Back to News
quantum-computing

Quantum Computing: Classical Physics Boosts Stability

Quantum Zeitgeist
Loading...
6 min read
0 likes
⚡ Quantum Brief
Heriot-Watt researchers developed a quantum-inspired numerical method approximating the Koopman-von Neumann equation, enabling unitary classical dynamics simulation—even for non-Hamiltonian systems where energy isn’t conserved. The approach links quantum mechanics with classical transfer operators, allowing data-driven approximation of system properties from observations, bypassing theoretical limitations of traditional Koopman and Perron-Frobenius operators. Demonstrated on oscillators and the Lotka-Volterra model, the method’s unitarity preserves norms and probabilities, preventing error accumulation in long-term predictions—critical for complex, dissipative systems. Unitary matrix representation enables potential quantum circuit encoding, mirroring speedups seen in factorization tasks, though scaling to high-dimensional systems remains a challenge due to qubit and gate requirements. Basis function selection and domain definition are key for accuracy, with higher-order functions improving precision but increasing computational cost, while data-driven techniques like dynamic mode decomposition refine operator approximations.
Quantum Computing: Classical Physics Boosts Stability

Summarize this article with:

Stefan Klus and colleagues at Heriot–Watt University present a numerical method for approximating the Koopman-von Neumann equation, a framework offering advantages over conventional transfer operators by remaining unitary even with non-Hamiltonian dynamics. The method links this quantum-inspired formulation with classical transfer operators. This enables the derivation of numerical techniques to approximate the operator and its properties from observed data. The research highlights the key role of selecting appropriate basis functions and domains for accurate operator definition and demonstrates the method’s efficacy using examples such as oscillators and the Lotka-Volterra model. Unitary representation enables quantum simulation of classical dynamics The Koopman-von Neumann operator is now successfully represented as a unitary matrix, potentially unlocking a pathway to quantum simulation of classical dynamical systems. Classical numerical simulations previously limited the complexity of systems that could be modelled. However, this new approach circumvents those limitations by using the principles of quantum mechanics. Remaining unitary even when modelling non-Hamiltonian dynamics, systems where energy is not conserved, the Koopman-von Neumann equation, a quantum-inspired formulation of classical mechanics, offers a sharp advantage over traditional transfer operators such as the Koopman and Perron-Frobenius operators. These conventional operators often suffer from issues related to spectral properties and can become ill-defined for dissipative systems, whereas unitarity guarantees the preservation of norms and probabilities during the evolution of the system’s wavefunction. Potential encoding as a quantum circuit is now possible with this unitary representation, mirroring established quantum speedups for calculations such as integer factorization, discrete logarithms, and related order-finding tasks. Unitary transformations can be efficiently implemented on a quantum computer, potentially offering exponential speedups for certain computations. The method’s efficacy was demonstrated using undamped and damped oscillators, as well as the more complex Lotka-Volterra model, a common tool for studying predator-prey dynamics. The undamped harmonic oscillator serves as a foundational test case, verifying the method’s ability to accurately represent periodic motion. The damped oscillator introduces dissipation, highlighting the Koopman-von Neumann operator’s advantage in handling non-Hamiltonian systems. The Lotka-Volterra model, with its non-linear interactions, provides a more challenging benchmark for assessing the method’s performance in a biologically relevant context. Careful selection of both basis functions and the domain of operation is required for accurate representation, as improper choices can lead to an ill-defined operator. The choice of basis functions directly impacts the accuracy of the approximation, with higher-order functions generally providing better accuracy but also increasing computational cost. The domain of operation must be chosen to adequately capture the relevant dynamics of the system. Further work will focus on optimising these choices for different system types. Building on extended dynamic mode decomposition, a machine learning technique, the research approximates the operator and its key properties from observed data, offering a pathway beyond purely theoretical calculations. Extended dynamic mode decomposition provides a data-driven approach to identifying dominant modes of behaviour in the system, which can then be used to construct an appropriate basis for the Koopman-von Neumann operator. Analysis of eigenvalues and eigenfunctions is now possible, revealing information about system timescales and behaviours. The eigenvalues of the Koopman-von Neumann operator correspond to the growth or decay rates of the system’s modes, while the eigenfunctions represent the spatial structures of these modes. While a practical quantum speedup over existing classical methods has not yet been demonstrated, and significant hurdles remain in scaling this technique to genuinely complex, high-dimensional scenarios, the potential for future acceleration is considerable. The primary challenge lies in the efficient implementation of the unitary operator on a quantum computer, which requires a significant number of qubits and gate operations for complex systems. Utilising quantum mechanics to accelerate modelling of complex dynamical systems Increasingly, scientists are focused on modelling complex systems, from weather patterns to financial markets, with greater accuracy and speed. This pursuit has led to exploration of quantum-inspired techniques, borrowing concepts from quantum mechanics to enhance classical computational methods. Despite doubts about immediately useful quantum advantage given limitations in building stable quantum computers, scientists are nonetheless pursuing these methods because they offer new ways to model complicated systems more efficiently than existing techniques. The motivation stems from the inherent limitations of classical computers in simulating high-dimensional, non-linear systems, and the potential of quantum mechanics to overcome these limitations. Refinement of modelling in areas like climate science and financial analysis could be achieved, even with classical hardware. A method for approximating the Koopman-von Neumann operator, a tool for modelling how systems evolve, has been established by representing it as a unitary matrix, preserving information during calculations unlike some traditional methods. Information preservation is crucial for long-term predictions, as it prevents the accumulation of errors that can occur in non-unitary systems. Linking this quantum-inspired operator to classical transfer operators allows scientists to derive numerical techniques to estimate its properties directly from observed data, bypassing purely theoretical approaches. This data-driven approach is particularly valuable for systems where the underlying equations of motion are unknown or too complex to solve analytically. Successful projection of the operator onto a smaller space is key, enabling its representation as a quantum circuit and opening avenues for potentially faster simulations of complex dynamics, while also allowing for efficient computation of the operator’s properties and enabling analysis of system behaviour and prediction of future states. Dimensionality reduction techniques, such as principal component analysis, can be used to project the operator onto a lower-dimensional subspace while preserving its essential features. The ability to efficiently compute the operator’s properties, such as its eigenvalues and eigenvectors, provides valuable insights into the system’s stability, bifurcations, and long-term behaviour. The researchers successfully established a method for approximating the Koopman-von Neumann operator by representing it as a unitary matrix. This approach offers a way to model complex systems more efficiently, potentially overcoming limitations of classical computers when simulating high-dimensional, non-linear dynamics. By linking this quantum-inspired operator to classical methods, scientists can now estimate its properties directly from data, which is particularly useful when underlying equations are unknown. The authors demonstrated this with examples including oscillators and the Lotka-Volterra model, and suggest careful selection of basis functions is crucial for accurate results. 👉 More information 🗞 Numerical approximation of the Koopman-von Neumann equation: Operator learning and quantum computing 🧠 ArXiv: https://arxiv.org/abs/2604.08414 Tags:

Read Original

Tags

energy-climate
government-funding
quantum-computing
quantum-simulation

Source Information

Source: Quantum Zeitgeist