Simulating Quantum Circuits with Tree Tensor Networks using Density-Matrix Renormalization Group Algorithm

Aditya Dubey, Zeki Zeybek, Peter Schmelcher · 2025

Read Simulating Quantum Circuits with Tree Tensor Networks using Density-Matrix Renormalization Group Algorithm by Aditya Dubey, Zeki Zeybek, Peter Schmelcher (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-16718.txt.

Introduction

Aditya Dubey, Zeki Zeybek, and Peter Schmelcher simulate quantum circuits on classical hardware in a 2025 paper. Checking real quantum machines against exact answers requires such simulators, and noise limits how far the machines themselves can be trusted.

Tensor networks store quantum states as webs of small arrays. A leading algorithm works with a chain shaped web, which fits states whose links run along a line. The authors extend that algorithm to tree shaped webs, which can mirror the layout of a circuit more closely. On random circuits, the tree form needed less memory. On an optimization algorithm's circuits, building the tree from the circuit's own graph sharply improved accuracy.

Catalog record

CREATORAditya Dubey, Zeki Zeybek, Peter Schmelcher
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-16718

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Simulating Quantum Circuits with Tree Tensor Networks using Density-Matrix Renormalization Group Algorithm Aditya Dubey, Zeki Zeybek, Peter Schmelcher Abstract Quantum computing offers the potential for computational abilities that can go beyond classical machines. However, they are still limited by several challenges such as noise, decoherence, and gate errors. As a result, efficient classical simulation of quantum circuits is vital not only for validating and benchmarking quantum hardware but also for gaining deeper insights into the behavior of quantum algorithms. A promising framework for classical simulation is provided by tensor networks. Recently, the Density-Matrix Renormalization Group (DMRG) algorithm was developed for simulating quantum circuits using matrix product states (MPS). Although MPS is efficient for representing quantum states with one-dimensional correlation structures, the fixed linear geometry restricts the expressive power of the MPS. In this work, we extend the DMRG algorithm for simulating quantum circuits to tree tensor networks (TTNs). The framework employs a variational compression scheme that optimizes the TTN to approximate the evolved quantum state. To benchmark the method, we simulate random circuits and the quantum approximate optimization algorithm (QAOA) with various two-qubit gate connectivities. For the random circuits, we devise tree-like gate layouts that are suitable for TTN and show that TTN requires less memory than MPS for the simulations. For the QAOA circuits, a naive TTN construction that exploits graph structure significantly improves the simulation fidelities. Our findings show that the DMRG algorithm with TTNs provides a promising framework for simulating quantum circuits, particularly when gate connectivities exhibit clustering or a hierarchical structure. I Introduction Quantum computing promises revolutionary computational capabilities, offering the potential to solve problems that are intractable for classical machines [ 1 , 2 , 3 ] . However, despite the allure of exponentially faster algorithms, quantum computers today are still hampered by several challenges [ 4 , 5 ] . Current devices are limited by noise, decoherence, and gate errors, all of which constrain their reliability. As a result, efficient classical simulation of quantum systems is vital for validating and benchmarking quantum hardware and gaining deeper insights into the behavior of quantum algorithms [ 6 , 7 , 8 ] . A promising framework for classical simulation is provided by tensor networks (TNs) [ 9 , 10 , 11 , 12 ] . They offer a way to manage the exponential complexity of quantum states by exploiting the structure inherent in many-body entanglement, rendering them ideal for the representation and manipulation of quantum states. Consequently, TN algorithms have become cornerstones in the numerical study of strongly correlated quantum systems, addressing problems ranging from ground state search [ 13 , 14 ] to real-time dynamics [ 15 , 16 , 17 ] . Tensor networks have also become a versatile tool in quantum information theory and quantum computing [ 18 ] . For example, they provide a framework for loading classical data into a quantum circuit [ 19 , 20 , 21 , 22 , 23 , 24 ] . TN techniques also provide valuable insights into the structure and performance of quantum algorithms [ 8 , 25 , 7 ] . Efficient classical simulation of quantum circuits is equally vital for the validation and benchmarking of quantum hardware. TN methods have proven particularly effective in this context by refuting quantum supremacy claims with classical simulation results that rival or surpass proposed quantum advantages [ 26 , 6 , 27 , 28 , 29 ] . A wide variety of tensor network architectures have been employed for simulating quantum circuits [ 6 , 27 , 30 , 31 , 32 ] . In particular, the DMRG algorithm using matrix product states was generalized to simulate quantum circuits with a finite fidelity [ 27 ] . The algorithm essentially takes the resulting state after the application of the gates and compresses it into an MPS. Although MPS is efficient for compressing quantum states that admit one-dimensional correlation structure, the fixed chain-like geometry restricts the expressive power of the MPS. Therefore, an appropriate kind of tensor network should be considered for the DMRG algorithm to represent states exhibiting multi-scale correlations. Tree tensor networks have emerged as a powerful tool in both many-body physics and quantum chemistry, with applications spanning ground state calculations [ 33 , 34 , 35 , 36 , 37 , 38 , 39 ] , real-time dynamics [ 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 ] , and the study of complex molecular systems [ 48 , 49 , 50 , 51 ] . However, TTNs have not been explored extensively in the context of quantum circuit simulation. Their structural flexibility allows for adapting the ansatz to match the underlying connectivity or correlation patterns of the quantum system. A direct application of gates on a TTN leads to a complex tensor network structure, which would then require transformation back into the TTN shape. However, naive transformation of a tensor network can result in an exponentially large memory requirement, scaling with the number of qubits. Ref. [ 30 ] addresses this challenge by introducing the concept of threading entanglement through the tree. This method is equivalent to approximating the true output quantum state using a truncation method based on singular value decomposition (SVD). The truncation is based on discarding small singular values which can introduce a notable error with each gate. These errors compound with subsequent gate applications, potentially leading to a significant loss of fidelity. In this work, we extend the Density-Matrix Renormalization Group (DMRG) algorithm, originally developed for simulating quantum circuits with MPS [ 27 ] , to TTNs. The central idea of this method is a variational compression scheme that optimizes the TTN to approximate the evolved quantum state. It iteratively optimizes each tensor to find the best possible TTN representation of the evolved quantum state for a given bond dimension. This variational approach is superior to the SVD-based compression method as it minimizes the distance to the target state and avoids the compounding errors associated with successive SVD truncations [ 9 , 27 ] . Furthermore, the framework preserves the shape of the TTN upon the application of gates and enables dynamical change of the TTN structure mid-simulation. The key implication of our work is that the DMRG algorithm with TTNs provides a promising framework for simulating quantum circuits, especially when gate connectivities exhibit some form of clustering or hierarchical structure. Therefore, the flexibility of TTNs allows them to adapt to a wide range of quantum circuit architectures. We focus on random circuits and QAOA circuits with various connectivities, including a special class of circuits referred to as tree-like circuits. For tree-like random circuits, TTNs require less memory than MPS for comparable accuracy due to their ability to better capture long-range correlations. In QAOA simulations for MaxCut on 3-regular and tree-like connected graphs, a naive TTN construction exploiting graph structure significantly improves fidelity compared to a blind approach. We achieve near-perfect fidelity for structured bridged 3-regular and tree-like connected graphs despite having small bond dimensions. The paper is organized as follows. In Sec. II.1 , we introduce the TTN ansatz and outline its key features. Section II.2 presents the DMRG algorithm as applied

Read the complete text