Hybrid Quantum Simulation with 14 Qubits Overcomes Limitations, Achieving Higher Ratios and Reducing Gate Counts

Quantum optimisation seeks to find the best solution from a vast number of possibilities, and researchers continually refine methods to accelerate this process, particularly for complex problems. Fei Li and Xiao-Wei Li, alongside their colleagues, now demonstrate a significant advance in this field with a new technique called Hybrid adaptive variational quantum dynamics simulation. This method overcomes a fundamental limitation of existing approaches, such as counterdiabatic driving, which struggle with complex optimisation landscapes and require increasingly intensive computational resources. The team’s simulation achieves superior results for challenging models, like the Sherrington-Kirkpatrick model, while dramatically reducing the number of computational steps needed, paving the way for more efficient and high-fidelity quantum optimisation.

Near-Term Quantum Simulation and State Preparation

Scientists are developing techniques to efficiently simulate quantum systems and prepare their ground states on near-term quantum computers. These methods address limitations of current hardware, including limited qubit counts and noise, to enable meaningful quantum computations. The core strategy involves reducing the complexity of quantum circuits while maintaining accuracy, allowing for more effective simulations and state preparation. One key approach centers on an adaptive product formula, which dynamically adjusts how the system’s energy is broken down during simulation. This intelligent adjustment minimizes the number of quantum gates needed, improving both efficiency and accuracy.

Another technique, Variational Quantum Imaginary-Time Evolution, efficiently prepares the ground state of a quantum system by combining imaginary time evolution with a variational approach, effectively targeting the lowest energy state. These methods leverage the power of the variational principle, allowing researchers to trade circuit depth for circuit width, optimizing performance on noisy quantum computers. Adaptive methods are crucial for dealing with complex quantum systems, as they dynamically adjust the simulation or state preparation process based on the system’s behavior. A primary goal is to minimize the number of quantum gates required, mitigating the effects of noise and decoherence. These techniques have potential applications in diverse fields, including materials science, drug discovery, quantum chemistry, and fundamental physics, offering a pathway to design new materials, identify drug candidates, and test fundamental theories.

Adaptive Hybrid Quantum Dynamics for Optimization

Scientists have developed a Hybrid adaptive variational quantum dynamics simulation (HAVQDS) to improve the performance of adiabatic quantum computing, particularly when tackling complex optimization problems. Traditional approaches face a trade-off where increasing evolution time weakens the driving force, wasting quantum resources. HAVQDS overcomes this limitation by combining adaptive real-time evolution with imaginary-time steps, effectively compressing circuits and suppressing unwanted excitations without increasing the number of quantum gates. The team implemented HAVQDS by building upon an existing adaptive variational quantum dynamics simulation, efficiently evolving the system in real time.

Interspersed with these real-time steps, they introduced imaginary-time evolution, optimizing the parameters of a quantum circuit to exponentially suppress excited-state components. This innovative approach bypasses the need for an explicit driving term, decoupling non-adiabatic suppression from the real-time driving Hamiltonian and reducing computational overhead. Evaluations using the Sherrington-Kirkpatrick model demonstrate that HAVQDS achieves higher approximation ratios compared to conventional adiabatic and counterdiabatic strategies. Importantly, the method requires significantly fewer CNOT gates, establishing HAVQDS as a high-fidelity and resource-efficient solution for quantum optimization on near-term devices.

Hybrid Algorithm Accelerates Quantum Dynamics Simulations

Scientists have developed a novel quantum dynamics simulation technique, HAVQDS, which significantly improves the performance of quantum optimization algorithms. This work addresses a fundamental limitation of counterdiabatic driving, a method designed to accelerate adiabatic quantum computing, which becomes ineffective in complex systems due to a trade-off between evolution time and driving strength. The team demonstrated that HAVQDS overcomes this limitation by combining adaptive real-time evolution with imaginary-time steps, effectively suppressing unwanted excitations without increasing the number of quantum gates. Experiments using the Sherrington-Kirkpatrick model reveal that HAVQDS achieves higher approximation ratios than both adiabatic and conventional counterdiabatic approaches, consistently delivering improved results in finding optimal solutions to this notoriously challenging problem.

Crucially, HAVQDS accomplishes this while reducing the number of CNOT gates required by one to two orders of magnitude, representing a substantial reduction in computational cost and resource requirements. The team identified a key dilemma in counterdiabatic acceleration, where increasing evolution time weakens the driving term. HAVQDS bypasses this trade-off by replacing the explicit driving term with a hybrid-time evolution strategy, decoupling the suppression of non-adiabatic transitions from the real-time driving Hamiltonian. This innovative approach allows for efficient and high-fidelity quantum optimization on near-term quantum devices.

Adaptive Quantum Dynamics Simulation Excels

Scientists have developed a novel hybrid approach to quantum optimization, HAVQDS, which overcomes limitations inherent in existing counterdiabatic driving techniques. Conventional methods encounter difficulties with complex problems because the required driving strength diminishes as the process evolves, hindering convergence. HAVQDS addresses this by combining adaptive real-time evolution, which compresses the computational circuit, with imaginary-time steps that effectively suppress unwanted excitations without adding to the computational cost. Through simulations on the Sherrington-Kirkpatrick model, HAVQDS demonstrably achieves higher approximation ratios than both standard adiabatic and counterdiabatic approaches. Importantly, this improved performance is accompanied by a significant reduction in required quantum resources, specifically a decrease of one to two orders of magnitude in the number of CNOT gates. This combination of enhanced accuracy and reduced computational burden positions HAVQDS as a promising algorithm for high-performance quantum optimization, particularly suited for implementation on near-term, noisy intermediate-scale quantum (NISQ) devices.