A method and system for characterizing power user price response behavior based on graph propagation

Abstract

The present invention relates to a method and system for characterizing electricity user price response behavior based on graph propagation, comprising: obtaining an electricity user price response behavior graph corresponding to a user's single-day time-of-use electricity price electricity consumption data; reconstructing the electricity user price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data using the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data; and correcting the pre-response behavior label based on the electricity user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price electricity consumption data and the real response behavior label. The technical solution provided by the present invention accurately characterizes the electricity user price response behavior by utilizing a graph propagation algorithm, the inherent characteristics of the electricity user price response behavior, and the similarity of the electricity price response behavior between electricity users, thereby solving the problem of a small amount of electricity user time-of-use electricity price electricity consumption data with real response behavior labels, thereby providing a basis for electricity users to participate in the formulation of policies for electricity price response and the prediction of implementation effects.

Description

Technical Field

[0001] The present invention relates to the field of user response behavior characterization, and in particular to a method and system for characterizing power user electricity price response behavior based on graph propagation. Background Art

[0002] With the development of smart grids, the response behavior of power users will have an increasingly greater impact on grid operations. It is particularly important to build an analysis model of power user response behavior and guide their positive interaction with the grid.

[0003] The patterns in electricity users' price-responsive behavior are often hidden in massive amounts of data. However, this data is often not rigorously categorized and often exhibits characteristics such as weak labeling, incompleteness, low value, and mixed categories. Traditional supervised learning algorithms, such as feature engineering and neural networks, have high sample data requirements and require large amounts of realistically labeled sample data for learning.

[0004] However, China is still in the development stage of its electricity market, and its electricity price mechanism is still imperfect. It is impossible to obtain a wealth of electricity price response samples from electricity users. That is, the sample information is insufficient and incomplete, and the number of samples with real labels is small. Therefore, it is impossible to accurately characterize the user's electricity price response behavior.

[0005] After preliminary search, no patents or literature have been found that can accurately characterize users' electricity price response behavior by utilizing a large number of insufficient, incomplete, and low-label electricity price response samples of electricity users. Summary of the Invention

[0006] In response to the deficiencies in the prior art, the purpose of the present invention is a method for characterizing electricity user electricity price response behavior based on graph propagation. The method constructs an electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data based on the intrinsic characteristics of the electricity user's electricity price response behavior and the similarity of the electricity price response behavior between electricity users. Based on the graph propagation algorithm and the electricity user electricity price response behavior graph, the operation of transferring the response behavior label of the electricity user's time-of-use electricity consumption data carrying the real response behavior label to the electricity user's time-of-use electricity consumption data not carrying the real response behavior label is completed, thereby achieving accurate characterization of the electricity user's electricity price response behavior, solving the problem of the lack of electricity user's time-of-use electricity consumption data with the real response behavior label, and providing a basis for electricity users to participate in the formulation of policies and prediction of implementation effects of electricity price response.

[0007] The purpose of the present invention is achieved by adopting the following technical solutions:

[0008] The present invention provides a method for characterizing electricity price response behavior of power users based on graph propagation, wherein the method comprises:

[0009] Obtain the electricity price response behavior diagram of the power user corresponding to the user's single-day time-of-use electricity price data;

[0010] Reconstructing a user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data using the response behavior label corresponding to the user's single-day time-of-use electricity price data; the response behavior label includes a real response behavior label and a pre-response behavior label;

[0011] Based on the electricity price response behavior graph of the power user corresponding to the reconstructed single-day time-of-use electricity consumption data and the actual response behavior label, the pre-response behavior label is corrected.

[0012] Preferably, the obtaining of the electricity price response behavior diagram of the power user corresponding to the user's single-day time-of-use electricity price data includes:

[0013] The initial power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is defined as G = (V, E, W), where V is the node set of the initial power user price response behavior graph, V = {v1…v L …,v N},i,j,L∈(1~N),v L is the node corresponding to the Lth user's single-day time-of-use electricity consumption data in the initial electricity user electricity price response behavior graph, N is the total number of nodes in the initial electricity user electricity price response behavior graph, E is the set of edges between each node in the node set of the initial electricity user electricity price response behavior graph, and e i,j ∈E,e i,j is the edge between the i-th node and the j-th node in the initial power user price response behavior graph, W is an N×N order weight matrix, w i,j ∈W, w i,j is the weight value of the edge between the i-th node and the j-th node in the initial power user electricity price response behavior graph;

[0014] The initial power user electricity price response behavior graph is thinned out using the Directed kNN thinning technology to obtain the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data.

[0015] Preferably, the process of obtaining the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes:

[0016] If the acquired user's single-day time-of-use electricity price data carries a response behavior label, then the response behavior label carried by the user's single-day time-of-use electricity price data is the real response behavior label; otherwise, the labeling system is used to add a pre-response behavior label to the user's single-day time-of-use electricity price data;

[0017] Among them, the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is {p L1 …p Lτ …p Lm}, p Lτ is the probability that the response level of the Lth user's single-day time-of-use electricity consumption data is τ, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, L∈(1~N), N is the total number of nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data;

[0018] If the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is p Lτ =1, the response behavior level displayed by the response behavior label of the Lth node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is τ, τ∈(1~m), and m is the total number of response behavior levels corresponding to the user's single-day time-of-use electricity consumption data.

[0019] Furthermore, the weight value w of the edge between the i-th node and the j-th node in the node set of the initial power user electricity price response behavior graph is obtained according to the following formula: i,j :

[0020] w i,j =w1sim i,j1 +w2sim i,j2 +w3sim i,j3

[0021] In the above formula, w1 is the preset weight corresponding to the first similarity, w2 is the preset weight corresponding to the second similarity, w3 is the weight corresponding to the third similarity, sim i,j1 is the first similarity between the i-th node and the j-th node in the initial power user price response behavior graph, sim i,j2 is the second similarity between the i-th node and the j-th node in the initial power user price response behavior graph, sim i,j3 is the third similarity between the i-th node and the j-th node in the initial power user electricity price response behavior graph, w1+w2+w3=1.

[0022] Furthermore, the first similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j1 :

[0023]

[0024] In the above formula, k∈(1~λ), λ is the total number of sampling points of the user's single-day time-of-use electricity price data, is the daily time-of-use electricity price data of the user at the kth sampling point corresponding to the i-th node in the initial power user electricity price response behavior graph, is the daily time-of-use electricity price data of the user at the kth sampling point corresponding to the jth node in the initial power user electricity price response behavior diagram, avg(d i ) is the average of the user's single-day time-of-use electricity price data of the λ sampling points corresponding to the i-th node in the initial power user electricity price response behavior diagram, avg(d j ) is the mean of the user's single-day time-of-use electricity price data of the λ sampling points corresponding to the jth node in the initial power user electricity price response behavior graph;

[0025] The second similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j2 :

[0026]

[0027] In the above formula, is the per-unit value of the user's single-day time-of-use electricity price data at the kth sampling point corresponding to the i-th node in the initial power user electricity price response behavior graph, is the per-unit value of the user's single-day time-of-use electricity price data at the kth sampling point corresponding to the jth node in the initial power user electricity price response behavior graph,

[0028] The third similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j3 :

[0029]

[0030] In the above formula, is the u-th dimension feature data value corresponding to the i-th node in the initial power user electricity price response behavior graph, is the u-th dimension feature data value corresponding to the j-th node in the initial power user electricity price response behavior graph, u∈(1~R), and R is the total number of dimensions of feature data corresponding to the nodes in the initial power user electricity price response behavior graph.

[0031] Preferably, the reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data by using the behavior label corresponding to the user's single-day time-of-use electricity price data includes:

[0032] If the i-th node and the j-th node in the electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula:

[0033] w' i,j =w i,j ×a

[0034] Among them, w i,j is the original weight value of the edge between the i-th node and the j-th node in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data, a is a coefficient greater than 1, i,j∈(1~N), N is the total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data.

[0035] Preferably, the reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data by using the behavior label corresponding to the user's single-day time-of-use electricity price data includes:

[0036] The weight value of the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data is modified according to the following formula:

[0037]

[0038] Among them, a is a coefficient greater than 1, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, c i is the response behavior level displayed in the response behavior label corresponding to the daily time-of-use electricity consumption data of the i-th user, c j is the response behavior level displayed in the response behavior label corresponding to the j-th user's single-day time-of-use electricity consumption data, i,j∈(1~N), N is the total number of nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data.

[0039] Preferably, the reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data by using the behavior label corresponding to the user's single-day time-of-use electricity price data includes:

[0040] If the difference between the response behavior level displayed by the response behavior label of the i-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data and the response behavior level displayed by the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is removed, where i,j∈(1,N), N is the total number of nodes in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data.

[0041] Preferably, the reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data by using the behavior label corresponding to the user's single-day time-of-use electricity price data includes:

[0042] If the difference between the response behavior level displayed in the response behavior label of the i-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data and the response behavior level displayed in the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is removed, where i,j∈(1,N), N is the total number of nodes in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data;

[0043] If the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, then the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula:

[0044] w' i,j =w i,j ×a

[0045] Among them, w i,j is the original weight value of the edge between the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data, and a is a coefficient greater than 1.

[0046] Preferably, the modifying of the pre-response behavior label based on the power user price response behavior graph and the actual response behavior label corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data includes:

[0047] S1: If the i-th node and the j-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data are not connected, then the propagation probability T of the response behavior label of the j-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data propagating to the response behavior label of the i-th node ij =0;

[0048] Otherwise, the weight value w' of the edge between the i-th node and the j-th node in the electricity price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data is i,j , calculate the propagation probability T of the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node ij :

[0049]

[0050] Where k∈V C , Vc is the set of nodes connected to the i-th node in the node set of the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data, w' i,k is the weight value of the edge connecting the i-th node in the node set of the electricity user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data, i,j∈(1,N), N is the total number of nodes in the electricity user price response behavior graph corresponding to the user's single-day time-of-use electricity price data;

[0051] S2: Update the response behavior label of each node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data according to the propagation probability of the response behavior label of other nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to the response behavior label of each node;

[0052] S3: Restore the actual response behavior labels of the nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to their initial values

[0053] S4: Repeat S2-S3 until the convergence condition is reached;

[0054] The convergence condition is that the changes in the probability distribution of the response behavior labels of all nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data are less than a set threshold.

[0055] Furthermore, the propagation probability of the response behavior label of each node connected to each node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data is propagated to the response behavior label of each node, and the probability distribution of the response behavior label of each node is updated, including:

[0056] The probability p of the response level being τ in the response behavior label of the i-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data is updated as follows: iτ :

[0057]

[0058] Where, T ik is the propagation probability of the response behavior label of the kth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node, p kτ is the probability that the response level is τ in the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity consumption data, τ∈(1~m), m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data.

[0059] The present invention provides a system for characterizing electricity user price response behavior based on graph propagation, wherein the improvement is that the system comprises:

[0060] An acquisition module is used to obtain a user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data;

[0061] A reconstruction module is used to reconstruct a power user price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data using the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data; the response behavior label includes a real response behavior label and a pre-response behavior label;

[0062] The correction module is used to correct the pre-response behavior label based on the power user price response behavior diagram corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data and the real response behavior label.

[0063] Compared with the closest prior art, the present invention has the following beneficial effects:

[0064] The technical solution provided by the present invention obtains a power user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data; uses the response behavior label corresponding to the user's single-day time-of-use electricity price data to reconstruct the power user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data; and based on the reconstructed power user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data and the actual response behavior label, corrects the pre-response behavior label. This solution solves the problem of a small amount of power user time-of-use electricity price data with actual response behavior labels, achieves accurate characterization of power user electricity price response behavior, and thus provides a basis for power users to participate in the formulation of policies for electricity price response and the prediction of implementation effects.

[0065] Compared with traditional supervised learning algorithms, the technical solution provided by the present invention has lower requirements on samples used to characterize the electricity price response behavior of power users and has higher credibility in characterizing the electricity price response behavior of power users.

[0066] The technical solution provided by the present invention is easy to understand, has strong scalability, is suitable for characterizing the response behavior of other types of power users, and has strong versatility, robustness and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a flow chart of a method for characterizing the electricity price response behavior of power users based on graph propagation;

[0068] Figure 2 It is a system structure diagram for describing the electricity price response behavior of power users based on graph propagation. DETAILED DESCRIPTION

[0069] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0070] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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 making creative efforts shall fall within the scope of protection of the present invention.

[0071] The present invention provides a method for describing the electricity price response behavior of power users based on graph propagation. Figure 1 As shown, the method includes:

[0072] Step 101 is used to obtain a user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data;

[0073] Step 102 is for reconstructing a power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data using the response behavior label corresponding to the user's single-day time-of-use electricity price data; the response behavior label includes a real response behavior label and a pre-response behavior label;

[0074] Step 103 is used to modify the pre-response behavior label based on the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data and the actual response behavior label.

[0075] Specifically, step 101 includes:

[0076] Step 101-1 is used to define the initial power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data as G=(V, E, W), where V is the node set of the initial power user price response behavior graph, V={v1…v L …,v N},i,j,L∈(1~N),v L is the node corresponding to the Lth user's single-day time-of-use electricity consumption data in the initial electricity user electricity price response behavior graph, N is the total number of nodes in the initial electricity user electricity price response behavior graph, E is the set of edges between each node in the node set of the initial electricity user electricity price response behavior graph, and e i,j ∈E,e i,j is the edge between the i-th node and the j-th node in the initial power user price response behavior graph, W is an N×N order weight matrix, w i,j ∈W, w i,j is the weight value of the edge between the i-th node and the j-th node in the initial power user electricity price response behavior graph;

[0077] Step 101 - 2 is for thinning the initial power user price response behavior graph using a Directed kNN thinning technique to obtain a power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data.

[0078] In a specific embodiment of the present invention, the specific operation of sparsifying the initial power user electricity price response behavior graph using the Directed kNN sparsification technology may be: if the i-th node in the initial power user electricity price response behavior graph is connected to the j-th node, and the weight value of the edge between the i-th node and the j-th node in the initial power user electricity price response behavior graph is one of the largest k in the set of weight values ​​of the edges between the i-th node and other nodes in the initial power user electricity price response behavior graph, then the edge between the i-th node and the j-th node in the initial power user electricity price response behavior graph is retained; otherwise, the edge between the i-th node and the j-th node in the initial power user electricity price response behavior graph is removed, where k is a preset value.

[0079] Specifically, the process of obtaining the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes:

[0080] If the acquired user's single-day time-of-use electricity price data carries a response behavior label, then the response behavior label carried by the user's single-day time-of-use electricity price data is the real response behavior label; otherwise, the labeling system is used to add a pre-response behavior label to the user's single-day time-of-use electricity price data;

[0081] Among them, the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is {p L1 …p Lτ …p Lm}, p Lτ is the probability that the response level of the Lth user's single-day time-of-use electricity consumption data is τ, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, L∈(1~N), N is the total number of nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data;

[0082] If the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is p Lτ =1, the response behavior level displayed by the response behavior label of the Lth node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is τ, τ∈(1~m), and m is the total number of response behavior levels corresponding to the user's single-day time-of-use electricity consumption data.

[0083] Furthermore, the weight value w of the edge between the i-th node and the j-th node in the node set of the initial power user electricity price response behavior graph is obtained according to the following formula: i,j :

[0084] w i,j =w1sim i,j1 +w2sim i,j2 +w3sim i,j3

[0085] In the above formula, w1 is the preset weight corresponding to the first similarity, w2 is the preset weight corresponding to the second similarity, w3 is the weight corresponding to the third similarity, sim i,j1 is the first similarity between the i-th node and the j-th node in the initial power user price response behavior graph, sim i,j2 is the second similarity between the i-th node and the j-th node in the initial power user price response behavior graph, sim i,j3 is the third similarity between the i-th node and the j-th node in the initial power user electricity price response behavior graph, w1+w2+w3=1.

[0086] Furthermore, the first similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j1 :

[0087]

[0088] In the above formula, k∈(1~λ), λ is the total number of sampling points of the user's single-day time-of-use electricity price data, is the daily time-of-use electricity price data of the user at the kth sampling point corresponding to the i-th node in the initial power user electricity price response behavior graph, is the daily time-of-use electricity price data of the user at the kth sampling point corresponding to the jth node in the initial power user electricity price response behavior diagram, avg(d i ) is the average of the user's single-day time-of-use electricity price data of the λ sampling points corresponding to the i-th node in the initial power user electricity price response behavior diagram, avg(d j ) is the mean of the user's single-day time-of-use electricity price data of the λ sampling points corresponding to the jth node in the initial power user electricity price response behavior graph;

[0089] The second similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j2 :

[0090]

[0091] In the above formula, is the per-unit value of the user's single-day time-of-use electricity price data at the kth sampling point corresponding to the i-th node in the initial power user electricity price response behavior graph, is the per-unit value of the user's single-day time-of-use electricity price data at the kth sampling point corresponding to the jth node in the initial power user electricity price response behavior graph,

[0092] The third similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j3 :

[0093]

[0094] In the above formula, is the u-th dimension feature data value corresponding to the i-th node in the initial power user electricity price response behavior graph, is the u-th dimension feature data value corresponding to the j-th node in the initial power user electricity price response behavior graph, u∈(1~R), and R is the total number of dimensions of feature data corresponding to the nodes in the initial power user electricity price response behavior graph.

[0095] In the best embodiment of the present invention, the characteristic data corresponding to the nodes in the initial power user electricity price response behavior diagram include: the mean, maximum value, minimum value, variance, load type, temperature factor and observation time of the user's single-day time-of-use electricity consumption data corresponding to each node in the initial power user electricity price response behavior diagram.

[0096] Specifically, step 102 is specifically used to:

[0097] If the i-th node and the j-th node in the electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula:

[0098] w' i,j =w i,j ×a

[0099] Among them, w i,j is the original weight value of the edge between the i-th node and the j-th node in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data, a is a coefficient greater than 1, i,j∈(1~N), N is the total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data.

[0100] Specifically, the step 102 is further specifically used to:

[0101] The weight value of the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data is modified according to the following formula:

[0102]

[0103] Among them, a is a coefficient greater than 1, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, c iis the response behavior level displayed in the response behavior label corresponding to the daily time-of-use electricity consumption data of the i-th user, c j is the response behavior level displayed in the response behavior label corresponding to the j-th user's single-day time-of-use electricity consumption data, i,j∈(1~N), N is the total number of nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data.

[0104] Specifically, the step 102 is further specifically used to:

[0105] If the difference between the response behavior level displayed by the response behavior label of the i-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data and the response behavior level displayed by the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is removed, where i,j∈(1,N), N is the total number of nodes in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data.

[0106] Specifically, the step 102 is further specifically used to:

[0107] If the difference between the response behavior level displayed in the response behavior label of the i-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data and the response behavior level displayed in the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is removed, where i,j∈(1,N), N is the total number of nodes in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data;

[0108] If the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, then the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula:

[0109] w' i,j =w i,j ×a

[0110] Among them, w i,jis the original weight value of the edge between the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data, and a is a coefficient greater than 1.

[0111] Specifically, step 103 includes:

[0112] Step 103-1 is used to calculate the propagation probability T of the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to the response behavior label of the i-th node if the i-th node and the j-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data are not connected. ij =0;

[0113] Otherwise, the weight value w' of the edge between the i-th node and the j-th node in the electricity price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data is i,j , calculate the propagation probability T of the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node ij :

[0114]

[0115] Where k∈V C , Vc is the set of nodes connected to the i-th node in the node set of the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data, w' i,k is the weight value of the edge connecting the i-th node in the node set of the electricity user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data, i,j∈(1,N), N is the total number of nodes in the electricity user price response behavior graph corresponding to the user's single-day time-of-use electricity price data;

[0116] Step 103-2 is used to update the response behavior label of each node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data based on the propagation probability of the response behavior label of other nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to the response behavior label of each node;

[0117] Step 103 - 3 , for restoring the actual response behavior labels of the nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to their initial values;

[0118] Step 103-4 is used to repeat steps 103-2 to 103-3 until the convergence condition is reached;

[0119] The convergence condition is that the changes in the probability distribution of the response behavior labels of all nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data are less than a set threshold.

[0120] Furthermore, the step 103-2 is specifically used to:

[0121] The probability p of the response level being τ in the response behavior label of the i-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data is updated as follows: iτ :

[0122]

[0123] Where, T ik is the propagation probability of the response behavior label of the kth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node, p kτ is the probability that the response level is τ in the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity consumption data, τ∈(1~m), m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data.

[0124] The present invention provides a system for describing the electricity price response behavior of power users based on graph propagation. Figure 2 As shown, the system includes:

[0125] An acquisition module is used to obtain a user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data;

[0126] A reconstruction module is used to reconstruct a power user price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data using the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data; the response behavior label includes a real response behavior label and a pre-response behavior label;

[0127] The correction module is used to correct the pre-response behavior label based on the power user price response behavior diagram corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data and the real response behavior label.

[0128] Specifically, the acquisition module includes:

[0129] A definition unit is used to define the initial power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data as G = (V, E, W), where V is the node set of the initial power user price response behavior graph, V = {v1…v L …,v N},i,j,L∈(1~N),v L is the node corresponding to the Lth user's single-day time-of-use electricity consumption data in the initial electricity user electricity price response behavior graph, N is the total number of nodes in the initial electricity user electricity price response behavior graph, E is the set of edges between each node in the node set of the initial electricity user electricity price response behavior graph, and e i,j ∈E,e i,j is the edge between the i-th node and the j-th node in the initial power user price response behavior graph, W is an N×N order weight matrix, w i,j ∈W, w i,j is the weight value of the edge between the i-th node and the j-th node in the initial power user electricity price response behavior graph;

[0130] The acquisition unit is configured to perform thinning on the initial power user price response behavior graph using a Directed kNN thinning technique to obtain the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data.

[0131] Specifically, the process of obtaining the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes:

[0132] If the acquired user's single-day time-of-use electricity price data carries a response behavior label, then the response behavior label carried by the user's single-day time-of-use electricity price data is the real response behavior label; otherwise, the labeling system is used to add a pre-response behavior label to the user's single-day time-of-use electricity price data;

[0133] Among them, the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is {p L1 …p Lτ …p Lm}, p Lτ is the probability that the response level of the Lth user's single-day time-of-use electricity consumption data is τ, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, L∈(1~N), N is the total number of nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data;

[0134] If the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is p Lτ =1, the response behavior level displayed by the response behavior label of the Lth node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is τ, τ∈(1~m), and m is the total number of response behavior levels corresponding to the user's single-day time-of-use electricity consumption data.

[0135] In the best embodiment of the present invention, the labeling system provides a suggested response behavior label, i.e., a pre-response behavior label, for each user's single-day time-of-use electricity consumption data according to preset labeling rules;

[0136] Among them, the preset labeling rules are determined based on human experience and actual working conditions.

[0137] In the best embodiment of the present invention, the response behavior label corresponding to the Lth user's single-day time-of-use electricity consumption data is {p L1 …p Lτ …p Lm Only one element has a value of 1, and the rest have values ​​of 0.

[0138] Furthermore, the weight value w of the edge between the i-th node and the j-th node in the node set of the initial power user electricity price response behavior graph is obtained according to the following formula: i,j :

[0139] w i,j =w1sim i,j1 +w2sim i,j2 +w3sim i,j3

[0140] In the above formula, w1 is the preset weight corresponding to the first similarity, w2 is the preset weight corresponding to the second similarity, w3 is the weight corresponding to the third similarity, sim i,j1 is the first similarity between the i-th node and the j-th node in the initial power user price response behavior graph, sim i,j2 is the second similarity between the i-th node and the j-th node in the initial power user price response behavior graph, sim i,j3 is the third similarity between the i-th node and the j-th node in the initial power user electricity price response behavior graph, w1+w2+w3=1.

[0141] Furthermore, the first similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j1 :

[0142]

[0143] In the above formula, k∈(1~λ), λ is the total number of sampling points of the user's single-day time-of-use electricity price data, is the daily time-of-use electricity price data of the user at the kth sampling point corresponding to the i-th node in the initial power user electricity price response behavior graph, is the daily time-of-use electricity price data of the user at the kth sampling point corresponding to the jth node in the initial power user electricity price response behavior diagram, avg(di ) is the average of the user's single-day time-of-use electricity price data of the λ sampling points corresponding to the i-th node in the initial power user electricity price response behavior diagram, avg(d j ) is the mean of the user's single-day time-of-use electricity price data of the λ sampling points corresponding to the jth node in the initial power user electricity price response behavior graph;

[0144] The second similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j2 :

[0145]

[0146] In the above formula, is the per-unit value of the user's single-day time-of-use electricity price data at the kth sampling point corresponding to the i-th node in the initial power user electricity price response behavior graph, is the per-unit value of the user's single-day time-of-use electricity price data at the kth sampling point corresponding to the jth node in the initial power user electricity price response behavior graph,

[0147] The third similarity sim between the i-th node and the j-th node in the initial power user price response behavior graph is obtained according to the following formula: i,j3 :

[0148]

[0149] In the above formula, is the u-th dimension feature data value corresponding to the i-th node in the initial power user electricity price response behavior graph, is the u-th dimension feature data value corresponding to the j-th node in the initial power user electricity price response behavior graph, u∈(1~R), and R is the total number of dimensions of feature data corresponding to the nodes in the initial power user electricity price response behavior graph.

[0150] Specifically, the reconstruction module is specifically used to:

[0151] If the i-th node and the j-th node in the electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula:

[0152] w' i,j =w i,j ×a

[0153] Among them, w i,jis the original weight value of the edge between the i-th node and the j-th node in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data, a is a coefficient greater than 1, i,j∈(1~N), N is the total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data.

[0154] Specifically, the reconstruction module is further configured to:

[0155] The weight value of the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data is modified according to the following formula:

[0156]

[0157] Among them, a is a coefficient greater than 1, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, c i is the response behavior level displayed in the response behavior label corresponding to the daily time-of-use electricity consumption data of the i-th user, c j is the response behavior level displayed in the response behavior label corresponding to the j-th user's single-day time-of-use electricity consumption data, i,j∈(1~N), N is the total number of nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data.

[0158] Specifically, the reconstruction module is further specifically used to:

[0159] If the difference between the response behavior level displayed by the response behavior label of the i-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data and the response behavior level displayed by the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is removed, where i,j∈(1,N), N is the total number of nodes in the electricity user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data.

[0160] In a preferred embodiment of the present invention, the preset threshold may be set to m / 2.

[0161] Specifically, the reconstruction module is further specifically used to:

[0162] If the difference between the response behavior level displayed in the response behavior label of the i-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data and the response behavior level displayed in the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data is removed, where i,j∈(1,N), N is the total number of nodes in the power user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data;

[0163] If the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, then the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula:

[0164] w' i,j =w i,j ×a

[0165] Among them, w i,j is the original weight value of the edge between the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data, and a is a coefficient greater than 1.

[0166] Specifically, the correction module includes:

[0167] The calculation unit is used to calculate the propagation probability T of the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to the response behavior label of the i-th node if the i-th node and the j-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data are not connected. ij =0;

[0168] If the i-th node in the power user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity consumption data is connected to the j-th node, and the response behavior label marked on the i-th node in the power user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity consumption data is the true response behavior label, then the propagation probability T of the response behavior label of the j-th node in the power user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity consumption data propagating to the response behavior label of the i-th node is ij =0;

[0169] Otherwise, the weight value w of the edge between the i-th node and the j-th node in the electricity price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data is i ' ,j , calculate the propagation probability T of the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node ij :

[0170]

[0171] Where k∈V C , Vc is the set of nodes connected to the i-th node in the node set of the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data, w' i,k is the weight value of the edge connecting the i-th node in the node set of the electricity user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data, i,j∈(1,N), N is the total number of nodes in the electricity user price response behavior graph corresponding to the user's single-day time-of-use electricity price data;

[0172] an updating unit, configured to update the response behavior label of each node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data according to a propagation probability of the response behavior label of other nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data being propagated to the response behavior label of each node;

[0173] A restoration unit, configured to restore the actual response behavior labels of the nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to initial values;

[0174] A convergence unit, used to repeat the operations of the update unit and the convergence unit until a convergence condition is reached;

[0175] The convergence condition is that the changes in the probability distribution of the response behavior labels of all nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data are less than a set threshold.

[0176] Furthermore, the updating unit is specifically configured to:

[0177] The probability p of the response level being τ in the response behavior label of the i-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data is updated as follows: iτ :

[0178]

[0179] Where, T ik is the propagation probability of the response behavior label of the kth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node, p kτ is the probability that the response level is τ in the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity consumption data, τ∈(1~m), m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data.

[0180] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0181] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0182] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0183] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1A step that specifies a function in one or more boxes.

[0184] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for characterizing the electricity price response behavior of power users based on graph propagation, characterized in that: The method comprises: Obtain the electricity price response behavior diagram of the power user corresponding to the user's single-day time-of-use electricity price data; Reconstructing a user electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data using the response behavior label corresponding to the user's single-day time-of-use electricity price data, wherein the response behavior label includes a real response behavior label and a pre-response behavior label; Based on the electricity price response behavior graph of the power user corresponding to the reconstructed single-day time-of-use electricity price data and the actual response behavior label, the pre-response behavior label is modified; The method of obtaining the electricity price response behavior diagram of the power user corresponding to the user's single-day time-of-use electricity price electricity consumption data includes: The initial electricity user price response behavior diagram corresponding to the user's single-day time-of-use electricity consumption data is defined as , where V is the node set of the initial power user price response behavior graph, , , For the The node corresponding to the user's single-day time-of-use electricity consumption data in the initial power user electricity price response behavior diagram is is the total number of nodes in the initial electricity user price response behavior graph, E is the set of edges between nodes in the node set of the initial electricity user price response behavior graph, , The first Node and The edges between nodes, for The weight matrix of order, , The first Node and The weight of the edge between the nodes; Using the Directed kNN sparsification technology to sparsify the initial power user price response behavior graph, to obtain the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data; The first node in the node set of the initial power user price response behavior graph is obtained according to the following formula: Node and The weight of the edge between nodes : In the above formula, is the preset weight corresponding to the first similarity, is the preset weight corresponding to the second similarity, is the weight corresponding to the third similarity, The first Node and The first similarity between nodes, The first Node and The second similarity between nodes, The first Node and The third similarity between nodes, ; The first step in the initial power user price response behavior diagram is obtained according to the following formula Node and The first similarity between nodes : In the above formula, , is the total number of sampling points for the user’s daily time-of-use electricity price data, The first The node corresponding to The daily time-of-use electricity consumption data of users at each sampling point, The first The node corresponding to The daily time-of-use electricity consumption data of users at each sampling point, The first The corresponding nodes The average of the daily time-of-use electricity consumption data of users at the sampling points, The first The corresponding nodes The average of the daily time-of-use electricity consumption data of users at the sampling points; The first step in the initial power user price response behavior diagram is obtained according to the following formula Node and The second similarity between nodes : ； In the above formula, The first The node corresponding to The per-unit value of the daily time-of-use electricity consumption data of users at each sampling point is: The first The node corresponding to The per-unit value of the daily time-of-use electricity consumption data of users at each sampling point is: The first step in the initial power user price response behavior diagram is obtained according to the following formula Node and The third similarity between nodes : in, Depend on Calculated, in the above formula, The first The node corresponding to dimensional feature data values, The first The node corresponding to dimensional feature data values, , is the total number of dimensions of the feature data corresponding to the nodes in the initial power user electricity price response behavior graph.

2. The method according to claim 1, wherein The process of obtaining the response behavior label corresponding to the user's single-day time-of-use electricity consumption data includes: If the acquired user's single-day time-of-use electricity price data carries a response behavior label, then the response behavior label carried by the user's single-day time-of-use electricity price data is the real response behavior label; otherwise, the labeling system is used to add a pre-response behavior label to the user's single-day time-of-use electricity price data; Among them, The response behavior label corresponding to the user's single-day time-of-use electricity consumption data is , For the The response level of the daily time-of-use electricity consumption data of a user is The probability of is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, , The total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data; Jordi The response behavior label corresponding to the user's single-day time-of-use electricity consumption data , then the electricity price response behavior diagram of the power user corresponding to the user's single-day time-of-use electricity price data is The response behavior label of each node shows the response behavior level. , , It is the total number of response behavior levels corresponding to the user's single-day time-of-use electricity consumption data.

3. The method according to claim 1, wherein The method of reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price electricity consumption data by using the behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes: If the i-th node and the j-th node in the electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula: in, is the original weight value of the edge between the i-th node and the j-th node in the electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data, a is a coefficient greater than 1, , The total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data.

4. The method according to claim 1, wherein The method of reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price electricity consumption data by using the behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes: The weight value of the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data is modified according to the following formula: Among them, a is a coefficient greater than 1, m is the total number of response behavior levels of the user's single-day time-of-use electricity consumption data, is the response behavior level displayed in the response behavior label corresponding to the daily time-of-use electricity consumption data of the i-th user, is the response behavior level displayed in the response behavior label corresponding to the daily time-of-use electricity consumption data of the jth user, , The total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data.

5. The method according to claim 1, wherein The method of reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price electricity consumption data by using the behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes: If the difference between the response behavior level displayed by the response behavior label of the i-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data and the response behavior level displayed by the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is removed, wherein, , The total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity consumption data.

6. The method according to claim 1, wherein The method of reconstructing the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price electricity consumption data by using the behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data includes: If the difference between the response behavior level displayed in the response behavior label of the i-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data and the response behavior level displayed in the response behavior label of the j-th node does not exceed the preset threshold, then the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is retained; otherwise, the edge between the i-th node and the j-th node in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data is removed, wherein, , The total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data; If the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price data are connected and have the same behavior label, then the weight value of the edge between the i-th node and the j-th node is amplified according to the following formula: in, is the original weight value of the edge between the i-th node and the j-th node in the node set of the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data, and a is a coefficient greater than 1.

7. The method according to claim 1, wherein The method of modifying the pre-response behavior label based on the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data and the actual response behavior label includes: S1: If the i-th node and the j-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data are not connected, then the propagation probability of the response behavior label of the j-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data propagating to the response behavior label of the i-th node ; Otherwise, the weight value of the edge between the i-th node and the j-th node in the electricity price response behavior graph corresponding to the reconstructed user's single-day time-of-use electricity price data is , calculate the propagation probability of the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data to the response behavior label of the i-th node : in, , is the set of nodes connected to the i-th node in the node set of the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data, is the weight value of the edge connecting the i-th node to the k-th node in the node set of the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data, , The total number of nodes in the electricity price response behavior graph of the power user corresponding to the user's single-day time-of-use electricity price data; S2: Update the response behavior label of each node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data according to the propagation probability of the response behavior label of other nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to the response behavior label of each node; S3: Restore the actual response behavior labels of the nodes in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to their initial values; S4: Repeat S2-S3 until the convergence condition is reached; The convergence condition is that the changes in the probability distribution of the response behavior labels of all nodes in the power user price response behavior graph corresponding to the user's single-day time-of-use electricity consumption data are less than a set threshold.

8. The method according to claim 7, wherein The method updates the probability distribution of the response behavior labels of each node by propagating the response behavior labels of the nodes connected to each node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data to the response behavior labels of each node, including: The response level in the response behavior label of the i-th node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity price data is updated as follows: Probability : Where, The first one in the electricity price response behavior diagram is the electricity price response behavior diagram of the power user corresponding to the reconstructed single-day time-of-use electricity price data of the user. The propagation probability of the response behavior label of the node to the response behavior label of the i-th node, The response level in the response behavior label of the jth node in the power user price response behavior graph corresponding to the reconstructed user single-day time-of-use electricity consumption data is The probability of , It is the total number of response behavior levels of the user's daily time-of-use electricity consumption data.

9. A system for implementing the method for characterizing the electricity price response behavior of power users based on graph propagation as claimed in claim 1, characterized in that: The system comprises: An acquisition module is used to obtain a user's electricity price response behavior graph corresponding to the user's single-day time-of-use electricity price data; A reconstruction module is used to reconstruct the power user price response behavior graph corresponding to the user's single-day time-of-use electricity price electricity consumption data using the response behavior label corresponding to the user's single-day time-of-use electricity price electricity consumption data; wherein the response behavior label includes a real response behavior label and a pre-response behavior label; The correction module is used to correct the pre-response behavior label based on the power user price response behavior diagram corresponding to the reconstructed user single-day time-of-use electricity price electricity consumption data and the real response behavior label.

Images