A molecular dynamics quantum variational simulation method and device

Abstract

The present application provides a molecular dynamics quantum variational simulation method and apparatus. In the molecular dynamics quantum variational simulation method, it includes obtaining a set of position basis vectors by measuring the initial quantum state of a target molecular system, then constructing a parameterized quantum circuit using the set of position basis vectors and the momentum basis vectors of the initial quantum state, and then iteratively updating the parameters in the parameterized quantum circuit until the target evolution time is reached to obtain the final quantum state of the target molecular system. The present application can handle large-scale molecular simulation problems on relatively shallow quantum circuits, ensuring that the circuit depth is only related to the system Hamiltonian and does not increase with the simulation time. Since the quantum circuit depth is independent of the evolution time, the method in the present application can be applied to simulate processes such as the structural evolution and phase transition of materials on long time scales.

Description

Technical Field

[0001] One or more embodiments of this specification relate to the field of molecular dynamics, and in particular, to a molecular dynamics quantum variational simulation method and apparatus. Background Art

[0002] Molecular dynamics simulation is one of the core issues in computational chemistry and has application scenarios in many fields such as chemistry, materials, and biology. For example, understanding the process of chemical reactions, the design and discovery of new materials and drugs, and the folding and binding of protein molecules all require in-depth understanding of the behavior of molecules under different conditions. Traditional classical algorithms usually assume that atomic nuclei are stable and their positions are given, so the nuclear motion is regarded as a parameter, thus simplifying the problem to solving the motion state of electrons. Information on dynamic coupling may be lost during the calculation process, resulting in failure to achieve ideal results when approximations fail under conditions such as high energy and high speed.

[0003] Applying quantum computing to molecular dynamics simulation can more accurately describe the structure of quantum systems. However, current quantum simulation algorithms require quantum circuits of considerable depth. The increase in the complexity of quantum circuits also increases the demand for computing resources, posing a huge challenge to quantum computers. Summary of the Invention

[0004] This application describes a molecular dynamics quantum variational simulation method and apparatus that can solve the above technical problems.

[0005] According to a first aspect, a molecular dynamics quantum variational simulation method is provided. The method includes:

[0006] Sampling the results obtained by measuring the initial quantum state of the target molecular system in the spatial basis vectors to obtain a set of position basis vectors, which is used to approximately describe the spatial distribution of the quantum state of the target molecular system;

[0007] Construct a parameterized quantum circuit in the unitary form through the set of position basis vectors and the momentum basis vectors of the initial quantum state of the target molecular system. The parameterized quantum circuit includes parameters to be optimized, and the parameterized quantum circuit is used to simulate the evolution of the target molecular system in momentum and position spaces;

[0008] At each iteration time interval, use the parameterized quantum circuit and the time-dependent Schrödinger equation to iteratively update the parameters to be optimized and the parameterized quantum circuit until the target evolution time is reached to obtain the final quantum state of the target molecular system.

[0009] Based on the above further embodiments, the parameterized quantum circuit includes a first quantum gate and a second quantum gate;

[0010] The first quantum gate constructs a rotation gate using momentum basis vectors, and the rotation angle in the first quantum gate belongs to the parameters to be optimized;

[0011] The second quantum gate constructs a rotation quantum gate using the set of position basis vectors, and the rotation angle in the second quantum gate belongs to the parameters to be optimized.

[0012] Based on the above further embodiments, within each iteration time step, using the parameterized quantum circuit and the time-dependent Schrödinger equation, iteratively update the parameters to be optimized and the parameterized quantum circuit, specifically including:

[0013] Before the start of each iteration time interval, use the parameterized quantum circuit to prepare the derivative state of the parameters to be optimized;

[0014] Through the derivative state of the parameters to be optimized and the time-dependent Schrödinger equation, obtain the change rate of the parameters to be optimized;

[0015] Utilize the change rate of the parameters to be optimized to iteratively update the parameters to be optimized, and update the parameterized quantum circuit through the updated parameters to be optimized.

[0016] Based on the above further embodiments, the obtaining the change rate of the parameters to be optimized through the derivative state of the parameters to be optimized and the time-dependent Schrödinger equation specifically includes:

[0017] Use the parameterized quantum circuit to prepare a parameterized quantum state;

[0018] Substitute the parameterized quantum state into the time-dependent Schrödinger equation to obtain a linear equation, which reflects the change of the parameters to be optimized over time. The linear equation includes an evolution matrix and an evolution vector. The evolution matrix is composed of the derivative states of the parameters to be optimized, and the evolution vector is composed of the derivative states of the parameters to be optimized and the Hamiltonian;

[0019] Measure the evolution matrix and the evolution vector using the derivative states of the parameters to be optimized;

[0020] Substitute the evolution matrix and the evolution vector into the linear equation to obtain the change rate of the parameters to be optimized.

[0021] Based on the above further embodiments, the linear equation is expressed as , the evolution matrix , the evolution vector ;

[0022] R and I respectively represent taking the real part and the imaginary part, H is the Hamiltonian, is The partial derivative state of the parameter to be optimized is The partial derivative state of the parameter to be optimized

[0023] Based on the above further embodiments, the parameterized quantum circuit , is the rotation angle in the first quantum gate , is the set of position basis vectors is the momentum basis vector is a position basis vector, and the position basis vector belongs to the set of position basis vectors

[0024] Based on the above further embodiments, using the change rate of the parameter to be optimized to iteratively update the parameter to be optimized specifically includes:

[0025] Using the gradient method, using the change rate of the parameter to be optimized to iteratively update the parameter to be optimized

[0026] According to a second aspect, there is provided a molecular dynamics quantum variational simulation device, the device includes:

[0027] A first processing module, configured to sample the result obtained by measuring the initial quantum state of the target molecular system in the spatial basis vectors to obtain a set of position basis vectors, and the set of position basis vectors is used to approximately describe the spatial distribution of the quantum state of the target molecular system;

[0028] A second processing module, configured to construct a unitary parameterized quantum circuit through the momentum basis vector and the set of position basis vectors of the initial quantum state of the target molecular system, and the parameterized quantum circuit includes parameters to be optimized, and the parameterized quantum circuit is used to simulate the evolution of the target molecular system in the momentum and position spaces;

[0029] A third processing module, configured to, within each iterative time step, use the parameterized quantum circuit and the time-dependent Schrödinger equation to iteratively update the parameter to be optimized and the parameterized quantum circuit until the target evolution time is reached to obtain the final quantum state of the target molecular system

[0030] Based on the above further embodiments, the parameterized quantum circuit includes a first quantum gate and a second quantum gate;

[0031] The first quantum gate uses the momentum basis vector to construct a rotation gate, and the rotation angle in the first quantum gate belongs to the parameter to be optimized;

[0032] The second quantum gate uses the set of position basis vectors to construct a rotation quantum gate, and the rotation angle in the second quantum gate belongs to the parameter to be optimized

[0033] Based on the above further embodiments, the third processing module is specifically configured to prepare a parameterized quantum state using the parameterized quantum circuit;

[0034] Substitute the parameterized quantum state into the time-dependent Schrödinger equation to obtain a linear equation, where the linear equation reflects the variation of the parameter to be optimized with time. The linear equation includes an evolution matrix and an evolution vector. The evolution matrix is composed of the partial derivative states of the parameter to be optimized, and the evolution vector is composed of the partial derivative states of the parameter to be optimized and the Hamiltonian;

[0035] Measure the evolution matrix and the evolution vector using the partial derivative states of the parameter to be optimized;

[0036] Substitute the evolution matrix and the evolution vector into the linear equation to obtain the rate of change of the parameter to be optimized.

[0037] Based on the above further embodiments, the linear equation is expressed as , where the evolution matrix , and the evolution vector ;

[0038] R and I respectively represent taking the real part and the imaginary part, H is the Hamiltonian, is the partial derivative state of the parameter to be optimized, is the partial derivative state of the parameter to be optimized.

[0039] Based on the above further embodiments, the parameterized quantum circuit , is the rotation angle in the first quantum gate, , is the set of position basis vectors, is the momentum basis vector, is a position basis vector, and the position basis vector belongs to the set of position basis vectors.

[0040] Using the gradient method, use the rate of change of the parameter to be optimized to iteratively update the parameter to be optimized.

[0041] According to a third aspect, there is provided a computer storage medium, on which a computer program is stored. When the computer program is executed by one or more processors, the molecular dynamics quantum variational simulation method described in any one of the above technical solutions is implemented.

[0042] According to a fourth aspect, there is provided an electronic device including a memory and one or more processors. A computer program is stored on the memory, and when the computer program is executed by the one or more processors, it implements the molecular dynamics quantum variational simulation method according to any one of the above technical solutions.

[0043] In the above systems and methods provided in the embodiments of the present specification, the architecture of quantum computing variational method is adopted to solve the molecular dynamics problem. A shallow quantum circuit can be used to simulate the evolution of a long-time actual quantum system, alleviating the demand for quantum device resources. By means of measurement sampling, the largest partial components of the nuclear wave function in the basis vectors are obtained to characterize the behavior of the nucleus, which can further avoid the influence of unwanted terms on the depth of the quantum circuit. The Hamiltonian and molecular structure are introduced into the design of the parameterized quantum circuit, taking into account the quantum effects between molecules and having higher calculation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and those of ordinary skill in the art can obtain other drawings according to these drawings without creative efforts.

[0045] Figure 1 A schematic flowchart showing the molecular dynamics quantum variational simulation method provided in the embodiments of the present specification;

[0046] Figure 2 A schematic diagram showing the quantum circuit structure provided in the embodiments of the present specification;

[0047] Figure 3 A schematic flowchart showing the molecular dynamics quantum variational simulation method provided in the embodiments of the present specification;

[0048] Figure 4 A schematic diagram showing the molecular dynamics quantum variational simulation device provided in the embodiments of the present specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] The solutions provided in the present specification will be described below with reference to the drawings.

[0050] In order to make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below with reference to the drawings.

[0051] In the description of the embodiments of the present application, words such as "exemplary", "for example", or "for illustration" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary", "for example", or "for illustration" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary", "for example", or "for illustration" is intended to present relevant concepts in a specific manner.

[0052] In the description of the embodiments of the present application, the term "and / or" is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, B exists alone, and both A and B exist simultaneously. Additionally, unless otherwise specified, the meaning of the term "plurality" refers to two or more.

[0053] Furthermore, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The terms "include", "comprise", "have" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0054] In the process of classical molecular dynamics simulation, the atomic nuclei are assumed to be stable, their positions are considered to be given, the nuclear motion is regarded as a parameter, and the problem is simplified to solving the motion state of electrons. The specific process is as follows: First, select an initial molecular configuration, determine the Hamiltonian. According to the Born - Oppenheimer approximation, the atomic nuclei in the Hamiltonian are regarded as stationary, and the molecular Hamiltonian is separated from the Hamiltonian as a function of the positions of the atomic nuclei and electrons. Solve the Schrödinger equation corresponding to the molecular Hamiltonian to obtain the electron orbit as a function of the positions of the atomic nuclei as parameters. Use the obtained electron density distribution to calculate the potential energy distribution where the atomic nuclei are located, calculate the classical motion of the atomic nuclei in the potential energy surface, update the positions of the atomic nuclei, and repeat the above steps until the dynamic simulation process of the molecular system is obtained at the target evolution time. This method can be used to simulate the dynamic behavior of a molecular system within a certain time, but the Born - Oppenheimer approximation method is limited to conditions such as high energy and high speed. In addition, traditional methods usually consider the coupling between the atomic nuclei and electrons to be static, that is, the positions of the atomic nuclei are regarded as fixed during the calculation process, and information on dynamic coupling may be lost during the calculation process. And due to the fact that long - time - scale simulations will lead to numerical instability or excessive computational costs, classical methods are usually limited to short - time - scale simulations.

[0055] The current technical solution for solving dynamic problems using quantum computing is the product formula method. Its main idea is to decompose the Hamiltonian of the many-body problem into local Hamiltonians. Its process includes decomposing the total Hamiltonian of the system into a sum of a series of commuting local Hamiltonians, using the Trotter-Suzuki decomposition to approximate the time evolution operator as a product form of local Hamiltonians, constructing a quantum circuit based on the decomposed Hamiltonian, with each local Hamiltonian corresponding to a local quantum gate in the circuit, and running the optimized quantum circuit on a quantum computer to simulate the evolution of the quantum state of the system over time until the target evolution time. However, this method usually requires very deep quantum circuits. In the method of considering the first-order slice in the product formula method, the number of quantum gates required is proportional to the square of the evolution time. The increase in the complexity of the quantum circuit also increases the demand for computing resources. Especially on current quantum computing devices, due to the existence of noise and a limited number of qubits, these additional gate operations may lead to the decoherence of the quantum state and error accumulation.

[0056] In view of this, the present invention proposes a molecular dynamics simulation method based on the variational method, which can solve the problem of the increase in the complexity of the quantum circuit with the increase in the evolution time. Figure 1 The conceptual schematic diagram of the dynamics simulation method of the present invention is shown. As shown in the figure, first, the initial quantum state of the target molecular system is set, and its wave function generally aggregates in space in the form of a wave packet. The initial quantum state of the target molecular system can be measured multiple times in the spatial basis, and the position eigenvalues obtained from the multiple measurements are statistically analyzed, and the position basis vectors with higher occurrence frequencies are selected to form the position basis vector set . A parameterized quantum circuit in the unitary format is designed , as shown in Figure 2 , the parameterized quantum circuit includes quantum gates for the momentum part , is the momentum basis vector, is the rotation angle in the quantum gate, and the quantum gates for the position part , is the position basis vector, . All the rotation angles are taken as the parameters to be optimized.

[0057] Subsequently, in an iterative time interval, the evolution matrix and the evolution vector are measured through the parameterized quantum circuit to obtain the change rate of the parameters to be optimized. The gradient method is used to iteratively update the parameters to be optimized, and the parameterized quantum circuit is updated using the updated parameters to be optimized until the target evolution time is reached, obtaining the final quantum state of the target molecular system, and the target evolution state information is obtained using the final quantum state of the target molecular system.

[0058] Based on the variational method process of classical approximation and wave packets, the present invention can handle large-scale molecular simulation problems on relatively shallow quantum circuits. At the same time, during the design stage of parameterized quantum circuits, the phases in the Hamiltonian evolution terms are taken into account to design parameterized unitary transformations, ensuring that the circuit depth is only related to the system Hamiltonian and does not increase with the simulation time. Since the quantum circuit depth is independent of the evolution time, it can be applied to simulate processes such as the structural evolution and phase transitions of materials on long time scales. In addition, the nuclear wave function is approximately regarded as a localized wave packet, and the most probable positions of the nuclei in the spatial basis are obtained through measurement sampling, thus avoiding the construction of quantum circuits in all spatial basis vectors and further reducing the number of layers of the quantum circuits.

[0059] The following will be described in detail in conjunction with Figure 3 the schematic flow diagram of the molecular dynamics quantum variational simulation method shown, specifically including:

[0060] 110. Sample the results obtained by measuring the initial quantum state of the target molecular system in the spatial basis vectors to obtain a set of position basis vectors, which is used to approximately describe the spatial distribution of the quantum state of the target molecular system.

[0061] Before measuring the target molecular system, first set the initial quantum state of the target molecular system. Generally, the wave function aggregates in space in the form of a wave packet. At the same time, set the target evolution time T, the time interval Δt, and the total number of evolution steps n = T / Δt.

[0062] Select a set of spatial basis vectors. Usually, these basis vectors are position eigenstates. Measure the initial quantum state multiple times in these spatial basis vectors. Each measurement may obtain a specific position eigenvalue, that is, a specific spatial position. Collect the results of multiple measurements and perform statistical analysis. The measurement results will give the probability distribution of finding the system in each position basis vector. According to the measurement results, select those position basis vectors with higher frequencies of occurrence to form a set of position basis vectors, which can be used to approximately describe the spatial distribution of the initial quantum state of the molecular system.

[0063] Specifically, in molecular dynamics simulation problems, due to the relatively complex interaction between atomic nuclei and electrons with each other, the Born-Oppenheimer approximation (BO approximation for short) is usually adopted. By regarding the atomic nuclei as stationary, the electronic Hamiltonian is separated, and the electron density distribution is solved, thereby in turn constituting the potential energy surface for the motion of the atomic nuclei. Since the atomic nuclei have a large mass and a relatively small range of motion, the nuclear wave function forms a localized wave packet in the spatial basis vectors below. In this embodiment, by repeatedly measuring the initial state of the atomic nuclei in the spatial basis vectors multiple times, a set of spatial basis vectors with larger components can be obtained , used to approximate the behavior of the original nuclear wave function. The potential energy at the positions where the probability of the nucleus appears is relatively large will largely determine how the state of the nucleus evolves.

[0064] 120. Construct a parameterized quantum circuit through a set of position basis vectors and the momentum basis vectors of the initial quantum state of the target molecular system. The parameterized quantum circuit includes parameters to be optimized and is used to simulate the evolution of the target molecular system in momentum and position spaces.

[0065] Specifically, for the Hamiltonian evolution in , the momentum term of can be represented in the momentum basis through a quantum Fourier transform, and the potential energy term can be approximated and restricted within the set of measurement results

[0066] to reduce the potential energy terms that need to be implemented. , which can ensure that the circuit depth is only related to the system Hamiltonian and does not increase with the simulation time.

[0067] Specifically, in this embodiment, the parameterized quantum circuit includes a first quantum gate and a second quantum gate. The first quantum gate constructs a rotation gate using the momentum basis vectors, and the rotation angle in the first quantum gate belongs to the parameters to be optimized. The second quantum gate constructs a rotation quantum gate using the set of position basis vectors, and the rotation angle in the second quantum gate belongs to the parameters to be optimized.

[0068] More specifically, as Figure 2 shown, the parameterized quantum circuit , is the rotation angle in the first quantum gate, , is the set of position basis vectors, is the momentum basis vector, is the position basis vector, and the position basis vector belongs to the set of position basis vectors.

[0069] 130. In each iteration time interval, use the parameterized quantum circuit and the time-dependent Schrödinger equation to iteratively update the parameters to be optimized and the parameterized quantum circuit until the target evolution time is reached, and obtain the final quantum state of the target molecular system.

[0070] Specifically, step 130 includes:

[0071] 131. Before the start of each iteration time interval, use the parameterized quantum circuit to prepare the partial derivative state preparation of the parameters to be optimized .

[0072] 132. Obtain the change rate of the parameters to be optimized through the partial derivative state of the parameters to be optimized and the time-dependent Schrödinger equation.

[0073] Specifically included in step 132 are:

[0074] 1321. Prepare a parameterized quantum state using a parameterized quantum circuit.

[0075] 1322. Substitute the parameterized quantum state into the time-dependent Schrödinger equation to obtain a linear equation, which reflects the change of the parameter to be optimized over time. The linear equation includes an evolution matrix and an evolution vector. The evolution matrix is composed of the partial derivative states of the parameter to be optimized, and the evolution vector is composed of the partial derivative states of the parameter to be optimized and the Hamiltonian.

[0076] 1323. Measure the evolution matrix and the evolution vector using the partial derivative states of the parameter to be optimized;

[0077] 1324. Substitute the evolution matrix and the evolution vector into the linear equation to obtain the rate of change of the parameter to be optimized.

[0078] Specifically, in this embodiment, the evolution of the quantum state over time is represented by a parameterized quantum circuit The quantum state , is the parameterized quantum circuit, is the initial state, and the equation satisfied by is calculated through the Schrödinger equation, and the structure of the parameterized quantum circuit is designed.

[0079] According to the time-dependent Schrödinger equation, substituting the parameterized quantum state, a linear equation for the change of the parameter over time can be obtained: , where ,

[0080] , and respectively represent taking the real part and the imaginary part, is the Hamiltonian, is The partial derivative state of the parameter to be optimized, is The partial derivative state of the parameter to be optimized. The partial derivative of the parameterized quantum state composed of single-qubit rotation gates is expressed as , which can be measured on the quantum circuit. W(θ) is a transformation matrix related to the derivative of the single-qubit rotation gate. For common rotation gates, it can be expressed as the evolution of the generator of the gate at a certain angle, and the process of taking the derivative can also be obtained by applying quantum gates on the quantum circuit and performing measurements.

[0081] 133. Use the rate of change of the parameter to be optimized to iteratively update the parameter to be optimized, and update the parameterized quantum circuit through the updated parameter to be optimized.

[0082] In this embodiment, the gradient method can be used to iteratively update the parameter to be optimized by using the change rate of the parameter to be optimized. For example, the formula can be used to update the parameter to be optimized .

[0083] After obtaining the quantum state of the molecular system at the final moment through the above steps, the observable to be measured can be measured to obtain information about the target quantum state of the molecular system.

[0084] This embodiment adopts the architecture of the quantum computing variational method to solve the molecular dynamics problem. A shallow quantum circuit can be used to simulate the evolution of a long-time actual quantum system, alleviating the demand for quantum device resources. By the method of measurement sampling, the largest partial components of the nuclear wave function in the basis vectors are obtained to characterize the behavior of the nucleus, which can further avoid the influence of unnecessary terms on the depth of the quantum circuit. Introducing the Hamiltonian and molecular structure into the design of the parameterized quantum circuit takes into account the quantum effects between molecules and has higher calculation accuracy.

[0085] Such as Figure 4 shown, a molecular dynamics quantum variational simulation device includes:

[0086] A first processing module, configured to obtain a set of position basis vectors according to the result obtained by measuring the initial quantum state of the target molecular system in the spatial basis vectors, where the set of position basis vectors is used to approximately describe the spatial distribution of the quantum state of the target molecular system;

[0087] A second processing module, configured to construct a parameterized quantum circuit through the momentum basis vector and the set of position basis vectors of the initial quantum state of the target molecular system, where the parameterized quantum circuit includes parameters to be optimized, and the parameterized quantum circuit is used to simulate the evolution of the target molecular system in the momentum and position spaces;

[0088] A third processing module, configured to iteratively update the parameter to be optimized and the parameterized quantum circuit by using the parameterized quantum circuit and the time-dependent Schrödinger equation at each iteration time step until the target evolution time is reached, and obtain the final quantum state of the target molecular system.

[0089] Based on the above further embodiment, the parameterized quantum circuit includes a first quantum gate and a second quantum gate;

[0090] The first quantum gate constructs a rotation gate using the momentum basis vector, and the rotation angle in the first quantum gate belongs to the parameter to be optimized;

[0091] The second quantum gate constructs a rotation quantum gate using the set of position basis vectors, and the rotation angle in the second quantum gate belongs to the parameter to be optimized.

[0092] Based on the above further embodiments, the third processing module is specifically configured to prepare a parameterized quantum state using the parameterized quantum circuit;

[0093] Substitute the parameterized quantum state into the time-dependent Schrödinger equation to obtain a linear equation, where the linear equation reflects the variation of the parameter to be optimized with time. The linear equation includes an evolution matrix and an evolution vector. The evolution matrix is composed of the partial derivative states of the parameter to be optimized, and the evolution vector is composed of the partial derivative states of the parameter to be optimized and the Hamiltonian;

[0094] Measure the evolution matrix and the evolution vector using the partial derivative states of the parameter to be optimized;

[0095] Substitute the evolution matrix and the evolution vector into the linear equation to obtain the rate of change of the parameter to be optimized.

[0096] Based on the above further embodiments, the linear equation is expressed as , where the evolution matrix , and the evolution vector ;

[0097] R and I respectively represent taking the real part and the imaginary part, H is the Hamiltonian, is the partial derivative state of the parameter to be optimized, is the partial derivative state of the parameter to be optimized.

[0098] Based on the above further embodiments, the parameterized quantum circuit , is the rotation angle in the first quantum gate, , is the set of position basis vectors, is the momentum basis vector, is a position basis vector, and the position basis vector belongs to the set of position basis vectors.

[0099] Using the gradient method, use the rate of change of the parameter to be optimized to iteratively update the parameter to be optimized.

[0100] The present invention also provides a computer storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is executed by one or more processors, the molecular dynamics quantum variational simulation method described in any one of the above technical solutions is implemented.

[0101] The present invention also provides an electronic device, including a memory and one or more processors. A computer program is stored on the memory, and when the computer program is executed by the one or more processors, it implements the molecular dynamics quantum variational simulation method according to any one of the above technical solutions.

[0102] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present application. It should be understood that the above are only specific embodiments of the present application and are not used to limit the protection scope of the present application. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solutions of the present application shall be included within the protection scope of the present application.

Claims

1. A quantum variational simulation method for molecular dynamics, characterized in that The method includes: Sampling the results obtained by measuring the initial quantum state of the target molecular system in the spatial basis vectors to obtain a set of position basis vectors, where the set of position basis vectors is used to approximately describe the spatial distribution of the quantum state of the target molecular system; Constructing a parameterized quantum circuit in unitary form through the set of position basis vectors and the momentum basis vectors of the initial quantum state of the target molecular system, where the parameterized quantum circuit includes parameters to be optimized, and the parameterized quantum circuit is used to simulate the evolution of the target molecular system in momentum and position spaces; Within each iteration time interval, using the parameterized quantum circuit and the time-dependent Schrödinger equation, iteratively update the parameters to be optimized and the parameterized quantum circuit until the target evolution time is reached to obtain the final quantum state of the target molecular system.

2. The method according to claim 1, wherein The parameterized quantum circuit includes a first quantum gate and a second quantum gate; The first quantum gate constructs a rotation gate using the momentum basis vectors, and the rotation angle in the first quantum gate belongs to the parameters to be optimized; The second quantum gate constructs a rotation gate using the set of position basis vectors, and the rotation angle in the second quantum gate belongs to the parameters to be optimized.

3. The method according to claim 1, characterized in that, The step of, within each iteration time step, using the parameterized quantum circuit and the time-dependent Schrödinger equation to iteratively update the parameters to be optimized and the parameterized quantum circuit specifically includes: Before the start of each iteration time interval, preparing the partial derivative state of the parameters to be optimized using the parameterized quantum circuit; Obtaining the change rate of the parameters to be optimized through the partial derivative state of the parameters to be optimized and the time-dependent Schrödinger equation; Using the change rate of the parameters to be optimized to iteratively update the parameters to be optimized, and updating the parameterized quantum circuit through the updated parameters to be optimized.

4. The method according to claim 3, characterized in that, The step of obtaining the change rate of the parameters to be optimized through the partial derivative state of the parameters to be optimized and the time-dependent Schrödinger equation specifically includes: Preparing a parameterized quantum state using the parameterized quantum circuit; Substituting the parameterized quantum state into the time-dependent Schrödinger equation to obtain a linear equation, where the linear equation reflects the change of the parameters to be optimized with time, and the linear equation includes an evolution matrix and an evolution vector, the evolution matrix is composed of the partial derivative states of the parameters to be optimized, and the evolution vector is composed of the partial derivative states of the parameters to be optimized and the Hamiltonian; Measuring the evolution matrix and the evolution vector using the partial derivative states of the parameters to be optimized; Substituting the evolution matrix and the evolution vector into the linear equation to obtain the change rate of the parameters to be optimized.

5. The method according to claim 4, wherein The linear equation is expressed as , the evolution matrix , the evolution vector ; R and I represent taking the real and imaginary parts respectively, and H is the Hamiltonian, is the partial derivative state of the parameter to be optimized, is the partial derivative state of the parameter to be optimized.

6. The method according to claim 2, wherein The parameterized quantum circuit , is the rotation angle in the first quantum gate, , is the set of position basis vectors, is the momentum basis vector, is a position basis vector, and the position basis vector belongs to the set of position basis vectors.

7. The method according to claim 2, wherein The step of using the change rate of the parameters to be optimized to iteratively update the parameters to be optimized specifically includes: Using the gradient method to iteratively update the parameters to be optimized using the change rate of the parameters to be optimized.

8. A quantum variational simulation device for molecular dynamics, characterized in that, The device includes: The first processing module is configured to sample the result obtained by measuring the initial quantum state of the target molecular system in the spatial basis vectors to obtain a set of position basis vectors, and the set of position basis vectors is used to approximately describe the spatial distribution of the quantum state of the target molecular system; The second processing module is configured to construct a parameterized quantum circuit in unitary format through the set of position basis vectors and the momentum basis vectors of the initial quantum state of the target molecular system. The parameterized quantum circuit includes parameters to be optimized, and the parameterized quantum circuit is used to simulate the evolution of the target molecular system in momentum and position spaces; The third processing module is configured to iteratively update the parameters to be optimized and the parameterized quantum circuit within each iteration time step by using the parameterized quantum circuit and the time-dependent Schrödinger equation until the target evolution time is reached, so as to obtain the final quantum state of the target molecular system.

9. A computer storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by one or more processors, the molecular dynamics quantum variational simulation method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that It includes a memory and one or more processors. A computer program is stored on the memory, and when the computer program is executed by the one or more processors, the molecular dynamics quantum variational simulation method according to any one of claims 1 to 7 is implemented.

Images