4

I had a question last week regarding the simulation of Hamiltonians composed of the sum of Pauli products: How can I simulate Hamiltonians composed of Pauli matrices? I'm having a follow-up question: still for those two Hamiltonians: $$ H_{1} = X_1+ Y_2 + Z_1\otimes Z_2 \\ H_{2} = X_1\otimes Y_2 + Z_1\otimes Z_2 $$ How can I perform the 'controlled version' of them? The thing really confused me is the 'tensor product term': for both $H_1$ and $H_2$, the two qubits are coupled, but if I want to do the controlled time-evolution simulation, should I couple the whole thing with the third qubit? If so, how to do that?

Thanks:)

ZR-
  • 2,388
  • 7
  • 19
  • 1
    The last post you used implicitly used exponentiation - do you want to use that approach or VQEs? (what's the end goal here might be a better question) – C. Kang Nov 12 '20 at 00:23
  • 1
    @C. Kang Thanks for the comment! I still prefer the exponentiation approach:) – ZR- Nov 12 '20 at 00:32

1 Answers1

3

So we know that $e^{i t H_2}$ has the following circuit:

enter image description here

From this answer along with page 13 from this paper we can try to build the controlled-version of $e^{i t H_2}$ as follow:

enter image description here

KAJ226
  • 13,822
  • 2
  • 10
  • 30
  • Thank you so much for the answer! Is there still a global phase gate needed on the control qubit? – ZR- Nov 12 '20 at 14:34
  • 1
    The global phase gate is not needed if we don't have $e^{i I t}$ from my understanding. So unless our Hamiltonian is something like $H = XY + ZZ + II$ then you won't be needing it. – KAJ226 Nov 12 '20 at 18:14
  • Got it. Thanks!! – ZR- Nov 12 '20 at 20:30