Molecular dynamics calculation methods and systems based on quantum simulation

Abstract

This invention discloses a molecular dynamics calculation method and system based on quantum simulation. The method first applies a discrete quantum Fourier transform to the atomic coordinates of the molecular system, mapping the original plane wave basis to a system containing only O(N) atoms. 2 The invention constructs a plane-wave bibasic quantum mechanics equation; then, it constructs and solves the electronic structure Hamiltonian on a quantum computer to obtain the ground state energy. This quantum energy is combined with the classical empirical potential energy according to predetermined weighting coefficients to calculate the total synthesized energy, from which the atomic forces are determined. The integration time step is then adaptively adjusted based on a comparison of the quantum energy change between consecutive time steps with a preset threshold. This adaptive mechanism shortens the time step to ensure accuracy when quantum effects are significant, and lengthens the step to improve efficiency when changes are gradual. This invention significantly reduces Hamiltonian complexity, balancing quantum accuracy and classical efficiency, and is applicable to various fields such as drug screening, materials design, and catalytic reaction simulation.

Description

Technical Field

[0001] This invention relates to the fields of quantum computing and molecular dynamics simulation technology, specifically to a molecular dynamics calculation method and system based on quantum simulation. Background Technology

[0002] Molecular dynamics simulations are crucial tools for studying materials, drugs, and chemical processes. Traditional molecular dynamics can be divided into classical simulations based on empirical force fields and ab initio simulations based on quantum mechanics. Classical simulations are computationally efficient but have limited accuracy; ab initio simulations obtain the ground-state energy of the system by solving the Schrödinger equation, achieving high accuracy, but their computational complexity increases exponentially with the system size, making them difficult to simulate large systems. Existing literature attempts to combine quantum computing with classical computing to obtain accurate energies. For example, CN114512195B discloses a hybrid method that solves for the ground-state energy on a quantum computer and integrates over a fixed time step, but it does not solve problems such as the high complexity of the Hamiltonian, adaptive integration, and the fusion of classical potential energy. Summary of the Invention

[0003] Technical objective: To address the efficiency bottleneck of existing ab initio molecular dynamics simulations in large-scale systems, this invention discloses a molecular dynamics calculation method and system based on quantum simulation.

[0004] Technical solution: To achieve the above technical objectives, the present invention adopts the following technical solution:

[0005] A molecular dynamics calculation method based on quantum simulation specifically includes the following steps:

[0006] Step 1: Perform a unitary transformation on the electronic structure Hamiltonian of the molecular system to reduce the number of non-zero terms in the basis of the Hamiltonian, thereby reducing the complexity of subsequent quantum calculations.

[0007] Step 2: Use a quantum algorithm on a quantum computing device to obtain the energy measure of the Hamiltonian after the unitary transformation, and obtain the quantum energy describing the current configuration of the system;

[0008] Step 3: Calculate the classical potential energy of the system based on the atomic coordinates of the system and the selected force field on a classical computing device, and combine the classical potential energy and quantum energy into the total energy according to the preset dimensionless weighting coefficients;

[0009] Step 4: Calculate the forces acting on each atom in the system based on the total energy, and update the velocity and position of each atom in the system using numerical integration.

[0010] Step 5: Determine the integration step size for the next time step based on the quantum energy changes of adjacent time steps. If the quantum energy change exceeds the threshold, reduce the step size; otherwise, maintain or increase the step size.

[0011] Step 6: Repeat steps 2 to 5 to obtain the time evolution of the molecular system.

[0012] Preferably, the unitary transform in step 1 is a discrete quantum Fourier transform, and the elements of the transform matrix U satisfy... ,in The number of basis functions. is the index of the basis function, and i is the imaginary unit.

[0013] Preferably, the quantum algorithm used to calculate energy in step 2 is a variable quantum eigenvalue solving algorithm.

[0014] Preferably, the total energy in step 3 is calculated using the following formula:

[0015] ,

[0016] in Represents the classical potential energy of the system. This represents the quantum energy of the system. These are the weighting coefficients.

[0017] Preferably, the method for determining the integration step size of the next time step in step 5 includes: calculating the difference in quantum energy between adjacent time steps. and compare it with a preset threshold. In comparison, when When the next time step is set to a fraction of the current time step, The next time step size will be set to a multiple of the current time step size or will remain unchanged.

[0018] A quantum simulation-based molecular dynamics computation system is provided to implement the quantum simulation-based molecular dynamics computation method described above. The system includes a quantum processing module, a classical processing module, a communication module, and a storage unit, wherein:

[0019] The quantum processing module is used to perform unitary transformations and run quantum algorithms to obtain the quantum energy of the system;

[0020] The classical processing module is used to calculate classical potential energy, combine total energy, calculate forces, and perform adaptive time-step integral updates based on the total energy.

[0021] The communication module is used to transmit coordinate information, energy data, and step size information between the quantum processing module and the classical processing module;

[0022] Storage units are used to store program instructions, parameters, and intermediate data.

[0023] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a quantum simulation-based molecular dynamics calculation method as described above.

[0024] A storage medium characterized in that it stores computer-executable instructions for performing a quantum simulation-based molecular dynamics calculation method as described above.

[0025] A computer program product containing instructions that, when executed on a computer, cause the computer to perform a quantum simulation-based molecular dynamics calculation method as described above.

[0026] Beneficial Effects: The molecular dynamics calculation method and system based on quantum simulation provided by this invention have the following beneficial effects:

[0027] 1. This invention introduces a dual-basis construction technique based on discrete quantum Fourier transform in step 2, transforming the electronic structure Hamiltonian, which originally exhibited high dimensionality and numerous coupling terms under plane wave basis, into one containing only approximately O(N) basis terms. 2 The sparsity of the terms significantly reduces the quantum circuit depth and number of measurements required by the quantum processing module to solve for electronic structures. This solves the technical problem in existing hybrid quantum-classical molecular dynamics where the quantum part becomes extremely complex with a single calculation, making it unusable for large systems. This allows high-precision quantum energy to be incorporated into actual molecular dynamics time evolution. This approach of first reducing the dimensionality of the basis space and then solving for quantum energy is not simply about connecting a quantum computer to classical molecular dynamics; rather, it involves structural modifications at the basis space level to match the computational complexity of the quantum end with the stepping speed of the classical end, achieving the effect of supporting large-scale system dynamics simulations with relatively small quantum resources.

[0028] 2. In steps 3, 4, and 5, this invention weighted and fused the electronic ground state energy obtained from quantum calculation with the classical potential energy to construct a unified total energy expression. In step 5, the atomic force is obtained by uniformly differentiating based on this total energy. This solves the technical problem in existing technologies where quantum energy and classical force fields are calculated separately, making it difficult to drive time progression within the same mechanical framework. This allows the dynamic updates of the system to be truly established on an energy surface that simultaneously incorporates quantum precision and classical efficiency. Through this integrated energy-level construction, this invention can significantly improve the physical reliability of trajectories in scenarios where local electronic effects determine the overall dynamic direction, such as protein-small molecule active sites, defect-dominated crystal materials, and surface catalytic reactions. It avoids the problems of underestimating the potential barrier, unstable adsorbed states, and failure to capture transition states caused by using only classical potential energy in traditional methods. Thus, it achieves the technical effect of improving precision in critical regions and maintaining efficiency in non-critical regions.

[0029] 3. In step 6, this invention introduces an adaptive time step adjustment mechanism triggered by quantum energy changes. This binds the time step value to the actual intensity of quantum state changes, solving the technical problem common in existing molecular dynamics where fixed time steps lead to either instability or low efficiency. When quantum energy fluctuations are large, the time step is automatically reduced to prevent quantum energy abrupt changes from being masked by large time step integrations. When quantum energy changes are gradual, the time step is maintained or appropriately increased, significantly reducing the number of quantum calls and classical integrations required for a single simulation. This mechanism, combined with the aforementioned bi-basic dimensionality reduction and quantum-classical energy fusion, enables this invention to obtain a smoother, more physically accurate dynamic trajectory with less computational overhead under the same hardware conditions. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0031] Figure 1 This is a flowchart of the method of the present invention;

[0032] Figure 2 This is a system block diagram of the present invention;

[0033] Figure 3 This is a schematic diagram of the protein-small molecule system in Example 1 of the present invention;

[0034] Figure 4 This is a schematic diagram of the material design system in Embodiment 2 of the present invention;

[0035] Figure 5 This is a schematic diagram of the catalytic reaction system in Example 3 of the present invention. Detailed Implementation

[0036] The present invention will now be described more clearly and completely by way of a preferred embodiment in conjunction with the accompanying drawings, but this does not limit the invention to the scope of the described embodiment.

[0037] like Figure 1 As shown, a molecular dynamics calculation method based on quantum simulation specifically includes the following steps:

[0038] Step 1: Perform a unitary transformation on the electronic structure Hamiltonian of the molecular system to reduce the number of non-zero terms in the basis of the Hamiltonian, thereby reducing the complexity of subsequent quantum calculations.

[0039] The atomic coordinates of the molecular system are taken as real vectors. The input is processed using the Discrete Quantum Fourier Transform (DQFT) in the quantum processing module to perform a Western transformation on the original plane-wave basis, yielding a plane-wave bibasis with diagonalized potential energy terms. DQFT is a discrete Fourier transform implemented in quantum computing that rearranges the amplitudes of quantum states by exponential factors; the specific transformation matrix satisfies... , where N is the number of basis functions. Through this mapping, the potential energy operator in the electronic structure Hamiltonian is approximately diagonalized in the bibasic basis, originally containing The Hamiltonian of the term is compressed to item.

[0040] Step 2: Use a quantum algorithm on a quantum computing device to obtain the energy measure of the Hamiltonian after the unitary transformation, and obtain the quantum energy describing the current configuration of the system.

[0041] Constructing the electronic structure Hamiltonian of the mapped system on a quantum computer The Variational Quantum Eigensolver (VQE) algorithm is employed, which uses parameterized quantum circuits to prepare test states and then adjusts the circuit parameters using a classical optimizer to minimize the expected value. Thus, the ground state energy of the system is obtained. ,in, For parameterized quantum state functions, it is represented by parameters. The state vector generated by the controlled quantum circuit. VQE is a hybrid quantum-classical algorithm that uses quantum circuits to generate parameterized states and classical optimization iteratively seeks the best solution. This invention uses it as an electronic structure solver.

[0042] Step 3: Calculate the classical potential energy of the system based on the atomic coordinates of the system and the selected force field on a classical computing device, and combine the classical potential energy and quantum energy into the total energy according to the preset dimensionless weighting coefficient.

[0043] The empirical potential energy of the system is calculated in the classical processing module. This includes force field terms such as bond stretching, angular deformation, and non-bonded interactions. It relates to quantum energy. With classical potential energy According to predetermined weighting coefficients Linear combination to construct the total energy function ,in This is a dimensionless constant, typically taken as 0.1 to 2.0; when As the value increases, the proportion of quantum energy rises, making it suitable for systems with strong quantum effects. When the value decreases, the proportion of classical potential energy increases, making it suitable for systems with weaker quantum effects. The force on each atom can be calculated from the total energy function. ,in Let be the spatial coordinates of the i-th atom. For position The gradient operator represents the partial derivative with respect to the coordinates of the i-th atom; the forces acting on it include classical force field interactions and correction terms obtained through the quantum energy gradient. .

[0044] Step 4: Calculate the forces acting on each atom in the system based on the total energy, and update the velocity and position of each atom in the system using numerical integration.

[0045] Step 5: Determine the integration step size for the next time step based on the quantum energy changes of adjacent time steps. If the quantum energy change exceeds the threshold, reduce the step size; otherwise, maintain or increase the step size.

[0046] An improved symplectic integral or velocity Verlet algorithm is used to update atomic velocities and positions. Let the current time step be... Step size is After completing the first integration step, the new quantum energy is obtained again using VQE. Calculate the difference between two quantum energies. Set a threshold. It is a positive real number used to measure the significance of changes in quantum energy; The typical range of values ​​is Hartree. If This indicates that the quantum energy of the system changes significantly, and the next step length should be reduced, for example... like If the step size is increased appropriately (e.g., multiplied by 1.2) or kept unchanged, the simulation efficiency can be improved. This adaptive strategy dynamically balances integration accuracy and efficiency.

[0047] Step 6: Repeat steps 2 to 5 to obtain the time evolution of the molecular system.

[0048] like Figure 2 As shown, a quantum simulation-based molecular dynamics calculation system is used to implement the quantum simulation-based molecular dynamics calculation method described above. It includes a quantum processing module, a classical processing module, a communication module, and a storage unit, wherein:

[0049] The quantum processing module is used to perform unitary transformations and run quantum algorithms to obtain the quantum energy of the system;

[0050] The classical processing module is used to calculate classical potential energy, combine total energy, calculate forces, and perform adaptive time-step integral updates based on the total energy.

[0051] The communication module is used to transmit coordinate information, energy data, and step size information between the quantum processing module and the classical processing module;

[0052] Storage units are used to store program instructions, parameters, and intermediate data.

[0053] This invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements any of the above-described quantum simulation-based molecular dynamics calculation methods. The memory can be of various types, such as random access memory, read-only memory, flash memory, etc. The processor can be of various types, such as a central processing unit, microprocessor, digital signal processor, or image processor, etc.

[0054] This invention also provides a storage medium storing computer-executable instructions for executing any of the quantum simulation-based molecular dynamics calculation methods described above. The storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0055] The present invention also provides a computer program product containing instructions that, when the instructions are executed on a computer, cause the computer to perform a quantum simulation-based molecular dynamics calculation method as described above, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0056] Example 1

[0057] This embodiment is used to verify the application of the present invention in biological macromolecular systems. It is assumed that the system consists of a protein receptor and multiple small molecule ligands, and its initial coordinates and velocities are known.

[0058] Perform DQFT on all atomic coordinates of the system to convert the plane wave basis into a plane wave bibasis, and obtain the Hamiltonian representation in the bibasis.

[0059] VQE is used in the quantum processing module to solve for bibasic Hamiltonians. The ground state energy is obtained. .

[0060] In the classical processing module, the classical potential energy of the system is calculated according to a specific force field (such as the AMBER or CHARMM force field). .

[0061] Using weighting coefficients By combining quantum energy and classical potential energy, the total energy can be calculated. .

[0062] according to Calculate the forces acting on each atom and update its position and velocity using the Verlet algorithm. Initial step size. Set to 1fs. Update after each step of the integration. And calculate .set up Hartree, if Then Halve the value, otherwise leave it unchanged. Iterate continuously to obtain the dynamic behavior of the protein-small molecule system. Results show that, compared to traditional ab initio molecular dynamics, this embodiment improves simulation efficiency by approximately 3 times while maintaining the accuracy of energy barriers and stable conformations.

[0063] like Figure 3 As shown, the protein molecule is represented by its overall outline, indicating the macromolecular object requiring molecular dynamics simulation. Several small ligands are shown at the active site or pocket region of the protein molecule, indicating that the quantum energy solution in this embodiment can perform refined calculations on the electronic structure near the active site to improve the accuracy of the interaction energy in this region. Multiple aqueous solvents are scattered around the protein molecule, representing a solvated environment, demonstrating that this invention can handle not only molecular systems under vacuum conditions but also systems containing solvent molecules that more closely resemble actual biological environments. By using the atomic coordinates of the protein-small molecule system as input to step 1, this invention can perform quantum-level energy estimation of the electronic structure near the ligand binding pocket on the quantum processing module, and then fuse it with the classical force field energy, so that the final force distribution can more accurately reflect the dynamic behavior of microscopic processes such as ligand entry, oscillation, and desorption. This figure is only a schematic diagram; the three-dimensional conformation, number of ligands, and amount of solvent in the actual protein can be adjusted according to the specific simulation object.

[0064] Example 2

[0065] This embodiment is used to verify the application of the present invention in materials science. Consider simulating the displacement evolution of a metallic crystal under external force loading.

[0066] Collect the initial coordinates and velocities of each atom in the crystal; establish a classical force field (e.g., the embedded atom method EAM) to describe the interactions within the crystal.

[0067] DQFT is performed on the wave function basis of the electrons in the system to construct a plane wave bibasis and obtain a simplified Hamiltonian.

[0068] Calculating the ground state energy of metallic crystals using VQE .

[0069] set up Combining classical potential energy and This forms the total energy. Because quantum effects are relatively weak in metallic systems, their weight is relatively small.

[0070] initial step size fs. Settings Hartree. During the simulation, when an external force causes a drastic change in energy... If the threshold is exceeded, the step size will automatically shorten; during the gradual plastic deformation phase, the step size will be appropriately increased. This strategy can capture energy spikes during the phase transition process while improving overall efficiency. The results verify the applicability of the invention.

[0071] like Figure 4 As shown, multiple crystal structures are arranged regularly on the left side of the figure, representing the basic unit of the material to be analyzed or optimized. These can be metallic crystals, semiconductor lattices, ceramic lattices, or quasicrystalline structures. A vertical loading direction is shown on the right side of the figure, indicating that stress, strain, or external loads need to be applied to the material in a specific direction in this embodiment to examine the microstructure evolution and phase transition behavior of the material under external conditions. The working method of this invention in this type of system is as follows: first, the atomic coordinates of the crystal structure are input into the calculation process; then, the quantum energy solution part performs a high-precision quantum solution for the electronic energy of the local bonding region or defect region; finally, the classical processing module completes the overall force and displacement integration, and combines the loading direction for dynamic stepping, thereby obtaining more realistic dynamic simulation results of the crystal response under loading at a lower quantum computing scale. This figure emphasizes that this invention is applicable not only to molecular or biological systems, but also to engineering material systems with periodicity, regularity, or loaded boundary conditions.

[0072] Example 3

[0073] This embodiment is used to verify the application of the present invention in catalytic chemistry. Taking the adsorption reaction on the surface of a metal oxide as an example, the reaction pathway and activation energy are simulated.

[0074] Atomic models of the catalyst surface and adsorbate are established, and classical force fields are used to describe the interaction between the surface and adsorbate.

[0075] After performing DQFT mapping on the electronic Hamiltonian of the system, the ground state energy was solved using VQE in the quantum processing module. To describe the breaking and formation of chemical bonds during the adsorption reaction, a... To increase the quantum energy weight.

[0076] An improved symplectic integral algorithm is used to calculate the forces and update the atomic coordinates. A threshold is set. Hartree. The energy change is small in the initial stage of the reaction, with a step size of 0.25 fs; as the system approaches the transition state... When the step size increases sharply, it automatically decreases, thus accurately capturing the reaction barrier. The simulation results agree well with high-precision ab initio data.

[0077] like Figure 5 As shown, the lower part of the figure represents a horizontal strip-shaped catalyst surface, used to indicate the reaction support for kinetic simulation. This surface can be a metal surface, an oxide surface, or the surface of other porous catalytic materials. Several adsorbate molecules are arranged above the catalyst surface, representing reactants, reaction intermediates, or cooperating molecules adsorbed onto the catalyst surface. A diagonal arrow extends upwards from one of the adsorbate molecule positions, with the tip labeled as a reaction intermediate state. This indicates that in the calculation process of this invention, the evolution process from the adsorbed state to the transition or intermediate state can be solved using quantum energy and energy fusion, ensuring that the generated force correctly guides the molecule's migration along the reaction path. At each time step, quantum-level energy estimation is performed on atoms or molecules near the catalytically active site, while the remaining regions far from the reaction center are still described by classical potential energy. When a quantum energy change exceeds the threshold ε, an adaptive time step mechanism is used to reduce the time step to capture the formation process of the reaction intermediate state, improving the resolution of surface reaction kinetics. This figure highlights the technical advantages of this invention in a scenario of local quantum refinement combined with overall classical advancement.

[0078] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A molecular dynamics calculation method based on quantum simulation, characterized in that, Specifically, the following steps are included: Step 1: Perform a unitary transformation on the electronic structure Hamiltonian of the molecular system to reduce the number of non-zero terms in the basis of the Hamiltonian, thereby reducing the complexity of subsequent quantum calculations. Step 2: Use a quantum algorithm on a quantum computing device to obtain the energy measure of the Hamiltonian after the unitary transformation, and obtain the quantum energy describing the current configuration of the system; Step 3: Calculate the classical potential energy of the system based on the atomic coordinates of the system and the selected force field on a classical computing device, and combine the classical potential energy and quantum energy into the total energy according to the preset dimensionless weighting coefficients; Step 4: Calculate the forces acting on each atom in the system based on the total energy, and update the velocity and position of each atom in the system using numerical integration. Step 5: Determine the integration step size for the next time step based on the quantum energy changes of adjacent time steps. If the quantum energy change exceeds the threshold, reduce the step size; otherwise, maintain or increase the step size. Step 6: Repeat steps 2 to 5 to obtain the time evolution of the molecular system.

2. The molecular dynamics calculation method based on quantum simulation according to claim 1, characterized in that, In step 1, the unitary transform is a discrete quantum Fourier transform, and the elements of the transform matrix U satisfy... ,in The number of basis functions. is the index of the basis function, and i is the imaginary unit.

3. The molecular dynamics calculation method based on quantum simulation according to claim 1, characterized in that, The quantum algorithm used to calculate energy in step 2 is a variable quantum eigenvalue solving algorithm.

4. The molecular dynamics calculation method based on quantum simulation according to claim 1, characterized in that, The total energy in step 3 is calculated using the following formula: ， in Represents the classical potential energy of the system. This represents the quantum energy of the system. These are the weighting coefficients.

5. The molecular dynamics calculation method based on quantum simulation according to claim 1, characterized in that, The method for determining the integration step size of the next time step in step 5 includes: calculating the difference in quantum energy between adjacent time steps. and compare it with a preset threshold. In comparison, when When the next time step is set to a fraction of the current time step, The next time step size will be set to a multiple of the current time step size or will remain unchanged.

6. A molecular dynamics calculation system based on quantum simulation, characterized in that, A molecular dynamics calculation method based on quantum simulation as described in any one of claims 1-5, comprising a quantum processing module, a classical processing module, a communication module, and a storage unit, wherein: The quantum processing module is used to perform unitary transformations and run quantum algorithms to obtain the quantum energy of the system; The classical processing module is used to calculate classical potential energy, combine total energy, calculate forces, and perform adaptive time-step integral updates based on the total energy. The communication module is used to transmit coordinate information, energy data, and step size information between the quantum processing module and the classical processing module; Storage units are used to store program instructions, parameters, and intermediate data.

7. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a quantum simulation-based molecular dynamics calculation method as described in any one of claims 1-5.

8. A storage medium, characterized in that, The device stores computer-executable instructions for performing a quantum simulation-based molecular dynamics calculation method as described in any one of claims 1-5.

9. A computer program product containing instructions, characterized in that, When the instructions are executed on a computer, the computer performs a molecular dynamics calculation method based on quantum simulation as described in any one of claims 1-5.

Images