I have two very simple (stupid) questions about quadratic forms.

  1. Having any matrices $A,B$ and vectors $x,y$ (real/complex, singular/regular, rectangular, infinite size, etc.) with appropriate size (so the difference exist but other than scalar) does this always hold (distributive law)? $$(Ax-By)^H(Ax-By)=x^HA^HAx-y^HB^HAx-x^HA^HBy+y^HB^HBy$$ ($()^H$ is hermitian transpose). I know that it is trivial and it holds for ''usual'' cases, I am just curious if it holds for any (imaginable) case.

  2. Is it true that having complex entries $y^HB^HAx\ne x^HA^HBy$ (after transposing, it becomes complex conjugate), but for any real case $y^HB^HAx = x^HA^HBy$?


  • $\begingroup$ I think yes and yes. $\endgroup$ – Peter Franek Dec 15 '14 at 17:33
  1. Yes, of course. Assuming you are comfortable with the following facts:

a) Matrix multiplication is associative

b) Matrix multiplication is distributive

c) $(AB)^H = B^HA^H$

d) $(A+B)^H = A^H+B^H$

Then what you have written is an immediate consequence of these properties.

  1. Yes. If $a$ is a real scalar, then $a^H = a$. Nothing changes if you write $a$ as a product of complex matrices.

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.