2
$\begingroup$

I have a state $$ |\tilde{\Phi_2}\rangle =\dfrac{1}{2^{3l/2}}\sum_{x=0}^{2^l-1}\sum_{y=0}^{2^l-1}\sum_{z=0}^{2^l-1}|0\rangle^{\otimes q}\otimes |x\rangle^{\otimes l}\otimes |y\rangle^{\otimes l}\otimes |z\rangle^{\otimes l} $$ Now I want to change the state $|z\rangle$ using the states $|x\rangle$ and $|y\rangle$. The controls originally $|x\rangle $ and $|y\rangle$ are originally in states $|x_0x_1...x_{l-1}\rangle$ and $|y_0x_1...y_{l-1}\rangle$, then if $$ |x_0\rangle=|y_0\rangle=|1\rangle$$ then the state $|z\rangle$ has to be changed. And the state $|z\rangle$ is changed to $|00....z_{l-3}\rangle$.

Since the $|0\rangle^{\otimes q}$ is unaltered we just operate an identity operator $I^{\otimes q}$.

The next thing I understand is that we have to break the summation into two parts where in the first part the summation indexes of $i,j$ go from $0$ to $\lfloor \dfrac{2^l-1}{2}\rfloor$.

But after that I am unable to write deduce further. Can somebody help me in writing the Dirac notation for this operator?

Edit: The transformation on the $z$ register is the right shift by $2$ operator $|z,0\rangle \to |z, \frac{z}{4}\rangle$

$\endgroup$

1 Answer 1

6
$\begingroup$

As you say, any register on which you do nothing, use the identity, $I$. This is also going to be the case on $x_1,x_2,\ldots$ and $y_1,y_2,\ldots$.

When you want to control something, use the projectors $P_0=|0\rangle\langle 0|$ and $P_1=|1\rangle\langle 1|$. So, controlled-$U$ looks like $P_0\otimes I+P_1\otimes U$. For controlled-controlled-$U$, you can make this notation a bit more concise: $$ I\otimes I\otimes I+P_1\otimes P_1\otimes(U-I). $$ This is essentially what you're going to be doing for your operation, where the two controls are $x_0$ and $y_0$.

So, the next question is what the $U$ looks like for transforming the register $z$. If I understand correctly, you basically want to introduce an additional register $|0\rangle^{\otimes(l-2)}$ and to perform the joint transformation $$ |z_0z_1z_2\ldots z_{l-1}\rangle|0\rangle^{\otimes(l-2)}\rightarrow|z_0z_1z_2\ldots z_{l-1}\rangle|z_0z_1z_2\ldots z_{l-3}\rangle $$ This is achieved in a straightforward manner. Let $C_i$ denote the controlled-not gate controlled off qubit $i$ of the first register, and targetting qubit $i$ on the second register (and identity on everything else). Then, $$ U=\prod_{i=0}^{l-3}C_i. $$ I remain unclear if that target register is supposed to be the one that starts $|0\rangle^{\otimes q}$, or not. I'm assuming not.

Overall, you would thus be left with $$ \prod_{i=0}^{l-3}I^{\otimes q}\otimes(I^{\otimes (4l-2)}+P_1\otimes I^{\otimes(l-1)}\otimes P_1\otimes I^{\otimes(l+i-1)}\otimes P_1\otimes I^{\otimes (l-1)}\otimes(X-I)\otimes I^{\otimes(l-3-i)}). $$ Note, the first two $P_1$s are the controls off $x_0$ and $y_0$. The third is the control off $z_i$, and the $(X-I)$ term should be acting on the target, $i$ of the extra register. Now you can express each of these operators in Dirac notation if you really want to.

$\endgroup$
5
  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – auden
    May 23, 2019 at 0:14
  • $\begingroup$ Yes sir now i think it is what i was looking for. Let me explain to you what i understood $\endgroup$
    – Upstart
    May 23, 2019 at 16:43
  • $\begingroup$ The operator can be $$ \prod_{i=0}^{l-3}I^{\otimes q}\otimes(I^{\otimes (4l-2)}+P_1\otimes I^{\otimes(l-1)}\otimes P_1\otimes I^{\otimes(l+i-1)}\otimes P_1\otimes I^{\otimes (l-1)} \otimes I^{\otimes 2}\otimes(X-I)\otimes I^{\otimes(l-3-i)})$$If i add the register $|0\rangle^{\otimes l}$ instead of $|0\rangle^{\otimes l-2}$ The change would just be that the control will still be $i$-th bit of $z$ but the target would be $i+3$-th qubit of the register $|0\rangle^{\otimes l}$ $\endgroup$
    – Upstart
    May 23, 2019 at 22:31
  • 1
    $\begingroup$ Yes, that would do. $\endgroup$
    – DaftWullie
    May 24, 2019 at 6:53
  • $\begingroup$ Sir i really appreciate the way you explained the answer inspite of my regular petty doubts. $\endgroup$
    – Upstart
    May 24, 2019 at 10:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.