Heisenberg evolution of matrix product operators by time-dependent variational principle

Kévin Hémery, Frank Pollmann, David J. Luitz

We review different matrix-product-state (MPS) approaches to study the spreading of operators in generic nonintegrable quantum systems. As a common ground to all methods, we quantify this spreading by means of the Frobenius norm of the commutator of a spreading operator with a local operator, which is usually referred to as the out-of-time-order correlation (OTOC) function. We compare two approaches based on matrix-product states in the Schr\"odinger picture: the time-dependent block decimation (TEBD) and the time-dependent variational principle (TDVP), as well as TEBD based on matrix-product operators directly in the Heisenberg picture. The results of all methods are compared to numerically exact results using Krylov space exact time evolution. We find that for the Schr\"odinger picture, the TDVP algorithm performs better than the TEBD algorithm.Moreover, the tails of the OTOC are accurately obtained both by TDVP MPS and TEBD MPO. They are in very good agreement with exact results at short times, and appear to be converged in bond dimensions even at longer times. However, the growth and saturation regimes are not well captured by either of the methods.

Abraham Klein, Ching‐Teh Li

Kévin Hémery, Frank Pollmann, David J. Luitz

We review different tensor network approaches to study the spreading of operators in generic nonintegrable quantum systems. As a common ground to all methods, we quantify this spreading by means of the Frobenius norm of the commutator of a spreading operator with a local operator, which is usually referred to as the out of time order correlation (OTOC) function. We compare two approaches based on matrix-product states in the Schrodinger picture: the time dependent block decimation (TEBD) and the time dependent variational principle (TDVP), as well as TEBD based on matrix-product operators directly in the Heisenberg picture. The results of all methods are compared to numerically exact results using Krylov space exact time evolution. We find that for the Schrodinger picture the TDVP algorithm performs better than the TEBD algorithm. Moreover the tails of the OTOC are accurately obtained both by TDVP MPS and TEBD MPO. They are in very good agreement with exact results at short times, and appear to be converged in bond dimension even at longer times. However the growth and saturation regimes are not well captured by both methods.

Carlos M. Mora

Christian B. Mendl

We devise a numerical scheme for the time evolution of matrix product operators by adapting the time-dependent variational principle for matrix product states [J. Haegeman et al, Phys. Rev. B 94, 165116 (2016)]. A simple augmentation of the initial operator $\mathcal{O}$ by the Hamiltonian $H$ helps to conserve the average energy $\mathrm{tr}[H \mathcal{O}(t)]$ in the numerical scheme and increases the overall precision. As demonstration, we apply the improved method to a random operator on a small one-dimensional lattice, using the spin-1 Heisenberg XXZ model Hamiltonian; we observe that the augmentation reduces the trace-distance to the numerically exact time-evolved operator by a factor of 10, at the same computational cost.

M. C. Cambiaggio, G.G. Dussel, Marcos Saraceno