Is the mathematical expression between the CNOT gates arranged in this way a tensor product or a concatenated product? That is \begin{matrix} \text{CNO}{{\text{T}}^{\otimes n}} & or & \prod\limits_{n}{\text{CNOT}} \\ \end{matrix}?
1 Answer
$CNOT^{\otimes n}$ is a matrix of size $4^n$ and corresponds to the following circuit :
$\prod_{i=1}^{n}{CNOT}$ is a matrix of size $2$ and corresponds to the following circuit :
The circuit you describe is more something like :
$\prod_{i=1}^{n}{I_{2^{i-1}}\otimes CNOT \otimes I_{2^{n-i}}}$
where $I_n$ is the identity matrix of size $n$
edit : explaination
You apply successively $CNOT$ gates on different qubits. In the circuit you present, when you apply a $CNOT$ gate on 2 qubits, you do not perform any transformation on other qubits, this is equivalent to perform an identity transformation (which does nothing).
The first $CNOT$ gate is applied on the two first qubits and an identity transformation is applied on all other qubits (n-1 qubits). The full transformation is then :
$CNOT \otimes I^{\otimes n-1} = CNOT \otimes I_{2^{n-1}}$
The second $CNOT$ gate is applied on the 2nd and 3rd qubits and an identity transformation is applied on the first qubit and all the qubits after the third. The full transformation is then :
$I \otimes CNOT \otimes I^{\otimes n-2} = I \otimes CNOT \otimes I_{2^{n-2}}$
And so on, if you apply a $CNOT$ gate on the qubits $i$ and $i+1$, you will apply an identity on all the qubits before the $i^{th}$ and all the qubits after the $i+1^{th}$, the transformation will be :
$I^{\otimes i-1} \otimes CNOT \otimes I^{\otimes n-i} = I_{2^{i-1}}\otimes CNOT \otimes I_{2^{n-i}}$
The overall circuit is the succession of all this transformations for i going from 1 to n and is therefore described by :
$\prod_{i=1}^{n}{I^{\otimes i-1}\otimes CNOT \otimes I^{\otimes n-i}} = \prod_{i=1}^{n}{I_{2^{i-1}}\otimes CNOT \otimes I_{2^{n-i}}}$
-
$\begingroup$ Will the result be the same if I change CNOT to SWAP? $\endgroup$ Commented Oct 6, 2022 at 10:32
-
$\begingroup$ How do I determine such a law? Could you please provide some necessary mathematical information? $\endgroup$ Commented Oct 6, 2022 at 10:48
-
1$\begingroup$ I edited my post to explain how to find the result. The result won't be the same if you change CNOT to SWAP. The matrices of this two gates are different and the tensor product of this matrices with idenitities will also be different. $\endgroup$ Commented Oct 6, 2022 at 12:30
-
$\begingroup$ Thanks. Another question has been asked about SWAP door and Toffoli door. I hope to get your reply.quantumcomputing.stackexchange.com/questions/28435/… $\endgroup$ Commented Oct 6, 2022 at 12:54