Power load incentive strategy optimization method and system based on responsiveness evaluation

Abstract

The invention discloses a power load incentive strategy optimization method and system based on response capability evaluation, and mainly relates to the technical field of power load strategy optimization. Comprising the following steps: collecting user behavior data of a target area through a data collection module to obtain a user multi-source data set; performing operation-level behavior feature point analysis based on the user multi-source data set to obtain a behavior feature point list; performing main line-branch line double-layer user response capability analysis on the user multi-source data set according to the behavior feature point list to obtain a user response capability feature set; and calling an intelligent incentive strategy unit to carry out strategy optimization on the user response capability feature set, and determining a target incentive strategy. The method has the beneficial effects that the technical problems that in the prior art, the excitation strategy is single, the user response condition cannot be dynamically captured, and the personalized adjustment capability is lacked are solved, and the technical effect of improving the optimization reliability of the power load excitation strategy is achieved.

Description

Technical Field

[0001] This invention relates to the field of power load strategy optimization technology, and more specifically to a method and system for optimizing power load incentive strategies based on response capability assessment. Background Technology

[0002] With the rapid popularization of user-side resources such as distributed energy, energy storage, and electric vehicles, the controllability of power grid operation and load fluctuations are increasing. Traditional power load management mainly guides users to smooth peak and valley loads through unified electricity prices and time-of-use billing, but this method has problems such as response lag and low control accuracy, and cannot meet the requirements of real-time and flexible control of new power systems.

[0003] In recent years, demand response technology has gradually become an important means of load regulation in power systems. However, user participation is influenced by various behavioral factors, such as electricity price sensitivity, comfort disturbances, usage preferences, and equipment flexibility. Existing demand response control methods mostly focus on price-driven quantitative regulation and do not model and predict user behavior characteristics, resulting in significant deviations between dispatch results and actual responses.

[0004] Therefore, there is an urgent need for an intelligent incentive control method for power load that combines behavioral modeling, data prediction, and incentive mechanisms. This method can predict user participation probability and generate personalized incentive signals by analyzing user behavior patterns and preference characteristics, thereby improving user participation rate and load regulation accuracy. Summary of the Invention

[0005] This application provides a method and system for optimizing power load incentive strategies based on response capability assessment, which is used to address the technical problems in the prior art, such as the single incentive strategy, inability to dynamically capture user response, and lack of personalized adjustment capabilities.

[0006] In view of the above problems, this application provides a method and system for optimizing power load incentive strategies based on response capability assessment.

[0007] The first aspect of this application provides a method for optimizing power load incentive strategies based on response capability assessment, the method comprising: The data acquisition unit collects user behavior data from the target area to obtain a multi-source user data set; based on the multi-source user data set, it performs operational-level behavioral feature point analysis to obtain a list of behavioral feature points; based on the list of behavioral feature points, it performs a main-branch two-layer user response capability analysis on the multi-source user data set to obtain a user response capability feature set; and it calls the intelligent incentive strategy unit to optimize the user response capability feature set and determine the target incentive strategy.

[0008] In one possible implementation, operational-level behavioral feature point analysis is performed based on the user multi-source data set to obtain a list of behavioral feature points. This includes: extracting operational points from the user multi-source data set to obtain an initial set of operational behavioral feature points; performing front-to-back reverse contrast degree identification on the initial set of operational behavioral feature points in conjunction with the user multi-source data set to obtain a front-to-back reverse contrast degree set; filtering the front-to-back reverse contrast degree set according to a preset contrast degree threshold; and mapping and filtering the initial set of operational behavioral feature points based on the filtering results to obtain a list of behavioral feature points.

[0009] In one possible implementation, a two-tiered (mainline-branchline) user response capability analysis is performed on the user multi-source data set based on the behavioral feature point list to obtain a user response capability feature set. This includes: retrieving data from the user multi-source data set according to the mainline window based on the behavioral feature point list to obtain a mainline-related data sequence set of behavioral feature points; retrieving data from the user multi-source data set according to the branchline window based on the behavioral feature point list to obtain a branchline-related data sequence set of behavioral feature points, wherein the branchline window is smaller than the mainline window; performing long-term trend feature analysis on the mainline-related data sequence set of behavioral feature points to obtain a mainline long-term trend feature set; extracting short-term concentrated data drift from the mainline-related data sequence set of behavioral feature points to obtain a branchline short-term concentrated data set; and mapping and integrating the mainline long-term trend feature set and the branchline short-term concentrated data set to obtain the user response capability feature set.

[0010] In one possible implementation, long-term trend feature analysis is performed on the set of behavioral feature point mainline associated data sequences to obtain a set of mainline long-term trend features, including: extracting a first behavioral feature point mainline associated data sequence from the set of behavioral feature point mainline associated data sequences; calling a multi-scale long-term trend feature analyzer to perform multi-scale feature analysis on the first behavioral feature point mainline associated data sequence to obtain a first multi-scale mainline long-term trend feature set; interactively enhancing the first multi-scale mainline long-term trend feature set to obtain a first mainline long-term trend feature; and adding the first mainline long-term trend feature to the mainline long-term trend feature set.

[0011] In one possible implementation, interactive enhancement is performed on the first multi-scale main line long-term trend feature set to obtain the first main line long-term trend feature, including: extracting the first multi-scale main line long-term trend feature and the second multi-scale main line long-term trend feature from the first multi-scale main line long-term trend feature set in ascending order of scale without replacement; calculating the feature similarity set of the first multi-scale main line long-term trend feature and the second multi-scale main line long-term trend feature, and normalizing the feature similarity set to construct a first adjacency interactive enhancement matrix; convolving the second multi-scale main line long-term trend feature with the first adjacency interactive enhancement matrix to obtain the first adjacency interactive enhanced main line long-term trend feature; interactively enhancing the third multi-scale main line long-term trend feature extracted from the first multi-scale main line long-term trend feature set in ascending order of scale without replacement with the first adjacency interactive enhanced main line long-term trend feature, and so on, until the last feature is reached to obtain the first main line long-term trend feature.

[0012] In one possible implementation, the intelligent incentive strategy unit is invoked to optimize the user responsiveness feature set and determine the target incentive strategy. This includes: invoking the intelligent incentive strategy unit to identify the user responsiveness feature set and obtain an initial incentive strategy; providing execution feedback on the initial incentive strategy in a preset feedback window to obtain feedback data; evaluating the strategy fitness based on the feedback data to obtain the fitness of the initial incentive strategy; and determining whether the fitness of the initial incentive strategy meets preset requirements. If so, the initial incentive strategy is adopted as the target incentive strategy.

[0013] In one possible implementation, if the fitness of the initial incentive strategy does not meet the preset requirements, the initial incentive strategy is randomly adjusted according to the size of the fitness of the initial incentive strategy to obtain multiple adjusted incentive strategies; the fitness of the multiple adjusted incentive strategies, the initial incentive strategy, and the fitness of the initial incentive strategy are respectively predicted to obtain multiple adjusted incentive strategy fitnesss; when there is an adjusted incentive strategy fitness among the multiple adjusted incentive strategy fitnesss that meets the preset requirements, the adjusted incentive strategy corresponding to the maximum value among the multiple adjusted incentive strategy fitnesss is taken as the target incentive strategy.

[0014] A second aspect of this application provides a power load incentive strategy optimization system based on response capability assessment, the system comprising: The data acquisition module is used to collect user behavior data in the target area through the data acquisition unit to obtain a multi-source user data set; the feature point analysis module is used to perform operational-level behavioral feature point analysis based on the multi-source user data set to obtain a list of behavioral feature points; the response capability analysis module is used to perform a main-branch two-layer user response capability analysis on the multi-source user data set according to the list of behavioral feature points to obtain a user response capability feature set; the incentive strategy optimization module is used to call the intelligent incentive strategy unit to optimize the user response capability feature set and determine the target incentive strategy.

[0015] A third aspect of this application provides an electronic device including a memory and a processor, the memory storing executable instructions, wherein when the processor executes the executable instructions stored in the memory, it implements any step of the first aspect disclosed in this application.

[0016] A fourth aspect of this application provides a computer-readable storage medium storing a computer program for performing any step of the first aspect disclosed in this application.

[0017] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application collects user behavior data from a target area using a data acquisition unit to obtain a multi-source user data set; performs operational-level behavioral feature point analysis based on the multi-source user data set to obtain a list of behavioral feature points; conducts a main-branch two-layer user response capability analysis on the multi-source user data set based on the behavioral feature point list to obtain a user response capability feature set; and calls an intelligent incentive strategy unit to optimize the user response capability feature set to determine the target incentive strategy. This achieves the technical effect of improving the adaptability of the intelligent incentive strategy to user response and enhancing the reliability of the strategy. Attached Figure Description

[0018] Appendix Figure 1 This is a schematic diagram of the power load incentive strategy optimization method based on response capability assessment provided in an embodiment of the present invention.

[0019] Appendix Figure 2 This is a schematic diagram of the power load incentive strategy optimization system structure based on response capability assessment provided in an embodiment of the present invention.

[0020] Figure 3 This is an internal structural diagram of an electronic device provided in an embodiment of this application.

[0021] The labels shown in the attached diagram: The system includes a data acquisition module 11, a feature point analysis module 12, a response capability analysis module 13, an excitation strategy optimization module 14, a bus 300, a receiver 301, a processor 302, a transmitter 303, a memory 304, and a bus interface 305. Detailed Implementation

[0022] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims. It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices.

[0023] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for optimizing power load incentive strategies based on response capability assessment, wherein the method includes: Step S100: Collect user behavior data in the target area through the data acquisition unit to obtain a multi-source user data set; Preferably, the data acquisition unit obtains multi-dimensional data from user-side smart meters (AMI), building energy management systems (BEMS), and IoT terminal devices, including: power consumption, active energy, voltage, and current; indoor temperature, equipment start / stop status, and comfort indicators; electricity price information, weather parameters, and holiday information; and basic user information such as type, contract capacity, and response level. The acquisition frequency can be set to 15 minutes or 1 minute. The acquired data is uploaded to the data center after being timestamped via NTP / PTP. Linear interpolation is used for missing data, and Z-score detection is used to remove outliers. After linear interpolation and outlier removal, the user multi-source data set is obtained.

[0024] Step S200: Perform operational-level behavioral feature point analysis based on the user multi-source data set to obtain a list of behavioral feature points; Furthermore, based on the aforementioned user multi-source data set, operational-level behavioral feature point analysis is performed to obtain a list of behavioral feature points. In this embodiment, step S200 further includes: Extract operation points from the user's multi-source data set to obtain an initial operation behavior feature point set; By combining the user's multi-source data set, the initial operation behavior feature point set is used to identify the before-and-after reverse contrast, and a before-and-after reverse contrast set is obtained. The set of before-and-after contrast degrees is filtered according to a preset contrast degree threshold, and the set of initial operation behavior feature points is mapped and filtered according to the filtering results to obtain a list of behavior feature points.

[0025] In one possible implementation, to analyze and evaluate user responsiveness, it is first necessary to extract user operation data. The collected operation points can be understood as key data nodes reflecting the load adjustment potential in user electricity consumption behavior, which need to be precisely located from multi-source data. For example, from smart meter data, extracting the time and corresponding parameters for a washing machine switching from standby (power < 5W) to running (power > 500W), and for an air conditioner lowering its set temperature from 28℃ to 24℃, completes the extraction of operation points. Preferably, extracting operation records from user APP data, such as clicking the "participate in off-peak" button and modifying electricity usage time reservations (e.g., changing the charging station reservation from 20:00 to 23:00), is also a way to obtain operation points. Furthermore, these scattered operation nodes are organized according to time-behavior type-data parameters to form an initial set of operation behavior feature points, such as: 2024-05-20-08:30: Air conditioner cooling operation, temperature change 4℃, power increase 300W.

[0026] At this point, many operation points will be extracted and need to be filtered. The criterion for filtering is whether the operation is representative, which means calculating the inverse change magnitude of the related data within a specific time period before and after a certain operation point. For example, for the initial operation point of the washing machine starting at 19:10 on May 20, 2024, we take the power consumption data for the hour before and after (18:10-19:10, 19:10-20:10). If the user's total power consumption is stable at 800-900W between 18:10 and 19:10, and the power consumption jumps directly to 1450W after the washing machine is started at 19:10, the inverse change magnitude is 550W. Combining this with the user's historical data, the power fluctuation between 18:10 and 20:10 in the past 7 days is ≤200W. The inverse difference magnitude before and after this operation point is calculated as 550W (actual change) - 200W (normal fluctuation) = 350W.

[0027] The set of reverse contrast degrees before and after is filtered according to the preset contrast degree threshold set by those skilled in the art. The reverse contrast degrees before and after that are greater than or equal to the preset contrast degree threshold are extracted. Based on the filtering results, the set of initial operation behavior feature points is mapped and filtered to obtain a list of behavior feature points.

[0028] Step S300: Perform a main-branch two-layer user response capability analysis on the user multi-source data set based on the behavioral feature point list to obtain a user response capability feature set; Furthermore, based on the list of behavioral feature points, a main-branch two-layer user response capability analysis is performed on the user multi-source data set to obtain a user response capability feature set. In this embodiment, step S300 further includes: Based on the list of behavioral feature points, data is retrieved from the user's multi-source data set according to the main window to obtain a set of behavioral feature point main line related data sequences. Based on the list of behavioral feature points, data is retrieved from the user's multi-source data set according to the branch window to obtain a set of behavioral feature point branch-related data sequences, wherein the branch window is smaller than the main window; Long-term trend feature analysis is performed on the set of data sequences associated with the main line of behavioral feature points to obtain the set of long-term trend features of the main line. Short-term concentrated data drift extraction is performed on the main line associated data sequence set of the behavioral feature points to obtain the branch short-term concentrated data set; Based on the long-term trend feature set of the main line and the short-term concentrated data set of the branch line, a user response capability feature set is obtained by mapping and integrating them.

[0029] In one embodiment, the main-branch two-layer user responsiveness analysis is a user behavior modeling method that combines long-term stable behavior with short-term dynamic changes. The main line characterizes users' long-term stable response preferences, such as long-term electricity price sensitivity, comfort tolerance, and load fluctuation trends. The branch line presents users' preferences within a short-term window, such as short-term load surges, changes in comfort due to sudden weather changes, and short-term equipment unavailability. The main line window and branch line window define the observation time ranges for long-term and short-term data, respectively. The main line window is typically set to several days to several weeks to analyze long-term behavioral trends, while the branch line window is typically several minutes to several hours to capture short-term behavior.

[0030] Based on the list of behavioral feature points, the system retrieves long-term and short-term data from multi-source user data, using these feature points as anchors, to form a set of main-line related data sequences and a set of branch-line related data sequences. Subsequently, the system performs long-term trend feature analysis on the main-line related data sequences. For example, it uses multi-scale decomposition analysis to examine user load change trends, response stability, and the long-term relationship between electricity price and response over longer periods, thereby obtaining a set of long-term trend features for the main lines.

[0031] The mean-shift algorithm is used to find the user preference information for each data sequence in the set of behavioral feature point main-line associated data sequences, which is the most representative short-term clustered data set. Specifically, the mean-shift algorithm is an adaptive clustering algorithm based on density gradient ascent, often used to find the center point of high-density regions in data distribution. The algorithm defines a kernel density estimate in the data space and continuously calculates the local mean, pushing points to move along the direction of the maximum density gradient, eventually aggregating them to the position representing the true cluster center of the data. The mean-shift algorithm is executed on each data sequence in the set of behavioral feature point main-line associated data sequences, and the high-density clustered region of the data in the long-term trend is determined by the kernel density estimate. Subsequently, the algorithm continuously calculates the local mean and updates the centroid position, causing the data points to gradually converge towards the high-density region, thereby obtaining the core behavioral preference pattern of the sequence, such as the peak response periods that users repeatedly experience in the long term, the typical air conditioning load adjustment range, and high-frequency load change points reflecting electricity price sensitivity. Furthermore, the data corresponding to these convergent cluster centers are used as lateral short-term aggregate data, which are then aggregated to obtain the lateral short-term aggregate data set, representing the most representative short-term behavioral segments. The mean-shift algorithm is then used to automatically identify user behavioral preferences over a recent period, i.e., the short term.

[0032] Then, based on each long-term trend feature in the main trend feature set, the corresponding short-term concentrated data in the branch short-term concentrated data set are identified, and mapping and integration are completed to obtain the user responsiveness feature set. The user responsiveness feature set includes not only the user's short-term actual preferences but also the user's long-term behavioral trends.

[0033] Furthermore, long-term trend feature analysis is performed on the set of data sequences associated with the main line of behavioral feature points to obtain a set of long-term trend features of the main line. In this embodiment, step S300 further includes: Extract the first behavioral feature point main line associated data sequence from the set of behavioral feature point main line associated data sequences; The multi-scale long-term trend feature analyzer is invoked to perform multi-scale feature analysis on the data sequence of the first behavioral feature point main line association, thereby obtaining the first multi-scale main line long-term trend feature set. Interactive enhancement is performed on the first multi-scale main line long-term trend feature set to obtain the first main line long-term trend feature, and the first main line long-term trend feature is added to the main line long-term trend feature set.

[0034] Furthermore, the first multi-scale main line long-term trend feature set is interactively enhanced to obtain the first main line long-term trend feature. In this embodiment, step S300 further includes: Extract the first and second multi-scale main line long-term trend features from the first multi-scale main line long-term trend feature set in ascending order of scale without replacement. Calculate the feature similarity set of the first multi-scale main line long-term trend feature and the second multi-scale main line long-term trend feature, and normalize the feature similarity set to construct the first adjacency interaction enhancement matrix. The first adjacency interaction enhancement matrix is ​​used to convolve the second multi-scale main line long-term trend feature to obtain the first adjacency interaction enhancement main line long-term trend feature. The first adjacency interaction is used to enhance the long-term trend features of the main line. The third multi-scale main line long-term trend features extracted from the first multi-scale main line long-term trend feature set in ascending order of scale are interactively enhanced without replacement. This process is repeated until the last feature is reached to obtain the first main line long-term trend features.

[0035] In one embodiment, multiple sample behavioral feature point main line correlation data sequences and multiple sample multi-scale main line long-term trend feature sets after multi-scale analysis are obtained as training data, wherein the multi-scale analysis is set by those skilled in the art. Then, the training data is used to supervise the training of a framework built on a feedforward neural network until training converges, obtaining the trained multi-scale long-term trend feature analyzer. Furthermore, the first and second multi-scale main line long-term trend features are extracted from the first multi-scale main line long-term trend feature set in ascending order of scale, without replacement, to facilitate feature interaction.

[0036] The feature similarity sets of the first and second multi-scale main line long-term trend features are calculated using the cosine similarity formula. Then, the feature similarity sets are normalized using the softmax formula. The normalized result is filled into an initially empty matrix to obtain the first adjacency interaction enhancement matrix. This first adjacency interaction enhancement matrix is ​​a weighted matrix constructed based on feature similarity. By encoding the interrelationships between features at different scales, it establishes connection weights between multi-scale features for subsequent feature convolution fusion. Furthermore, multiple sample adjacency interaction enhancement matrices, multiple sample multi-scale main line long-term trend features, and multiple sample adjacency interaction enhancement main line long-term trend features are obtained as convolution training data. The framework constructed based on a convolutional neural network is supervised and trained using this training data until convergence, resulting in the trained first adjacency interaction enhancement main line long-term trend features.

[0037] Based on the principle of obtaining the first adjacency interaction-enhanced long-term trend feature, this method uses the first adjacency interaction-enhanced long-term trend feature to interactively enhance the third-level multi-scale long-term trend feature extracted from the first multi-scale long-term trend feature set in ascending order of scale without replacement. This process is repeated until the last feature is reached, thus obtaining the first long-term trend feature. This achieves the technical effect of integrating detailed information from smaller-scale features while preserving the original trend pattern.

[0038] Step S400: Invoke the intelligent incentive strategy unit to optimize the user response capability feature set and determine the target incentive strategy.

[0039] Furthermore, the intelligent incentive strategy unit is invoked to optimize the user response capability feature set and determine the target incentive strategy. In this embodiment, step S400 further includes: The intelligent incentive strategy unit is invoked to identify the user response capability feature set and obtain an initial incentive strategy. The initial incentive strategy is executed and feedback is obtained in a preset feedback window; Based on the feedback data, a strategy fitness assessment is performed to obtain the initial incentive strategy fitness. Determine whether the fitness of the initial incentive strategy meets the preset requirements. If so, use the initial incentive strategy as the target incentive strategy.

[0040] Furthermore, if the fitness of the initial incentive strategy does not meet the preset requirements, the initial incentive strategy is randomly adjusted according to the fitness of the initial incentive strategy to obtain multiple adjusted incentive strategies. Multiple adjustment incentive strategies, the initial incentive strategy, and the fitness of the initial incentive strategy are used to predict the fitness of multiple adjustment incentive strategies. When there is an adjustment incentive strategy fitness among the multiple adjustment incentive strategy fitness that meets the preset requirements, the adjustment incentive strategy corresponding to the maximum value among the multiple adjustment incentive strategy fitness is taken as the target incentive strategy.

[0041] In one embodiment, the preset feedback window refers to a specific time period after the policy is issued, used to collect real user response data, such as actual power reduction, response latency, execution success rate, equipment performance, and user feedback. The intelligent incentive strategy unit is invoked to generate an initial incentive strategy based on user response capability characteristics, including main long-term trend characteristics and branch short-term concentrated data. For example, it infers the incentive amount, load adjustment range, and execution period that the user is most likely to accept under the current electricity price and status. Subsequently, the system implements the strategy within the preset feedback window and collects feedback data in real time, including actual response power, execution latency, user feedback comfort, and equipment performance.

[0042] Based on this feedback data, the fitness of the initial incentive strategy is calculated, such as peak shaving error rate, response probability deviation, and load adjustment success rate. If the strategy fitness meets a preset threshold, such as achieving the peak shaving target or the response deviation being below a set range, the initial incentive strategy is directly determined as the target incentive strategy.

[0043] If the fitness does not meet the requirements, a strategy adjustment mechanism is activated. Random perturbations are performed based on the initial fitness value, adjusting incentive parameters such as amount, duration, execution period, and incentive type in various random ways to generate multiple candidate strategies. Subsequently, a fitness prediction model is invoked to quickly predict and evaluate the fitness of the multiple candidate strategies and the initial strategy, determining which adjusted strategies can meet the preset requirements. When a strategy among the multiple candidate strategies has a fitness value that meets the threshold, the one with the highest fitness is selected as the final target incentive strategy.

[0044] Example 2, based on the same inventive concept as the power load incentive strategy optimization method based on response capability assessment in the foregoing examples, as shown in the appendix. Figure 2 As shown, this application provides a power load incentive strategy optimization system based on response capability assessment. The system and method embodiments in this application are based on the same inventive concept. The system includes: Data acquisition module 11 is used to collect user behavior data in the target area through the data acquisition unit to obtain a multi-source user data set; Feature point analysis module 12 is used to perform operational-level behavioral feature point analysis based on the user multi-source data set to obtain a list of behavioral feature points; The response capability analysis module 13 is used to perform a main-branch two-layer user response capability analysis on the user multi-source data set based on the behavioral feature point list, and obtain a user response capability feature set. The incentive strategy optimization module 14 is used to call the intelligent incentive strategy unit to optimize the user response capability feature set and determine the target incentive strategy.

[0045] Furthermore, the feature point analysis module 12 is used to perform the following steps: Extract operation points from the user's multi-source data set to obtain an initial operation behavior feature point set; By combining the user's multi-source data set, the initial operation behavior feature point set is used to identify the before-and-after reverse contrast, and a before-and-after reverse contrast set is obtained. The set of before-and-after contrast degrees is filtered according to a preset contrast degree threshold, and the set of initial operation behavior feature points is mapped and filtered according to the filtering results to obtain a list of behavior feature points.

[0046] Furthermore, the response capability analysis module 13 is used to perform the following steps: Based on the list of behavioral feature points, data is retrieved from the user's multi-source data set according to the main window to obtain a set of behavioral feature point main line related data sequences. Based on the list of behavioral feature points, data is retrieved from the user's multi-source data set according to the branch window to obtain a set of behavioral feature point branch-related data sequences, wherein the branch window is smaller than the main window; Long-term trend feature analysis is performed on the set of data sequences associated with the main line of behavioral feature points to obtain the set of long-term trend features of the main line. Short-term concentrated data drift extraction is performed on the main line associated data sequence set of the behavioral feature points to obtain the branch short-term concentrated data set; Based on the long-term trend feature set of the main line and the short-term concentrated data set of the branch line, a user response capability feature set is obtained by mapping and integrating them.

[0047] Furthermore, the response capability analysis module 13 is used to perform the following steps: Extract the first behavioral feature point main line associated data sequence from the set of behavioral feature point main line associated data sequences; The multi-scale long-term trend feature analyzer is invoked to perform multi-scale feature analysis on the data sequence of the first behavioral feature point main line association, thereby obtaining the first multi-scale main line long-term trend feature set. Interactive enhancement is performed on the first multi-scale main line long-term trend feature set to obtain the first main line long-term trend feature, and the first main line long-term trend feature is added to the main line long-term trend feature set.

[0048] Furthermore, the response capability analysis module 13 is used to perform the following steps: Extract the first and second multi-scale main line long-term trend features from the first multi-scale main line long-term trend feature set in ascending order of scale without replacement. Calculate the feature similarity set of the first multi-scale main line long-term trend feature and the second multi-scale main line long-term trend feature, and normalize the feature similarity set to construct the first adjacency interaction enhancement matrix. The first adjacency interaction enhancement matrix is ​​used to convolve the second multi-scale main line long-term trend feature to obtain the first adjacency interaction enhancement main line long-term trend feature. The first adjacency interaction is used to enhance the long-term trend features of the main line. The third multi-scale main line long-term trend features extracted from the first multi-scale main line long-term trend feature set in ascending order of scale are interactively enhanced without replacement. This process is repeated until the last feature is reached to obtain the first main line long-term trend features.

[0049] Furthermore, the incentive strategy optimization module 14 is used to perform the following steps: The intelligent incentive strategy unit is invoked to identify the user response capability feature set and obtain an initial incentive strategy. The initial incentive strategy is executed and feedback is obtained in a preset feedback window; Based on the feedback data, a strategy fitness assessment is performed to obtain the initial incentive strategy fitness. Determine whether the fitness of the initial incentive strategy meets the preset requirements. If so, use the initial incentive strategy as the target incentive strategy.

[0050] Furthermore, if the fitness of the initial incentive strategy does not meet the preset requirements, the initial incentive strategy is randomly adjusted according to the fitness of the initial incentive strategy to obtain multiple adjusted incentive strategies. Multiple adjustment incentive strategies, the initial incentive strategy, and the fitness of the initial incentive strategy are used to predict the fitness of multiple adjustment incentive strategies. When there is an adjustment incentive strategy fitness among the multiple adjustment incentive strategy fitness that meets the preset requirements, the adjustment incentive strategy corresponding to the maximum value among the multiple adjustment incentive strategy fitness is taken as the target incentive strategy.

[0051] Example 3, as Figure 3 The diagram shown is a schematic representation of the structure of an exemplary electronic device of this application. Figure 3 In this document, the bus architecture is represented by bus 300. Bus 300 may include any number of interconnected buses and bridges, and bus 300 connects various circuits including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 305 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, i.e., a transceiver, providing a unit for communicating with various other devices over a transmission medium.

[0052] The memory 304, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the power load incentive strategy optimization method based on response capability assessment in this embodiment. The processor 302 executes various functional applications and data processing of the computer device by running the software programs, instructions, and modules stored in the memory 304, thereby realizing the aforementioned power load incentive strategy optimization method based on response capability assessment.

[0053] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0054] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0055] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for optimizing power load incentive strategy based on response capability assessment, characterized in that, The method includes: The user behavior data in the target area is collected by the data acquisition unit to obtain a multi-source user data set; Based on the aforementioned user multi-source data set, operational-level behavioral feature point analysis is performed to obtain a list of behavioral feature points. Based on the list of behavioral feature points, perform a two-layer main-branch user response capability analysis on the user multi-source data set to obtain a user response capability feature set. The intelligent incentive strategy unit is invoked to optimize the user response capability feature set and determine the target incentive strategy. 2.The method of claim 1, wherein, Based on the aforementioned user multi-source data set, operational-level behavioral feature point analysis is performed to obtain a list of behavioral feature points, including: Extract operation points from the user's multi-source data set to obtain an initial operation behavior feature point set; By combining the user's multi-source data set, the initial operation behavior feature point set is used to identify the before-and-after reverse contrast, and a before-and-after reverse contrast set is obtained. The set of before-and-after contrast degrees is filtered according to a preset contrast degree threshold, and the set of initial operation behavior feature points is mapped and filtered according to the filtering results to obtain a list of behavior feature points. 3.The method of claim 1, wherein, Based on the list of behavioral feature points, a two-tiered (main line-branch line) user response capability analysis is performed on the user multi-source data set to obtain a user response capability feature set, including: Based on the list of behavioral feature points, data is retrieved from the user's multi-source data set according to the main window to obtain a set of behavioral feature point main line related data sequences. Based on the list of behavioral feature points, data is retrieved from the user's multi-source data set according to the branch window to obtain a set of behavioral feature point branch-related data sequences, wherein the branch window is smaller than the main window; Long-term trend feature analysis is performed on the set of data sequences associated with the main line of behavioral feature points to obtain the set of long-term trend features of the main line. Short-term concentrated data drift extraction is performed on the main line associated data sequence set of the behavioral feature points to obtain the branch short-term concentrated data set; Based on the long-term trend feature set of the main line and the short-term concentrated data set of the branch line, a user response capability feature set is obtained by mapping and integrating them. 4.The method of claim 3, wherein, Long-term trend feature analysis is performed on the set of data sequences related to the main line of behavioral feature points to obtain a set of long-term trend features of the main line, including: Extract the first behavioral feature point main line associated data sequence from the set of behavioral feature point main line associated data sequences; The multi-scale long-term trend feature analyzer is invoked to perform multi-scale feature analysis on the data sequence of the first behavioral feature point main line association, thereby obtaining the first multi-scale main line long-term trend feature set. Interactive enhancement is performed on the first multi-scale main line long-term trend feature set to obtain the first main line long-term trend feature, and the first main line long-term trend feature is added to the main line long-term trend feature set.

5. The power load incentive strategy optimization method based on response capability assessment as described in claim 4, characterized in that, Interactive enhancement is performed on the first multi-scale main line long-term trend feature set to obtain the first main line long-term trend features, including: Extract the first and second multi-scale main line long-term trend features from the first multi-scale main line long-term trend feature set in ascending order of scale without replacement. Calculate the feature similarity set of the first multi-scale main line long-term trend feature and the second multi-scale main line long-term trend feature, and normalize the feature similarity set to construct the first adjacency interaction enhancement matrix. The first adjacency interaction enhancement matrix is ​​used to convolve the second multi-scale main line long-term trend feature to obtain the first adjacency interaction enhancement main line long-term trend feature. The first adjacency interaction is used to enhance the long-term trend features of the main line. The third multi-scale main line long-term trend features extracted from the first multi-scale main line long-term trend feature set in ascending order of scale are interactively enhanced without replacement. This process is repeated until the last feature is reached to obtain the first main line long-term trend features.

6. The power load incentive strategy optimization method based on response capability assessment as described in claim 1, characterized in that, The intelligent incentive strategy unit is invoked to optimize the user response capability feature set and determine the target incentive strategy, including: The intelligent incentive strategy unit is invoked to identify the user response capability feature set and obtain an initial incentive strategy. The initial incentive strategy is executed and feedback is obtained in a preset feedback window; Based on the feedback data, a strategy fitness assessment is performed to obtain the initial incentive strategy fitness. Determine whether the fitness of the initial incentive strategy meets the preset requirements. If so, use the initial incentive strategy as the target incentive strategy.

7. The power load incentive strategy optimization method based on response capability assessment as described in claim 6, characterized in that, If the fitness of the initial incentive strategy does not meet the preset requirements, the initial incentive strategy is randomly adjusted according to the fitness of the initial incentive strategy to obtain multiple adjusted incentive strategies. Multiple adjustment incentive strategies, the initial incentive strategy, and the fitness of the initial incentive strategy are used to predict the fitness of multiple adjustment incentive strategies. When there is an adjustment incentive strategy fitness among the multiple adjustment incentive strategy fitness that meets the preset requirements, the adjustment incentive strategy corresponding to the maximum value among the multiple adjustment incentive strategy fitness is taken as the target incentive strategy.

8. A power load incentive strategy optimization system based on response capability assessment, characterized in that, The system is used to implement the power load incentive strategy optimization method based on response capability assessment as described in any one of claims 1-7, and the system comprises: The data acquisition module is used to collect user behavior data in the target area through the data acquisition unit to obtain a multi-source user data set; The feature point analysis module is used to perform operational-level behavioral feature point analysis based on the user multi-source data set to obtain a list of behavioral feature points. The response capability analysis module is used to perform a main-branch two-layer user response capability analysis on the user multi-source data set based on the list of behavioral feature points, and to obtain a user response capability feature set. The incentive strategy optimization module is used to call the intelligent incentive strategy unit to optimize the user response capability feature set and determine the target incentive strategy.

9. An electronic device, characterized in that, The electronic device includes: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the power load incentive strategy optimization method based on response capability assessment as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the power load incentive strategy optimization method based on response capability assessment as described in any one of claims 1-7.

Images