but none of them figured it out.
Can anyone help me out?
On
This is in fact true. The reason is as follows. We work over the complex numbers and write $A^* := \left( \overline A \right)^T$. Now, since $A$ is real and skew-symmetric, we have $A^* = A^T = -A$. This implies $AA^* = A^*A$, i.e. $A$ is normal. It is well known that a complex matrix which is normal can be diagonalized over $\mathbb C$ via a unitary transformation, i.e. there is a complex matrix $U$ with the properties $U^*U = I$ (identity matrix), or equivalently $U^{-1} = U^*$, and $U^*AU = D$ where $D$ is a complex diagonal matrix. Using all these facts, we get $$D^* = \left( U^*AU \right)^* = U^*A^*U = U^* A^T U = -U^*AU = -D.$$ This shows that the diagonal entries of $D$ (i.e. the eigenvalues of $A$) are purely imaginary. Moreover, we get $$U^*(I + A)U = U^*U + U^*AU = I + D.$$ Here, the r.h.s is a diagonal matrix with nonzero entries on the diagonal, since in the sum $I + D$ there's no cancellation between the diagonal entries of $I$ and $D$. So the r.h.s. is invertible, and thus also the l.h.s. This implies that $I + A$ is invertible. Since $-A$ is also real and skew-symmetric, we get the invertibility of $I-A$ by the same argument.
From http://en.wikipedia.org/wiki/Skew-symmetric_matrix#Spectral_theory the eigenvalues of a real skew-symmetric matrix are imaginary.
If $I-A$ were singular, there exists a vector $x$ such that $(I-A)x=0 \Rightarrow Ax=x$, which implies 1 is an eigenvalue of $A$. Contradiction.