Quantum computing method and device based on quantum state cooperative manipulation and computer equipment

Abstract

This invention belongs to the field of quantum computing and relates to a quantum computing method, device, and computer equipment based on quantum state collaborative manipulation. The method includes: constructing and initializing a data foundation; decomposing an external computing task into multiple sub-tasks using a classical intelligent decision-making model, and dynamically allocating the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task; building a hierarchical asymmetric entangled topology and constructing an entangled resource pool to achieve dynamic scheduling of idle resources; performing dynamic dimensional basis vector transformation through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation; performing real-time monitoring and hybrid verification and error correction using a classical dual-model collaboration; outputting the computation results encoded in the quantum state and completing the sedimentation of the entire process of computation data, forming a closed loop of computation-data-optimization-re-computation. It enables end-to-end integrated collaboration, dynamic and efficient utilization of resources, and full-process data closed-loop optimization.

Description

Technical Field

[0001] This invention relates to the field of quantum computing technology, and in particular to quantum computing methods, devices, and computer equipment based on the cooperative manipulation of quantum states. Background Technology

[0002] Quantum computing, relying on the core quantum mechanical properties of quantum superposition and quantum entanglement, has shown parallel computing potential that classical computers cannot match in fields such as large number factorization, cryptography decryption, new material simulation, and quantum chemical calculation. It is the core development direction of the next generation of information processing technology. Building an ultra-high efficiency quantum computing power foundation is the core key to realizing the leap of quantum computing from laboratory prototypes to industrial-scale applications.

[0003] Existing solutions all focus on localized optimizations of a single aspect of quantum computing, failing to address core issues such as the separation of quantum and classical layers, the disconnect between quantum and classical collaboration, and the inefficient management of massive quantum data. They have not yet formed an integrated computing power infrastructure architecture and have several technical bottlenecks and gaps. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a quantum computing method based on quantum state cooperative manipulation, employing the following technical solution, including the following steps: Construct a data base and initialize the data base; The classical quantum intelligent decision-making model, based on a 6-layer residual network trained with 28-dimensional quantum operation features, decomposes the external computing task into multiple sub-tasks and dynamically allocates the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task. The multi-dimensional encoding scheme uses 32 core parameter templates from 12 types of quantum hardware platforms and is adapted across platforms by using dedicated spectrum analysis and hardware feature intelligent recognition equipment. The 12 types of quantum hardware platforms include superconducting, ion trap, photonic, and neutral atom platforms. The 32 core parameter templates include the energy level relaxation rate, spin-flip probability, orbital angular momentum diffusion coefficient, coherence time, coupling strength, impulse response frequency, phase adjustment range, qubit topological connection constraints, gate operation error threshold, readout fidelity, temperature drift coefficient, and magnetic field noise tolerance parameters of the quantum hardware platform. Based on the computational power weights of subtasks, a hierarchical asymmetric entanglement topology is constructed, and an entanglement resource pool is built to achieve dynamic scheduling of idle resources. Dynamic dimensional basis transformation is achieved through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation; Real-time monitoring is performed using a quantum layer, and hybrid verification and error correction are performed using a classical dual-model approach. The output is encoded in the quantum state to calculate the results and complete the full accumulation of computational data, forming a closed loop of computation-data-optimization-re-computation.

[0005] To address the aforementioned technical problems, the present invention also provides a quantum computing device based on quantum state cooperative manipulation, employing the following technical solution, including: An initialization module is used to construct the data base and initialize the data base; The decomposition module is used to decompose external computing tasks into multiple sub-tasks using a classical quantum intelligent decision-making model. This model is based on a 6-layer residual network trained with 28-dimensional quantum operation features. It dynamically allocates the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task. The multi-dimensional encoding scheme uses 32 core parameter templates from 12 types of quantum hardware platforms and is adapted across platforms using dedicated spectrum analysis and hardware feature intelligent recognition devices. The 12 types of quantum hardware platforms include superconducting, ion trap, photonic, and neutral atom platforms. The 32 core parameter templates include the energy level relaxation rate, spin-flip probability, orbital angular momentum diffusion coefficient, coherence time, coupling strength, impulse response frequency, phase adjustment range, qubit topological connection constraints, gate operation error threshold, readout fidelity, temperature drift coefficient, and magnetic field noise tolerance parameters of the quantum hardware platform. The module is used to build a hierarchical asymmetric entanglement topology based on the computing power weight of subtasks, and to build an entanglement resource pool to realize the dynamic scheduling of idle resources. The conversion module is used to perform dynamic dimensional basis vector conversion through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation; The verification module is used for real-time monitoring using a quantum layer and hybrid verification and error correction using a classical dual-model approach. The output module is used to output the computation results encoded in the quantum state and to complete the full accumulation of computation data throughout the entire process, forming a digital intelligent iterative closed loop of computation-data-optimization-re-computation.

[0006] To address the aforementioned technical problems, the present invention also provides a computer device that employs the technical solution described below, comprising a memory and a processor. The memory stores computer-readable instructions, and the processor executes the computer-readable instructions to implement the steps of the aforementioned quantum computing method based on quantum state cooperative manipulation.

[0007] To address the aforementioned technical problems, the present invention also provides a computer-readable storage medium, which employs the technical solution described below. The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the aforementioned quantum computing method based on quantum state cooperative manipulation.

[0008] Compared with the prior art, the present invention has the following main advantages: (1) End-to-end integrated collaboration: Breaking through the limitations of traditional quantum-classical separation design, the classical intelligent decision-making model and the quantum layer real-time monitoring and dual-model collaborative error correction are used to achieve deep integration of task decomposition, resource allocation and error correction, eliminating the bottleneck of collaborative disconnection and significantly improving the overall computing efficiency and stability.

[0009] (2) Resources can be used dynamically and efficiently: By introducing a hierarchical asymmetric entanglement topology and entanglement resource pool, idle entanglement resources can be dynamically scheduled according to the subtask computing power weight, which solves the problem of inefficient management of massive quantum data, avoids resource waste and local overload, and provides a flexible and scalable computing power base for large-scale parallel computing.

[0010] (3) Full-process data closed-loop optimization is possible: It not only completes the output of calculation results, but also realizes the complete accumulation and feedback of calculation data, forming a digital intelligent iterative closed loop of calculation-data-optimization-re-computation. This enables the system to have continuous self-evolution capability, continuously optimize the coding scheme, resource allocation and error correction strategy, and accelerate the leap from prototype to industrial application.

[0011] (4) Strong cross-platform compatibility: It has 12 built-in quantum hardware platform parameter templates and is equipped with dedicated spectrum analysis and hardware feature intelligent recognition equipment. It can quickly adapt to various types of hardware such as superconductors, ion traps, photonic quantum, and neutral atoms, reducing the cross-platform code reconstruction rate from 70% to below 10%.

[0012] (5) High resource utilization: The hierarchical asymmetric entanglement topology is adopted. High-weight subtasks are configured with high-density entangled links, and low-weight subtasks are matched with sparse links. The overall quantum resource utilization rate is increased to more than 90%, and the information carrying capacity of a single quantum bit is increased by 2-3 orders of magnitude compared with the traditional single-dimensional scheme. Attached Figure Description

[0013] To more clearly illustrate the solutions in this invention, the accompanying drawings used in the description of the embodiments of this invention will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0014] Figure 1 This is a flowchart of an embodiment of the quantum computing method based on quantum state cooperative manipulation of the present invention; Figure 2 This is an exemplary architecture diagram illustrating the application of the method of the present invention; Figure 3 This is a schematic diagram of the entire closed-loop process of the method of the present invention; Figure 4 This is a schematic diagram of the structure of an embodiment of the quantum computing device based on quantum state cooperative manipulation of the present invention; Figure 5 This is a schematic diagram of another embodiment of the quantum computing device based on quantum state cooperative manipulation of the present invention; Figure 6 yes Figure 5 Implementation case diagram; Figure 7 This is a schematic diagram of the structure of an embodiment of the computer device of the present invention. Detailed Implementation

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains; the terminology used herein in the specification is for the purpose of describing particular embodiments only and is not intended to limit the invention; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings are used to distinguish different objects and not to describe a particular order.

[0016] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0017] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0018] It should be noted that the quantum computing method based on quantum state cooperative manipulation provided in the embodiments of the present invention is generally executed by a server / terminal device, and correspondingly, the quantum computing device based on quantum state cooperative manipulation is generally set in the server / terminal device.

[0019] It should be understood that the number of terminal devices, networks, and servers is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be used.

[0020] Example 1 Please refer to Figure 1The diagram illustrates a flowchart of an embodiment of the quantum computing method based on quantum state cooperative manipulation according to the present invention. The quantum computing method based on quantum state cooperative manipulation includes the following steps: Step S1: Construct the data base and initialize it.

[0021] In this embodiment, the electronic device (e.g., a server / terminal device) running on the quantum computing method based on quantum state cooperative manipulation can receive quantum computing requests based on quantum state cooperative manipulation via wired or wireless connections. It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra-wideband) connections, and other currently known or future-developed wireless connection methods.

[0022] In this embodiment, step S1 may specifically include the following steps: S11, Create and configure multiple data partitions for Ceph distributed storage.

[0023] Hardware platform: Deploy at least 3 Ceph storage nodes, each node is configured with 4 10TB NVMe SSD solid-state drives, 16-core CPU, 64GB memory, and the nodes are interconnected via 10 Gigabit Ethernet.

[0024] Storage mode: A hybrid mode of Ceph RGW object storage and CephFS file storage is adopted.

[0025] Five data partitioning strategies: Quantum state parameter library: Stores real-time parameters of qubits (energy level relaxation rate, spin-flip probability, orbital angular momentum diffusion coefficient, etc.). Employs Ceph RGW object storage with a 3-replica strategy for permanent data retention.

[0026] Model training library: Stores training data, model parameters, and iteration history of the intelligent decision-making model. It also uses CephRGW object storage with a 3-replica strategy for permanent retention.

[0027] The manipulation and computation database stores task breakdown results, computing power allocation schemes, entangled topology parameters, intermediate computation results, etc. It employs CephFS file storage with a 2-replica strategy for permanent retention.

[0028] Cross-platform adaptation library: Stores 32 core parameter templates for 12 types of hardware platforms, including superconductors, ion traps, and quantum photonics. Uses CephFS file storage with a 2-copy strategy for permanent retention.

[0029] Error correction log repository: Stores quantum state fluctuation data, error types, error correction commands and effects, etc. It uses CephFS file storage with a 2-replica strategy, and data retention for 1-3 years.

[0030] Configuration command example (based on Ceph CLI): ceph osd pool create quantum_params_pool 128 128; ceph osd pool set quantum_params_pool size 3; radosgw-admin bucket create --bucket=quantum-params-bucket.

[0031] Performance metrics: Ceph RGW read / write speed ≥ 1 GB / s, CephFS read / write speed ≥ 800 MB / s, network latency ≤ 0.5 ms.

[0032] The purpose of step S11 is to complete the creation and strategy configuration of the five major data partitions of Ceph distributed storage, realize the classification and hierarchical storage management of quantum computing data across the entire chain, and provide high-throughput and high-reliability data read and write services for subsequent steps.

[0033] S12: Load the quantum computing intelligent decision-making model, read the latest model parameters from the Ceph model training library, and complete the inference engine initialization.

[0034] Hardware platform: Kunpeng 920 server (48 cores, 192GB memory), equipped with a dedicated AI acceleration card (such as Ascend 910), and connected to the quantum layer via a 100Gbps high-speed fiber optic cable.

[0035] Model architecture: A 6-layer residual network (ResNet) based on 28-dimensional feature input, with hidden layer dimensions of 256, 128, 64, 32, 16, and 8, and the output layer is a multi-task head (task decomposition, computational weights, encoding scheme, etc.).

[0036] Feature composition (28 dimensions): Quantum state parameters: energy level relaxation rate (1-dimensional) Spin-flip probability (1-dimensional) orbital angular momentum diffusion coefficient It has 12 dimensions in total, including 1-dimensional quantum bit health (1-dimensional) and coherence time T2 (1-dimensional).

[0037] Task characteristics: 8 dimensions including task type encoding (one-hot, 8 categories), data scale (logarithmic transformation), real-time requirement level, etc.

[0038] Resource status: 8 dimensions including available qubits, entangled link idle rate, and dimensional availability.

[0039] Initialization process: Read the latest model weight file (.h5 or .onnx format) from the Ceph model training library.

[0040] Load the model using ONNX Runtime or TensorRT to optimize the inference engine.

[0041] Performing a single dummy inference operation to verify successful model loading takes ≤10ns.

[0042] The model enters hot standby mode, with inference latency ≤1μs.

[0043] The residual network based on 28-dimensional features can capture the complex nonlinear relationship between the dynamic changes of quantum states and task requirements. Compared with the traditional greedy algorithm, the task decomposition accuracy is improved from 65%~70% to over 95%.

[0044] The purpose of step S12 is to provide intelligent decision-making capabilities for subsequent task decomposition, computing power allocation, and dimensional transformation prediction by loading the classical quantum intelligent decision-making model.

[0045] S13 sends a detection command to the multi-dimensional quantum state storage module, collects the initial parameters of the quantum state, verifies whether the health of the quantum layer is ≥99%, and ensures that the computing power base enters the standby mode in a reliable state.

[0046] The hardware components include: a cross-shaped aluminum-based Josephson junction superconducting quantum bit chip (5×5μm). 2 It is divided into three independent control regions: energy level, spin, and orbital angular momentum.

[0047] The three-dimensional parameter sensors include: energy level sensors: SQUID-based structure, measuring energy level relaxation rate. Spin sensor: Micro magneto-optical trap structure, measuring spin-flip probability. Orbital angular momentum sensor: holographic grating interferometer structure, used to measure orbital angular momentum diffusion coefficient. .

[0048] The parameters collected include: Energy level relaxation rate (unit: μs) -1 ), reflecting the quantum bit from |1 The state decays to |0 The velocity of the state; Spin flip probability (dimensionless, 0~1), representing the probability of the spin direction changing randomly per unit time; : Orbital angular momentum diffusion coefficient (unit: rad² / s), which describes the random walk diffusion rate of the orbital angular momentum quantum number.

[0049] The formula for calculating health is: .

[0050] in: Quantum layer health level, with a value ranging from 0 to 1 or 0% to 100%. The closer the value is to 1, the healthier the quantum layer is. Energy level relaxation rate (unit: μs) -1 ), indicating that the quantum bit changes from the excited state |1 Decay to ground state | 0 The rate; Maximum allowable value of energy level relaxation rate (unit: μs) -1 If the value exceeds this, the qubit is considered unusable; Spin flip probability (dimensionless, 0~1), representing the probability that the spin direction will randomly flip per unit time; Orbital angular momentum diffusion coefficient (unit: rad) 2 / s), describing the random walk diffusion rate of the orbital angular momentum quantum number; Measurement time (unit: ns), fixed at 10 ns; : Natural exponential function, used to calculate the exponential decay factor; The product of the mass factors of all qubits. For the first The mass factor of a quantum bit (dimensionless, usually determined by the coherence time) is... Determined by parameters, when When the value is 1, otherwise it decays linearly.

[0051] The verification logic is as follows: if H ≥ 99%, then send a Ready signal to the classical end; otherwise, mark the unhealthy qubit and remove it from the entanglement resource pool.

[0052] Data writing: The initial parameters collected are written to the Ceph quantum state parameter library in real time with a delay of ≤1ns.

[0053] Ensure the quantum layer operates in a healthy state, avoid computational errors caused by hardware defects, and achieve a system failure-free operation rate of ≥99.9%.

[0054] The purpose of step S13 is to perform initial state detection and verification of the quantum layer.

[0055] The purpose of step S1 is to establish a unified data storage architecture for the quantum computing system, load the classical edge intelligent decision-making model, detect and verify the initial health state of the quantum layer, and put the computing power base into a ready-to-operate state. This step is the data and model foundation for all subsequent steps, ensuring that the data partitioning is reasonable, the model is ready, and the quantum layer is usable.

[0056] Step S2 involves using a classic edge intelligent decision-making model to break down the external computing task into multiple sub-tasks and dynamically allocating the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task.

[0057] In this embodiment, step S2 may specifically include the following steps: S21 receives external computing tasks from the classic control terminal and extracts the characteristics of task type, scale, complexity, and real-time requirements.

[0058] Task type identification: Predefined task types: Large number factorization (Shor's algorithm), quantum chemical simulation (VQE), new material simulation (DFT), quantum machine learning (QNN), cryptography breaking (Grover), optimization problem (QAOA), etc., totaling 8 categories.

[0059] A pattern matching algorithm is used to compare the task description string with the feature library, with a processing time of ≤10ns.

[0060] Scale and complexity quantification: Large number factorization: number of bits N bits (e.g., 1024).

[0061] Quantum chemical simulation: number of molecular orbitals N orbitals Number of electrons N electrons .

[0062] Complexity levels: low (1~10), medium (11~100), high (>100) qubit requirements.

[0063] Real-time requirements are divided into three levels: real-time (<1ms), near real-time (1~10ms), and batch processing (>10ms).

[0064] Data output: The feature vector (8-dimensional) is sent to the quantum intelligent decision-making module.

[0065] The purpose of step S21 is to receive external computing tasks (such as 1024-bit large number factorization, quantum chemical simulation) from the classical control terminal, extract core features such as task type, scale, complexity, and real-time requirements, and provide input for subsequent decomposition.

[0066] S22, the quantum computing intelligent decision-making model, decomposes the computing task into multiple sub-tasks that can be parallelized or serialized based on historical data, and calculates the computational density, resource requirements and priority of each sub-task.

[0067] Decomposition Algorithm: Sequence-to-Sequence (Seq2Seq) model based on attention mechanism, input task feature sequence, output subtask sequence.

[0068] The formula for operational density is: .

[0069] in: : No. The computational density of each subtask (dimensionless), the larger the value, the more intensive the demand for quantum computing resources of the subtask; : No. The total number of quantum gate operations (in terms of times) required for each subtask is estimated through historical statistics or theoretical methods. : Assigned to the The time window (unit: ns or μs) of each subtask is estimated by the classic client based on the task's real-time requirements; A base-2 logarithmic function used for compressed mapping of the number of qubits; : No. Number of qubits required for each subtask (unit: qubits); : The number of reference qubits (unit: qubits), which is a fixed value of 10 for normalization.

[0070] Resource requirements include the number of qubits, entanglement depth, and dimensionality preference.

[0071] Priority: Calculate the critical path based on the task dependency graph. Priority = 1 / (latest start time - earliest start time + 1).

[0072] Historical data utilization: Read historical breakdown records of similar tasks from the Ceph manipulation and computation database, and initialize the model output by k-nearest neighbor (k=5) weighted average.

[0073] The computation density formula can quantify the pressure of subtasks on quantum resources, guide the weight allocation of subsequent asymmetric entanglement topologies, and avoid resource idleness or overload.

[0074] The purpose of step S22 is to decompose complex tasks into sub-tasks using a quantum computing intelligent decision-making model.

[0075] S23, the quantum computing intelligent decision-making model, combines the real-time state of the quantum layer to dynamically allocate the number of qubits, computing power weights, and multi-dimensional encoding schemes for each subtask.

[0076] Quantum bit allocation: A constrained linear programming approach is used, with the objective of minimizing the total number of quantum bits while satisfying the Q-value of each subtask. i Lower limit.

[0077] The formula for computing power weight is: .

[0078] in: : The computational weight of the i-th subtask (0~1), the higher the weight, the more entanglement resources are allocated and the faster the update cycle.

[0079] The Sigmoid function maps the input to (0,1).

[0080] Subtask computation density.

[0081] : The maximum computational density among all subtasks.

[0082] The amount of quantum resources currently in use (such as the number of qubits already occupied, the number of entangled links).

[0083] Total sub-resource quantity.

[0084] Subtask importance coefficient (derived from priority, 0.5~1.5).

[0085] Multi-dimensional encoding scheme selection: The optimal scheme is selected from 5 predefined combinations (energy level and spin, energy level and orbital angular momentum, spin and orbital angular momentum, energy level and spin and orbital angular momentum, and single dimension as a backup), based on the availability score of each dimension (read in real time from the quantum state parameter library).

[0086] Response time: Total time for model inference and resource allocation ≤ 300ns, resource matching degree ≥ 90%.

[0087] The computing power weighting formula dynamically balances task demand, resource availability, and importance, preventing low-weight tasks from monopolizing high-density entangled resources.

[0088] The purpose of step S23 is to achieve the optimal match between task requirements and quantum resources by dynamically allocating quantum resources.

[0089] S24 sends computing power configuration instructions and multi-dimensional encoding instructions to the quantum control module, and simultaneously writes the task decomposition results, computing power allocation scheme and encoding scheme into the Ceph control and operation database in real time.

[0090] Instruction format: A fixed-length binary protocol is adopted. The first 4 bytes are the instruction type (0x01 computing power configuration, 0x02 encoding configuration), followed by the subtask ID, quantum bit index list, weight value, and encoding scheme ID.

[0091] Command delivery channel: via 100Gbps high-speed fiber optic cable, using RDMA (Remote Direct Memory Access) technology, with a command delivery latency of ≤500ns.

[0092] Database writes: Using CephFS's asynchronous IO interface, data write latency is ≤50ns, and automatic 2-replica synchronization is triggered.

[0093] High-speed command issuance ensures that the quantum layer can respond promptly to the intelligent decisions of the classical layer, avoiding control latency from becoming a bottleneck.

[0094] The purpose of step S24 is to ensure data traceability through instruction issuance and data writing.

[0095] The purpose of step S2 is to connect classical tasks and quantum execution through intelligent task decomposition and computing power configuration, which is the key to improving the efficiency of quantum-classical collaboration.

[0096] Step S3: Based on the computing power weight of the subtasks, construct a hierarchical asymmetric entanglement topology and build an entanglement resource pool to achieve dynamic scheduling of idle resources.

[0097] In this embodiment, step S3 may specifically include the following steps: S31, the quantum manipulation module receives the computing power configuration instructions from the classical end, reads the computing power weight of each subtask from the Ceph manipulation and computation database, and converts it into entangled topology parameters.

[0098] Weighting rules: High weight ( High-density entangled links, update cycle ns, pulse power dBm, configured with 3:1 redundant links (one spare link for every three working links).

[0099] Medium weight ( Medium-density entangled link, update cycle ns, pulse power dBm, no redundancy.

[0100] low weight ( ): Sparse entangled links, update cycle s, pulse power dBm.

[0101] The conversion function is: .

[0102] in, The update period of the entangled link (unit: ns or μs) represents the time interval for refreshing and maintaining the entangled link. : Baseline update period (unit: μs), fixed value 1 μs; : Natural exponential function; : Attenuation coefficient (dimensionless), with a fixed value of 2.3, used to adjust the intensity of the influence of the weight on the update cycle; : No. The computational weights of each subtask (same as above) range from 0 to 1. This ensures... hour ns.

[0103] Redundancy coefficient: That is, the number of redundant links is 3 when the weight is high.

[0104] when hour, (Sparse update); when hour, (High-density updates). This exponential decay relationship ensures that high-weight tasks receive extremely short update cycles, thus prioritizing computing resources.

[0105] Asymmetric allocation avoids entanglement updates to all qubits at the same frequency, minimizing the entanglement overhead of idle resources (low-weighted subtasks).

[0106] The purpose of step S31 is to convert the weights into entangled topology parameters (such as link density, update cycle, pulse power and redundancy configuration).

[0107] S32 uses a multi-channel phase controller to synchronously adjust the coupling pulses of each qubit, builds the corresponding entangled links, constructs an entangled resource pool, reserves 5% to 10% of the links as backup, and verifies the coupling strength and synchronization of all links.

[0108] Hardware: 8-channel phase controller, each channel can independently adjust the phase from 0 to 360° with an accuracy of ±0.1°; programmable quantum pulse generator output frequency from 1 to 10 GHz.

[0109] Entangled link construction: For high-weighted sub-tasks, cross-resonance gates are used to achieve entanglement between superconducting qubits, with a pulse amplitude of Ω = 2π × 20 MHz and a duration of 30 ns.

[0110] For medium to low weights, use weaker microwave pulses (Ω=2π×10MHz) or directly use existing entanglement.

[0111] Entangled resource pool: 5% to 10% of the total number of sub-bits are allocated as backup links (which do not participate in the main operation).

[0112] The links within the pool remain in a low-power standby state (pulse power 1dBm) and can be activated at any time.

[0113] Verification method: Measuring Bell state fidelity: Perform quantum state tomography to calculate the fidelity F with the ideal Bell state.

[0114] All working links must have an F ≥ 99.5% accuracy; otherwise, recalibrate.

[0115] Data writing: Write the entangled topology (adjacency matrix) into the Ceph manipulation and computation database.

[0116] The entangled resource pool enables dynamic scheduling of idle resources. When a subtask completes ahead of schedule or fails, it can quickly switch to a backup link to avoid overall computational interruption.

[0117] The purpose of step S32 is to establish a hierarchical entangled link and resource pool.

[0118] S33 synchronizes the state data of idle links and available qubit resources in the entanglement resource pool to the Ceph quantum state parameter library.

[0119] Data format: JSON format, containing the ID of each quantum bit, the current entangled link status (idle / busy / faulty), and the dimension availability score.

[0120] Synchronization frequency: Synchronization is triggered once every time an entanglement topology adjustment is completed (≤200ns).

[0121] Synchronization latency: ≤30ns, achieved through Ceph's asynchronous write and memory-mapped file technology.

[0122] Real-time resource status is a key input for predictive dimensionality transformation and dynamic entanglement optimization in classic edge models.

[0123] The purpose of step S33 is to provide real-time data for predicting and dynamically optimizing subsequent dimensional transformations on the classic client through resource status synchronization.

[0124] The purpose of step S3 is to build a hierarchical asymmetric entanglement topology (high / medium / low density links) based on the computing power weight of the subtasks, and to construct an entanglement resource pool to realize the dynamic scheduling of idle resources, thereby increasing the utilization rate of quantum resources to over 90%.

[0125] Step S4 involves performing dynamic dimensional basis transformation through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation.

[0126] In this embodiment, step S4 may specifically include the following steps: S41 generates composite quantum pulses and single pulses according to the encoding instructions of the classical end, and simultaneously realizes the synchronous operation of multiple quantum states.

[0127] Composite Pulse Design: For energy level and spin two-dimensional coding, design a pulse that simultaneously modulates the microwave frequency (for energy level transitions) and the magnetic field gradient (for spin flipping).

[0128] The formula for the Gaussian envelope of a pulse is: .

[0129] in: Envelope amplitude of a composite quantum pulse (in MHz) as a function of time. change; Peak amplitude of the pulse (unit: MHz), ranging from 10 to 100 MHz; : Natural exponential function, used to construct Gaussian envelope; Time variable (unit: ns); Pulse center time (unit: ns), i.e., the time when the pulse amplitude reaches its peak. The moment; : Pulse width parameter (unit: ns), controls the broadening of the Gaussian envelope, typical value is about 5 ns; The cosine function represents the carrier oscillation of a pulse. : Driving angular frequency (unit: rad / ns), usually close to the energy level transition frequency of a quantum bit; The initial phase of the pulse (in radians) is used for fine adjustments such as spin manipulation.

[0130] This formula describes a Gaussian-modulated cosine pulse that can simultaneously excite energy level transitions (via...). ) and spin flip (via phase) (In conjunction with a magnetic field), it enables synchronous control of multiple dimensions with a single pulse.

[0131] Synchronization mechanism: Utilizing the multi-channel output capability of the programmable quantum pulse generator, synchronous trigger signals are sent to both the energy level control region and the spin control region simultaneously, with time jitter <1ps.

[0132] Operation time: Single-step operation (such as CNOT gate equivalent operation) takes ≤20ns, while existing single-dimensional operation requires 100ns.

[0133] Composite pulses combine multi-dimensional operations into a single step, fundamentally reducing the number of quantum gates. For example, in a 16-qubit QFT, the number of operation steps is reduced from O(N²) = 256 to approximately O(N) = 16.

[0134] The purpose of step S41 is to generate a composite quantum pulse according to the encoding instructions of the classical end. A single pulse can simultaneously realize the synchronous operation of multiple quantum states (such as the simultaneous execution of energy level flip and spin flip), without the need for additional quantum gates, thus greatly reducing the operation steps and time.

[0135] S42 acquires quantum state parameters in real time and writes them into the Ceph quantum state parameter library.

[0136] Sensor operating modes: SQUID (Spirit Level Sensor): Detects quantum states from |1 to |1| by measuring changes in magnetic flux. To |0 The transition, the output signal amplitude is proportional to .

[0137] Spin sensor (micro magneto-optical trap): Utilizes the Faraday rotation effect to measure spin polarization and calculates it through photon counting. .

[0138] Orbital angular momentum sensor (holographic grating interferometry): inferring from changes in the contrast of interference fringes .

[0139] Sampling rate: 1GHz, meaning data is collected every 1ns with no data loss (using high-speed ADC and FPGA buffer).

[0140] Data writing: Data is transferred from the FPGA to the Ceph storage node using DMA (Direct Memory Access), with a latency of ≤1ns.

[0141] High-frequency real-time parameter acquisition is a prerequisite for classic edge models to perform dimensional availability scoring and predictive dimensional transformation, ensuring the timeliness of dynamic adjustments.

[0142] The function of step S42 is to acquire quantum state parameters (energy level relaxation rate) in real time at a sampling rate of 1 GHz. Spin flip probability Orbital angular momentum diffusion coefficient (and written into the Ceph quantum state parameter library to provide real-time feedback for classical end-models).

[0143] S43, the classical quantum intelligent decision-making model reads quantum state parameters in real time from the Ceph quantum state parameter library, calculates the availability score of each dimension, and predicts the decoherence trend of the dimension based on historical data.

[0144] The availability rating formula is: .

[0145] in: Dimension The availability score (0~100 points) indicates that the dimension is more suitable for the current computing task; Dimension The relaxation rate of the energy level (in μs) -1 This reflects the rate of energy decay in that dimension; Maximum allowable relaxation rate (in μs) -1 ), take a fixed value of 0.1 μs -1 ; Dimension The spin-flip probability under the given condition (dimensionless, 0~1). Dimension Orbital angular momentum diffusion coefficient (unit: rad² / s); : Expected quantum operation time (in ns), taken as a fixed value of 10 ns; : Natural exponential function, used to calculate the exponential decay caused by decoherence; Dimension The coherence time (in μs) represents the length of time that a quantum state remains coherent. Reference coherence time (unit: μs), taken as a fixed value of 30 μs.

[0146] Decoherence trend prediction: Using a Long Short-Term Memory (LSTM) network to analyze the past 1000 time points. Sequence modeling to predict the next 10 ns If the predicted value will drop below the threshold of 50 points within the next 5ns, the pre-conversion strategy will be triggered.

[0147] The predictive mechanism avoids computational errors caused by dimensional decoherence, improving system stability by 30% compared to the passive response approach.

[0148] The purpose of step S43 is to achieve a prediction accuracy of ≥92% through dimensional availability scoring and phase regression intervention.

[0149] S44, when a dimension score is about to drop to the dimension score threshold, the classical quantum intelligent decision-making model retrieves the health dimension combination from multiple predefined dimension combinations and sends a dimension conversion command to the quantum manipulation module. The quantum layer then uses the natural evolution characteristics of quantum states to complete the dimension conversion.

[0150] Predefined dimension combination library: Combination 1: energy level and spin; Combination 2: energy level and orbital angular momentum; Combination 3: spin and orbital angular momentum; Combination 4: energy level, spin and orbital angular momentum; Combination 5: single dimension (specified).

[0151] Natural evolutionary transformation mechanisms: Utilizing the Zeeman effect: When the applied magnetic field changes, the energy level and spin state will undergo coupled evolution. By precisely controlling the magnetic field ramp time (3ns), the quantum state can smoothly transition from pure energy level encoding to a mixed encoding of energy level and spin.

[0152] Transformation path: Hamiltonian from Become ,in: Initial Hamiltonian (unit: joule), describing the energy of a quantum bit without an external magnetic field; Reduced Planck constant (unit: J·s) ,in It is Planck's constant; The energy level transition angular frequency of a quantum bit (unit: rad / s), corresponding to |0 Ω·cm|. State and |1 Energy difference between states; Pauli Z-matrix , used to describe the quantum state of an energy level; : Total Hamiltonian after applying a time-varying magnetic field (unit: joules); Bohr magneton (unit: J / T), a physical constant, approximately equal to J / T; The magnetic field strength (unit: Tesla, T) that varies with time is a control parameter for dimension transformation; Pauli X matrix , is used to describe spin flips or transitions between energy levels.

[0153] When an external magnetic field is applied When the Hamiltonian gradually increases from 0, it starts from... (Pure energy level encoding) Smooth transition to including The form of the term allows the quantum state to evolve naturally between energy levels and spin dimensions, thereby achieving dimension transformation without additional quantum gates.

[0154] Response time: ≤3ns from the issuance of the conversion command to the completion of the conversion, and conversion error ≤0.5% (compensated by pulse shaping).

[0155] The dimension conversion without additional quantum gates avoids the 0.8% to 1.2% error introduced by the additional quantum gates required in existing technologies, while compressing the conversion time from the 100ns level to 3ns.

[0156] The purpose of step S44 is to achieve dynamic dimensional transformation without the need for additional quantum gates.

[0157] S45, the classical end model, reads the resource pool status from the Ceph quantum state parameter library based on the real-time changes of quantum state parameters and the computation progress, and dynamically optimizes the link parameters of the entangled topology.

[0158] Optimization goal: Maximize resource utilization This also satisfies the computational density requirements of each subtask.

[0159] Adjust strategy: If a subtask lags behind in its computation progress, its weight is temporarily increased (by increasing the update frequency and pulse power).

[0160] If an increase in the error rate of a link is detected, the system will automatically switch to a backup link in the entanglement resource pool.

[0161] Control algorithm: Proportional-Integral-Derivative (PID) controller; input is the deviation between the real-time calculated progress and the target progress; output is the weight increment. .

[0162] Response time: ≤50ns from reading data to sending adjustment command.

[0163] Dynamic optimization maintains quantum resource utilization at over 90% (compared to <60% in traditional schemes), reducing the demand for qubits by 70% for the same task.

[0164] The purpose of step S45 is to dynamically optimize the link parameters (such as update period, pulse power, etc.) of the entangled topology, and optimize the response time to ≤50ns.

[0165] The purpose of step S4 is: in the core quantum computing power output stage, through multi-dimensional quantum state combination encoding, composite quantum pulse synchronous control, and dynamic dimensional basis vector conversion, the efficiency of quantum computing power is greatly improved, and the single-step operation time is ≤20ns.

[0166] Step S5: Real-time monitoring is performed using a quantum layer, and hybrid verification and error correction are performed using a classical dual-model approach.

[0167] In this embodiment, step S5 may specifically include the following steps: S51, real-time monitoring of lightweight auxiliary qubits.

[0168] Auxiliary qubit: Aluminum-based superconducting qubit, T2≥30μs, coupling strength adjustable from 1 to 10MHz.

[0169] Monitoring scheme: Connect the auxiliary bit to all operational bits through an adjustable coupler, periodically execute the CNOT gate (once every 10ns), and measure the phase shift of the auxiliary bit.

[0170] Threshold setting: An early warning is triggered when the phase deviation is ≥5° or the amplitude deviation is ≥10%.

[0171] Warning delay: ≤1ns, output directly via hardware comparator.

[0172] Data writing: Writes fluctuation data to the Ceph error correction log library in real time.

[0173] Compared to pure quantum error correction which requires more than 20 auxiliary qubits, this scheme only requires 1, reducing resource consumption by 75%.

[0174] The function of step S51 is to sample the operational quantum state at a frequency of 100MHz using a lightweight auxiliary quantum bit to monitor quantum state fluctuations in real time; when the fluctuation exceeds a preset threshold, an early warning signal is immediately sent to the FPGA processor and the classical terminal.

[0175] The S52 FPGA processor calls a lightweight neural network to perform preliminary identification of fluctuating data; the classical quantum intelligent decision-making model combines historical data to complete error identification through the collaboration of the two models.

[0176] Lightweight Neural Networks (FPGA): Architecture: 3-layer fully connected network (64 nodes in the input layer, 32 nodes in the hidden layer, and 10 error categories in the output layer).

[0177] Input: Fluctuation data (phase, amplitude) for the past 10 time points.

[0178] Output: Error probability vector.

[0179] Inference time: ≤2ns (hardware parallelization).

[0180] Classical quantum intelligent decision-making model: Read real-time fluctuation data and historical error data (past 1000 records) from the Ceph error correction log library.

[0181] The timing features are extracted using a Transformer encoder, and the final judgment is made in combination with the FPGA output.

[0182] Unknown error clustering: The DBSCAN algorithm is used to cluster the error feature space. Samples that are far from all known clusters are labeled as "unknown type" and given a category label.

[0183] Identification time: ≤5ns (FPGA and classic terminal in parallel).

[0184] The dual-model approach combines the advantages of low latency in hardware with the high precision of classical large-scale models, and its ability to predict unknown errors is something that pure quantum error correction and a single classical model cannot achieve.

[0185] The function of step S52 is as follows: the FPGA processor calls the lightweight neural network to perform preliminary identification of the fluctuation data; the classical quantum intelligent decision-making model combines historical data to complete error identification through dual-model collaboration, which can identify more than 98% of known errors and achieve more than 85% clustering prediction for new unknown errors.

[0186] The S53 classical terminal model generates a personalized error correction strategy based on the error type and historical error correction data. It outputs error correction pulses through an error correction pulse generator, and the quantum core control module executes the error correction operation to restore the quantum state to a healthy range.

[0187] Error correction strategy library: Pre-stores error correction pulse sequences for 20 common errors (phase flip, bit flip, amplitude damping, etc.).

[0188] Personalized adjustment: For unknown errors, Bayesian optimization is used to search for the optimal pulse parameters (amplitude, width, phase).

[0189] Command issuance: Error correction commands are sent to the error correction pulse generator via a 10Gbps bus with a delay of ≤2ns.

[0190] Pulse output: Error correction pulse generator response time ≤ 5ns, output frequency 1~20GHz, amplitude accuracy ±0.01dB.

[0191] Error correction operation: The quantum core control module executes pulses to project the quantum state back to the ideal state, with an error correction success rate of ≥99%.

[0192] Personalized error correction strategies are more adaptable than fixed-pattern error correction, and are especially suitable for high-speed dynamic scenarios such as quantum machine learning.

[0193] The purpose of step S53 is to generate and execute an error correction strategy.

[0194] S54 writes error type, error correction command, and error correction effect data throughout the entire process to the Ceph error correction log library and model training library in real time.

[0195] Data format: Structured log, including timestamp, qubit ID, error type (encoding), error correction pulse parameters, and post-correction fidelity.

[0196] Write latency: ≤5ns, optimized using Ceph's asynchronous batch write.

[0197] Iterative triggering: Every 1000 new data points accumulated, an incremental learning task is triggered in the model training library to update the error recognition weights of the classic end model.

[0198] This forms a closed loop of error correction, data processing, and optimization, improving the model's error recognition accuracy by 0.5% to 1% with each iteration.

[0199] The purpose of step S54 is to provide data support for subsequent iterative optimization of the model through error correction data writing and iteration.

[0200] The purpose of step S5 is to adopt a hybrid error correction method that combines real-time monitoring of the quantum layer with collaboration between classical end dual models, throughout the entire operation, to achieve low-latency and high-accuracy error correction of quantum errors, with an error identification accuracy of ≥98% and an error correction success rate of ≥99%.

[0201] Step S6: Output the computation results encoded in the quantum state and complete the full accumulation of computation data throughout the entire process, forming a closed loop of digital intelligence iteration of computation-data-optimization-re-computation.

[0202] In this embodiment, step S6 may specifically include the following steps: S61 decodes the computation results encoded in the quantum state into classical data format and outputs them to the target terminal through a cross-platform adapter interface.

[0203] Cross-platform adaptation interface: It has built-in parameter templates for 12 mainstream quantum hardware platforms, and is equipped with dedicated spectrum analysis and intelligent hardware feature recognition equipment. It can quickly adapt to various types of hardware such as superconductors, ion traps, photonic quantum, and neutral atoms, reducing the cross-platform code reconstruction rate from 70% to below 10%.

[0204] Decoding method: Perform inverse quantum Fourier transform (IQFT) or direct projection measurement to map multidimensional quantum states to classical bit strings.

[0205] Decoding error: ≤0.3%, noise is suppressed by taking the mode of multiple measurements.

[0206] Output latency: ≤50ns, directly forwarded via cross-platform adaptation interface.

[0207] Data writing: Synchronously writes the calculation results to the Ceph control and computation database.

[0208] Low-latency output meets the needs of industrial-grade real-time computing (such as cryptography breaking).

[0209] The purpose of step S61 is to decode and output the calculation results.

[0210] S62, the classic client model, organizes all the data from the entire computation process and categorizes it for writing to various Ceph data partitions.

[0211] Data processing: Apache Arrow format columnar storage is used with a compression ratio of 5:1.

[0212] Write strategy: Concurrent writes to five data partitions, each with an independent thread.

[0213] Total time: ≤30ns, thanks to Ceph's high throughput and NVMe SSD.

[0214] The data from the entire process provides a complete traceability for subsequent model iterations and system fault analysis.

[0215] The purpose of step S62 is to organize and write the data for the entire process (such as task breakdown, computing power configuration, entangled topology, dimension transformation, error correction operations, and calculation results).

[0216] S63, the classic end-user model, uses the current computation data and historical data to iteratively optimize the quantum computing-specific intelligent decision-making model.

[0217] Optimization algorithm: Incremental learning (online gradient descent), processing 1000 new samples in each iteration.

[0218] Optimized latency: ≤20ns, using GPU acceleration (Kunpeng 920 + Ascend 910).

[0219] Accuracy improvement: 0.5%~1% improvement with each iteration, infinitely approaching 100%.

[0220] Model parameter update: Write the optimized model parameters into the Ceph model training library and notify all inference nodes to hot-load.

[0221] Continuous iteration enables the system to adapt to new types of quantum computing tasks and hardware aging drift.

[0222] The purpose of step S63 is to iteratively optimize the intelligent decision-making model.

[0223] S64, the quantum manipulation module sends resource reclamation commands to each module of the quantum layer, releases all qubits and entangled resources, restores the initial state of the quantum layer, Ceph completes persistent data backup, and the computing power base is reset.

[0224] Resource recycling: Reset all qubits to |0 using a pulse sequence state.

[0225] Shut down the coupling pulses of all entangled links and disconnect the auxiliary bit connections.

[0226] Time taken ≤20ns.

[0227] Persistent backup: Ceph performs a snapshot and asynchronously replicates it to a remote disaster recovery node.

[0228] Reset signal: Sends "Ready" to the classic end, waiting for the next task.

[0229] Ensure that the computing power base can be reused an unlimited number of times without residual interference.

[0230] The purpose of step S64 is to perform resource recovery and reset.

[0231] The purpose of step S6 is to achieve efficient output of calculation results and complete the finalization of all calculation data.

[0232] Figure 2 This is an exemplary architectural diagram illustrating the application of the method of the present invention. For example... Figure 2 As shown, this embodiment overcomes the shortcomings of traditional quantum computing, which separates the quantum layer, classical layer, and data layer, and relies on localized optimization. It adopts a three-layer integrated closed-loop architecture: a quantum layer computing core, a classical layer intelligent control system, and a Ceph distributed data platform. This three-layer architecture achieves bidirectional real-time data interaction and command transmission via a 10Gbps / 100Gbps high-speed bus. Without any technology stacking, it realizes end-to-end collaboration in computing power output, intelligent control, and data iteration, constructing an industrial-grade quantum computing power platform. The overall architecture is divided into three levels: a quantum layer core module, a classical layer support module, and a data layer storage module. Each level's sub-modules have clear division of labor and deep interoperability.

[0233] Figure 3 This is a schematic diagram of the entire closed-loop process of the method of this invention. For example... Figure 3 As shown, a closed loop is formed through base visualization, task decomposition and computing power configuration, entangled topology construction, multi-dimensional collaborative computing, hybrid verification and error correction, and output and data iteration.

[0234] This invention can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This invention can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0235] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0236] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0237] Example 2 Further reference Figure 2 As a response to the above Figure 1 The present invention provides an embodiment of a quantum computing device based on quantum state cooperative manipulation, which is similar to the method shown. Figure 1 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0238] like Figure 2 As shown, the quantum computing device 70 based on quantum state cooperative manipulation described in this embodiment includes: an initialization module 71, a disassembly module 72, a construction module 73, a conversion module 74, a verification module 75, and an output module 76. Wherein: Initialization module 71 is used to construct the data base and initialize the data base; The decomposition module 72 is used to decompose external computing tasks into multiple sub-tasks through a classical edge intelligent decision-making model, and dynamically allocate the number of qubits, computing power weight and multi-dimensional encoding scheme to each sub-task. Module 73 is used to build a hierarchical asymmetric entanglement topology based on the computing power weight of subtasks, and to build an entanglement resource pool to realize the dynamic scheduling of idle resources. The conversion module 74 is used to perform dynamic dimension basis conversion through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation; Verification module 75 is used for real-time monitoring using a quantum layer and hybrid verification and error correction using a classical dual-model approach. Output module 76 is used to output the computation results encoded in the quantum state and complete the full accumulation of computation data throughout the entire process, forming a digital intelligent iterative closed loop of computation-data-optimization-re-computation.

[0239] Example 3 Figure 5 This is a schematic diagram of another embodiment of the quantum computing device based on quantum state cooperative manipulation of the present invention. Figure 5 As shown, the device in this embodiment includes: 1. Quantum Layer Core Module (Computing Power Output Layer).

[0240] The computing power output layer, as the core of the system's computing power, is responsible for quantum state encoding, manipulation, entanglement construction, and real-time error correction. It is the carrier of quantum computing power output and includes three main sub-modules: Multi-dimensional quantum state storage module: Equipped with a cross-shaped aluminum-based Josephson junction superconducting quantum bit chip, it divides the energy level, spin, and orbital angular momentum into three independent controllable dimensions, and supports five types of multi-dimensional combination encoding; It is equipped with a 1GHz ultra-high precision three-dimensional parameter sensor to collect the core parameters of the quantum state in real time, providing accurate hardware data support for upper-level decision-making. The information carrying capacity of a single quantum bit is increased by 2-3 orders of magnitude compared with the traditional single-dimensional scheme.

[0241] The quantum core control module consists of a programmable quantum pulse generator, an 8-channel phase controller, and a real-time timing regulator. It can generate composite quantum pulses, realize multi-dimensional quantum state synchronous control, dynamic construction of entangled topology and parameter optimization. The single-step quantum operation takes ≤20ns, which greatly reduces the computational latency.

[0242] Quantum-classical hybrid error correction unit: It is equipped with only one lightweight superconducting auxiliary qubit, paired with an FPGA processor and a dedicated error correction pulse generator. It monitors quantum state fluctuations in real time throughout the entire operation process, achieving low latency and high accuracy hybrid error correction. Compared with traditional pure quantum error correction schemes, the auxiliary qubit resource consumption is reduced by 75%.

[0243] 2. Classic layer support module (intelligent control layer).

[0244] The intelligent control layer, as the system's decision-making center, undertakes external computing tasks, performs intelligent decomposition and computing power scheduling, and links the quantum layer and data layer. It comprises three main sub-modules: Classical quantum intelligent decision-making module: Based on the Kunpeng 920 server, a dedicated intelligent decision-making model is deployed. The model is trained with 28-dimensional quantum computing features and adopts a 6-layer residual network structure. It can realize intelligent task decomposition, computing power weight allocation, dimensional dephase intervention judgment, dynamic optimization of entanglement topology, and error correction strategy generation. The task decomposition accuracy is ≥95% and the inference latency is ≤1μs.

[0245] Cross-platform adaptation interface: Built-in parameter templates for 12 mainstream quantum hardware platforms. When paired with dedicated spectrum analysis and intelligent hardware feature recognition equipment, it can quickly adapt to various types of hardware such as superconductors, ion traps, and photonic quantum, reducing the cross-platform code reconstruction rate from 70% to below 10% and solving the problem of poor quantum hardware compatibility.

[0246] Intelligent Iteration Unit: Relying on full-process computational data, it completes incremental iterative optimization of the decision-making model, continuously improves the accuracy of system decision-making, error correction, and resource scheduling, and realizes system self-evolution.

[0247] 3. Data layer storage module (data base layer).

[0248] Based on the Ceph distributed storage architecture, this system adopts a hybrid mode of Ceph RGW object storage and CephFS file storage, dividing the data into five data partitions and employing differentiated replication and data retention strategies to solve the problems of slow read / write speeds, poor scalability, and single points of failure inherent in traditional centralized storage. Quantum state parameter library (object storage, 3 copies, permanent retention): stores real-time quantum state parameters and hardware state data; Control and computation database (file storage, 2 copies, permanent retention): stores task breakdown, computing power configuration, and intermediate computation data; Error correction log repository (file storage, 2 copies, retention period 1-3 years): stores quantum error data and error correction records; Cross-platform adaptation library (file storage, 2 copies, permanent retention): stores various quantum hardware parameter templates; Model training library (object storage, 3 copies, permanent retention): Stores decision model training data and iteration parameters.

[0249] The module has an overall read / write speed of ≥1GB / s and supports PB / EB-level linear expansion, providing complete data support for the entire system's computation, model iteration, and fault tracing.

[0250] Figure 6 yes Figure 5 Implementation case illustration. (Example) Figure 6 As shown, the working process of this device is as follows: Following a closed-loop workflow of base initialization → task decomposition and computing power configuration → entangled topology construction → multi-dimensional collaborative computing → hybrid verification and error correction → result output and data iteration, the six steps are interconnected and linked throughout the process. The time consumption of all processes has been experimentally verified and can be directly implemented, forming a digital intelligent iterative closed loop of "computation-data-optimization-recomputation".

[0251] Data base initialization steps: Total time ≤ 100ns. Complete system hardware and software initialization to put the computing base into a ready-to-operate state. First, deploy the Ceph distributed storage cluster and complete the strategy configuration for the five data partitions; then, load the quantum-specific intelligent decision-making model on the classical server and read historical training parameters to complete the inference engine initialization; finally, the quantum layer sensors collect initial quantum state parameters, verify that the quantum layer hardware health is ≥ 99%, and the system is ready to operate after verification.

[0252] Task intelligent decomposition and computing power configuration steps: total time ≤ 500ns. The classical end receives external computing tasks such as large number factorization and quantum chemical simulation, and extracts core features such as task type, complexity, and real-time requirements; the intelligent decision model combines historical computing data to decompose complex tasks into 10-20 sub-tasks, quantizing the computing density and priority of each sub-task; combined with the real-time hardware status of the quantum layer, it dynamically allocates the number of qubits, computing power weights (0-1), and multi-dimensional encoding schemes, and finally sends down control commands, while simultaneously storing task data in a distributed database, achieving optimal matching between task requirements and quantum resources, with a task decomposition accuracy of ≥ 95%.

[0253] Asymmetric entanglement topology construction steps: Total time ≤ 200ns. Abandoning the traditional symmetric fully connected entanglement layout, a hierarchical asymmetric entanglement construction is achieved based on subtask computational power weights: high-weight subtasks are configured with high-density, high-frequency, and redundant entangled links, while medium- and low-weight subtasks are matched with sparse links; simultaneously, a 5%-10% entanglement reserve resource pool is constructed, idle quantum resources are dynamically scheduled, and after verifying that the fidelity of all entangled links meets the standards, the quantum resource state data is updated synchronously, increasing the overall quantum resource utilization rate to over 90%.

[0254] Multi-dimensional collaborative computing steps: Core computing power efficiency is improved by 80 times compared to existing technologies. In the core computing power output stage of the system, through multi-dimensional quantum state combination encoding, synchronous manipulation of composite quantum pulses, and dynamic dimensional basis vector transformation, the efficiency of quantum computing power is greatly improved, with a single-step operation time of ≤20ns. Furthermore, the quantum layer generates composite quantum pulses according to classical end instructions, and a single pulse synchronously completes multi-dimensional quantum state manipulation of energy level, spin, and orbital angular momentum without the need for additional quantum gates; sensors collect quantum state parameters at a high frequency of 1GHz, and the classical end model calculates the availability of each dimension and predicts decoherence trends in real time; when the dimensional performance decays to the threshold, the natural evolution characteristics of quantum states are used to complete a 3ns ultra-fast dimensional transformation, while dynamically optimizing entanglement topology parameters to continuously ensure computing power output efficiency.

[0255] Quantum-classical hybrid error correction step: This step is implemented throughout the entire computation process, with a single error correction cycle time ≤10ns. It employs a hybrid error correction approach of "real-time monitoring at the quantum layer + dual-model collaboration at the classical end," achieving low-latency, high-accuracy error correction of quantum errors. The error identification accuracy is ≥98%, and the error correction success rate is ≥99%. Lightweight auxiliary qubits monitor quantum state fluctuations in real time, triggering an early warning immediately when thresholds are exceeded. A lightweight FPGA neural network performs initial error identification, forming a dual-model collaborative judgment with the classical quantum decision model. This accurately identifies known quantum errors and clusters to predict novel unknown errors. Personalized error correction pulses are generated based on the error type to quickly repair the quantum state, and all error correction data is stored in a database to support model iteration.

[0256] Results output and data accumulation steps: total time ≤ 100ns. After the operation is completed, a digital intelligent iterative closed loop of "operation-data-optimization-reoperation" is formed. It supports the quantum layer to decode multi-dimensional quantum state encoded data into classical data and output it externally; the classical layer uniformly organizes the data of the entire operation process and classifies and accumulates it into various data partitions of Ceph; based on the newly added operation data, the intelligent decision model is incrementally iterated, slightly improving the model's decision accuracy; finally, the quantum layer reclaims all quantum resources, the computing power base is reset, and it awaits the next round of computing tasks, completing the complete closed loop.

[0257] Example 4 As a response to the above Figure 1 The present invention provides another embodiment of a quantum computing device based on quantum state cooperative manipulation, which is similar to the method shown. Figure 1 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0258] The quantum computing device based on quantum state cooperative manipulation described in this embodiment includes: The integrated architecture consists of a quantum layer core module and a classical layer support module, which achieve real-time data interaction and command issuance through a 10Gbps / 100Gbps high-speed bus.

[0259] The quantum layer core module is the core of quantum computing power output of the computing power base. It is responsible for multi-dimensional quantum state encoding, storage, manipulation and real-time error correction. It includes three sub-modules: multi-dimensional quantum state storage module, quantum core manipulation module and quantum-classical hybrid verification unit, realizing a triple improvement in quantum bit information carrying capacity, manipulation speed and error correction efficiency.

[0260] The multi-dimensional quantum state storage module is used to realize the storage of multi-dimensional quantum states and real-time parameter acquisition, providing accurate quantum state data for quantum manipulation and classical decision-making; Hardware components: a cross-shaped aluminum-based Josephson junction superconducting quantum bit chip (5×5μm²) and a three-dimensional parameter sensor; the chip is divided into three independent control regions: energy level, spin, and orbital angular momentum. The three-dimensional parameter sensor includes a SQUID-based structure energy level sensor, a micro magneto-optical trap structure spin sensor, and a holographic grating interference structure orbital angular momentum sensor. It supports five types of multi-dimensional combination encoding, including energy levels and spins, and the information carrying capacity of a single quantum bit is increased by 2-3 orders of magnitude compared with single-dimensional encoding; the sampling rate is 1 GHz, the measurement accuracy is ±0.1%, and the parameter acquisition delay is ≤1 ns.

[0261] The quantum core control module is used to receive classical terminal commands, generate composite quantum pulses, and realize the synchronous control of multi-dimensional quantum states and the dynamic adjustment of entanglement topology. Hardware components: programmable quantum pulse generator, 8-channel phase controller and real-time timing regulator, with modules connected by 50Ω low-noise coaxial cable; Key parameters: signal transmission delay 3±0.5ns, single-step operation time ≤20ns; programmable quantum pulse generator output frequency 1-10GHz, frequency resolution 0.1Hz; 8-channel phase controller phase adjustment range 0-360°, accuracy ±0.1°; real-time timing regulator supports 1000 pre-stored instructions, execution interval ≤10ns.

[0262] The quantum-classical hybrid verification unit is used to realize real-time monitoring of quantum states and low-latency, high-accuracy quantum error correction, and is the core of the stability guarantee for the computing power base. Hardware components: one lightweight aluminum-based superconducting auxiliary quantum bit, a Xilinx Kintex UltraScale KU115 FPGA processor and an error correction pulse generator, which are connected to the quantum core control module and the classical quantum intelligent decision-making module via a 10Gbps high-speed bus; Key parameters: transmission delay ≤2ns, auxiliary quantum bit T2 ≥30μs, coupling strength 1-10MHz (adjustable); FPGA processor main frequency 1GHz, 10MB cache, built-in quantum error feature library; error correction pulse generator response time ≤5ns, output frequency 1-20GHz, amplitude accuracy ±0.01dB.

[0263] The classical layer support module serves as the intelligent control and data core of the computing power foundation. It is responsible for intelligent decision-making in quantum computing, massive data storage management, and cross-platform adaptation. It includes three sub-modules: the classical quantum intelligent decision-making module, the Ceph distributed data storage module, and the cross-platform adaptation interface. This module fills the gaps in existing technologies, such as inefficient classical decision-making, weak data management, and poor cross-platform compatibility.

[0264] The classical quantum intelligent decision-making module serves as the intelligent control hub of the computing power base, completing intelligent decision-making throughout the entire process, including task decomposition, computing power allocation, dimensional transformation prediction, entanglement topology optimization, and error correction strategy generation. The classic quantum intelligent decision-making module hardware consists of a Kunpeng 920 server (48 cores, 192GB memory), equipped with a quantum computing-specific intelligent decision-making model, and connected to the quantum layer via a 100Gbps high-speed optical fiber. Model and core parameters: The model is trained based on 28-dimensional features (quantum state parameters, task computing power requirements, resource status, etc.), and the hidden layer is a 6-layer residual network structure; the model inference latency is ≤1μs, the decision accuracy is ≥95%, the instruction issuance latency is ≤500ns, and the task decomposition accuracy is ≥95%.

[0265] Ceph distributed data storage module serves as the unified data core of the computing power foundation, enabling efficient and reliable storage and management of massive amounts of data across the entire quantum computing chain, and providing complete data support for quantum layer optimization and classical edge decision-making; Ceph distributed data storage module hardware composition: 3 or more storage nodes, each node is configured with 4 10TB SSD solid-state drives, 16-core CPU, 64GB memory, and the nodes are interconnected via 10 Gigabit Ethernet; The core parameters of the Ceph distributed data storage module are: network latency ≤ 0.5ms, a hybrid mode of Ceph RGW object storage and CephFS file storage, Ceph RGW read / write speed ≥ 1GB / s, CephFS read / write speed ≥ 800MB / s; supports PB / EB level linear expansion, 3 copies of core data, 2 copies of log data, 100% data retention rate, and no single point of failure.

[0266] The cross-platform adaptation interface is used to enable rapid identification and adaptive control logic of different quantum hardware platforms, and to complete the conversion of general quantum computing logic to the physical parameters of the target platform. Cross-platform adaptation interface hardware components: physical parameter database and QX-2000 hardware identification chip; the physical parameter database stores 32 core parameter templates for 12 types of quantum hardware platforms such as superconductivity, ion trap, and photonic quantum; Cross-platform adaptation interface core parameters: QX-2000 chip spectrum analysis resolution 1kHz, hardware recognition time ≤100μs; general logic to physical parameter conversion error ≤1.5%, cross-platform migration code reconstruction rate ≤10%.

[0267] The system data interaction and command issuance logic is as follows: The quantum layer core module and the classical layer support module achieve bidirectional real-time data interaction through a 10Gbps / 100Gbps high-speed bus, specifically: Classical layer → Quantum layer: Issues instructions for computing power configuration, encoding, dimension transformation, entanglement topology optimization, error correction, etc., with instruction issuance latency ≤500ns; Quantum layer → Classical layer: Upload data such as quantum state parameters, entangled link status, quantum error warning, and computation progress, with a maximum data upload frequency of 1GHz; The data layer (Ceph distributed storage) acts as an intermediate hub, enabling unified storage, retrieval, and updating of all data, and providing data support for the collaboration between the quantum layer and the classical layer.

[0268] The beneficial effects of implementing this embodiment are: it enables end-to-end integrated collaboration, allows for dynamic and efficient utilization of resources, and enables closed-loop optimization of data throughout the entire process.

[0269] Example 5 To address the aforementioned technical problems, embodiments of the present invention also provide a computer device. Please refer to [link / reference needed]. Figure 7 , Figure 7 This is a basic structural block diagram of the computer device in this embodiment.

[0270] The aforementioned computer device 8 includes a memory 81, a processor 82, and a network interface 83 that are interconnected via a system bus. It should be noted that only the computer device 8 with components 81, 82, and 83 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described herein is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0271] The aforementioned computer devices can be desktop computers, laptops, handheld computers, and cloud servers, among other computing devices. These devices can facilitate human-computer interaction with users through keyboards, mice, remote controls, touchpads, or voice-activated devices.

[0272] The aforementioned memory 81 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the aforementioned memory 81 may be an internal storage unit of the aforementioned computer device 8, such as the hard disk or memory of the computer device 8. In other embodiments, the aforementioned memory 81 may also be an external storage device of the aforementioned computer device 8, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 8. Of course, the aforementioned memory 81 may also include both the internal storage unit and its external storage device of the aforementioned computer device 8. In this embodiment, the aforementioned memory 81 is typically used to store the operating system and various application software installed on the aforementioned computer device 8, such as computer-readable instructions based on quantum computing methods of quantum state cooperative manipulation. In addition, the aforementioned memory 81 can also be used to temporarily store various types of data that have been output or will be output.

[0273] In some embodiments, the processor 82 described above may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 82 is typically used to control the overall operation of the computer device 8. In this embodiment, the processor 82 is used to execute computer-readable instructions stored in the memory 81 or to process data, for example, to execute computer-readable instructions of the quantum computing method based on quantum state cooperative manipulation.

[0274] The network interface 83 may include a wireless network interface or a wired network interface, which is typically used to establish a communication connection between the computer device 8 and other electronic devices.

[0275] Example 5 The present invention also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the quantum computing method based on quantum state cooperative manipulation as described above.

[0276] The beneficial effects of implementing the above embodiments are as follows: (1) End-to-end integrated collaboration: Breaking through the limitations of traditional quantum-classical separation design, the classical intelligent decision-making model and the quantum layer real-time monitoring and dual-model collaborative error correction are used to achieve deep integration of task decomposition, resource allocation and error correction, eliminating the bottleneck of collaborative disconnection and significantly improving the overall computing efficiency and stability.

[0277] (2) Resources can be used dynamically and efficiently: By introducing a hierarchical asymmetric entanglement topology and entanglement resource pool, idle entanglement resources can be dynamically scheduled according to the subtask computing power weight, which solves the problem of inefficient management of massive quantum data, avoids resource waste and local overload, and provides a flexible and scalable computing power base for large-scale parallel computing.

[0278] (3) Full-process data closed-loop optimization is possible: It not only completes the output of calculation results, but also realizes the complete accumulation and feedback of calculation data, forming a digital intelligent iterative closed loop of calculation-data-optimization-re-computation. This enables the system to have continuous self-evolution capability, continuously optimize the coding scheme, resource allocation and error correction strategy, and accelerate the leap from prototype to industrial application.

[0279] (4) Strong cross-platform compatibility: It has 12 built-in quantum hardware platform parameter templates and is equipped with dedicated spectrum analysis and hardware feature intelligent recognition equipment. It can quickly adapt to various types of hardware such as superconductors, ion traps, photonic quantum, and neutral atoms, reducing the cross-platform code reconstruction rate from 70% to below 10%.

[0280] (5) High resource utilization: The hierarchical asymmetric entanglement topology is adopted. High-weight subtasks are configured with high-density entangled links, and low-weight subtasks are matched with sparse links. The overall quantum resource utilization rate is increased to more than 90%, and the information carrying capacity of a single quantum bit is increased by 2-3 orders of magnitude compared with the traditional single-dimensional scheme.

[0281] 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 the present invention, 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 device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0282] Obviously, the embodiments described above are merely some embodiments of the present invention, not all embodiments. The accompanying drawings show preferred embodiments of the present invention, but do not limit the patent scope of the present invention. The present invention can be implemented in many different forms; rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the patent protection scope of this invention.

Claims

1. A quantum computing method based on quantum state cooperative manipulation, characterized in that, Includes the following steps: Construct a data base and initialize the data base; The classical quantum intelligent decision-making model, based on a 6-layer residual network trained with 28-dimensional quantum operation features, decomposes the external computing task into multiple sub-tasks and dynamically allocates the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task. The multi-dimensional encoding scheme uses 32 core parameter templates from 12 types of quantum hardware platforms and is adapted across platforms by using dedicated spectrum analysis and hardware feature intelligent recognition equipment. The 12 types of quantum hardware platforms include superconducting, ion trap, photonic, and neutral atom platforms. The 32 core parameter templates include the energy level relaxation rate, spin-flip probability, orbital angular momentum diffusion coefficient, coherence time, coupling strength, impulse response frequency, phase adjustment range, qubit topological connection constraints, gate operation error threshold, readout fidelity, temperature drift coefficient, and magnetic field noise tolerance parameters of the quantum hardware platform. Based on the computational power weights of subtasks, a hierarchical asymmetric entanglement topology is constructed, and an entanglement resource pool is built to achieve dynamic scheduling of idle resources. Dynamic dimensional basis transformation is achieved through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation; Real-time monitoring is performed using a quantum layer, and hybrid verification and error correction are performed using a classical dual-model approach. The output is encoded in the quantum state to calculate the results and complete the full accumulation of computational data, forming a closed loop of computation-data-optimization-re-computation.

2. The quantum state cooperation-manipulation-based quantum computing method according to claim 1, characterized in that, The steps of constructing the data foundation and initializing the data foundation specifically include: Create and configure multiple data partitions for Ceph distributed storage, using a hybrid mode of Ceph RGW object storage and CephFS file storage. The multiple data partitions include: a quantum state parameter library, a model training library, a manipulation and computation database, a cross-platform adaptation library, and an error correction log library. The quantum computing intelligent decision-making model is loaded, the latest model parameters are read from the Ceph model training library, and the inference engine is initialized. The model is based on a 6-layer residual network with 28-dimensional feature input, and the hidden layer dimensions are 256, 128, 64, 32, 16, and 8 respectively. The output layer is a multi-task head. A detection command is sent to the multi-dimensional quantum state storage module to collect the initial parameters of the quantum state. The multi-dimensional quantum state storage module is equipped with a cross-shaped aluminum-based Josephson junction superconducting quantum bit chip, which is divided into three independent control dimensions: energy level, spin, and orbital angular momentum. The initial parameters of the quantum state are collected by a 1GHz ultra-high precision three-dimensional parameter sensor, which includes an energy level sensor, a spin sensor, and an orbital angular momentum sensor. To verify whether the quantum layer's health level is ≥99%, the health level calculation formula is as follows: wherein, is the energy level relaxation rate, is the spin flip probability, is the orbital angular momentum diffusion coefficient, is the measurement time, is the mass factor of the i-th qubit; ensures that the computing power base enters the mode to be operated in a reliable state.

3. The quantum computing method based on quantum state cooperative manipulation according to claim 1, characterized in that, The steps of decomposing the external computing task into multiple sub-tasks using a classical edge intelligent decision-making model, and dynamically allocating the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task, specifically include: Receive external computing tasks from the classic control terminal and extract characteristics such as task type, scale, complexity, and real-time requirements; The quantum computing intelligent decision-making model, based on historical data, decomposes the computational task into multiple subtasks that can be performed in parallel or sequentially, and calculates the computational density, resource requirements, and priority of each subtask. The formula for calculating the computational density is as follows: ,in, Let i be the computation density of the i-th subtask. The total number of quantum gate operations. For the allocated time window, For the required number of qubits, The number of reference qubits; The quantum computing intelligent decision-making model, combined with the real-time state of the quantum layer, dynamically allocates the number of qubits, computing power weight, and multi-dimensional encoding scheme to each subtask. The formula for calculating the computing power weight is as follows: ,in, Let σ(·) be the computational power weight of the i-th subtask, and let σ(·) be the Sigmoid function. The amount of quantum resources already used. For total sub-resource quantity, This represents the importance coefficient of the subtask. The computing power configuration instructions and multi-dimensional encoding instructions are sent to the quantum control module, and the task decomposition results, computing power allocation scheme and encoding scheme are written to the Ceph control and operation database in real time.

4. The quantum computing method based on quantum state cooperative manipulation according to claim 1, characterized in that, The steps of constructing a hierarchical asymmetric entanglement topology based on the computing power weights of subtasks and building an entanglement resource pool to achieve dynamic scheduling of idle resources specifically include: The quantum manipulation module receives computing power configuration instructions from the classical end, reads the computing power weights of each subtask from the Ceph manipulation and computation database, and converts them into entangled topology parameters, which include an update period. The calculation formula is: ,in, To address the update cycle of the entangled link, As the baseline update cycle, The attenuation coefficient is... Weight based on computing power; By synchronously adjusting the coupling pulses of each qubit using a multi-channel phase controller, corresponding entangled links are established. At the same time, an entangled resource pool is constructed, reserving 5% to 10% of backup links, and the coupling strength and synchronization of all links are verified. Synchronize the state data of idle links and available qubit resources in the entanglement resource pool to the Ceph quantum state parameter library.

5. The quantum computing method based on quantum state cooperative manipulation according to claim 1, characterized in that, The steps for dynamic dimensional basis transformation through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation specifically include: Based on the classical encoding instructions, composite quantum pulses and single pulses are generated, simultaneously achieving synchronous operations on multiple quantum states. The envelope formula of the composite quantum pulse is: Where Ω(t) is the envelope amplitude of the composite quantum pulse. The peak amplitude is denoted as t, and t is a time variable. Where σ is the pulse center time, and σ is the pulse width parameter. The driving angular frequency is φ, and the initial phase is φ. Quantum state parameters are acquired in real time and written into the Ceph quantum state parameter library; The classical quantum intelligent decision-making model reads quantum state parameters in real time from the Ceph quantum state parameter library, calculates the availability score for each dimension, and predicts the decoherence trend of each dimension based on historical data. The formula for calculating the availability score is as follows: ,in, Score the usability of dimension d. Let be the energy level relaxation rate in dimension d. Let be the spin-flip probability in dimension d. Let be the orbital angular momentum diffusion coefficient in dimension d. For the expected quantum operation time, For the coherence time in dimension d, The reference coherence time; When a dimension score is about to fall to the dimension score threshold, the classical quantum intelligent decision-making model retrieves the health dimension combination from multiple predefined dimension combinations and sends a dimension conversion command to the quantum manipulation module. The quantum layer uses the natural evolution characteristics of quantum states to complete the dimension conversion. The classical end model reads the resource pool status from the Ceph quantum state parameter library based on the real-time changes in quantum state parameters and the computation progress, and dynamically optimizes the link parameters of the entangled topology.

6. The quantum computing method based on quantum state cooperative manipulation according to claim 1, characterized in that, The steps of using a quantum layer for real-time monitoring and employing a classical dual-model approach for hybrid verification and error correction specifically include: Real-time monitoring of lightweight auxiliary qubits, wherein the auxiliary qubits are aluminum-based superconducting qubits with a coherence time T2≥30μs and a coupling strength adjustable from 1 to 10MHz; The FPGA processor calls a lightweight neural network to perform preliminary identification of fluctuating data. The lightweight neural network is a 3-layer fully connected network with 64 nodes in the input layer, 32 nodes in the hidden layer, and 10 error categories in the output layer. The classical quantum intelligent decision-making model is combined with historical data to complete error identification through the collaboration of the two models. The classic end model generates a personalized error correction strategy based on the error type and historical error correction data. It outputs error correction pulses through an error correction pulse generator, and the quantum core control module executes the error correction operation to restore the quantum state to a healthy range. Error types, correction commands, and correction effects are recorded throughout the entire error correction process and written to the Ceph error correction log library and model training library in real time.

7. The quantum computing method based on quantum state cooperative manipulation according to any one of claims 1 to 6, characterized in that, The steps of encoding the output in the quantum state and completing the full accumulation of computational data to form a closed loop of computation-data-optimization-re-computation specifically include: The computation results encoded in the quantum state are decoded into classical data format and output to the target terminal through a cross-platform adaptation interface. The cross-platform adaptation interface has built-in 12 types of quantum hardware platform parameter templates, which can quickly adapt to superconducting, ion trap, photonic quantum and neutral atom hardware. The classic client model organizes all the data from the entire computation process and categorizes it for writing to various Ceph data partitions. The classical edge model uses the data from the current operation and historical data to iteratively optimize the quantum computing-specific intelligent decision-making model; The quantum manipulation module sends resource reclamation commands to each module of the quantum layer, releasing all qubits and entangled resources, restoring the quantum layer to its initial state, Ceph completes persistent data backup, and the computing power base is reset.

8. A quantum computing device based on quantum state cooperative manipulation, characterized in that, include: An initialization module is used to construct the data base and initialize the data base; The decomposition module is used to decompose external computing tasks into multiple sub-tasks using a classical quantum intelligent decision-making model. This model is based on a 6-layer residual network trained with 28-dimensional quantum operation features. It dynamically allocates the number of qubits, computing power weights, and multi-dimensional encoding schemes to each sub-task. The multi-dimensional encoding scheme uses 32 core parameter templates from 12 types of quantum hardware platforms and is adapted across platforms using dedicated spectrum analysis and hardware feature intelligent recognition devices. The 12 types of quantum hardware platforms include superconducting, ion trap, photonic, and neutral atom platforms. The 32 core parameter templates include the energy level relaxation rate, spin-flip probability, orbital angular momentum diffusion coefficient, coherence time, coupling strength, impulse response frequency, phase adjustment range, qubit topological connection constraints, gate operation error threshold, readout fidelity, temperature drift coefficient, and magnetic field noise tolerance parameters of the quantum hardware platform. The module is used to build a hierarchical asymmetric entanglement topology based on the computing power weight of subtasks, and to build an entanglement resource pool to realize the dynamic scheduling of idle resources. The conversion module is used to perform dynamic dimensional basis vector conversion through multi-dimensional quantum state combination encoding and composite quantum pulse synchronous manipulation; The verification module is used for real-time monitoring using a quantum layer and hybrid verification and error correction using a classical dual-model approach. The output module is used to output the computation results encoded in the quantum state and complete the full accumulation of computation data throughout the entire process, forming a digital intelligent iterative closed loop of computation-data-optimization-re-computation.

9. A computer device, characterized in that, The device includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor executes the computer-readable instructions to implement the steps of the quantum computing method based on quantum state cooperative manipulation as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the quantum computing method based on quantum state cooperative manipulation as described in any one of claims 1 to 7.

Images