Heterogeneous data packet scheduling method and system based on time-varying features and priorities

Abstract

The invention discloses a heterogeneous data packet scheduling method and system based on time-varying features and priorities, and relates to the field of data processing, and the method comprises the steps: collecting multivariate heterogeneous data, and converting the multivariate heterogeneous data into carbon emission data in a uniform format by adopting a functional data analysis method; performing time-varying feature extraction and data grouping on the carbon emission data by using a CNN-LSTM model based on an attention mechanism; according to a data scheduling demand estimation model, respectively calculating a demand value of each data packet; and calculating the similarity between each data packet and the scheduling link according to the demand value, and scheduling the data packets. The mass heterogeneous data can be effectively and orderly managed and utilized by performing packet scheduling on the mass heterogeneous data, and the intelligent level of a dual-carbon digital monitoring center is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more particularly to a method and system for scheduling heterogeneous data packets based on time-varying characteristics and priorities. Background Art

[0002] Against the backdrop of global climate change and increasing pressure on resources and the environment, a "dual carbon" digital monitoring center that uses digital means to monitor and calculate carbon emissions in real time has emerged. The digital monitoring center integrates data collection, storage, processing and analysis to provide comprehensive and accurate carbon emissions data support for governments, enterprises and the public.

[0003] However, with the continuous development of the "dual carbon" digital intelligent monitoring center, the types of carbon emission data collected are becoming increasingly rich, covering electricity, industry, transportation, construction and other fields, forming a huge amount of heterogeneous data. These data not only differ in format, structure and source, but also show different dynamic characteristics over time. Faced with such a complex data environment, how to effectively and orderly manage and utilize these data has become an urgent problem to be solved. The disordered accumulation of a large amount of heterogeneous data not only increases the difficulty and cost of data processing, but may also affect the accuracy of data analysis and decision-making results.

[0004] Therefore, how to group and schedule massive heterogeneous data and improve the intelligence level of the "dual carbon" digital monitoring center is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0005] In view of this, the present invention provides a heterogeneous data grouping scheduling method and system based on time-varying characteristics and priorities, which combines the time-varying characteristics of heterogeneous data and realizes the reasonable division of data through dynamic grouping strategies, thereby improving the intelligence level of the "dual-carbon" digital intelligent monitoring center.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention discloses a heterogeneous data packet scheduling method based on time-varying characteristics and priorities, and the specific steps are as follows:

[0008] Collecting multivariate heterogeneous data and converting the multivariate heterogeneous data into carbon emission data in a unified format using a functional data analysis method;

[0009] Using a CNN-LSTM model based on an attention mechanism, time-varying feature extraction and data grouping are performed on the carbon emission data;

[0010] According to the data scheduling demand estimation model, the demand value of each data packet is calculated respectively; according to the demand value, the similarity between each data packet and the scheduling link is calculated, and the data packet is scheduled.

[0011] Furthermore, the functional data analysis method is specifically as follows:

[0012] The multivariate heterogeneous data is smoothly expanded using a basis function to obtain a fitting function;

[0013] Smoothing the fitting function by adding a penalty term;

[0014] The coefficient matrix of the fitting function is fitted using the least squares method, and the penalty factor coefficient of the penalty term is solved using the generalized cross-validation method to obtain a representation function of the discrete data sequence of multivariate heterogeneous data.

[0015] Furthermore, the fitting function is expressed as:

[0016]

[0017] Among them, F(x) represents the fitting function, x represents the xth sampling point of the scattered data sequence, φ i (x) represents the i-th basis function, i = 1, 2, 3...n, n is the total number of basis functions, c i represents the coefficient vector of the basis function;

[0018] The specific formula of the penalty term ∈(x) is:

[0019] ∈(x)=λC T RC;

[0020] Among them, λ represents the penalty factor, C represents the coefficient vector c i The corresponding coefficient matrix, R represents the matrix related to the second-order derivative of the basis function;

[0021] The expression function f(x) is expressed as follows:

[0022] f(x)=F(x)+∈(x).

[0023] Furthermore, the basis function adopts a B-spline basis function, which is expressed as:

[0024]

[0025] Among them, B i,k (x) represents the B-spline basis function, t represents the time parameter, t i ,t i+1 are the time parameters of the i-th and i+1-th basis functions respectively, and k represents the order of the basis function;

[0026] The B-spline basis function is fitted using a recursive formula, which is:

[0027]

[0028] Furthermore, the least squares method is expressed as:

[0029]

[0030] The solution of the coefficient matrix is ​​expressed as:

[0031] C=(G T G+λR) -1 G T y;

[0032] Among them, y x The original data collected at the x-th sampling point, G is the value of each sampling point of the basis function matrix G = [φ(1), φ(2)…φ(x)], φ(x) is the superposition vector of all basis functions at the x-th sampling point;

[0033] The generalized cross validation formula is:

[0034]

[0035] Among them, GCV represents the value of generalized cross validation, which is used to evaluate the goodness of fit of the model, tr() represents the trace of the matrix, which is defined as the sum of the diagonal elements of the matrix; the value of the penalty factor coefficient is determined by minimizing the GCV value.

[0036] Furthermore, the CNN-LSTM model includes an input layer, a convolutional layer, a multi-layer LSTM layer, an attention layer and an output layer connected in sequence;

[0037] The input layer samples the carbon emission data according to the set batch size, time step T, and feature dimension D. A data sample is represented by a real number sequence matrix X∈R T×D , where X represents the data sample, R T×D represents a real number sequence matrix, x t ″ represents the data vector at time t′ in the real number sequence matrix;

[0038] The convolution layer performs convolution, pooling and dimensionality reduction operations on the input data vector in sequence. The feature after convolution is represented as: h t′ =f(W·x t ″+b); where h t′ represents the output feature, f() represents the nonlinear activation function, W represents the weight matrix, and b represents the bias;

[0039] The multi-layer LSTM layer realizes the state transfer of time series data through a gating mechanism, including a forget gate, a selective memory gate, and an output gate. The forget gate determines the degree of forgetting of the previous hidden layer state, which is expressed as: t′ =σ(W f ·[h t′-1 ,x t ″]+b f ), where f t′ represents the value of the forget gate at time t′, σ represents the activation function, W f represents the forget gate weight matrix, h t′-1 represents the hidden layer state at time t′-1, b f represents the bias of the forget gate;

[0040] The attention layer performs weighted summation on the output of the last hidden layer in the multi-layer LSTM layer by weighted average, and the formula is: Among them, c t′ , α t′ 、h t ″ respectively represent the output and weight of the attention layer at time t′ and the output of the multi-layer LSTM layer;

[0041] The output layer outputs the final grouping result according to the output of the attention layer.

[0042] Furthermore, the data scheduling demand estimation model is expressed as:

[0043] Z = f × I × Q;

[0044] Where Z represents the demand estimate, f represents the frequency of data grouping, I represents the data scheduling amount, and Q represents the data scheduling density;

[0045] The calculation formula of the similarity is:

[0046]

[0047] Wherein, RTT represents the response time factor of the link; the data packet selects R DPLS The link with the smallest value is used for transmission.

[0048] The present invention also discloses a heterogeneous data packet scheduling system based on time-varying characteristics and priorities, comprising:

[0049] Data preprocessing module: collects multivariate heterogeneous data and converts the multivariate heterogeneous data into carbon emission data in a unified format using a functional data analysis method;

[0050] Feature extraction and grouping module: using a CNN-LSTM model based on an attention mechanism to extract time-varying features and group the carbon emission data;

[0051] Scheduling module: Calculate the demand value of each data group according to the data scheduling demand estimation model; Calculate the similarity between each data group and the scheduling link according to the demand value, and schedule the data group.

[0052] Through the above technical solutions, it can be known that compared with the prior art, the present invention discloses a method and system for heterogeneous data group scheduling based on time-varying characteristics and priorities. In the data preprocessing stage, considering the diversity and complexity of heterogeneous data, this paper uses functional data analysis to preprocess the data to ensure the consistency and availability of the data, laying a solid foundation for subsequent data analysis and scheduling; in the data grouping stage, according to the time-varying characteristics of the data and business needs, the massive data is reasonably divided into different groups, providing a clear framework for subsequent scheduling optimization; in the specific scheduling process, the data priority, response time factor, load factor and link performance characteristics are comprehensively considered, and the data priority-oriented data scheduling link similarity (DPLS) is calculated, and the scheduling of data packets is realized accordingly. By grouping and scheduling massive heterogeneous data, the present invention can effectively and orderly manage and utilize massive heterogeneous data, and improve the intelligence level of the "dual carbon" digital monitoring center. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0054] Figure 1 It is a schematic diagram of the overall process of an embodiment of the present invention.

[0055] Figure 2 Schematic diagram of the CNN-LSTM model structure based on the attention mechanism of an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] The embodiment of the present invention discloses a heterogeneous data packet scheduling method based on time-varying characteristics and priorities, such as Figure 1 As shown, the specific steps are as follows:

[0058] Collect multivariate heterogeneous data and use functional data analysis methods to convert multivariate heterogeneous data into carbon emission data in a unified format;

[0059] The CNN-LSTM model based on the attention mechanism is used to extract time-varying features and group data of carbon emission data;

[0060] According to the data scheduling demand estimation model, the demand value of each data group is calculated respectively; the similarity between each data group and the scheduling link is calculated according to the demand value, and the data group is scheduled.

[0061] Specifically, the data collected by the "dual carbon" digital monitoring center comes from a wide range of sources, covering electricity, industry, transportation, construction and other fields. These data have significant differences in format, structure and sampling frequency, forming heterogeneous data sets. In addition, due to the differences in sampling frequency and period, direct integration is difficult. The functional data analysis method can unify heterogeneous data into a functional expression to facilitate subsequent comprehensive analysis and processing. By converting discrete data into continuous functions, the unification and comparability of data of different frequencies at the functional level can be achieved, thereby solving the data fusion problem caused by inconsistent sampling frequencies. In theory, methods such as function fitting and interpolation can be used to increase the number of sampling points and achieve expansion, which will help improve the accuracy and robustness of massive data analysis in the "dual carbon" digital monitoring center.

[0062] In a specific embodiment, the functional data analysis method is specifically:

[0063] The basis function is used to smoothly expand the multivariate heterogeneous data to obtain the fitting function;

[0064] The fitting function is smoothed by adding a penalty term;

[0065] The coefficient matrix of the fitting function is fitted using the least squares method, and the penalty factor coefficient of the penalty term is solved using the generalized cross-validation method to obtain the representation function of the discrete data sequence of multivariate heterogeneous data.

[0066] In a specific embodiment, the fitting function is expressed as:

[0067]

[0068] Among them, F(x) represents the fitting function, x represents the xth sampling point of the scattered data sequence, φ i (x) represents the i-th basis function, i = 1, 2, 3...n, n is the total number of basis functions, c i represents the coefficient vector of the basis function;

[0069] The specific formula of the penalty term ∈(x) is:

[0070] ∈(x)=λC T RC;

[0071] Among them, λ represents the penalty factor, C represents the coefficient vector c i The corresponding coefficient matrix, R represents the matrix related to the second-order derivative of the basis function;

[0072] Represents the function f(x), the formula is:

[0073] f(x)=F(x)+∈(x).

[0074] In a specific embodiment, the basis function adopts a B-spline basis function, and the node sequence is set to be non-decreasing in the interval [a, b]. Then, the B-spline basis function with an order of k at the node is expressed as:

[0075]

[0076] Among them, B i,k (x) represents the B-spline basis function, t represents the time parameter, t i ,t i+1 are the time parameters of the i-th and i+1-th basis functions respectively, and k represents the order of the basis function;

[0077] The B-spline basis function is fitted using a recursive formula, which is:

[0078]

[0079] Specifically, the essence of the basis function smoothing method is to linearly combine multiple basis functions to obtain the functional expression of the massive multi-source heterogeneous discrete data of the original "Dual Carbon" digital intelligent monitoring center. The properties of the basis function are very important for data fitting. Commonly used basis functions include polynomial, Fourier and B-spline basis functions. Since the time series data of the "Dual Carbon" digital monitoring center is non-periodic, the B-spline basis function is usually used, and its fitting function is expressed by a recursive formula.

[0080] When fitting the time series data of the "dual carbon" digital intelligent monitoring center using the basis function smoothing method, the fitting function should express the main characteristics of the original discrete data as much as possible. In order to avoid excessive fluctuations or local changes caused by overfitting of the curve, the roughness penalty factor is introduced to control the degree of curve fitting, so that the function expression of discrete data is more reasonable. The roughness penalty method is based on the basis function method and controls the curvature of the fitting curve by adding a penalty term. The square integral of the second-order derivative is usually used as a penalty term to measure the roughness of the curve: This results in a matrix associated with the second-order derivatives of the basis functions.

[0081] In a specific embodiment, the least squares method is expressed as:

[0082]

[0083] The solution of the coefficient matrix is ​​expressed as:

[0084] C=(G T G+λR) -1 G T y;

[0085] Among them, y x is the original data collected at the x-th sampling point, G is the basis function matrix representing the value of each sampling point G = [φ(1), φ(2)…φ(x)], φ(x) is the superposition vector of all basis functions at the x-th sampling point;

[0086] The generalized cross-validation formula is:

[0087]

[0088] Among them, GCV represents the value of generalized cross validation, which is used to evaluate the goodness of fit of the model, tr() represents the trace of the matrix, which is defined as the sum of the diagonal elements of the matrix; the value of the penalty factor coefficient is determined by minimizing the GCV value.

[0089] The functional data analysis (FDA) method converts the massive multi-source heterogeneous data of the "dual carbon" digital intelligent monitoring center into a continuous function through basis function smoothing, thereby solving the problem of inconsistent frequencies, and achieving data smoothing by adding penalty terms to improve data accuracy. The least squares method calculates coefficients and optimizes the fitting effect. This process unifies the data format and reduces the data dimension, providing a high-quality data foundation for subsequent carbon emission monitoring and energy efficiency evaluation, and promoting the development of digital intelligent applications.

[0090] In a specific embodiment, in order to realize the grouping and scheduling of massive heterogeneous data in the "dual carbon" digital intelligent monitoring center, a CNN-LSTM model based on the attention mechanism is used to group data based on preprocessed data. This model combines the advantages of convolutional neural networks (CNN) in feature extraction with the powerful ability of long short-term memory networks (LSTM) in processing time series data. By introducing the attention mechanism, the model can pay more attention to the key information in the data, thereby improving the accuracy and efficiency of time-varying feature extraction. This method can better complete the grouping task of massive heterogeneous data and provide strong support for subsequent data scheduling. Figure 2 As shown, the CNN-LSTM model includes an input layer, a convolutional layer, a multi-layer LSTM layer, an attention layer, and an output layer connected in sequence;

[0091] The input layer samples the carbon emission data according to the set batch size, time step T, and feature dimension D. A data sample is represented as a real number sequence matrix X∈RT×D , where X represents the data sample, R T×D represents a real number sequence matrix, x t ″ represents the data vector at time t′ in the real number sequence matrix;

[0092] The convolution layer performs convolution, pooling and dimensionality reduction operations on the input data vector in sequence, which can extract the spatial connection between different time feature values ​​in the data, thereby making up for the deficiency of LSTM in capturing the spatial components of the data. When the data enters the CNN layer, convolution, pooling and node expansion (dimensionality reduction) operations are performed in sequence. For sequence data, the model uses one-dimensional convolution, and the convolution kernel only performs convolution in a single time domain direction. The feature after convolution is expressed as: h t′ =f(W·x t ″+b); where h t′ represents the output feature, f() represents the nonlinear activation function, W represents the weight matrix, and b represents the bias;

[0093] The multi-layer LSTM layer is composed of multiple LSTM layers. The LSTM layer has a memory function and realizes the state transfer of time series data through a gating mechanism, including a forget gate, a selective memory gate, and an output gate. The forget gate determines the degree of forgetting of the previous hidden layer state, which is expressed as: f t′ =σ(W f ·[h t′-1 ,x t ″]+b f ), where f t′ represents the value of the forget gate at time t′, σ represents the activation function, W f represents the forget gate weight matrix, h t′-1 represents the hidden layer state at time t′-1, b f Represents the bias of the forget gate. The memory gate is selected to control the degree of adding new information. The unit state is updated through the forget gate and memory gate outputs. The output gate controls the output degree of the current unit. The last hidden layer output of the multi-layer LSTM is used as the grouping result of heterogeneous data and enters the attention layer for further processing.

[0094] Attention layer: The attention layer enhances the role of important time steps in LSTM, thereby further reducing the heterogeneous data grouping error of the model. This layer performs weighted summation of the output of the last hidden layer in the multi-layer LSTM layer by weighted average. The formula is: Among them, c t′ , α t′ 、h t ″ respectively represent the output and weight of the attention layer at time t′ and the output of the multi-layer LSTM layer;

[0095] The output layer outputs the final grouping result based on the output of the attention layer.

[0096] In a specific embodiment, after completing the data grouping of the "dual carbon" digital intelligent monitoring center, differentiated priorities are set for different data groups, and the data scheduling demand estimation model, RTT factor and congestion factor are combined to calculate the data scheduling link similarity (DPLS) to realize data packet scheduling.

[0097] The data scheduling demand estimation model is expressed as:

[0098] Z = f × I × Q;

[0099] Wherein, Z represents the demand estimate, f represents the frequency of data grouping, i.e., the frequency of the original data being divided into the current grouping, I represents the data scheduling amount, i.e., the total amount of data in the current grouping, and Q represents the data scheduling density, i.e., the proportion of the amount of data that needs to be scheduled in the current grouping to the total amount of data;

[0100] The similarity calculation formula is:

[0101]

[0102] RTT represents the response time factor of the link; data packet selection R DPLS The link with the smallest value is used for transmission to ensure the priority, delay and load balance of data transmission.

[0103] The embodiment of the present invention further discloses a heterogeneous data packet scheduling system based on time-varying characteristics and priorities, including:

[0104] Data preprocessing module: collects multivariate heterogeneous data and uses functional data analysis methods to convert multivariate heterogeneous data into carbon emission data in a unified format;

[0105] Feature extraction and grouping module: Use the CNN-LSTM model based on the attention mechanism to extract time-varying features and group data of carbon emission data;

[0106] Scheduling module: Calculate the demand value of each data group according to the data scheduling demand estimation model; calculate the similarity between each data group and the scheduling link according to the demand value, and schedule the data group.

[0107] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0108] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A heterogeneous data packet scheduling method based on time-varying characteristics and priorities, characterized in that: The specific steps are as follows: Collecting multivariate heterogeneous data and converting the multivariate heterogeneous data into carbon emission data in a unified format using a functional data analysis method; Using a CNN-LSTM model based on an attention mechanism, time-varying feature extraction and data grouping are performed on the carbon emission data; According to the data scheduling demand estimation model, the demand value of each data packet is calculated respectively; according to the demand value, the similarity between each data packet and the scheduling link is calculated, and the data packet is scheduled.

2. A method for scheduling heterogeneous data packets based on time-varying characteristics and priorities according to claim 1, characterized in that: The functional data analysis method is specifically as follows: The multivariate heterogeneous data is smoothly expanded using a basis function to obtain a fitting function; Smoothing the fitting function by adding a penalty term; The coefficient matrix of the fitting function is fitted using the least squares method, and the penalty factor coefficient of the penalty term is solved using the generalized cross-validation method to obtain a representation function of the discrete data sequence of multivariate heterogeneous data.

3. A method for scheduling heterogeneous data packets based on time-varying characteristics and priorities according to claim 2, characterized in that: The fitting function is expressed as: Among them, F(x) represents the fitting function, x represents the xth sampling point of the scattered data sequence, φ i (x) represents the i-th basis function, i = 1, 2, 3...n, n is the total number of basis functions, c i represents the coefficient vector of the basis function; The specific formula of the penalty term ∈(x) is: ∈(x)=λC T RC; Among them, λ represents the penalty factor, C represents the coefficient vector c i The corresponding coefficient matrix, R represents the matrix related to the second-order derivative of the basis function; The expression function f(x) is expressed as follows: f(x)=F(x)+∈(x).

4. A method for scheduling heterogeneous data packets based on time-varying characteristics and priorities according to claim 3, characterized in that: The basis function adopts B-spline basis function, which is expressed as: Among them, B i,k (x) represents the B-spline basis function, t represents the time parameter, t i ,t i+1 are the time parameters of the i-th and i+1-th basis functions respectively, and k represents the order of the basis function; The B-spline basis function is fitted using a recursive formula, which is:

5. The method for scheduling heterogeneous data packets based on time-varying characteristics and priorities according to claim 3, characterized in that: The least squares method is expressed as: The solution of the coefficient matrix is ​​expressed as: C=(G T G+λR) -1 G T y; Among them, y x The original data collected at the x-th sampling point, G is the value of each sampling point of the basis function matrix G = [φ(1), φ(2)…φ(x)], φ(x) is the superposition vector of all basis functions at the x-th sampling point; The generalized cross validation formula is: Among them, GCV represents the value of generalized cross validation, which is used to evaluate the goodness of fit of the model, tr() represents the trace of the matrix, which is defined as the sum of the diagonal elements of the matrix; the value of the penalty factor coefficient is determined by minimizing the GCV value.

6. The method for scheduling heterogeneous data packets based on time-varying characteristics and priorities according to claim 1, characterized in that: The CNN-LSTM model includes an input layer, a convolutional layer, a multi-layer LSTM layer, an attention layer and an output layer connected in sequence; The input layer samples the carbon emission data according to the set batch size, time step T, and feature dimension D. A data sample is represented by a real number sequence matrix X∈R T×D , where X represents the data sample, R T×D represents a real number sequence matrix, x t ″ represents the data vector at time t′ in the real number sequence matrix; The convolution layer performs convolution, pooling and dimensionality reduction operations on the input data vector in sequence. The feature after convolution is represented as: h t′ =f(W·x t ″+b); where h t′ represents the output feature, f() represents the nonlinear activation function, W represents the weight matrix, and b represents the bias; The multi-layer LSTM layer realizes the state transfer of time series data through a gating mechanism, including a forget gate, a selective memory gate, and an output gate. The forget gate determines the degree of forgetting of the previous hidden layer state, which is expressed as: t′ =σ(W f ·[h t′-1 ,x t ″]+b f ), where f t′ represents the value of the forget gate at time t′, σ represents the activation function, W f represents the forget gate weight matrix, h t′-1 represents the hidden layer state at time t′-1, b f Represents the bias of the forget gate; The attention layer performs weighted summation on the output of the last hidden layer in the multi-layer LSTM layer by weighted average, and the formula is: Among them, c t′ , α t′ 、h t ″ respectively represent the output and weight of the attention layer at time t′ and the output of the multi-layer LSTM layer; The output layer outputs the final grouping result according to the output of the attention layer.

7. The method for scheduling heterogeneous data packets based on time-varying characteristics and priorities according to claim 1, characterized in that: The data scheduling demand estimation model is expressed as: Z = f × I × Q; Where Z represents the demand estimate, f represents the frequency of data grouping, I represents the data scheduling amount, and Q represents the data scheduling density; The calculation formula of the similarity is: Wherein, RTT represents the response time factor of the link; the data packet selects R DPLS The link with the smallest value is used for transmission.

8. A heterogeneous data packet scheduling system based on time-varying characteristics and priorities, applying a heterogeneous data packet scheduling method based on time-varying characteristics and priorities as described in any one of claims 1 to 7, characterized in that: include: Data preprocessing module: collects multivariate heterogeneous data and converts the multivariate heterogeneous data into carbon emission data in a unified format using a functional data analysis method; Feature extraction and grouping module: using a CNN-LSTM model based on an attention mechanism to extract time-varying features and group the carbon emission data; Scheduling module: Calculate the demand value of each data group according to the data scheduling demand estimation model; Calculate the similarity between each data group and the scheduling link according to the demand value, and schedule the data group.

Images