4
$\begingroup$

Assuming we have two sets of $n$ qubits. The first set of $n$ qubits is in state $|a\rangle$ and second set in $|b\rangle$. Is there a fixed procedure that generates a superposed state of the two $|a\rangle + |b\rangle$ ?

$\endgroup$
7
$\begingroup$

Depending on what precisely your assumptions are about a, b, I think this is essentially impossible, and is something called the "no superposing theorem". Please see this paper.

$\endgroup$
4
$\begingroup$

Your question is not quite correctly defined.

First of all, $|a\rangle + |b\rangle$ is not a state. You need to normalize it by considering $\frac{1}{|||a\rangle + |b\rangle||}(|a\rangle + |b\rangle)$.
Secondly, in fact, you don't have access to the states $|a\rangle$ and $|b\rangle$ but to the states up to some global phase, i.e. you can think that the first register is in the vector-state $e^{i\phi}|a\rangle$ and the second register is in the vector-state $e^{i\psi}|b\rangle$ with inaccessible $\phi, \psi$. Since you don't have access to $\phi, \psi$, you can't define sum $|a\rangle + |b\rangle$. But you can ask to construct normalized state $|a\rangle + e^{ it}|b\rangle$ for some $t$. This question is correct. Though, as DaftWullie pointed out in his answer, it is impossible to construct such a state.

$\endgroup$
  • $\begingroup$ Can't you just set $t = 0 \mod 2\pi$? Or is $t$ something out of your control? $\endgroup$ – wizzwizz4 Apr 17 at 15:01
  • 2
    $\begingroup$ @wizzwizz4 Yes, $t$ is unknown. Again, you can't define sum $e^{i\phi}|a\rangle + e^{i\psi}|b\rangle$ to be equal to some exact state (that depends only on $|a\rangle$ and $|b\rangle$). But you can define this sum to be of some type of state. This type of state can be defined as $|a\rangle + e^{ it}|b\rangle$ since $e^{i\phi}|a\rangle + e^{i\psi}|b\rangle = e^{i\phi} (|a\rangle + e^{ i(\psi - \phi)}|b\rangle) \propto |a\rangle + e^{ i(\psi - \phi)}|b\rangle$. $\endgroup$ – Danylo Y Apr 17 at 15:30

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.