A Renewable Energy Data Clustering Method and System Based on Optical Quantum Computers

Abstract

This invention discloses a renewable energy data clustering method and system based on optical quantum computing. First, a renewable energy data clustering optimization problem based on a quadratic unconstrained binary optimization (QUBO) model is constructed based on a discrete optimization objective function. Then, the objective function is matrix-transformed, and the kernel method is used to optimize the clustering of the objective function terms, converting them into matrix form to construct a new matrix-transformed clustering model. Finally, the new QUBO optimization problem is solved using CIM, and the result is transformed into the optimal clustering result. This invention's method, by constructing a QUBO model and combining it with a kernel method to improve data similarity measurement, transforms the renewable energy data clustering problem into a form suitable for CIM solving, thus avoiding the computational bottleneck of classical computational methods in high-dimensional data processing. This method not only preserves the nonlinear characteristics of the data but also achieves efficient clustering optimization computation, significantly improving computational efficiency and clustering accuracy.

Description

Technical Field

[0001] This invention belongs to the technical field of data clustering and quantum computing, and mainly relates to a method and system for clustering renewable energy data based on optical quantum computers. Background Technology

[0002] Clustering technology is a fundamental tool for power system data analysis and aggregation control, and clustering analysis of renewable energy data is a key step in its pattern recognition and feature extraction. However, clustering technology is often formulated as a combinatorial optimization problem, which faces excessive computational complexity when dealing with large-scale, high-dimensional data, especially when solved on classical computers based on integrated circuits, where it is easily limited by the curse of dimensionality.

[0003] When data dimensionality is high, the computational complexity of traditional clustering methods increases significantly, and the computation time rises sharply, making it difficult to meet real-time requirements. To address this issue, researchers have proposed various clustering methods based on classical computers, such as K-means and K-medoids. Although these methods perform well on small-scale datasets, their computational complexity increases rapidly when dealing with large-scale datasets, making them unsuitable for NP-hard problems. Summary of the Invention

[0004] This invention addresses the limitations of existing technologies in handling large-scale data and their low processing efficiency. It proposes a renewable energy data clustering method and system based on optical quantum computing. First, a renewable energy data clustering optimization problem based on a quadratic unconstrained binary optimization (QUBO) model is constructed using a discrete optimization objective function. Then, the objective function is matrix-transformed, and the kernel method is used to optimize the clustering of the objective function terms, converting them into matrix form to construct a new matrix-transformed clustering model. Finally, CIM is used to solve the new QUBO optimization problem, transforming it into the optimal clustering result. This invention's method, by constructing a QUBO model and combining it with a kernel method to improve data similarity metrics, transforms the renewable energy data clustering problem into a form suitable for CIM solving, thus avoiding the computational bottleneck of classical computational methods in high-dimensional data processing. This method not only preserves the nonlinear characteristics of the data but also achieves efficient clustering optimization computation, significantly improving computational efficiency and clustering accuracy.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a renewable energy data clustering method based on optical quantum computers, comprising the following steps:

[0006] S1. Based on the discrete optimization objective function, construct a clustering optimization problem for renewable energy data based on a quadratic unconstrained binary optimization (QUBO) model. The specific clustering optimization problem is as follows:

[0007]

[0008] Where g∈{1,2,...,G} is the cluster index, G is the number of clusters; i,j∈{1,2,...,N}, N is the number of samples; d i,j This represents the Euclidean distance between sample i and sample j; For binary decision variables, This indicates that sample i is assigned to the g-th cluster;

[0009] S2. Matrix the objective function in step S1 to make it suitable for the calculation of the optical coherence Ising machine (CIM);

[0010] S3. Using kernel methods, cluster the objective function terms and convert them into matrix form to construct a new matrix-based clustering model;

[0011] S4. Use CIM to solve the new QUBO optimization problem constructed in step S3, and transform it into the optimal clustering result.

[0012] As an improvement to the present invention, the discrete optimization objective function in step S1 is specifically as follows:

[0013]

[0014] Constraints Augmentation to penalty item H cons The objective function is reconstructed into a quadratic unconstrained bivariate optimization (QUBO) form, specifically:

[0015]

[0016] Where, λ i This is the penalty coefficient.

[0017] As another improvement of the present invention, the penalty coefficient λ in step S1 i satisfy:

[0018]

[0019] in, Let d be the Euclidean distance between all sample pairs. i,j The maximum value.

[0020] As another improvement of the present invention, step S2, which matrixifies the objective function, specifically includes the following steps:

[0021] S21. Regarding the objective function term H obj Perform matrix transformation:

[0022]

[0023] in, A binary decision variable vector; Let d be the upper triangular part of matrix D, where the element is d. i,j ; It is the identity matrix; Represents the Kronecker product;

[0024] S22. Ignoring the constant term, the constraint term H... cons Perform matrix transformation:

[0025]

[0026] in, It is a diagonal matrix, and its diagonal elements are the penalty coefficients λ. i ; It is an upper triangular matrix, with 0 elements on the main diagonal and 1 elements on the rest of the upper triangular matrix;

[0027] S23. Combining the objective function term from step S21 with the constraint term from step S22, the matrix form of the objective function is as follows:

[0028]

[0029] As another improvement of the present invention, the objective function term optimized by kernel clustering in step S2 is specifically as follows:

[0030]

[0031] Among them, g i,j It is a non-linear kernel similarity index;

[0032] Convert the objective function term of the kernel method into matrix form:

[0033]

[0034] in, It is a diagonal matrix containing only the diagonal elements of G; matrix Its element is g i,j It contains all the similarity information used for clustering; Let G be an upper triangular matrix with 0 as its main diagonal element and the same as G as the other upper triangular elements.

[0035] As another improvement of the present invention, step S4 specifically includes the following steps:

[0036] S41. The QUBO objective function is transformed into an equivalent quadratic optimization problem, the mathematical form of which is as follows:

[0037] f QUBO (x)=x T Qx

[0038] Where Q is a symmetric weight matrix;

[0039] S42. By substitution, the QUBO problem is transformed into the Ising Hamiltonian form. Based on the QUBO matrix output from steps S2 and S3, the coupling coefficient J in CIM is constructed. i,j and bias term h i ;

[0040] S43. CIM physically realizes the spin state using a degenerate optical parametric oscillator. During the iteration process, CIM adjusts the variable states according to the principle of minimizing system energy to bring the system to a steady-state solution, and the final output spin configuration s i ∈{-1,+1}, The optimal solution corresponds to the optimal value of the QUBO variable;

[0041] S44, Optimal spin configuration for each bit of the CIM output. i Mapping variables ∈{-1,+1} back to the original binary decision variables The optimal solution calculated by CIM is mapped back to the clustering label to achieve optimal data classification.

[0042] To achieve the above objectives, the present invention also adopts the following technical solution: a renewable energy data clustering system based on an optical quantum computer, comprising a computer program, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.

[0043] Compared with existing technologies, this invention offers the following advantages: By revealing the connection between the QUBO model and CIM, this invention proposes a clustering method and system for renewable energy data based on optical quantum computers. This method effectively overcomes the computational complexity problem of classical computers when processing large-scale data, thus significantly reducing the computational burden. Furthermore, by introducing a kernel method, this invention further improves clustering accuracy, better capturing the nonlinear structure of the data and overcoming the limitations of traditional methods in processing nonlinear data. This invention not only preserves the nonlinear characteristics of the data but also achieves efficient clustering optimization calculations, significantly improving computational efficiency and clustering accuracy. Attached Figure Description

[0044] Figure 1 This is a flowchart of the steps of the renewable energy data clustering method based on optical quantum computers of the present invention;

[0045] Figure 2 This is a comparison chart of clustering results of the classic computer and different CIM algorithms under different numbers of variables in the test examples of this invention. Detailed Implementation

[0046] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0047] Example 1

[0048] A clustering method for renewable energy data based on optical quantum computers, such as Figure 1 As shown, it includes the following steps:

[0049] Step S1: Construct a clustering optimization problem for renewable energy data based on a quadratic unconstrained binary optimization (QUBO) model.

[0050] This embodiment addresses the clustering optimization problem of renewable energy power output data, using photovoltaic (PV) time-series power output data as the research object. Assume there are N sets of power output characteristic data (i.e., N samples), which need to be divided into G clusters. A discrete optimization objective function H is defined. obj This is used to minimize the distance between data points within the same cluster, and its expression is as follows:

[0051]

[0052] Where g∈{1,2,...,G} is the clustering index; y i ,y j Let ∈{1,2,..,G}(i,j∈{1,2,...,N}) be an integer variable representing the cluster number to which sample i belongs. If y i =y j This indicates that sample i and sample j are assigned to the same cluster; d i,j This represents the Euclidean distance between sample i and sample j, used to quantify the similarity between data.

[0053] Because the feasible solution space of the objective function is highly non-convex, traditional methods struggle to find its optimal solution. To overcome this challenge, integer variables y are typically used... i Convert to G binary variables, that is:

[0054]

[0055] in, For binary decision variables, This indicates that sample i is assigned to the g-th cluster. Furthermore, since each sample can only belong to one cluster, the following constraints must be satisfied:

[0056]

[0057] The clustering problem can then be transformed into the following form:

[0058]

[0059] Constraints As a penalty term, the objective function is reconstructed into a quadratic unconstrained bivariate optimization (QUBO) form:

[0060]

[0061] Where, λ i This is the penalty coefficient.

[0062] To ensure the constraint is effective, λ i It should meet the following requirements:

[0063]

[0064] in, Let d be the Euclidean distance between all sample pairs. i,j The maximum value.

[0065] Step S2: Matrix the QUBO objective function to make it suitable for calculations of the Optical Coherence Ising Machine (CIM).

[0066] S21. Regarding the objective function term H obj The matrix transformation is as follows:

[0067]

[0068] in, A binary decision variable vector; Let d be the upper triangular part of matrix D, where the element is d. i,j ; It is the identity matrix; Represents the Kronecker product;

[0069] S22. Ignoring the constant term, the constraint term H... cons The matrix transformation is as follows:

[0070]

[0071] in, It is a diagonal matrix, and its diagonal elements are the penalty coefficients λ. i ; It is an upper triangular matrix, with 0 elements on the main diagonal and 1 elements on the rest of the upper triangular matrix;

[0072] S23. Combining the objective function terms and constraint terms, the matrix form of the QUBO objective function is as follows:

[0073]

[0074] Step S3: Combine kernel methods to improve clustering accuracy and construct a new matrix-based clustering model.

[0075] S31. Using a nonlinear kernel similarity index g i,j Instead of the original Euclidean distance d i,j as follows:

[0076]

[0077] Where c∈{1,2,...,N} is the column index, and r∈{1,2,...,N} is the row index. Represents a nonlinear Gaussian kernel function;

[0078] Define matrix Its element is g i,j It contains all the similarity information used for clustering; based on this, the clustering optimization objective function of the kernel method can be expressed as follows:

[0079]

[0080] S32. The objective function term of the kernel method is further converted into matrix form as follows:

[0081]

[0082] in, It is a diagonal matrix containing only the diagonal elements of G; Let G be an upper triangular matrix with 0 elements on its main diagonal and the same other elements on its upper triangular matrix as G.

[0083] Since the constraints remain unchanged, the QUBO objective function based on the kernel method is expressed as follows:

[0084]

[0085] This formula is a QUBO form that can be solved by an optical quantum computer (CIM).

[0086] Step S4: Solve the QUBO optimization problem using CIM and transform it into the optimal clustering result.

[0087] This invention uses the optical coherent Ising machine to solve the QUBO objective function and obtain the optimal clustering result. The CIM is calculated using the Ising model, and its objective is to minimize the total system energy.

[0088]

[0089] Among them, s i∈{-1,+1} represents the spin variable, corresponding to the QUBO variable x. i binary mapping; J i,j h is the interaction coupling coefficient, which defines the influence relationship between variables; i For external field effects, adjust the bias value of each variable.

[0090] Since the QUBO objective function can be transformed into an equivalent quadratic optimization problem, its mathematical form is as follows:

[0091] f QUBO (x)=x T Qx

[0092] Where Q is a symmetric weight matrix, which corresponds to the QUBO objective function term based on the kernel method in S32. part.

[0093] By substituting variables, the QUBO problem is transformed into an Ising-Hamiltonian form, making it applicable to CIM. The QUBO matrix output from steps S2 and S3 is used to construct the coupling coefficients J in CIM. i,j and bias term h i .

[0094] CIM physically realizes the spin state using a degenerate optical parametric oscillator. During the iteration process, CIM continuously adjusts the variable states according to the principle of minimizing system energy, until the system reaches a steady-state solution. The final output spin configuration is the optimal solution, corresponding to the optimal values ​​of the QUBO variables.

[0095] Because the CIM outputs s i For variables ∈{-1,+1}, it is necessary to map them back to the original binary decision variables. The optimal solution calculated by CIM is mapped back to the cluster labels to achieve optimal data classification. Finally, the clustering effect can be evaluated based on the calculated silhouette coefficient to ensure that the classification quality meets application requirements.

[0096] Test case

[0097] This test case was conducted on a 400-bit CIM provided by QBoson. Photovoltaic output data was obtained from the Yulara solar system, collected at five-minute intervals, and analyzed using hourly averages.

[0098] This test case compares different clustering algorithms, including K-means (MATLAB), K-medoids (MATLAB), and simulated quantum annealing (SQA, OpenJij) running on a classical computer (i5-13400F 2.5GHz, 32GB RAM). Four different test case sizes were set up, dividing 50 / 60 / 70 / 80 photovoltaic power output data points into 2 / 3 / 4 / 5 clusters, corresponding to 100 / 180 / 280 / 400 binary variables, respectively. The silhouette coefficient proposed by Rousseeuw was used to evaluate the clustering performance.

[0099] The steps of the renewable energy data clustering method based on optical quantum computers according to the present invention are as follows: To verify the ability of CIM to overcome the curse of dimensionality and compare the computational efficiency of different methods, the clustering results of classical computers and different CIM algorithms under different numbers of variables in this test case are compared. Figure 2 As shown.

[0100] The computation time of different clustering algorithms under different variable scales is as follows: Figure 2 As shown in (a). Figure 2 As shown in (a), the computation time of CIM is almost unaffected by the size of the input variables, while the computation time of all other methods running on classical computers increases significantly with the increase of variable size. This phenomenon clearly demonstrates that classical computers are susceptible to the curse of dimensionality when solving NP-hard clustering problems, while CIM is almost unaffected. In all test cases of this test case, the computation time of CIM remained relatively stable at around 3ms, with a maximum of no more than 3.105ms.

[0101] Silhouette coefficients of different clustering algorithms under different variable scales, such as Figure 2 As shown in (b). Figure 2 As can be seen in (b), the clustering effect of this method is comparable to that of K-medoids and K-means, indicating that this method can guarantee the effectiveness and accuracy of clustering, thus verifying its applicability in renewable energy data analysis.

[0102] The test comparison above shows that this method performs well in both computational efficiency and clustering effect, proving its feasibility in large-scale quantum computing applications.

[0103] Therefore, this method proposes a renewable energy data clustering framework based on optical quantum computing, which utilizes QUBO modeling and CIM computation to achieve efficient clustering analysis. The method constructs a clustering model combining kernel methods, improves data similarity measurement through matrix transformation, and adapts to CIM computation. Optical quantum computing overcomes the computational complexity problem encountered by traditional methods in high-dimensional data processing. Furthermore, the proposed QUBO modeling scheme provides valuable guidance for the application of CIM in combinatorial optimization problems, contributing to further expanding the application of optical quantum computing in large-scale renewable energy data analysis.

[0104] It should be noted that the above content merely illustrates the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. For those skilled in the art, various improvements and modifications can be made without departing from the principle of the present invention, and all such improvements and modifications fall within the scope of protection of the claims of the present invention.

Claims

1. A clustering method for renewable energy data based on optical quantum computers, characterized in that, Includes the following steps: S1. Based on the discrete optimization objective function, construct a renewable energy data clustering optimization problem based on a quadratic unconstrained binary optimization QUBO model. The specific clustering optimization problem is as follows: ； in, For clustering index, The number of clusters; , The number of samples; Indicates sample and samples The Euclidean distance between them; For binary decision variables, Indicates sample Assigned to the One cluster; The discrete optimization objective function is specifically as follows: ； Constraints Expanded to include penalties The objective function is reconstructed into a quadratic unconstrained bivariate optimization (QUBO) form, specifically: ； in, This is the penalty coefficient; S2. Matrix the objective function in step S1 to make it suitable for the calculation of the Optical Coherence Ising Machine (CIM). S3. Using kernel methods, cluster the objective function terms and convert them into matrix form to construct a new matrix-based clustering model; S4. Use CIM to solve the new QUBO optimization problem constructed in step S3, and transform it into the optimal clustering result.

2. The renewable energy data clustering method based on optical quantum computers as described in claim 1, characterized in that: The penalty coefficient in step S1 satisfy: ； in, Euclidean distance between all sample pairs The maximum value.

3. The renewable energy data clustering method based on optical quantum computers as described in claim 2, characterized in that: Step S2, which matrixizes the objective function, specifically includes the following steps: S21. Regarding the objective function term Perform matrix transformation: ； in, A binary decision variable vector; For matrix The upper triangular part, where the element is ; It is the identity matrix; Represents the Kronecker product; S22. After ignoring the constant term, the constraint term... Perform matrix transformation: ； in, It is a diagonal matrix, and its diagonal elements are penalty coefficients. ; It is an upper triangular matrix, with 0 elements on the main diagonal and 1 elements on the rest of the upper triangular matrix; S23. Combining the objective function term from step S21 with the constraint term from step S22, the matrix form of the objective function is as follows: 。 4. The renewable energy data clustering method based on optical quantum computers as described in claim 2, characterized in that: The objective function term optimized by kernel clustering in step S2 is specifically: ； in, It is a non-linear kernel similarity index; Convert the objective function term of the kernel method into matrix form: ； in, For only containing A diagonal matrix with diagonal elements; a matrix Its elements are It contains all the similarity information used for clustering; for An upper triangular matrix whose main diagonal elements are 0, and whose other upper triangular elements are 0 and 0. same.

5. The renewable energy data clustering method based on optical quantum computers as described in claim 4, characterized in that: Step S4 specifically includes the following steps: S41. The QUBO objective function is transformed into an equivalent quadratic optimization problem, the mathematical form of which is as follows: ； in, It is a symmetric weight matrix; S42. By substituting variables, the QUBO problem is transformed into the Ising Hamiltonian form. Based on the QUBO matrix output from steps S2 and S3, the coupling coefficients in the CIM are constructed. and bias terms ; S43. CIM physically realizes the spin state using a degenerate optical parametric oscillator. During the iteration process, CIM adjusts the variable states according to the principle of minimizing system energy to bring the system to a steady-state solution, and the final output spin configuration is... The optimal solution corresponds to the optimal value of the QUBO variable; S44. Optimal spin configuration for each bit of the CIM output. Map the variables back to the original binary decision variables The optimal solution calculated by CIM is mapped back to the clustering label to achieve optimal data classification.

6. A renewable energy data clustering system based on optical quantum computers, comprising a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-5 above.

Images