MHB Quantum Computing: Positive Operators are Hermitian

Click For Summary
The discussion focuses on proving that a positive operator is necessarily Hermitian, as outlined in Exercise 2.24 of Nielsen and Chuang's text on quantum computation. The solution starts by expressing an arbitrary operator A as the sum of two Hermitian operators, B and C. It establishes that if A is positive, then the expectation value of C must be zero, leading to the conclusion that C equals zero. Consequently, A is shown to be Hermitian since it can be represented as A = B + iC, where C is zero. The conversation suggests a transition to discussions in quantum mechanics or computer science forums for further exploration.
Ackbach
Gold Member
MHB
Messages
4,148
Reaction score
94
Exercise 2.24 on page 71 of Nielsen and Chuang's Quantum Computation and Quantum Information asks the reader to show that a positive operator is necessarily Hermitian. There is a hint given; namely, that you first show an arbitrary operator can be written $A=B+iC$, where $B$ and $C$ are Hermitian. N.B., Nielsen and Chuang are pretty much always working in finite-dimensional Hilbert spaces. Recall that a positive operator $A$ is one such that $\langle x|A|x\rangle\ge 0$ for all vectors $|x\rangle$. An Hermitian operator $A$ is one such that $A=A^{\dagger}$. Here is my solution to the problem:

I claim that $A=B+iC$, where
\begin{align*}
B&=\frac{A+A^{\dagger}}{2} \\
C&=\frac{A-A^{\dagger}}{2i}.
\end{align*}
We can easily verify that $A=B+iC$. Note that
\begin{align*}
B^{\dagger}&=\frac{A^{\dagger}+A}{2}=B \\
C^{\dagger}&=-\frac{A^{\dagger}}{2i}+\frac{A}{2i}=C.
\end{align*}
Hence, $B$ and $C$ are both Hermitian. Now we assume that $A$ is positive, and that $B$ and $C$ are defined as above. For a positive operator, we must have $\langle x|A|x\rangle\ge 0$ for all vectors $|x\rangle$. Since $C$ is Hermitian, it is normal, and hence is diagonalizable. That is, it has a representation
$$C=\sum_i \lambda_i|i\rangle\langle i|,$$
where the $\{|i\rangle\}$ is an orthonormal basis of the space $V$. If it is an orthonormal basis, then we can write
$$|x\rangle=\sum_jx_j|j\rangle.$$
Hence,
\begin{align*}
C|x\rangle&=\sum_i \lambda_i|i\rangle\langle i|\sum_jx_j|j\rangle \\
&=\sum_{i,j}\lambda_ix_j|i\rangle\langle i|j\rangle \\
&=\sum_i\lambda_i x_i|i\rangle.
\end{align*}
Since
$$\langle x|=\sum_j x_j^*\langle j|,$$
we have that
\begin{align*}
\langle x|C|x\rangle&=\sum_j x_j^*\langle j|\sum_i\lambda_i x_i|i\rangle \\
&=\sum_{j,i}\lambda_ix_j^*x_i\langle j|i\rangle \\
&=\sum_i\lambda_i|x_i|^2.
\end{align*}
Since $C$ is Hermitian, its eigenvalues are real. Hence, $\langle x|C|x\rangle$ is real. By the same token, $\langle x|B|x\rangle$ is real. In order to be able even to write
$$\langle x|A|x\rangle=\langle x|(B+iC)|x\rangle=\langle x|B|x\rangle+i\langle x|C|x\rangle\ge 0,$$
the portion $\langle x|C|x\rangle$ must be either pure imaginary or zero. It is not pure imaginary. Hence, it must be zero. Therefore, $A$ is Hermitian.
 
Last edited by a moderator:
Physics news on Phys.org
You don't need to introduce a basis. Since ##\hat{C}=\hat{C}^{\dagger}## you have
$$\langle x|\hat{C} x \rangle=\langle \hat{C}^{\dagger} x| x\rangle = \langle \hat{C} x|x \rangle=\langle x|\hat{C} x \rangle^* \; \Rightarrow \; \langle x|\hat{C} x \rangle \in \mathbb{R}.$$
Since by assumption ##\hat{A}## is positive semidefinite, from that you must have ##\langle x|\hat{C} x \rangle=0## for all ##|x \rangle \in \mathcal{H}##, and thus ##\hat{C}=0##.
 
  • Like
Likes PeroK and Ackbach
Time reversal invariant Hamiltonians must satisfy ##[H,\Theta]=0## where ##\Theta## is time reversal operator. However, in some texts (for example see Many-body Quantum Theory in Condensed Matter Physics an introduction, HENRIK BRUUS and KARSTEN FLENSBERG, Corrected version: 14 January 2016, section 7.1.4) the time reversal invariant condition is introduced as ##H=H^*##. How these two conditions are identical?

Similar threads

  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 21 ·
Replies
21
Views
3K
  • · Replies 1 ·
Replies
1
Views
966
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
974
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 4 ·
Replies
4
Views
1K