1
$\begingroup$

In section $5.2$ of the QAOA chapter in Qiskit textbook, section $5.2$, state preparation uses the gate $U_{k,l}(\gamma) = e^{\frac{i \gamma}{2} (1-Z_k Z_l)}$. Later, in section $5.3$, this gate is given in terms of basic gates as $U_{k,l}(\gamma)=C_{u1}(−2 \gamma)_{k,l}u_1(\gamma)_k u_1(\gamma)_l$. Why is this the correct implementation?

$\endgroup$
0
$\begingroup$

Qiskit gates $u_1$ and $C_{u1}$ are

$$ u_1(\theta) = \begin{pmatrix} 1 & 0 \\ 0 & e^{i\theta} \end{pmatrix} \,\,\, C_{u1}(\theta) = \begin{pmatrix} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & e^{i\theta} \end{pmatrix} $$

see here and here. Exponential of a normal matrix $A$ is the matrix $e^A$ with the same eigenbasis as $A$ and eigenvalues that are exponentials of the eigenvalues of $A$. Therefore, $u_1(\theta) = e^{\frac{i\theta}{2}(I-Z)}$. Also, note that

$$ I - Z_1 - Z_2 + Z_1Z_2 = \begin{pmatrix} 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 4 \end{pmatrix}, $$

so $C_{u1}(\theta) = e^{\frac{i\theta}{4}(I - Z_1 - Z_2 + Z_1Z_2)}$. Finally, diagonal matrices commute, so

$$ U_{k,l}(\gamma) = e^{\frac{i \gamma}{2}(I-Z_k Z_l)} = e^{-\frac{i\gamma}{2}(I-Z_k-Z_l+Z_k Z_l)}e^{\frac{i\gamma}{2}(I-Z_k)}e^{\frac{i\gamma}{2}(I-Z_l)} = C_{u1}(-2\gamma)_{k,l} u_1(\gamma)_k u_1(\gamma)_l. $$

$\endgroup$
1
  • $\begingroup$ Hi Adam, thanks for a very clear explanation. $\endgroup$ – creet Feb 21 at 17:09

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.