Non-commutative optimization problems with differential constraints

TL;DR

This paper develops a hierarchy of semidefinite programming relaxations for non-commutative polynomial optimization problems with differential constraints, enabling effective bounds on quantum spin system observables during evolution.

Contribution

It introduces a method to convert differential operator constraints into standard NPO problems, facilitating SDP-based solutions for quantum dynamics.

Findings

Hierarchy provides accurate bounds even in the thermodynamic limit.

Low hierarchy levels yield good approximations for long evolution times.

Method applies to quantum spin systems under Hamiltonian evolution.

Abstract

Non-commutative polynomial optimization (NPO) problems seek to minimize the state average of a polynomial of some operator variables, subject to polynomial constraints, over all states and operators, as well as the Hilbert spaces where those might be defined. Many of these problems are known to admit a complete hierarchy of semidefinite programming (SDP) relaxations. In this work, we consider a variant of NPO problems where a subset of the operator variables satisfies a system of ordinary differential equations. We find that, under mild conditions of operator boundedness, for every such problem one can construct a standard NPO problem with the same solution. This allows us to define a complete hierarchy of SDPs to tackle the original differential problem. We apply this method to bound averages of local observables in quantum spin systems subject to a Hamiltonian evolution (i.e., a quench). We find that, even in the thermodynamic limit of infinitely many sites, low levels of the hierarchy provide very good approximations for reasonably long evolution times.