2
$\begingroup$

Suppose $$\mathcal{H}=\bigotimes_{i\in I} \mathcal{H}_i$$ is a tensor product of Hilbert spaces, where $I$ is some index set. Given a $J\subset I$, let $$\mathcal{H}_J=\bigotimes_{i\in J} \mathcal{H}_i,$$ and let $\bar{J}=I\setminus J$. Given a state $\rho$ on $\mathcal{H}$ (i.e. a linear operator mapping $\mathcal{H}$ to itself which satisfies $\operatorname{tr}(\rho)=1$), we define its reduced state in $J$ as $$\rho_J = \operatorname{tr}_{\bar{J}}(\rho);$$ in other words we just take the partial trace over $\mathcal{H}_{\bar{J}}$.

Given a collection of subsets $J_1,J_2,\dots\subset I$, and states $\sigma_{J_1},\sigma_{J_2},\dots$ on $\mathcal{H}_1,\mathcal{H}_2,\dots$, when does there exist a state $\rho$ on $\mathcal{H}$ whose reduced states on each of $J_1,J_2,\dots$ are given by $\sigma_{J_1},\sigma_{J_2},\dots$?

Clearly this is possible when the subsets $J_1,J_2,\dots$ are pairwise disjoint. Clearly also, this is only possible if the reduced states on the intersections of the subsets agree. However, I know that the answer is further constrained by various inequalities such as monogamy of entanglement.

I am also interested in the more general analogous question for states on Von Neumann algebras and the induced states on their subalgebras.

$\endgroup$
2
  • 1
    $\begingroup$ This is called the "quantum marginal problem", and there is a wealth of papers on this (very hard) problem. I doubt a single answer can really answer this question without narrowing it down or further specifying what type of criterion you would like. If dimensions are small, it can be solved via semidefinite programming. $\endgroup$ Commented Jun 7, 2021 at 0:38
  • $\begingroup$ Thank you, I had a feeling this is a hard problem, but didn’t know what it’s called. It is very helpful to be able to have a name to google now. Do you know of any useful books or reviews that might bring me up to date on what is known? For context, I am interested in applying this question to QFTs, with the subsystems corresponding to regions of spacetime. $\endgroup$ Commented Jun 7, 2021 at 0:59

0

You must log in to answer this question.