2025-11-22T01:07:16.112219

Trotterization is substantially efficient for low-energy states

Mizuta, Kuwahara
Trotterization is one of the central approaches for simulating quantum many-body dynamics on quantum computers or tensor networks. In addition to its simple implementation, recent studies have revealed that its error and cost can be reduced if the initial state is closed in the low-energy subspace. However, the improvement by the low-energy property rapidly vanishes as the Trotter order grows in the previous studies, and thus, it is mysterious whether there exists genuine advantage of low-energy initial states. In this Letter, we resolve this problem by proving the optimal error bound and cost of Trotterization for low-energy initial states. For generic local Hamiltonians composed of positive-semidefinite terms, we show that the Trotter error is at most linear in the initial state energy $Δ$ and polylogarithmic in the system size $N$. As a result, the computational cost becomes substantially small for low-energy states with $Δ\in o(Ng)$ compared to the one for arbitrary initial states, where $g$ denotes the energy per site and $Ng$ means the whole-system energy. Our error bound and cost of Trotterization achieve the theoretically-best scaling in the initial state energy $Δ$. In addition, they can be partially extended to weakly-correlated initial states having low-energy expectation values, which are not necessarily closed in the low-energy subspace. Our results will pave the way for fast and accurate simulation of low-energy states, which are one central targets in condensed matter physics and quantum chemistry.
academic

Trotterization is substantially efficient for low-energy states

Basic Information

  • Paper ID: 2504.20746
  • Title: Trotterization is substantially efficient for low-energy states
  • Authors: Kaoru Mizuta, Tomotaka Kuwahara
  • Classification: quant-ph cond-mat.other math-ph math.MP
  • Publication Date: April 30, 2025
  • Paper Link: https://arxiv.org/abs/2504.20746

Abstract

Trotterization is one of the core methods for simulating quantum many-body dynamics on quantum computers or tensor networks. Beyond its straightforward implementation, recent studies have shown that its error and cost can be reduced if the initial state lies in the low-energy subspace. However, in previous research, the improvements from low-energy properties diminish rapidly with increasing Trotter order, leaving it an open question whether low-energy initial states provide genuine advantages. In this paper, the authors address this question by proving optimal error bounds and costs for Trotterization of low-energy initial states. For general local Hamiltonians composed of positive semidefinite terms, the authors prove that the Trotter error scales at most linearly with the initial state energy Δ and polylogarithmically with system size N. Consequently, for low-energy states with Δ ∈ o(Ng), the computational cost becomes substantially smaller compared to arbitrary initial states, where g denotes the energy per site and Ng represents the total system energy.

Research Background and Motivation

Problem Background

  1. Importance of Hamiltonian Simulation: Time evolution simulation of quantum many-body systems is a core area of quantum computing with applications ranging from condensed matter physics to quantum chemistry.
  2. Trotterization Method: As one of the most renowned quantum algorithms, Trotterization is feasible on near-term quantum computers due to its simple circuit structure and exhibits better gate complexity on large-scale quantum computers.

Core Problem

The key open question in existing research is: whether Trotterization can provide provable fundamental advantages when dynamics begin from certain interesting classes of initial states?

Limitations of Existing Approaches

  1. Overestimation: In general cases, gate complexity is determined by worst-case error across all possible initial states, which typically overestimates error and cost.
  2. Limited Low-Energy Advantage: Previous studies showed that advantages for low-energy states exist only within very restricted energy ranges Δ ∈ o(N^(1/(p+1))g) and shrink with increasing Trotter order p.
  3. Theoretical Gap: Whether provable advantages exist for general low-energy initial states Δ ∈ o(Ng) remains an open question.

Core Contributions

  1. Optimal Error Bounds: First proof of optimal Trotter error bounds for low-energy initial states, with error scaling linearly in initial state energy Δ and polylogarithmically in system size N.
  2. Theoretical Optimality: Error bounds and costs achieve theoretically optimal scaling in initial state energy Δ, recovering optimal results for arbitrary initial states when Δ = Ng.
  3. Broad Low-Energy Advantage: Demonstrates cost reduction in the broadest low-energy regime Δ ∈ o(Ng), independent of Trotter order p.
  4. Exponential Speedup: For extremely low-energy initial states Δ ∈ polylog(N)g, achieves polylog(N) gate complexity at fixed time t and error ε.
  5. Extension to Weakly Correlated States: Partially extends results to weakly correlated initial states with low energy expectation values.

Methodology Details

Task Definition

Consider time evolution simulation of an N-qubit Hamiltonian H on lattice Λ = {1,2,···,N}:

  • Input: k-local Hamiltonian H, low-energy initial state |ψ⟩, evolution time t, allowed error ε
  • Output: Quantum algorithm implementing e^(-iHt)|ψ⟩ within error ε
  • Objective: Minimize the number of quantum gates (Trotter number r)

Hamiltonian Setup

The Hamiltonian H is written as: H=∑X⊂Λ;∣X∣≤khX,hX≥0H = \sum_{X⊂Λ;|X|≤k} h_X, \quad h_X ≥ 0

where each h_X acts nontrivially on domain X, with positive semidefiniteness ensured through shifting h_X → h_X + ∥h_X∥.

Low-Energy Subspace Projection

Define the subspace projection for energies below Δ: Π≤Δ=∑n;En≤Δ∣En⟩⟨En∣\Pi_{≤Δ} = \sum_{n;E_n≤Δ} |E_n⟩⟨E_n|

Consider initial states completely within this subspace: Π_{≤Δ}|ψ⟩ = |ψ⟩.

Core Theoretical Results

Theorem 1 (Main Result): For any ϵ ∈ (0,1), there exists: Δ′=Δ+O(glog⁡(N/ϵ))Δ' = Δ + O(g\log(N/ϵ))

such that the Trotter error bound for low-energy initial states is:

O((gt)^p Δ't + ϵ) & \text{if } Γ ∈ O(1) \\ O(\{gt\log(N/ϵ)\}^p Δ't + ϵ) & \text{otherwise} \end{cases}$$ ### Technical Innovations 1. **Commutator Scaling Refinement**: First accurate capture of commutator scaling properties in low-energy space, avoiding the overestimation of 1-norm scaling O((Δ't)^(p+1)) from previous studies. 2. **Improved Leakage Assessment**: Precise evaluation of leakage from the low-energy subspace, improving leakage Δ' - Δ from scaling proportionally with N to O(g log(N/ϵ)). 3. **Locality of Nested Commutators**: Exploits locality of nested commutators in local Hamiltonians, proving the key inequality: $$∥Π_{≤Δ'}[H_γ_q, ···, [H_γ_1, H_γ_0]]Π_{≤Δ'}∥ ≤ q!(2kg)^q Δ'$$ ## Experimental Setup ### Numerical Verification The authors conducted numerical verification on two frustration-free Hamiltonians: 1. **AKLT Hamiltonian**: $$H_{AKLT} = \sum_{i=1}^{N-1} P[(S_i + S_{i+1})^2 = 2(2+1)]$$ 2. **Majumdar-Ghosh Hamiltonian**: $$H_{MG} = \sum_{i=1}^{N-2} P[(S_i + S_{i+1} + S_{i+2})^2 = \frac{3}{2}(\frac{3}{2}+1)]$$ ### Experimental Parameters - Trotter order: p = 1, 2 - Evolution time: t = 0.1 - Energy bounds: Δ = 0.5, 1.0 - System size: varying N values ## Experimental Results ### Main Findings **System Size Dependence**: - Without energy constraints, Trotter error scales approximately linearly with system size (consistent with commutator scaling) - Error growth for low-energy initial states is significantly slower, consistent with the logarithmic dependence predicted by Theorem 1 **Energy Bound Dependence**: - In the regime Δ ≪ ∥H∥, error exhibits approximately linear relationship with energy bound Δ - The slope is nearly independent of system size N, supporting theoretical predictions ### Improvement in Trotter Number For the case Γ ∈ O(1), the Trotter number is: $$r \sim gt\left(\frac{Δt + gt\log(N/ε)}{ε}\right)^{1/p}$$ Compared to arbitrary initial states with r ∈ O(gt(Ngt/ε)^(1/p)), this is substantially reduced for Δ ∈ o(Ng). ## Related Work ### Hamiltonian Simulation Algorithms - **Trotterization**: The most classical approach with simple circuit structure - **Linear Combination of Unitaries (LCU)**: More efficient in certain cases - **Quantum Singular Value Transform (QSVT)**: Latest unified framework ### Initial State-Dependent Improvements - **Haar Random States**: Known to exhibit better scaling - **Highly Entangled States**: Recently shown to accelerate Hamiltonian simulation - **Low-Energy States**: First optimal theoretical bounds provided in this work ## Conclusions and Discussion ### Main Conclusions 1. First proof of optimal error bounds and costs for Trotterization of low-energy initial states 2. Cost reduction achieved in the broadest low-energy regime Δ ∈ o(Ng) 3. Exponential speedup achievable for extremely low-energy states ### Limitations 1. **Positive Semidefiniteness Requirement**: Requires shifting to ensure all local terms are positive semidefinite 2. **Complete Low-Energy Assumption**: Main results require initial states to lie completely in the low-energy subspace 3. **Limited Numerical Verification**: Verification only on specific frustration-free models ### Future Directions 1. **Extension to Other Trotter Variants**: Such as randomized compiled Trotterization (qDRIFT) and multi-product formulas 2. **Time-Dependent Trotterization**: Applications to adiabatic state preparation 3. **Ground State Energy-Based Low-Energy States**: Seeking advantages based on ground state energy rather than shifted energy ## In-Depth Evaluation ### Strengths 1. **Theoretical Breakthrough**: First optimal theoretical bounds for Trotterization of low-energy states, resolving an important open problem in the field 2. **Technical Innovation**: Accurate capture of commutator scaling in low-energy subspace represents significant technical advancement 3. **Practical Value**: Results apply to core objectives in condensed matter physics and quantum chemistry—low-energy state simulation 4. **Rigor**: Mathematical proofs are rigorous with optimality guarantees ### Weaknesses 1. **Scope of Applicability**: Primarily applicable to approximately frustration-free systems; limited applicability to strongly correlated systems 2. **Experimental Verification**: Numerical verification mainly on small systems and special models; lacks verification on large-scale practical systems 3. **Implementation Complexity**: While theoretically optimal, efficient verification of low-energy conditions in practical implementation requires further exploration ### Impact 1. **Theoretical Significance**: Establishes theoretical framework for initial state-dependent Hamiltonian simulation 2. **Practical Value**: Provides theoretical guidance for quantum simulation in quantum chemistry and condensed matter physics 3. **Methodological Contribution**: Developed commutator scaling techniques extensible to other initial state families ### Applicable Scenarios 1. **Quantum Chemistry**: Dynamics simulation of molecular ground and low-excited states 2. **Condensed Matter Physics**: Quantum simulation of low-temperature phenomena and phase transitions 3. **Quantum Algorithms**: Quantum algorithm design requiring high-precision low-energy state evolution ## References This paper cites 37 important references covering cutting-edge work in Trotterization theory, quantum simulation algorithms, and commutator scaling theory, providing solid theoretical foundations for the research. --- **Summary**: This paper makes important theoretical contributions to the field of quantum Hamiltonian simulation, providing for the first time optimal theoretical bounds for Trotterization of low-energy initial states. While further verification is needed for practical applications, its theoretical significance and guidance value for future research are substantial.