suppose that $C$ is full column rank matrix, can we say that the following equality is true:

$$\operatorname{rank}\biggl(\begin{bmatrix} A & 0\\B & C \end{bmatrix}\biggr)= \operatorname{rank}\biggl(\begin{bmatrix} A \\B \end{bmatrix}\biggr) +\operatorname{rank}(C)$$

I have an idea about it:

$$\begin{bmatrix} A & 0\\B & C \end{bmatrix}=\begin{bmatrix} A & 0\\B & 0 \end{bmatrix}+\begin{bmatrix} 0 & 0\\0 & C \end{bmatrix}$$.

Now using the inequality, $\text{rank}(A+B)\leq \text{rank}(A)+\text{rank}(B)$ results in

$$\operatorname{rank}\biggl(\begin{bmatrix} A & 0\\B & C \end{bmatrix}\biggr)\leq \operatorname{rank}\biggl(\begin{bmatrix} A \\B \end{bmatrix}\biggr) +\operatorname{rank}(C).$$

The equality holds whenever ${R}\left(\begin{bmatrix} A \\B \end{bmatrix}\right) \cap R(C)=0$ and $C\left(\begin{bmatrix} A \\B \end{bmatrix}\right) \cap C(C)=0$.

But how to conclude it in this case?!

  • $\begingroup$ See this $\endgroup$ – user376343 Sep 17 '18 at 14:55

You cannot say the equality is true. For example,


  • $\begingroup$ Yes, You are right. But under what conditions we can say that?! $\endgroup$ – im hs Sep 17 '18 at 12:54
  • $\begingroup$ It should be true since I see this in many papers and books! but without any proof and refer! it seems that it should be trivial!! $\endgroup$ – im hs Sep 17 '18 at 12:54
  • $\begingroup$ and how about if $A$ is invertible? $\endgroup$ – im hs Sep 17 '18 at 12:56
  • $\begingroup$ @imhs please refer us to an example of a paper or book saying such a thing. "should be true" says nothing if you are given a counter example. $\endgroup$ – GuySa Sep 17 '18 at 13:30
  • $\begingroup$ @GuySa please look at this for example: robust model-based fault diagnosis for dynamic systems by Chen and Ron Patton on page 75. $\endgroup$ – im hs Sep 17 '18 at 16:57

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.