The paper shows that many static properties of quantum impurity models can be computed on a classical computer in polynomial time. In plain terms, the authors give classical algorithms that estimate the ground-state energy to additive precision δ in time poly(n, δ^{-1}), and that estimate the thermal partition function at inverse temperature β to relative precision δ in time poly(n, β, δ^{-1}). At the same time, they prove that dynamical quantities — for example nonequilibrium Green’s functions — are as hard as general quantum computation, so those remain a natural target for quantum advantage.

Quantum impurity models describe a small, strongly interacting subsystem (the impurity) coupled to a large noninteracting environment (the bath). These models are widely used as building blocks in electronic-structure and condensed-matter methods such as dynamical mean-field theory (DMFT). Historically, impurity problems have been tackled by many numerical tools — numerical renormalization group, matrix-product-state methods, and quantum Monte Carlo — but their provable computational complexity was not settled before.

To get polynomial-time classical algorithms the authors use a structural result they call a compression lemma. After a change of basis (a Bogoliubov transformation), they show that the ground state — and the relevant part of the thermal state — has almost all its weight on a much smaller subspace of the full Hilbert space. That smaller subspace can be found and its basis computed in time proportional to its dimension, and diagonalizing the Hamiltonian there gives the desired estimates. The runtimes stated in the paper are polynomial in the total system size n and the inverse precision parameters. The algorithms do have an exponential dependence on the impurity size, but the work assumes the impurity is constant in size, so that dependence is not a practical obstacle in the setting studied.