Hybrid computing task processing method and device and related equipment

Abstract

The invention provides a hybrid computing task processing method and device and related equipment, and relates to the technical field of quantum computing, and the method comprises the steps: carrying out the grouping of Fouli items of a target molecule based on the opinion relationship between Fouli operator strings, and obtaining at least two measurement item groups; packaging each measurement item group into a quantum measurement task; constructing a directed acyclic graph (DAG) on the basis of iterative logic of a variable component child solicitation solver (VQE) algorithm; and in the process of carrying out the classic calculation task, obtaining a quantum measurement task of a next iteration round based on the DAG, and compiling a corresponding quantum circuit. The assembly line work of quantum and classical tasks is realized, and the quantum and classical computing resource utilization rate is improved. The method can be applied to chemical calculation scenes such as drug discovery and material design.

Description

Technical Field

[0001] This application relates to the field of quantum computing technology, and in particular to a hybrid computing task processing method, apparatus and related equipment. Background Technology

[0002] In existing hybrid quantum-classical computing, algorithms such as the variational quantum eigensolver (VQE) are widely used in quantum chemistry to solve for the ground-state energy of the molecular Hamiltonian. In VQE algorithms, the Hamiltonian, describing the total energy of a molecule, is typically decomposed into a linear combination of numerous Pauli operators. The expectation of the Hamiltonian is then calculated by measuring these Pauli operators, and the solution is iteratively obtained using a classical optimizer. However, in current techniques, quantum and classical tasks are strictly sequential, leading to low utilization of both quantum and classical computing resources and high latency. Summary of the Invention

[0003] This application provides a hybrid computing task processing method, apparatus, and related equipment to improve the utilization rate of quantum and classical computing resources.

[0004] In a first aspect, embodiments of this application provide a method for processing hybrid computing tasks, the method comprising:

[0005] The Pauli terms of the target molecule are grouped based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When the measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule.

[0006] Each of the measurement items is grouped and encapsulated into a quantum measurement task, which is used to determine the expected value of all Pauli terms in the corresponding group;

[0007] Based on the iterative logic of the VQE algorithm of the variable quantum eigenvalue solver, a directed acyclic graph (DAG) is constructed. The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters based on the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit.

[0008] During the classical computation task, the quantum measurement task for the next iteration is obtained based on the DAG, and the corresponding quantum circuit is compiled.

[0009] Optionally, the grouping of the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups includes:

[0010] The Hamiltonian corresponding to a single or multiple target molecules is analyzed separately. Based on the commutation relationship between Pauli operator strings in all Pauli terms of each Hamiltonian, the groups are obtained to obtain multiple initial measurement terms corresponding to each target molecule.

[0011] When there are multiple target molecules, the initial measurement terms corresponding to each target molecule are jointly analyzed. Groups containing Pauli terms corresponding to the exact same Pauli operator string in the initial measurement term groups of different target molecules are merged into common measurement term groups. Groups containing unique Pauli terms in the initial measurement term groups of each target molecule are determined as independent measurement term groups.

[0012] The step of grouping and encapsulating each of the measurement items into a quantum measurement task includes:

[0013] The common measurement items are grouped and encapsulated into a first quantum measurement task, and the independent measurement items are grouped and encapsulated into a second quantum measurement task. The measurement results of the first quantum measurement task are reused in the classical computing task.

[0014] Optionally, the analysis of Hamiltonians corresponding to one or more target molecules is performed separately, and the groups are formed according to the commutation relations between Pauli operator strings in all Pauli terms of each Hamiltonian to obtain multiple initial measurement term groups corresponding to each target molecule, including:

[0015] The Hamiltonian of each target molecule is decomposed into a linear combination of multiple Pauli terms, where each Pauli term consists of a string of Pauli operators and corresponding real coefficients.

[0016] Traverse all the Pauli terms obtained from the decomposition, calculate the commutative between any two Pauli terms corresponding to the Pauli operator strings, and determine the Pauli terms corresponding to the two Pauli operator strings whose commutative is zero as mutually commutative Pauli terms.

[0017] Using each Pauli term as a node, establish edge connections between the nodes corresponding to the mutually commutative Pauli terms to obtain a Pauli term relationship graph;

[0018] The Pauli term relationship graph is colored using a graph coloring algorithm. At least one Pauli term corresponding to a node with the same color is assigned to the same group to obtain multiple initial measurement term groups corresponding to the first target molecule. The first target molecule is any one of the target molecules.

[0019] Optionally, the iterative logic of the VQE algorithm based on the variable quantum eigenvalue solver constructs a directed acyclic graph (DAG), including:

[0020] Based on the VQE algorithm flow, a workflow containing multiple iterations and the dependencies between tasks are determined. The workflow is used to indicate that the goal of the quantum measurement task is to determine the expected value under a given parameterized quantum state, and the goal of the classical computation task is to summarize the results of the quantum measurement tasks in the current iteration to calculate the total energy expectation and update the parameters. The dependencies are used to indicate that the current round of classical computation task depends on the completion results of all quantum measurement tasks in the previous round, and the execution of the next round of quantum measurement task depends on the updated parameters of the current round of classical computation task.

[0021] Using the quantum measurement task and the classical computing task as nodes and the dependencies as directed edges, a Directed Acyclic Graph (DAG) is constructed so that the DAG maps the workflow.

[0022] Optionally, during the classical computation task, obtaining the quantum measurement task for the next iteration based on the DAG and compiling the corresponding quantum circuit includes:

[0023] While performing the classical computation task in the current iteration, the DAG is used to parse and determine all quantum measurement tasks to be executed in the next iteration and their corresponding dependencies, thus obtaining the set of quantum measurement tasks for the next iteration.

[0024] Within the time window of the classical computation task in the current iteration round, the quantum circuit corresponding to each quantum measurement task in the set of quantum measurement tasks in the next iteration round is pre-compiled to determine the target quantum circuit, which is used to execute the corresponding quantum measurement task under a given parameterized quantum state;

[0025] The target quantum circuit is associated with the corresponding quantum measurement task in the set of quantum measurement tasks for the next iteration round, and submitted to the task waiting queue. The task waiting queue is used to temporarily store the target quantum circuit and the associated quantum measurement task. After the classical computation task of the current iteration round is completed, the quantum measurement task of the next iteration round is directly scheduled and executed.

[0026] Optionally, the method further includes:

[0027] Periodically obtain calibration reports for quantum hardware, the calibration reports containing at least one of physical qubit fidelity, coherence time, and readout error rate, the quantum hardware being used to perform the quantum measurement task;

[0028] The calibration report is subjected to performance analysis to obtain analysis results, which are used to indicate the real-time performance of the physical qubits planned to be used in the quantum hardware;

[0029] If the real-time performance of the first physical qubit is lower than a preset threshold, the logical qubit originally mapped to the first physical qubit is remapped to the second physical qubit. The first physical qubit is the physical qubit planned to be used in the quantum hardware, and the second physical qubit is a physical qubit in the quantum hardware other than the first physical qubit that meets the quantum circuit connectivity requirements.

[0030] Secondly, embodiments of this application provide a hybrid computing task processing apparatus, the apparatus comprising:

[0031] A grouping module is used to group the Pauli terms of a target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When a measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule.

[0032] An encapsulation module is used to group and encapsulate each of the measurement items into a quantum measurement task, the quantum measurement task being used to determine the expected value of all Pauli terms in the corresponding group;

[0033] The construction module is used to construct a directed acyclic graph (DAG) based on the iterative logic of the variable quantum eigenvalue solver (VQE) algorithm. The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters based on the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit.

[0034] The first acquisition module is used to acquire the quantum measurement task for the next iteration based on the DAG during the classical computing task, and to compile the corresponding quantum circuit.

[0035] Thirdly, embodiments of this application provide an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described in the first aspect.

[0036] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0037] Fifthly, embodiments of this application provide a computer program product including computer instructions that, when executed by a processor, implement the steps of the method described in the first aspect.

[0038] In this embodiment, Pauli terms are first grouped according to the commutation relation of Pauli operator strings, grouping simultaneously measurable Pauli terms together to reduce the number of quantum measurement rounds. Each measurement term group is then encapsulated as an independent quantum measurement task, and parallel measurement of each group improves the measurement efficiency and accuracy of the expected value of the Pauli term. Next, a Directed Acyclic Graph (DAG) is constructed based on VQE iterative logic to clarify the dependency between classical computation tasks and quantum measurement tasks, ensuring orderly and coordinated execution of multi-round iterative tasks. Finally, the time-consuming window of classical computation is used to pre-compile the next round of quantum circuitry, hiding the preparation delay of the quantum measurement task. This achieves pipelined operation of quantum and classical tasks, improving the utilization rate of quantum and classical computing resources. It can be applied to chemical computation scenarios such as drug discovery and materials design. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is one of the flowcharts of a hybrid computing task processing method provided in the embodiments of this application;

[0041] Figure 2 This is a deployment diagram of the scheduling system provided in the embodiments of this application;

[0042] Figure 3 This is a schematic diagram of the task dependency of a single iteration of VQE provided in an embodiment of this application;

[0043] Figure 4 This is a second flowchart of a hybrid computing task processing method provided in the embodiments of this application;

[0044] Figure 5 This is a schematic diagram of Hamiltonian decomposition and Pauli grouping provided in the embodiments of this application;

[0045] Figure 6 This is a schematic diagram of cross-task batch processing provided in an embodiment of this application;

[0046] Figure 7 This is a schematic diagram of dynamic adjustment of quantum hardware provided in an embodiment of this application;

[0047] Figure 8 This is a schematic diagram of the structure of a hybrid computing task processing device provided in an embodiment of this application;

[0048] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0049] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] See Figure 1 , Figure 1 This is one of the flowcharts of a hybrid computing task processing method provided in the embodiments of this application, such as... Figure 1 As shown. The hybrid computing task processing method provided in this application embodiment can be applied to a scheduling system, which may include a preprocessing module, a task planning module, and a dynamic scheduling engine, such as... Figure 2 As shown, the scheduling system is deployed on a classical control server to collaboratively manage computing tasks on quantum processing units (QPUs) and classical processors (central processing units (CPUs) / graphics processing units (GPUs)), thereby achieving efficient execution of the entire hybrid computing workflow.

[0051] The method includes the following steps:

[0052] Step 101: Group the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When a measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule.

[0053] In this step, the preprocessing module receives the Hamiltonian of one or more target molecules. The Hamiltonian is a core physical quantity describing the energy state of the target molecule, and its mathematical form can be a linear combination of multiple Pauli operator strings. For example, the Hamiltonian H of target molecule i is... i It can be represented as:

[0054]

[0055] Among them, c i P are real coefficients. i It is a Pauli operator string.

[0056] The preprocessing module groups the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings. This can be achieved by calculating the commutators of any two Pauli operator strings; if the commutator is zero, the Pauli terms corresponding to the two strings are considered commutative and can be measured in the same round of quantum measurement, thus grouping them into the same measurement term group. Non-commutative Pauli terms cannot be measured precisely simultaneously and are therefore placed in different measurement term groups. The preprocessing module can construct a relation graph and use algorithms such as graph coloring to assign all mutually commutative Pauli terms to the same measurement term group. This results in at least two measurement term groups, the number of which depends on the complexity of the commutation relations between the Pauli operator strings. Furthermore, the Pauli terms within each group can be measured simultaneously, reducing the number of rounds of quantum measurement and improving quantum computing efficiency.

[0057] For example, with any two Pauli operator strings P i and P j For example, P i and P j The commutator being zero can be expressed as:

[0058] .

[0059] [P i ,P j That is, P. i and P j The exchange of children.

[0060] Step 102: Group and encapsulate each measurement item into a quantum measurement task, which is used to determine the expected value of all Pauli terms in the corresponding group.

[0061] In this step, the preprocessing module assigns all mutually commutative Pauli terms to the same measurement term group, resulting in at least two measurement term groups. Then, it encapsulates each measurement term group into a quantum measurement task. Each quantum measurement task corresponds to only one measurement term group, and the Pauli terms within each group are mutually commutative. This reduces the total number of quantum measurement rounds and avoids measurement interference caused by non-commutative Pauli terms from different groups being mixed into the same task. This ensures that the measurement results of each quantum measurement task meet chemical accuracy requirements and improves the accuracy of subsequent total energy calculations.

[0062] Step 103: Based on the iterative logic of the VQE algorithm of the variable quantum eigenvalue solver, construct a directed acyclic graph (DAG). The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters according to the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit.

[0063] In this step, the task planning module abstracts the complete VQE algorithm flow into a workflow. The operations in the workflow can be divided into two categories: one is the quantum measurement task (denoted as Q-Task) generated in step 102, whose goal is to measure quantum states in a given parameterized quantum state. Next, measure the expected value of all Pauli items in the corresponding measurement item group. Another type is the classic computational task (denoted as C-Task). The core task of C-Task is to summarize the results of all Q-Tasks and calculate the expected value of the total energy. And update the parameter θ. The calculation of the total expected energy follows the formula:

[0064]

[0065] The task planning module automatically constructs a Directed Acyclic Graph (DAG) based on the iterative logic of the VQE algorithm. The DAG is used to explicitly define the dependencies between classical computation tasks and quantum measurement tasks. For example, ... Figure 3 As shown, taking the task dependency of a single iteration of VQE as an example, Q-Task1 to Q-TaskN can be N quantum measurement tasks encapsulated by grouping N measurement terms, where N is a positive integer. Each Q-Task is used to obtain the expected value of the corresponding group's inner Pauli term. C-Task (calculating total energy) is a classical computational task used to summarize the expected values ​​of all Q-Task outputs, substitute them into the linear combination formula of the target molecule's Hamiltonian, and calculate the expected value of the total energy. C-Task (updating parameters) is the C-Task used to update the parameters. Based on the expected value of the total energy, it updates the parameters of the parameterized quantum state through classical optimization algorithms such as gradient descent. The updated parameters will be used as the input for the next round of Q-Tasks. This makes the C-Task used to update parameters dependent on the completion of all Q-Tasks in the previous round. In this way, the DAG accurately maps the computational flow of the quantum chemistry algorithm, laying the foundation for subsequent scheduling optimization.

[0066] Step 104: During the classical computation task, obtain the quantum measurement task for the next iteration based on the DAG, and compile the corresponding quantum circuit.

[0067] In this step, during the classical computation tasks, the dynamic scheduling engine utilizes the computation time of the classical optimization step (i.e., C-Task) in the VQE algorithm to hide the preparation delay of the quantum measurement tasks (i.e., Q-Task). For example, based on the constructed DAG, all quantum measurement tasks to be executed in the next iteration are resolved. Simultaneously with the classical computation, the quantum circuits corresponding to these quantum measurement tasks are pre-compiled. These quantum circuits can be the specific execution carriers of the quantum measurement tasks on quantum hardware; the compilation process needs to be adapted to the quantum hardware architecture, which is time-consuming. This allows the circuit preparation work for the next round of quantum measurement tasks to be completed in parallel with the current round of classical computation. After the current round of classical computation is completed, the next round of quantum measurement can be executed directly, avoiding the idle waiting delay of waiting for quantum circuit compilation after classical computation.

[0068] In one embodiment, such as Figure 4 As shown, when a C-Task begins execution in a classical optimizer (such as a CPU), its core mathematical computation is to solve for the energy gradient. And update the parameters using methods such as gradient descent:

[0069]

[0070] The above calculation process is a classic optimization process, which is time-consuming. The scheduling engine can utilize this time window to pre-fetch and analyze all Q-Tasks that the next iteration (round k+1) depends on in the task graph. The dynamic scheduling engine pre-processes the required parameters for these Q-Tasks based on the new parameters. The quantum circuit (or its predicted value) is compiled and submitted to the waiting queue of a quantum processor (e.g., a QPU). In this way, when classical optimization is completed, the subsequent quantum measurement task is ready, realizing pipelined operation of quantum and classical tasks, avoiding the idle waiting delay of waiting for quantum circuit compilation after classical calculation, thereby improving the utilization rate of quantum and classical computing resources.

[0071] In this embodiment, Pauli terms are first grouped according to the commutation relation of Pauli operator strings, grouping simultaneously measurable Pauli terms together to reduce the number of quantum measurement rounds. Each measurement term group is then encapsulated as an independent quantum measurement task, and parallel measurement of each group improves the measurement efficiency and accuracy of the expected value of the Pauli term. Next, a Directed Acyclic Graph (DAG) is constructed based on VQE iterative logic to clarify the dependency between classical computation tasks and quantum measurement tasks, ensuring orderly and coordinated execution of multi-round iterative tasks. Finally, the time-consuming window of classical computation is used to pre-compile the next round of quantum circuitry, hiding the preparation delay of the quantum measurement task. This achieves pipelined operation of quantum and classical tasks, improving the utilization rate of quantum and classical computing resources. It can be applied to chemical computation scenarios such as drug discovery and materials design.

[0072] Optionally, step 101, grouping the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups, includes:

[0073] The Hamiltonian corresponding to a single or multiple target molecules is analyzed separately. Based on the commutation relationship between Pauli operator strings in all Pauli terms of each Hamiltonian, the groups are obtained to obtain multiple initial measurement terms corresponding to each target molecule.

[0074] When there are multiple target molecules, the initial measurement terms corresponding to each target molecule are jointly analyzed. Groups containing Pauli terms corresponding to the exact same Pauli operator string in the initial measurement term groups of different target molecules are merged into common measurement term groups. Groups containing unique Pauli terms in the initial measurement term groups of each target molecule are determined as independent measurement term groups.

[0075] Step 102, grouping and encapsulating each measurement item into a quantum measurement task, includes:

[0076] The common measurement items are grouped and encapsulated into a first quantum measurement task, and the independent measurement items are grouped and encapsulated into a second quantum measurement task. The measurement results of the first quantum measurement task are reused in the classical computing task.

[0077] In this embodiment, the preprocessing module analyzes the Hamiltonians corresponding to single or multiple target molecules based on the commutation relations between Pauli operator strings. It groups the Hamiltonians according to the commutation relations between all Pauli terms, which can be achieved by calculating the commutators of any two Pauli operator strings. If the commutator is zero, the Pauli terms corresponding to the two Pauli operator strings are considered commutative and can be measured in the same round of quantum measurement, thus grouping them into the same initial measurement term group. Non-commutative Pauli terms cannot be measured precisely simultaneously and are therefore placed in different initial measurement term groups. For each target molecule, at least two measurement term groups can be obtained. The number of groups depends on the complexity of the commutation relations between the Pauli operator strings, and the Pauli terms within each group can be measured simultaneously, reducing the number of rounds of quantum measurement and improving quantum computing efficiency. See the following description for details:

[0078] Optionally, the analysis of Hamiltonians corresponding to one or more target molecules is performed separately, and the groups are formed according to the commutation relations between Pauli operator strings in all Pauli terms of each Hamiltonian to obtain multiple initial measurement term groups corresponding to each target molecule, including:

[0079] The Hamiltonian of each target molecule is decomposed into a linear combination of multiple Pauli terms, where each Pauli term consists of a string of Pauli operators and corresponding real coefficients.

[0080] Traverse all the Pauli terms obtained from the decomposition, calculate the commutative between any two Pauli terms corresponding to the Pauli operator strings, and determine the Pauli terms corresponding to the two Pauli operator strings whose commutative is zero as mutually commutative Pauli terms.

[0081] Using each Pauli term as a node, establish edge connections between the nodes corresponding to the mutually commutative Pauli terms to obtain a Pauli term relationship graph;

[0082] The Pauli term relationship graph is colored using a graph coloring algorithm. At least one Pauli term corresponding to a node with the same color is assigned to the same group to obtain multiple initial measurement term groups corresponding to the first target molecule. The first target molecule is any one of the target molecules.

[0083] In this example, taking any target molecule, i.e., the first target molecule, as an example, the Hamiltonian of the first target molecule is first decomposed into a linear combination of multiple Pauli terms. Each Pauli term can include a Pauli operator string and real coefficients. Then, all Pauli terms are traversed, and the commutator of the Pauli operator strings corresponding to any two Pauli terms is calculated. If the commutator is zero, it means that the two Pauli terms are commutative. Each Pauli term is treated as a node in a graph, and edges are connected to the nodes corresponding to mutually commutative Pauli terms, transforming the abstract commutation relationship into a visual graph structure. The graph is then colored using a graph coloring algorithm, and nodes of the same color do not conflict with each other. The Pauli terms corresponding to nodes with the same color are grouped into the same group, finally obtaining multiple initial measurement term groups for the first target molecule. In some optional examples, each initial measurement term group can be encapsulated as a quantum measurement task (Q-Task), for example, as shown in the example. Figure 5 As shown, the Hamiltonian H of the first target molecule can be decomposed into a linear combination of 5 Pauli terms. Each Pauli term consists of a Pauli operator string and a real coefficient c. For example, the Pauli operator string corresponding to P1 is XZII with coefficient c1; the Pauli operator string corresponding to P2 is YIYI with coefficient c2; the Pauli operator string corresponding to P3 is ZZII with coefficient c3; the Pauli operator string corresponding to P4 is IXYZ with coefficient c4; and the Pauli operator string corresponding to P5 is IIZZ with coefficient c5. A commutativity graph is constructed with P1, P2, P3, P4, and P5 as nodes. Then, a graph coloring algorithm is used to assign commutative nodes to the same color, resulting in 3 initial measurement term groups: Group 1: includes P1 (XZII) and P5 (IIZZ); Group 2: includes P2 (YIYI) and P3 (ZZII); Group 3: includes P4 (IXYZ). P4 is in a separate group and is not commutative with other Pauli terms. Then, each group is encapsulated into a quantum measurement task; for example, group 1 is encapsulated as Q-Task1, group 2 as Q-Task2, and group 3 as Q-Task3. This reduces the number of rounds of quantum measurement and improves the efficiency of quantum computing.

[0084] Furthermore, in chemical calculation scenarios commonly used in drug discovery and materials design, such as molecular screening, the system may receive calculation requests for multiple molecules simultaneously. When multiple target molecules exist (e.g., target molecule A and target molecule B), the task planning module and dynamic scheduling engine collaborate to jointly analyze the multiple initial measurement item groups corresponding to each target molecule. Groups containing Pauli terms corresponding to the exact same Pauli operator string from the initial measurement item groups of different target molecules are merged into a common measurement item group. Groups containing unique Pauli terms from the initial measurement item groups of each target molecule are identified as independent measurement item groups. See the following description for details:

[0085] During the joint analysis, the task planning module does not process requests for each molecule independently, but rather considers the Hamiltonian (e.g., H) of all molecules to be computed. A H B (etc.) A joint analysis is performed. By establishing a global Pauli operator string hash table, the task planning module can efficiently identify Pauli operator strings that are completely identical in the Hamiltonian expressions of different molecules, i.e., grouping them into common measurement terms, denoted as... Then, the dynamic scheduling engine reconstructs the original task list based on the identification results. For all tasks, the common... The dynamic scheduling engine will create a unique, shared Q-Task. However, for each molecule's unique Pauli term (such as P...),... A_uniqe P B_uniqe (etc.), the dynamic scheduling engine will create separate Q-Tasks for them, such as Figure 6 As shown. During scheduled execution, the shared Q-Task only needs to be executed once on the quantum processor. Its measurement result (i.e., expected value) It will be stored in a shared cache associated with the Pauli string.

[0086] Thus, when the classical computational task (C-Task) for molecules A and B requires... When calculating the total energy of each component based on its expected value, the results can be read from a shared cache without repeatedly submitting measurement requests to the quantum processor. Through this mechanism of identification, merging, caching, and distribution, multiple independent computational processes are intelligently transformed into a collaborative, partially shared computational graph, thereby significantly reducing the total number of quantum measurements and greatly improving the computational throughput of molecular screening.

[0087] Optionally, step 103, constructing a directed acyclic graph (DAG) based on the iterative logic of the variational quantum eigenfunction solver (VQE) algorithm, includes:

[0088] Based on the VQE algorithm flow, a workflow containing multiple iterations and the dependencies between tasks are determined. The workflow is used to indicate that the goal of the quantum measurement task is to determine the expected value under a given parameterized quantum state, and the goal of the classical computation task is to summarize the results of the quantum measurement tasks in the current iteration to calculate the total energy expectation and update the parameters. The dependencies are used to indicate that the current round of classical computation task depends on the completion results of all quantum measurement tasks in the previous round, and the execution of the next round of quantum measurement task depends on the updated parameters of the current round of classical computation task.

[0089] Using the quantum measurement task and the classical computing task as nodes and the dependencies as directed edges, a Directed Acyclic Graph (DAG) is constructed so that the DAG maps the workflow.

[0090] In this embodiment, the workflow involving multiple iterations and the dependencies between tasks are clearly defined according to the VQE algorithm flow. The goal of the quantum measurement task is to determine the expected value under a given parameterized quantum state, while the goal of the classical computation task is to summarize the results of the quantum measurement tasks in the current iteration to calculate the total expected energy and update the parameters. Furthermore, the current round of classical computation task depends on the completion results of all quantum measurement tasks in the previous round, and the execution of the next round of quantum measurement task depends on the updated parameters of the current round of classical computation task. By using the quantum measurement tasks and classical computation tasks as nodes of a Directed Acyclic Graph (DAG) and the dependencies between tasks as directed edges, the DAG is constructed. This achieves a precise mapping of the DAG to the multi-round iterative workflow of the VQE algorithm, providing a clear dependency basis for subsequent quantum circuit pre-compilation and task scheduling optimization.

[0091] Optionally, step 104, during the classical computation task, involves obtaining the quantum measurement task for the next iteration based on the DAG and compiling the corresponding quantum circuit, including:

[0092] While performing the classical computation task in the current iteration, the DAG is used to parse and determine all quantum measurement tasks to be executed in the next iteration and their corresponding dependencies, thus obtaining the set of quantum measurement tasks for the next iteration.

[0093] Within the time window of the classical computation task in the current iteration round, the quantum circuit corresponding to each quantum measurement task in the set of quantum measurement tasks in the next iteration round is pre-compiled to determine the target quantum circuit, which is used to execute the corresponding quantum measurement task under a given parameterized quantum state;

[0094] The target quantum circuit is associated with the corresponding quantum measurement task in the set of quantum measurement tasks for the next iteration round, and submitted to the task waiting queue. The task waiting queue is used to temporarily store the target quantum circuit and the associated quantum measurement task. After the classical computation task of the current iteration round is completed, the quantum measurement task of the next iteration round is directly scheduled and executed.

[0095] In this embodiment, after constructing the DAG, all quantum measurement tasks and their corresponding dependencies to be executed in the next iteration round are resolved from the DAG, forming the set of Q-Tasks for the next round. Then, by using the time window of the current classical computing task, the quantum circuits (i.e., target quantum circuits, used to execute the corresponding Q-Tasks under a given parameterized quantum state) corresponding to each Q-Task in the set are pre-compiled. Finally, the pre-compiled circuits are associated with the corresponding Q-Tasks and stored in the task waiting queue for temporary storage. Once the current classical computing task is completed, the next round of Q-Tasks can be scheduled and executed directly without waiting for the circuits to be compiled. This hides the preparation delay of quantum tasks and improves the overall iteration efficiency of VQE.

[0096] Optionally, the method further includes:

[0097] Periodically obtain calibration reports for quantum hardware, the calibration reports containing at least one of physical qubit fidelity, coherence time, and readout error rate, the quantum hardware being used to perform the quantum measurement task;

[0098] The calibration report is subjected to performance analysis to obtain analysis results, which are used to indicate the real-time performance of the physical qubits planned to be used in the quantum hardware;

[0099] If the real-time performance of the first physical qubit is lower than a preset threshold, the logical qubit originally mapped to the first physical qubit is remapped to the second physical qubit. The first physical qubit is the physical qubit planned to be used in the quantum hardware, and the second physical qubit is a physical qubit in the quantum hardware other than the first physical qubit that meets the quantum circuit connectivity requirements.

[0100] In this embodiment, the dynamic scheduling engine suppresses quantum hardware noise and ensures the accuracy of VQE computational chemistry by monitoring hardware status and performing bit remapping. During real-time monitoring of the hardware status, the dynamic scheduling engine can periodically request the quantum hardware control system via the Application Programming Interface (API) to obtain an updated calibration report of the quantum hardware. The calibration report includes key performance indicators such as the fidelity of each physical qubit, coherence time (T1, T2), and readout error rate.

[0101] The scheduling engine internally maintains a configurable performance threshold (e.g., single-bit gate fidelity > 99.9%, readout error rate < 1%). Before submitting each Q-Task, the engine performs a performance analysis on the calibration report to obtain the analysis results, which indicate the real-time performance of the physical qubits planned to be used in the quantum hardware. If the performance of one or more bits falls below the preset threshold, the engine initiates a dynamic adjustment strategy. If the engine determines that the performance of some planned physical qubits is poor, it immediately triggers a quantum circuit recompilation process. During recompilation, the scheduling engine instructs the compiler to remap the logical qubits originally planned to be mapped to these "bad" bits (i.e., the first physical qubit, denoted as Q2) to the currently optimal performing "good" bits (i.e., the second physical qubit, denoted as Q3) that meet the circuit connectivity requirements, such as... Figure 7 As shown, this avoids known, localized noise sources at the execution level.

[0102] In some examples, the overall noise level of all available bits may be high, or a better remapping solution may not be found. In such cases, to suppress noise through statistical methods, the scheduling engine adopts another strategy: dynamically increasing the number of measurements (shots) for the Q-Task. By collecting more measurement samples, the statistical average of the Pauli term's expected value can be made closer to its true value, thereby offsetting the impact of random noise to some extent and improving the accuracy of the final energy calculation.

[0103] See Figure 8 , Figure 8 This is a schematic diagram of the structure of a hybrid computing task processing device provided in an embodiment of this application, as shown below. Figure 8 As shown, the hybrid computing task processing device 800 includes:

[0104] Grouping module 801 is used to group the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When the measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule.

[0105] The encapsulation module 802 is used to group and encapsulate each of the measurement items into a quantum measurement task, the quantum measurement task being used to determine the expected value of all Pauli terms in the corresponding group;

[0106] The construction module 803 is used to construct a directed acyclic graph (DAG) based on the iterative logic of the variable quantum eigenvalue solver (VQE) algorithm. The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters based on the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit.

[0107] The first acquisition module 804 is used to acquire the quantum measurement task for the next iteration based on the DAG during the classical computing task and compile the corresponding quantum circuit.

[0108] Optionally, the grouping module 801 is specifically used for:

[0109] The Hamiltonian corresponding to a single or multiple target molecules is analyzed separately. Based on the commutation relationship between Pauli operator strings in all Pauli terms of each Hamiltonian, the groups are obtained to obtain multiple initial measurement terms corresponding to each target molecule.

[0110] When there are multiple target molecules, the initial measurement terms corresponding to each target molecule are jointly analyzed. Groups containing Pauli terms corresponding to the exact same Pauli operator string in the initial measurement term groups of different target molecules are merged into common measurement term groups. Groups containing unique Pauli terms in the initial measurement term groups of each target molecule are determined as independent measurement term groups.

[0111] Package module 802 is specifically used for:

[0112] The common measurement items are grouped and encapsulated into a first quantum measurement task, and the independent measurement items are grouped and encapsulated into a second quantum measurement task. The measurement results of the first quantum measurement task are reused in the classical computing task.

[0113] Optionally, the analysis of Hamiltonians corresponding to one or more target molecules is performed separately, and the groups are formed according to the commutation relations between Pauli operator strings in all Pauli terms of each Hamiltonian to obtain multiple initial measurement term groups corresponding to each target molecule, including:

[0114] The Hamiltonian of each target molecule is decomposed into a linear combination of multiple Pauli terms, where each Pauli term consists of a string of Pauli operators and corresponding real coefficients.

[0115] Traverse all the Pauli terms obtained from the decomposition, calculate the commutative between any two Pauli terms corresponding to the Pauli operator strings, and determine the Pauli terms corresponding to the two Pauli operator strings whose commutative is zero as mutually commutative Pauli terms.

[0116] Using each Pauli term as a node, establish edge connections between the nodes corresponding to the mutually commutative Pauli terms to obtain a Pauli term relationship graph;

[0117] The Pauli term relationship graph is colored using a graph coloring algorithm. At least one Pauli term corresponding to a node with the same color is assigned to the same group to obtain multiple initial measurement term groups corresponding to the first target molecule. The first target molecule is any one of the target molecules.

[0118] Optionally, module 803 is constructed specifically for:

[0119] Based on the VQE algorithm flow, a workflow containing multiple iterations and the dependencies between tasks are determined. The workflow is used to indicate that the goal of the quantum measurement task is to determine the expected value under a given parameterized quantum state, and the goal of the classical computation task is to summarize the results of the quantum measurement tasks in the current iteration to calculate the total energy expectation and update the parameters. The dependencies are used to indicate that the current round of classical computation task depends on the completion results of all quantum measurement tasks in the previous round, and the execution of the next round of quantum measurement task depends on the updated parameters of the current round of classical computation task.

[0120] Using the quantum measurement task and the classical computing task as nodes and the dependencies as directed edges, a Directed Acyclic Graph (DAG) is constructed so that the DAG maps the workflow.

[0121] Optionally, the first acquisition module 804 is specifically used for:

[0122] While performing the classical computation task in the current iteration, the DAG is used to parse and determine all quantum measurement tasks to be executed in the next iteration and their corresponding dependencies, thus obtaining the set of quantum measurement tasks for the next iteration.

[0123] Within the time window of the classical computation task in the current iteration round, the quantum circuit corresponding to each quantum measurement task in the set of quantum measurement tasks in the next iteration round is pre-compiled to determine the target quantum circuit, which is used to execute the corresponding quantum measurement task under a given parameterized quantum state;

[0124] The target quantum circuit is associated with the corresponding quantum measurement task in the set of quantum measurement tasks for the next iteration round, and submitted to the task waiting queue. The task waiting queue is used to temporarily store the target quantum circuit and the associated quantum measurement task. After the classical computation task of the current iteration round is completed, the quantum measurement task of the next iteration round is directly scheduled and executed.

[0125] Optionally, the device further includes:

[0126] The second acquisition module is used to periodically acquire calibration reports of quantum hardware. The calibration reports include at least one of physical qubit fidelity, coherence time, and readout error rate. The quantum hardware is used to perform the quantum measurement task.

[0127] An analysis module is used to perform performance analysis on the calibration report and obtain analysis results, which are used to indicate the real-time performance of the physical qubits planned to be used in the quantum hardware;

[0128] The mapping module is used to remap the logical qubit originally mapped to the first physical qubit to the second physical qubit when the real-time performance of the first physical qubit is lower than a preset threshold. The first physical qubit is the physical qubit planned to be used in the quantum hardware, and the second physical qubit is a physical qubit in the quantum hardware other than the first physical qubit that meets the quantum circuit connectivity requirements.

[0129] The hybrid computing task processing device 800 is capable of implementing the various processes of the above-described embodiments of the hybrid computing task processing method. The technical features correspond one-to-one and can achieve the same technical effects. To avoid repetition, it will not be described again here.

[0130] This application also provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor. When the program is executed by the processor, it implements the various processes of the above-described hybrid computing task processing method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0131] For details, see Figure 9 This application also provides an electronic device, including a bus 901, a transceiver 902, an antenna 903, a bus interface 904, a processor 905, and a memory 906.

[0132] The processor 905 is configured to perform the following steps:

[0133] The Pauli terms of the target molecule are grouped based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When the measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule.

[0134] Each of the measurement items is grouped and encapsulated into a quantum measurement task, which is used to determine the expected value of all Pauli terms in the corresponding group;

[0135] Based on the iterative logic of the VQE algorithm of the variable quantum eigenvalue solver, a directed acyclic graph (DAG) is constructed. The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters based on the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit.

[0136] The transceiver 902 is used to perform the following steps:

[0137] During the classical computation task, the quantum measurement task for the next iteration is obtained based on the DAG, and the corresponding quantum circuit is compiled.

[0138] Optionally, the grouping of the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups includes:

[0139] The Hamiltonian corresponding to a single or multiple target molecules is analyzed separately. Based on the commutation relationship between Pauli operator strings in all Pauli terms of each Hamiltonian, the groups are obtained to obtain multiple initial measurement terms corresponding to each target molecule.

[0140] When there are multiple target molecules, the initial measurement terms corresponding to each target molecule are jointly analyzed. Groups containing Pauli terms corresponding to the exact same Pauli operator string in the initial measurement term groups of different target molecules are merged into common measurement term groups. Groups containing unique Pauli terms in the initial measurement term groups of each target molecule are determined as independent measurement term groups.

[0141] The step of grouping and encapsulating each of the measurement items into a quantum measurement task includes:

[0142] The common measurement items are grouped and encapsulated into a first quantum measurement task, and the independent measurement items are grouped and encapsulated into a second quantum measurement task. The measurement results of the first quantum measurement task are reused in the classical computing task.

[0143] Optionally, the analysis of Hamiltonians corresponding to one or more target molecules is performed separately, and the groups are formed according to the commutation relations between Pauli operator strings in all Pauli terms of each Hamiltonian to obtain multiple initial measurement term groups corresponding to each target molecule, including:

[0144] The Hamiltonian of each target molecule is decomposed into a linear combination of multiple Pauli terms, where each Pauli term consists of a string of Pauli operators and corresponding real coefficients.

[0145] Traverse all the Pauli terms obtained from the decomposition, calculate the commutative between any two Pauli terms corresponding to the Pauli operator strings, and determine the Pauli terms corresponding to the two Pauli operator strings whose commutative is zero as mutually commutative Pauli terms.

[0146] Using each Pauli term as a node, establish edge connections between the nodes corresponding to the mutually commutative Pauli terms to obtain a Pauli term relationship graph;

[0147] The Pauli term relationship graph is colored using a graph coloring algorithm. At least one Pauli term corresponding to a node with the same color is assigned to the same group to obtain multiple initial measurement term groups corresponding to the first target molecule. The first target molecule is any one of the target molecules.

[0148] Optionally, the iterative logic of the VQE algorithm based on the variable quantum eigenvalue solver constructs a directed acyclic graph (DAG), including:

[0149] Based on the VQE algorithm flow, a workflow containing multiple iterations and the dependencies between tasks are determined. The workflow is used to indicate that the goal of the quantum measurement task is to determine the expected value under a given parameterized quantum state, and the goal of the classical computation task is to summarize the results of the quantum measurement tasks in the current iteration to calculate the total energy expectation and update the parameters. The dependencies are used to indicate that the current round of classical computation task depends on the completion results of all quantum measurement tasks in the previous round, and the execution of the next round of quantum measurement task depends on the updated parameters of the current round of classical computation task.

[0150] Using the quantum measurement task and the classical computing task as nodes and the dependencies as directed edges, a Directed Acyclic Graph (DAG) is constructed so that the DAG maps the workflow.

[0151] Optionally, during the classical computation task, obtaining the quantum measurement task for the next iteration based on the DAG and compiling the corresponding quantum circuit includes:

[0152] While performing the classical computation task in the current iteration, the DAG is used to parse and determine all quantum measurement tasks to be executed in the next iteration and their corresponding dependencies, thus obtaining the set of quantum measurement tasks for the next iteration.

[0153] Within the time window of the classical computation task in the current iteration round, the quantum circuit corresponding to each quantum measurement task in the set of quantum measurement tasks in the next iteration round is pre-compiled to determine the target quantum circuit, which is used to execute the corresponding quantum measurement task under a given parameterized quantum state;

[0154] The target quantum circuit is associated with the corresponding quantum measurement task in the set of quantum measurement tasks for the next iteration round, and submitted to the task waiting queue. The task waiting queue is used to temporarily store the target quantum circuit and the associated quantum measurement task. After the classical computation task of the current iteration round is completed, the quantum measurement task of the next iteration round is directly scheduled and executed.

[0155] Optionally, the transceiver 902 is further configured to perform the following steps:

[0156] Periodically obtain calibration reports for quantum hardware, the calibration reports containing at least one of physical qubit fidelity, coherence time, and readout error rate, the quantum hardware being used to perform the quantum measurement task;

[0157] The processor 905 is also configured to perform the following steps:

[0158] The calibration report is subjected to performance analysis to obtain analysis results, which are used to indicate the real-time performance of the physical qubits planned to be used in the quantum hardware;

[0159] If the real-time performance of the first physical qubit is lower than a preset threshold, the logical qubit originally mapped to the first physical qubit is remapped to the second physical qubit. The first physical qubit is the physical qubit planned to be used in the quantum hardware, and the second physical qubit is a physical qubit in the quantum hardware other than the first physical qubit that meets the quantum circuit connectivity requirements.

[0160] exist Figure 9In this document, a bus architecture (represented by bus 901) is used. Bus 901 can include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 905 and memory represented by memory 906. Bus 901 can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 904 provides an interface between bus 901 and transceiver 902. Transceiver 902 can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 905 is transmitted over a wireless medium via antenna 903, which further receives data and transmits it to processor 905.

[0161] Processor 905 manages bus 901 and general processing, and also provides various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 906 can be used to store data used by processor 905 during operation.

[0162] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, this computer program implements the various processes of the above-described hybrid computing task processing method embodiments and achieves the same technical effects. To avoid repetition, it will not be described again here. The computer-readable storage medium may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0163] This application also provides a computer program product, including computer instructions. When executed by a processor, these computer instructions implement the various processes of the above-described hybrid computing task processing method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0164] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0165] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0166] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A hybrid computing task processing method, characterized in that, The method includes: The Pauli terms of the target molecule are grouped based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When the measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule. Each of the measurement items is grouped and encapsulated into a quantum measurement task, which is used to determine the expected value of all Pauli terms in the corresponding group; Based on the iterative logic of the VQE algorithm of the variable quantum eigenvalue solver, a directed acyclic graph (DAG) is constructed. The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters based on the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit. During the classical computation task, the quantum measurement task for the next iteration is obtained based on the DAG, and the corresponding quantum circuit is compiled.

2. The method according to claim 1, characterized in that, The method of grouping the Pauli terms of the target molecule based on the commutation relations between Pauli operator strings yields at least two measurement term groups, including: The Hamiltonian corresponding to a single or multiple target molecules is analyzed separately. Based on the commutation relationship between Pauli operator strings in all Pauli terms of each Hamiltonian, the groups are obtained to obtain multiple initial measurement terms corresponding to each target molecule. When there are multiple target molecules, the initial measurement terms corresponding to each target molecule are jointly analyzed. Groups containing Pauli terms corresponding to the exact same Pauli operator string in the initial measurement term groups of different target molecules are merged into common measurement term groups. Groups containing unique Pauli terms in the initial measurement term groups of each target molecule are determined as independent measurement term groups. The step of grouping and encapsulating each of the measurement items into a quantum measurement task includes: The common measurement items are grouped and encapsulated into a first quantum measurement task, and the independent measurement items are grouped and encapsulated into a second quantum measurement task. The measurement results of the first quantum measurement task are reused in the classical computing task.

3. The method according to claim 2, characterized in that, The analysis of Hamiltonians corresponding to one or more target molecules is performed separately. Based on the commutation relations between Pauli operator strings in all Pauli terms of each Hamiltonian, groups are formed to obtain multiple initial measurement term groups corresponding to each target molecule, including: The Hamiltonian of each target molecule is decomposed into a linear combination of multiple Pauli terms, where each Pauli term consists of a string of Pauli operators and corresponding real coefficients. Traverse all the Pauli terms obtained from the decomposition, calculate the commutative between any two Pauli terms corresponding to the Pauli operator strings, and determine the Pauli terms corresponding to the two Pauli operator strings whose commutative is zero as mutually commutative Pauli terms. Using each Pauli term as a node, establish edge connections between the nodes corresponding to the mutually commutative Pauli terms to obtain a Pauli term relationship graph; The Pauli term relationship graph is colored using a graph coloring algorithm. At least one Pauli term corresponding to a node with the same color is assigned to the same group to obtain multiple initial measurement term groups corresponding to the first target molecule. The first target molecule is any one of the target molecules.

4. The method according to claim 1, characterized in that, The iterative logic of the VQE algorithm based on the variational quantum eigenvalue solver constructs a directed acyclic graph (DAG), including: Based on the VQE algorithm flow, a workflow containing multiple iterations and the dependencies between tasks are determined. The workflow is used to indicate that the goal of the quantum measurement task is to determine the expected value under a given parameterized quantum state, and the goal of the classical computation task is to summarize the results of the quantum measurement tasks in the current iteration to calculate the total energy expectation and update the parameters. The dependencies are used to indicate that the current round of classical computation task depends on the completion results of all quantum measurement tasks in the previous round, and the execution of the next round of quantum measurement task depends on the updated parameters of the current round of classical computation task. Using the quantum measurement task and the classical computing task as nodes and the dependencies as directed edges, a Directed Acyclic Graph (DAG) is constructed so that the DAG maps the workflow.

5. The method according to claim 1, characterized in that, During the classical computation task, obtaining the quantum measurement task for the next iteration based on the DAG and compiling the corresponding quantum circuit includes: While performing the classical computation task in the current iteration, the DAG is used to parse and determine all quantum measurement tasks to be executed in the next iteration and their corresponding dependencies, thus obtaining the set of quantum measurement tasks for the next iteration. Within the time window of the classical computation task in the current iteration round, the quantum circuit corresponding to each quantum measurement task in the set of quantum measurement tasks in the next iteration round is pre-compiled to determine the target quantum circuit, which is used to execute the corresponding quantum measurement task under a given parameterized quantum state; The target quantum circuit is associated with the corresponding quantum measurement task in the set of quantum measurement tasks for the next iteration round, and submitted to the task waiting queue. The task waiting queue is used to temporarily store the target quantum circuit and the associated quantum measurement task. After the classical computation task of the current iteration round is completed, the quantum measurement task of the next iteration round is directly scheduled and executed.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Periodically obtain calibration reports for quantum hardware, the calibration reports containing at least one of physical qubit fidelity, coherence time, and readout error rate, the quantum hardware being used to perform the quantum measurement task; The calibration report is subjected to performance analysis to obtain analysis results, which are used to indicate the real-time performance of the physical qubits planned to be used in the quantum hardware; If the real-time performance of the first physical qubit is lower than a preset threshold, the logical qubit originally mapped to the first physical qubit is remapped to the second physical qubit. The first physical qubit is the physical qubit planned to be used in the quantum hardware, and the second physical qubit is a physical qubit in the quantum hardware other than the first physical qubit that meets the quantum circuit connectivity requirements.

7. A hybrid computing task processing device, characterized in that, The device includes: A grouping module is used to group the Pauli terms of a target molecule based on the commutation relations between Pauli operator strings to obtain at least two measurement term groups. Each measurement term group includes at least one Pauli term. When a measurement term group includes multiple Pauli terms, the commutator of any two Pauli terms in the same measurement term group is zero. The Pauli terms of the target molecule are determined according to the Hamiltonian of the target molecule. An encapsulation module is used to group and encapsulate each of the measurement items into a quantum measurement task, the quantum measurement task being used to determine the expected value of all Pauli terms in the corresponding group; The construction module is used to construct a directed acyclic graph (DAG) based on the iterative logic of the variable quantum eigenvalue solver (VQE) algorithm. The DAG is used to clarify the dependency between the classical computing task and the quantum measurement task. The classical computing task is used to determine the total energy expectation and update the parameters based on the results of each set of quantum measurement tasks. The parameters are used to determine the state of the qubit. The first acquisition module is used to acquire the quantum measurement task for the next iteration based on the DAG during the classical computing task, and to compile the corresponding quantum circuit.

8. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes computer instructions that, when executed by a processor, implement the method as described in any one of claims 1 to 6.

Images