Let $\mathbf{v}_1,\mathbf{v}_2,\ldots,\mathbf{v}_n$ column vectors, each with the same $n$ components. So: \begin{equation} \mathbf{v}_i = \left[\begin{array}{c}v_i\\ v_i \\\vdots \\ v_i\end{array}\right] \end{equation} Also let \begin{equation} \mathbf{b} = \left[\begin{array}{c}b\\ b \\\vdots \\ b\end{array}\right] \end{equation} and \begin{equation} \mathbf{C} = D(c_1,c_2,\ldots,c_n) \end{equation} where $D(c_1,c_2,\ldots,c_n)$ denotes the diagonal square matrix with $c_1,c_2,\ldots,c_n$ on the diagonal.

All the $v$, $b$ and $c$ are known to be positive.

How would you solve the equation \begin{equation} (\,(\mathbf{v}_1,\mathbf{v}_2,\ldots,\mathbf{v}_n)+\mathbf{C}\,)\,\mathbf{x} = \mathbf{b} \end{equation} Any pointers in the right direction would be appreciated. I'm essentially looking for a formula for each of the $x_i$ (each component of the solution $\mathbf{x}$).

Is the determinant of $(\,(\mathbf{v}_1,\mathbf{v}_2,\ldots,\mathbf{v}_n)+\mathbf{C}\,)$ at least positive? That would at least guarantee that a solution always exists.


If you define $e=[1,1,1,\ldots,1]$ and $u=[v_1,v_2,\ldots,v_n]$, then the coefficient matrix can be written as $$ C + eu^T.$$

Since $C$ is diagonal, it is invertible if all $c_i\neq0$. With the assumption of invertibility, you can apply Sherman-Morrison formula to obtain the inverse of the coefficient matrix, also assuming that $1+uC^{-1}e\neq 0$. The solution of a linear system is fairly straightforward once you know the inverse of a matrix.

  • $\begingroup$ Ah! I had seen this formula a while back, but had forgotten its name. Thank you! $\endgroup$ – cont-math Sep 18 '12 at 1:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.