University College London
Browse

Data repository for paper: "Variational approach to open quantum systems with long-range competing interactions"

Download (36.2 MB)
dataset
posted on 2025-11-24, 12:28 authored by Dawid HryniukDawid Hryniuk, Marzena SzymanskaMarzena Szymanska
<p dir="ltr">Simulation data for paper: "Variational approach to open quantum systems with long-range competing interactions". Full article accessible under <a href="https://arxiv.org/abs/2510.01543" rel="noreferrer" target="_blank">https://arxiv.org/abs/2510.01543<br></a></p><h4><b>Overview: </b>This datasets contains simulation data presented in figures 1-6 in the Main Text and figures 2-61 in the Supplementary Material. This data was obtained from t-VMC+MPO (the method introduced in the above article) and t-MPS simulations. t-VMC+MPO is a method for efficient and scalable simulation of the dynamics of open quantum lattices. It relies on solving the variational equations of motion efficiently by means of time-dependent variational Monte Carlo (t-VMC), while employing compact matrix product operator (MPO) trial states for the many-body density matrix. Among its other advantages, this method is suitable for simulating large driven-dissipative lattices with multiple long-ranged interactions.</h4><p dir="ltr"><b>Implementation: </b>A Julia implementation of t-VMC+MPO (termed t-VMPOMC), which was used to generate this dataset, is available at the public GitHub repository <a href="https://github.com/dhryniuk/t-VMPOMC" rel="noreferrer" target="_blank">https://github.com/dhryniuk/t-VMPOMC</a>. This repository contains example simulation scripts aimed to make the method accessible for first-time users. </p><p dir="ltr"><b>Directory and file organisation:</b> For the Main text, the data is divided first by figure/subplot index, and second by simulation type (please refer to the relevant figure in the article). For the Supplementary Material, the data is divided first by the optimisation hyperparameter label, for which convergence was analysed, and second by model. The folder names specify the model via initialism: I - Ising, C - competing, XYZ - the XYZ model, e.g. the folder name "CI_N200_strong" will contain simulation data for an spin chain of N=200 sites with strong long-ranged competing Ising interactions. Relevant parameter and hyperparameter values distinguishing individual simulations are included in the folder titles. Remaining parameter values not specified can be found in tables 1 and 2 in Supplementary Note 3 on pages 57-59. </p><p dir="ltr">Explanation on t-VMPOMC simulations data file names and structure: "obs.out" - m_x, m_y, m_z, S_2 expectation values per iteration; "corrD.out" - D_local C_xx, C_yy, C_zz correlation functions per iteration; "times.out" - recorded simulation time t per iteration, "iter_times.out" - wall time per iteration; "C.out" - cost function value per iteration; "lambda.out" - smallest eigenvalue of reduced density matrix per iteration; "Schmidt_values.out" - Schmidt singular values per iteration.<a href="https://arxiv.org/abs/2510.01543" rel="noreferrer" target="_blank"><br></a></p>

Funding

DTP 2018-19 University College London

Engineering and Physical Sciences Research Council

Find out more...

EPSRC Centre for Doctoral Training in Delivering Quantum Technologies

Engineering and Physical Sciences Research Council

Find out more...

History