Researchers Improve Molecular Simulation with Hamiltonian-Aware Quantum Algorithm

Constructing compact ansätze, the initial configurations used in Variational Quantum Eigensolver calculations, has been a key limitation in solving molecular energy problems on current quantum computers. Runhong He of the Chinese Academy of Science and colleagues from East China Normal University and MindSpore Quantum Special Interest Group have achieved a sharp advance with the Hamiltonian-Aware ADAPT-VQE algorithm, which overcomes issues of selecting appropriate operators and the build-up of redundant calculations in previous methods. The algorithm is a new computational technique to improve the modelling of molecules using quantum computers, offering a pathway towards more accurate and efficient quantum simulations.

The advance addresses shortcomings in current methods by intelligently choosing the necessary steps for calculations, resulting in more precise outcomes with reduced computational demands. This refined approach constructs more efficient ‘ansätze’ by considering how strongly each step impacts the final energy of the molecule. Current approaches, like the Variational Quantum Eigensolver (VQE), seek the lowest energy state of a molecule using both classical and quantum computers, but rely on ‘ansätze’, essentially recipes or a set of tools, which can become inefficient as molecular complexity increases. The VQE method decomposes the molecular Hamiltonian into a sum of Pauli strings, and the ansatz is used to approximate the ground state wavefunction. However, the size of the ansatz, and therefore the number of quantum gates required, grows rapidly with system size, posing a significant challenge for near-term quantum devices. Runhong He and colleagues’ Hamiltonian-Aware ADAPT-VQE algorithm overcomes issues with selecting appropriate operators and avoiding redundant calculations; it is like assembling a puzzle where you intelligently choose pieces, discarding those that don’t fit, rather than trying every possible combination. This refined approach constructs more efficient ansätze by prioritising steps that strongly impact molecular energy, and the team now aims to demonstrate its scalability for increasingly complex systems. The algorithm builds upon the Adaptive Derivative-Assembled Pseudo-Trotter (ADAPT) method, a technique designed to dynamically construct the ansatz during the optimisation process, but addresses its inherent limitations.

Hamiltonian guidance improves convergence and accuracy in variational quantum eigensolver

A 30% reduction in energy error was observed with the Hamiltonian-Aware ADAPT-VQE algorithm when compared to baseline algorithms on strongly correlated molecular systems, achieving a threshold previously unattainable without encountering energy plateaus. This advance addresses limitations in existing Adaptive Derivative-Assembled Pseudo-Trotter (ADAPT) ansätze, specifically issues with selecting appropriate operators and the accumulation of ineffective calculations during quantum simulations. The ADAPT method typically relies on selecting operators based on their contribution to the Hamiltonian, but this can lead to the inclusion of operators that have a negligible impact on the energy, increasing the size of the ansatz without improving accuracy. By incorporating Hamiltonian information into operator selection, the algorithm prioritised physically meaningful excitations without increasing computational demands, thereby avoiding the stagnation observed in prior methods like Param-ADAPT-VQE. The Hamiltonian-Aware approach introduces a novel excitation operator selection criterion, evaluating operators based on their expected contribution to the energy gradient. This allows the algorithm to focus on the most important excitations, leading to a more efficient and accurate calculation. Furthermore, a novel pruning method systematically removes redundant operators, ensuring a streamlined and efficient process for convergence and enabling large-scale VQE implementation in quantum chemistry. This pruning is achieved by monitoring the contribution of each operator to the energy gradient and removing those that fall below a certain threshold.

Energy plateaus, a common issue in variational quantum algorithms, were avoided during application of the algorithm to strongly correlated molecules such as stretched H2O, a standard test case for demanding quantum chemistry calculations. These plateaus occur when the optimisation algorithm gets stuck in a local minimum, preventing it from finding the true ground state energy. A problem-adaptive method that discriminates and prunes redundant excitation operators accomplished this, balancing their removal with the need for convergence, and is applicable to ansätze of any size. The algorithm dynamically adjusts the pruning threshold during the optimisation process, ensuring that important operators are not removed prematurely. Numerical experiments on molecular systems demonstrate superior performance compared to baseline methods in terms of energy error, ansatz size, and measurement cost. Incorporating Hamiltonian information into the excitation operator selection prioritises physically meaningful operators without increasing computational demands. The reduction in measurement cost is particularly important for near-term quantum devices, where measurement errors can significantly impact the accuracy of the results.

Optimised quantum calculation starting points enhance molecular simulation efficiency

Constructing efficient ‘ansätze’, the initial configurations for quantum calculations, remains a key challenge in simulating molecules with limited quantum resources. The difficulty arises from the exponential scaling of the Hilbert space with the number of electrons and orbitals, requiring a careful selection of basis states to represent the molecular wavefunction. Researchers at Chinese Academy of Sciences and East China Normal University have demonstrably improved upon existing Adaptive Derivative-Assembled Pseudo-Trotter (ADAPT) methods, though further investigation is needed to determine how well this Hamiltonian-Aware ADAPT-VQE algorithm scales with molecular complexity. Sensibly, acknowledging concerns about scalability with increasingly complex molecules is important given the nascent stage of quantum computing, and future work will focus on addressing these limitations. The current limitations of quantum hardware, such as qubit coherence times and gate fidelities, further exacerbate the challenges of simulating large molecules.

The method prioritises physically relevant calculations by incorporating information about a molecule’s overall energy into the selection process, without increasing computational demands, and offers a strong model for large-scale VQE implementation. This opens avenues for exploring increasingly complex molecular systems and potentially accelerating discoveries in quantum chemistry. The algorithm represents a step towards more efficient molecular simulations on quantum computers, allowing for the exploration of increasingly complex molecular systems and accelerating discoveries in quantum chemistry by overcoming limitations found in earlier methods. These limitations involved selecting appropriate computational steps and avoiding the build-up of redundant calculations. The streamlined process and reduced computational demands make it particularly suitable for near-term quantum devices with limited qubit counts, and dynamic adjustment of the tolerance threshold during optimisation further enhances its efficiency. The tolerance threshold controls the balance between accuracy and computational cost, allowing the algorithm to adapt to the specific requirements of the problem. Future research will likely focus on extending the algorithm to larger molecular systems and exploring its potential for use in other quantum chemistry applications, such as excited state calculations and molecular dynamics simulations.

The researchers developed a new algorithm, Hamiltonian-Aware ADAPT-VQE, which improves the efficiency of variational quantum eigensolver calculations for molecules. It addresses previous issues with selecting appropriate computational steps and avoiding redundant calculations, resulting in reduced energy error and measurement cost. This method prioritises physically meaningful calculations without increasing computational demands, offering a robust approach for large-scale VQE implementation. The authors indicate future work will focus on extending the algorithm to larger molecular systems and exploring its use in other quantum chemistry applications.