My problem is the following:
Given a column vector $\mathbf{e''}\in\mathbb{R}^{3}$ with $||\mathbf{e''}||=1$ and a $3\times 3$ matrix $B$ such that $rank\{B\}=3$, I would like to find a vector $\mathbf{v}\in\mathbb{R}^3$ garantying that $rank\{A(\mathbf{v})\}=3$, where $A(\mathbf{v})=(\mathbf{e''}\mathbf{e''}^T-I_{3\times 3})B+\mathbf{e''}\mathbf{v}^T$.
One approach would be to try to maximize $|det\{A(\mathbf{v})\}|$ with respect to $\mathbf{v}$, but I think this would involve non-linear optimization. As my problem seems to relate to rank-one updates, I would rather find a closed-form formula if it exists. I found this thread, which is quite close but only provides a solution to rank-one updates of non-singular matrices. This is not the case here as $rank\{(\mathbf{e''}\mathbf{e''}^T-I_{3\times 3})B\}\leq rank\{\mathbf{e''}\mathbf{e''}^T-I_{3\times 3}\}=2$.
Can someone think of a way to solve this ?
Thanks