So we assume $M_{2\times 2}(\Bbb{C}) $ has inner product $\langle A,B\rangle=$trace${(A^T\bar{B})}$

How to find an orthogonal basis for $W=\{\begin{bmatrix}a&b \\c&d\end{bmatrix}|a+b+c=0,a=d\}$ defined under $M_{2\times 2}(\Bbb{C}) $

In my opinion, the subspace is spanned by the vectors that satisfy $a=-b-c,\ b=b,\ c=c,\ d=-b-c$. So they will be $(-1,1,0,-1),(-1,0,1,-1)$. To find orthogonal basis we apply Gram–Schmidt process. But the problem is the order. Let's say $v_2=u_2-\frac{\langle v_1,u_2\rangle}{\langle v_1,v_1\rangle}v_1$. In complex inner product space, $\langle v_1,u_2\rangle\ne\langle u_2,v_1\rangle $. Which order should we put?

In addition, if we want to find any orthogonal projection (i.e. $P_w(\begin{bmatrix}1&0\\0&0\end{bmatrix})$) onto $W$, we just find the orthonormal basis of $W$ and then apply $P_w(x)=\sum^k_{i=1} \langle x,w_i\rangle w_i$.

Am I missing anything?

  • $\begingroup$ I think you are on track here in terms of how one would approach this problem. But make sure you understand why that recipe for projection works; for example, before asking for the "orthogonal" projection, you must indicate your inner product (which you rightly employ in the definition of $P_W$). $\endgroup$ – knsam Apr 10 '16 at 13:14

If a+ b+ c= 0, then c= -a- b. If, further, d= a, then any such matrix can be written $\begin{bmatrix}a & b \\ -a- b & a\end{bmatrix}= \begin{bmatrix}a & 0 \\- a & a\end{bmatrix}+ \begin{bmatrix}0 & b \\ -b & 0\end{bmatrix}= a\begin{bmatrix}1 & 0 \\ -1 & 1\end{bmatrix}+ b\begin{bmatrix}0 & 1 \\ -1 & 0\end{bmatrix}$

So those two matrices form a basis. As for "orthogonality" how are you defining "orthogonal" for matrices? What inner product are you using?


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.