Let $E$ be a subspace of $\mathbb{R}^4$ that is composed of vectors that support the following system :

$$\begin{cases} x - 2y + z + t = 0 \\2x+y-2z-t = 0 \end{cases}$$

So, I've used Gaussian elimination so I've found the following basis for $E$: $$ B = \{(3,4,5,0), (1,3,0,5)\}. $$

Well now i defined $W= \{ w \in W \mid \langle v, w \rangle = 0 \}$ . What do I have to do next? I used the following system :

$$ \begin{cases}3x_1 + 4x_2 + 5x_3 = 0 \\ 1x_1+3x_2+5x_4 = 0 \end{cases} $$

by using the inner product, where $x_1,x_2,x_3,x_4$ are the elements of an arbitrary vector in $W$. Is this approach correct? If i now find the basis $W$ from this system will i find the orthogonal complement basis of $E$?


Your approach is correct, but you probably mean inner product instead of cross product?

There's also a much easier way to see this. Note that if you take $$ v_1 = \left( \begin{matrix} 1 \\ -2 \\ 1 \\ 1\end{matrix} \right), v_2 = \left( \begin{matrix} 2 \\ 1 \\ -2 \\ -1\end{matrix} \right), $$ ( the components of the vectors are simply the coefficients in the system of equations determining $E$) we have $w \in E$ if and only if $\langle w, v_1 \rangle =0$ and $\langle w, v_2 \rangle = 0$. So in fact $E = \text{span}(v_1,v_2)^\perp$. But then the orthogonal complement is easily found to be $E^\perp = \text{span}(v_1,v_2)$. So you don't actually have to calculate or do any Gaussian elimination at all.

  • $\begingroup$ Thank you for your answer! $\endgroup$ – Eduard6421 Jun 14 '17 at 18:51
  • $\begingroup$ @Eduard6421 You're welcome! $\endgroup$ – Demophilus Jun 14 '17 at 19:02

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.