Exercise 2.5#
In the textbook, we define an operation as
To check whether eq. (1) defines an inner product, we need to check the following properties of inner product.
\((\cdot, \cdot)\) is linear in the second argument. Suppose we have \(N\) vectors and the \(k-\)th vector is defined by
(2)#\[ \mathbf{z}_{k} = (z_{k1}, z_{k2}, \dotsc, z_{kn}) \]Then we can compute
(3)#\[ \sum_{k=1}^{N} \lambda_k\mathbf{z}_{k} = \left(\sum_{k=1}^{N}\lambda_kz_{k1}, \sum_{k=1}^{N}\lambda_kz_{k2}, \dotsc, \sum_{k=1}^{N}\lambda_kz_{kn}\right) \]From eq. (1), we can compute
(4)#\[ \left((y_1, \dotsc, y_n), \sum_{k} \lambda_k\mathbf{z}_{k}\right) = \sum_{i=1}^n y_i^{*}\sum_{k=1}^{N}\lambda_kz_{k1} = \sum_{i=1}^n\sum_{k=1}^{N} y_i^{*}\lambda_kz_{k1} \]Meanwhile, we can compute
(5)#\[ \sum_{k=1}^{N}\lambda_k\left((y_1, \dotsc, y_n), (z_{k1}, z_{k2}, \dotsc, z_{kn})\right) = \sum_{k=1}^{N}\lambda_k \sum_{i=1}^{n}y_i^{*}z_i = \sum_{i=1}^n\sum_{k=1}^{N} y_i^{*}\lambda_kz_{k1} \]From eq. (4) and eq. (5), we conclude that the given definition of \((\cdot, \cdot)\) is linear in the second argument.
\((|v\rangle, |w\rangle) = (|w\rangle, |v\rangle)^*\). From eq. (1) we can compute
(6)#\[ ((z_1, \dotsc, z_n), (y_1, \dotsc, y_n)) = \sum_{i}z_i^{*}y_i = \sum_{i}(y_i^*z_i)^* = \left(\sum_{i}y_i^*z_i\right)^* \]where we use relation \(\left(\sum_{i} z_i\right)^* = \sum_i z^*_i\) and \((z_1z_2)^* = z^*_1 z^*_2\) (see appendix). Compare with eq. (1) we could see that
(7)#\[ ((z_1, \dotsc, z_n), (y_1, \dotsc, y_n)) = \left(\sum_{i}y_i^*z_i\right)^* =((y_1, \dotsc, y_n), (z_1, \dotsc, z_n))^* \]Therefore, we could conclude that under the definition in eq. (1), we have \((|v\rangle, |w\rangle) = (|w\rangle, |v\rangle)^*\)
\((|v\rangle, |v\rangle) \geq 0\), and \((|v\rangle, |v\rangle) = 0\) iff \(|v\rangle = 0\). For \((|v\rangle, |v\rangle)\) we have
(8)#\[ ((y_1, \dotsc, y_n), (y_1, \dotsc, y_n)) = \sum_{i}y_i^{*}y_i =\sum_{i} |y_i|^2 \]Since for any complex number \(|y_i|^2 \geq 0\), we should have \(((y_1, \dotsc, y_n), (y_1, \dotsc, y_n)) \geq 0\). Moreover,
If \(((y_1, \dotsc, y_n), (y_1, \dotsc, y_n)) = 0\), from eq. (8) we know that can only have \(|y_1| = \dotsc = |y_n| = 0\), or equivalently, \(|y\rangle = 0\)
If \(|y_1| = \dotsc = |y_n| = 0\), from eq. (8) we know that \(((y_1, \dotsc, y_n), (y_1, \dotsc, y_n)) = 0\).
Therefore, we can conclude that with eq. (1), we have \((|v\rangle, |v\rangle) \geq 0\), and \((|v\rangle, |v\rangle) = 0\) iff \(|v\rangle = 0\)
From the proof above, we can conclude that eq. (1) defines an inner product on \(\mathbb{C}^n\).
Appendix#
We use the following two relations above to prove \((|v\rangle, |w\rangle) = (|w\rangle, |v\rangle)^*\),
To prove these two relations, we need to write complex number in the form \(z_j = x_j + iy_j\). Then we can compute
and also,
Compare above two equations, we have
To prove the second relation, we need to firstly prove \((z_1z_2)^* = z^*_1 z_2^*\). For \((z_1z_2)^*\) we have
Then we have,
For \(z^*_1 z_2^*z_3^*\) and \((z_1z_2z_3)^*\), we can make use of the relation \((z_1z_2)^* = z^*_1 z_2^*\) to prove. That is,
Similarly, we can prove by induction that if
We should have