Resource scheduling method, device and equipment for virtual power plant, medium and product

Abstract

The invention discloses a resource scheduling method and device for a virtual power plant, equipment, a medium and a product, and the method comprises the steps: calculating a user scheduling strategy according to collected user behavior data; wherein the user scheduling strategy is used for representing a resource allocation tendency to a target user; calculating a load prediction value of the target user according to the user behavior data and the user scheduling strategy; performing curve smoothing processing on the load prediction value by adopting a preset interpolation algorithm to obtain a processed load prediction value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; according to the processed load prediction value and the load demand of the target user, generating a resource scheduling strategy of the virtual power plant for the user; wherein the resource scheduling strategy comprises a resource scheduling value. By adopting the method, the dynamic change of the user behavior can be effectively captured, and the resource scheduling flexibility and accuracy of the virtual power plant are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system scheduling, and in particular to a resource scheduling method, device, equipment, medium and product for a virtual power plant. Background Art

[0002] Virtual power plants optimize energy utilization by aggregating distributed energy resources, but existing scheduling methods have significant shortcomings in processing user behavior: fixed rule-based methods, such as time-of-use electricity pricing, simplify user behavior into static patterns and cannot adapt to dynamic changes. Forecasting-based methods, such as LSTM (Long Short-Term Memory Network for Load Forecasting), rely on historical data, and prediction errors accumulate when user behavior fluctuates, and lack real-time feedback. Centralized control methods, such as central EMS (Energy Management System), although capable of unified scheduling, are complex and subject to high communication latency, making them difficult to scale. In real-world cases, sudden increases in user air conditioning load or charging delays often lead to scheduling failures. Therefore, a new collaborative scheduling method is urgently needed that can capture dynamic changes in user behavior in real time and enable two-way interaction between users and the system. Summary of the Invention

[0003] The purpose of the embodiments of the present invention is to provide a resource scheduling method, device, equipment, medium and product for a virtual power plant, which can effectively capture the dynamic changes in user behavior and improve the flexibility and accuracy of resource scheduling of the virtual power plant.

[0004] To achieve the above objectives, an embodiment of the present invention provides a resource scheduling method for a virtual power plant, comprising: Calculating a user scheduling strategy based on the collected user behavior data; wherein the user scheduling strategy is used to characterize the resource allocation tendency for the target user; Calculating a load forecast value of the target user based on the user behavior data and the user scheduling policy; A preset interpolation algorithm is used to perform curve smoothing processing on the load forecast value to obtain a processed load forecast value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; Based on the processed load forecast value and the load demand of the target user, a resource scheduling strategy of the virtual power plant for the user is generated; wherein the resource scheduling strategy includes a resource scheduling value.

[0005] As an improvement to the above solution, generating a user scheduling strategy based on the collected user behavior data includes: Collecting user behavior data; wherein the user behavior data includes the load demand, response delay and preference window of the target user; Normalizing the load demand, the response delay, and the preference window and converting them into feature vectors; The feature vector, the feature vectors of the target user's neighbor users, and the adjacency matrix are input into a preset user behavior chain feedback model for processing, and a user scheduling policy for the target user is output; wherein the adjacency matrix is ​​used to characterize the degree of mutual influence between the target user and the neighbor users, and the neighbor users are other users with behavioral associations with the target user; the user behavior chain feedback model is constructed and trained based on deep reinforcement learning.

[0006] As an improvement to the above solution, the method of using a preset interpolation algorithm to perform curve smoothing processing on the load forecast value to obtain the processed load forecast value includes: According to the actual load value of the target user at time (t-1) and the load forecast value at time (t+1), a preset linear interpolation algorithm is used to calculate the load forecast value at time t, which is recorded as a first load forecast value; According to the actual load values ​​of neighboring users at time t, the load prediction value of the target user at time t is calculated using the local weighted regression interpolation algorithm as the second load prediction value; wherein the neighboring users are other users with behavioral associations with the target user; A load prediction value is calculated according to the first load prediction value, the second load prediction value and a preset adaptive weight coefficient.

[0007] As an improvement to the above solution, the method further includes: Using a roulette algorithm, calculating the scheduling selection probability of the target user; Determining the policy scheduling priority of the target user according to the scheduling selection probability; wherein the scheduling selection probability is positively correlated with the policy scheduling priority; The resource scheduling policy is executed for the target user according to the policy scheduling priority.

[0008] As an improvement to the above solution, the roulette algorithm is used to calculate the scheduling selection probability of the target user, including: Calculating the utility value of the target user; wherein the utility value is used to evaluate the value of the user's behavior; Calculating a scheduling selection probability of the target user according to the utility value; A timing scheduling control method is introduced to adjust the scheduling selection probability according to the historical behavior weight of the target user to obtain the final scheduling selection probability; wherein the historical behavior weight is used to characterize the responsiveness and reliability of the target user to the resource scheduling strategy within the historical period.

[0009] As an improvement to the above solution, after generating the resource scheduling strategy of the virtual power plant for the user based on the processed load forecast value and the load demand of the target user, the method further includes: Calculating a scheduling error; wherein the scheduling error is used to represent the difference between the load forecast value and the actual load of all users; Through a global-local dual-level game strategy, the utility function values ​​of the global game and the local game are calculated, as well as the global and local adjustment factors. The goal of the global game layer is to optimize the resource allocation of the entire virtual power plant to balance the grid load and user demand, while the goal of the local game layer is to optimize the resource scheduling strategy for each user. Calculating a game theory adjustment term according to the utility function value of the global game, the utility function value of the local game, the global adjustment factor, and the local adjustment factor; The resource scheduling value is modified according to the game theory adjustment item to form a final resource scheduling strategy.

[0010] An embodiment of the present invention further provides a resource scheduling device for a virtual power plant, comprising: A user scheduling strategy generation module is used to generate a user scheduling strategy based on the collected user behavior data; wherein the user scheduling strategy is used to represent the resource allocation tendency for the target user; A load forecast value calculation module, configured to calculate the load forecast value of the target user based on the user behavior data and the user scheduling policy; A load forecast value processing module, configured to perform curve smoothing processing on the load forecast value using a preset interpolation algorithm to obtain a processed load forecast value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; A resource scheduling strategy generation module is used to generate a resource scheduling strategy of the virtual power plant for the user based on the processed load forecast value and the load demand of the target user; wherein the resource scheduling strategy includes a resource scheduling value.

[0011] An embodiment of the present invention also provides a resource scheduling device for a virtual power plant, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the resource scheduling method for a virtual power plant as described in any one of the above items.

[0012] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the resource scheduling method of the virtual power plant as described in any one of the above items.

[0013] An embodiment of the present invention also provides a computer program product, which includes a computer program or computer instructions. When the computer program or the computer instructions are executed by a processor, the resource scheduling method of the virtual power plant as described in any one of the above items is implemented.

[0014] Compared with the prior art, the resource scheduling method, device, equipment, medium and product of the virtual power plant disclosed in the present invention, by collecting rich user behavior data and combining with improved calculation models and algorithms, can deeply explore the user's electricity consumption patterns and needs, making the calculated user scheduling strategy and load forecast value more accurate. The smoothing of the load forecast curve can further improve the stability and reliability of the load forecast, thereby generating a more accurate resource scheduling strategy, avoiding excessive or insufficient allocation of resources. The virtual power plant can better balance the supply and demand of electricity and improve the scheduling efficiency and accuracy of the virtual power plant. At the same time, by guiding users to adjust their electricity consumption behavior, the optimal configuration of electricity resources in time and space is achieved, and the stability and reliability of the power system are improved. By formulating personalized user scheduling strategies, the characteristics and needs of users' electricity consumption behavior are taken into account. On the premise of meeting the normal electricity consumption of users, users are guided to participate in power resource scheduling, which can reduce electricity costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a resource scheduling method for a virtual power plant provided by an embodiment of the present invention; Figure 2 1 is a flow chart of a resource scheduling method for a virtual power plant preferably used in an embodiment of the present invention; Figure 3 This is a structural diagram of a resource scheduling device for a virtual power plant provided by an embodiment of the present invention; Figure 4 It is a structural diagram of a resource scheduling device for a virtual power plant provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0017] In the description of this application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application.

[0018] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include one or more of such features. Throughout this application, unless otherwise specified, "plurality" means two or more.

[0019] See also Figure 1 , is a flow chart of a resource scheduling method for a virtual power plant provided by an embodiment of the present invention. The embodiment of the present invention provides a resource scheduling method for a virtual power plant, including steps S11 to S14: S11. Calculating a user scheduling strategy based on the collected user behavior data; wherein the user scheduling strategy is used to characterize a resource allocation tendency for a target user; S12. Calculating a load forecast value of the target user based on the user behavior data and the user scheduling policy; S13. Using a preset interpolation algorithm to perform curve smoothing processing on the load forecast value to obtain a processed load forecast value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; S14. Generate a resource scheduling strategy of the virtual power plant for the target user based on the processed load forecast value and the load demand of the target user; wherein the resource scheduling strategy includes a resource scheduling value.

[0020] In this embodiment of the present invention, user behavior data, including that of target users and neighboring users, is collected in real time through smart terminal devices. A user scheduling policy calculation model is constructed, using preprocessed user behavior data as input. The model is trained to explore the relationship between user electricity usage patterns and resource allocation preferences. Based on the trained model and combined with the real-time collected user behavior data, a user scheduling policy is calculated for each target user. This policy quantifies the resource allocation preferences for the target user, such as allocation priority and the proportion of schedulable resources.

[0021] Features relevant to load forecasting are extracted from user behavior data. Combined with the user scheduling strategy, the target user's load forecast value is predicted and smoothed using an interpolation algorithm. Finally, the difference between the processed load forecast value and the target user's actual load demand is analyzed to determine the direction and extent of resource scheduling. Based on the analysis results and the overall resource status of the virtual power plant, a resource scheduling strategy is generated for the target user. This resource scheduling strategy includes specific resource scheduling values, such as adjusting the user's power consumption and the amount of allocated power resources. Simultaneously, the resource scheduling strategy is distributed to the target user's smart terminal device via the communication network to guide the user in adjusting their electricity usage behavior.

[0022] By adopting the technical means of the embodiments of the present invention, by collecting rich user behavior data and combining it with improved calculation models and algorithms, it is possible to deeply explore the user's electricity consumption patterns and needs, so that the calculated user scheduling strategies and load forecast values ​​are more accurate. The smoothing of the load forecast curve can further improve the stability and reliability of the load forecast, thereby generating a more accurate resource scheduling strategy, avoiding excessive or insufficient allocation of resources. The virtual power plant can better balance the supply and demand of electricity and improve the scheduling efficiency and accuracy of the virtual power plant. At the same time, by guiding users to adjust their electricity consumption behavior, the optimal configuration of electricity resources in time and space is achieved, and the stability and reliability of the power system are improved. By formulating personalized user scheduling strategies, the characteristics and needs of users' electricity consumption behavior are taken into account. Under the premise of meeting the normal electricity consumption of users, users are guided to participate in the scheduling of electricity resources, which can reduce electricity costs.

[0023] As a preferred implementation, the embodiment of the present invention is further implemented on the basis of the above embodiment. Step S11, i.e., generating a user scheduling strategy based on the collected user behavior data, includes steps S111 to S113: S111. Collect user behavior data; wherein the user behavior data includes the load demand, response delay, and preference window of the target user; S112, normalizing the load demand, the response delay, and the preference window, and converting them into a feature vector; S113. Input the feature vector, the feature vectors of the target user's neighbor users, and the adjacency matrix into a preset user behavior chain feedback model for processing, and output a user scheduling policy for the target user; wherein the adjacency matrix is ​​used to characterize the degree of mutual influence between the target user and the neighbor users, and the neighbor users are other users with behavioral associations with the target user; the user behavior chain feedback model is constructed and trained based on deep reinforcement learning.

[0024] In this embodiment of the present invention, a user behavior data collection module collects user electricity usage data in real time, including information such as user load demand, usage preferences, and response delays. By building a user behavior chain feedback model, it is possible to effectively analyze and predict user electricity usage behavior, and based on this, generate preliminary scheduling strategies.

[0025] Specifically, the system collects users’ load demands in real time through smart terminal devices (such as smart meters). , response delay , Preferences Window For the collected data, after standardization, user behavior is converted into feature vectors, which are further used for model training and strategy optimization. Feature vector The calculation formula is as follows: in, , , , are the mean and standard deviation of the load data, and is the weight coefficient.

[0026] The user behavior chain feedback model is built based on deep reinforcement learning (DRL), which learns the dynamic characteristics of user behavior through neural networks. It is obtained through neural network training and is used to Dynamically adjust scheduling strategies.

[0027] Scheduling policy for each user Based on its current state and influence from neighbors (adjacency matrix The dynamic update equation of user behavior is: in: is a user In time User scheduling policy. Is the feedback function, which is learned through neural network and represents the user's state The scheduling strategy is dynamically adjusted based on the status of neighboring users. is a user The neighbor set of user A collection of other users with behavioral associations. is the adjacency matrix, representing the user With other users The strength of the behavioral association between them reflects the degree of mutual influence between users. is a user In time The state vector contains information such as user load demand, response delay, and preference. is a user In time The state vector represents the mutual influence between users, and the feedback function The impact of these user behaviors will be taken into consideration when adjusting scheduling strategies.

[0028] In this model, the user scheduling strategy is implemented through a deep reinforcement learning model. The model is learned by the user, and makes scheduling decisions based on the user's current state and the behavior of its neighbors (calculated through a weighted adjacency matrix). The model's goal is to dynamically adjust and optimize user behavior strategies, enabling the virtual power plant to adapt to changing power demand and user behavior.

[0029] Through the recursive feedback mechanism of the neural network, the system can optimize the scheduling strategy at each time step and dynamically adjust the relationship between user behavior and load scheduling, thereby improving the scheduling efficiency and stability of the overall power resources.

[0030] As a preferred embodiment, step S12, i.e., calculating the load forecast value of the target user based on the user behavior data and the user scheduling policy, includes: Generate a benchmark load forecast value based on the user behavior data using a traditional time series model or a machine learning model; The user scheduling strategy is used as a dynamic adjustment factor to correct the reference load forecast value to obtain the load forecast value of the target user.

[0031] As an example, if If the user accepts delayed charging, the charging load forecast value for the t+1 period will be reduced; If the user responds to the real-time electricity price, the weight of the adjustable load will be increased.

[0032] Preferably, the correction formula is as follows: in, is the load forecast value at time t+1, is the baseline load forecast value, is the strategy influence coefficient, , which is the dynamic adjustment factor converted from the user scheduling strategy, and is used to indicate the degree of user response.

[0033] As a preferred embodiment, the embodiment of the present invention is further implemented on the basis of the above embodiment, step S13, that is, using a preset interpolation algorithm to perform curve smoothing processing on the load forecast value to obtain the processed load forecast value, includes steps S131 to S133: S131. Calculate the load forecast value at time t using a preset linear interpolation algorithm based on the actual load value of the target user at time (t-1) and the load forecast value at time (t+1), and record it as a first load forecast value. S132. Calculate the load prediction value of the target user at time t using the local weighted regression interpolation algorithm based on the actual load values ​​of neighboring users at time t, as a second load prediction value; wherein the neighboring users are other users with behavioral associations with the target user; S133. Calculate a load prediction value according to the first load prediction value, the second load prediction value, and a preset adaptive weight coefficient.

[0034] In an embodiment of the present invention, after completing the user behavior modeling, the system enters the prediction and optimization stage. The linear interpolation optimization module receives the user feature vector and improves the accuracy and stability of the virtual power plant resource scheduling through a series of refined prediction and optimization steps. In order to effectively cope with load fluctuations and optimize resource scheduling, the system combines the traditional linear interpolation method with the innovative local weighted polynomial regression interpolation technology. This innovative solution can not only smooth out load fluctuations, but also has an adaptive adjustment mechanism that can adjust the load forecast in real time based on feedback. In this way, the system can maintain a high load forecast accuracy during normal operation, and respond quickly to sudden fluctuations to ensure the stability and efficiency of scheduling.

[0035] Traditional linear interpolation methods can smooth load fluctuations, but their accuracy is insufficient when load demand changes dramatically or unexpected events occur. To address this issue, this paper combines local weighted polynomial regression interpolation technology to perform local weighted regression fitting on historical load data, generating more accurate load forecast values.

[0036] First, the load forecast value at time t is calculated using the traditional linear interpolation algorithm, which is recorded as the first load forecast value. The traditional linear interpolation formula is as follows: in, For users At time step The first load forecast value, is the time step ( The actual load value, is the time step load forecast value. is the time interval.

[0037] To further improve the accuracy, the present invention introduces a local weighted regression interpolation method to calculate the load forecast value of the target user at time t as the second load forecast value. The local weighted regression equation is: in, For users At time step The second load forecast value; For users The neighbor set of user A collection of other users with behavioral associations; is the weighting coefficient, which is dynamically adjusted based on the time distance and data importance, and usually adopts Gaussian function or exponential decay function; For users At time step The load value.

[0038] The load prediction value is calculated according to the first load prediction value, the second load prediction value and the preset adaptive weight coefficient using the following dynamic fusion formula: in, is the adaptive weight coefficient. Optionally, when the load is stable, =0.7; when the load fluctuates greatly, =0.3.

[0039] By combining linear interpolation and local weighted regression interpolation, the system can fully utilize the simplicity of linear interpolation and the accuracy of local regression when facing complex load fluctuations, thereby improving the stability and accuracy of prediction.

[0040] Preferably, considering the unpredictability of load fluctuations, traditional interpolation methods have limited ability to cope with extreme fluctuations. The present invention adopts an adaptive adjustment mechanism to dynamically adjust the weighting coefficients in the regression model based on real-time feedback and scheduling errors. Especially when the load fluctuation is large, the system will automatically increase the weighting of historical data to ensure the accuracy of load forecasting. The adaptive adjustment formula is as follows: in, is the standard deviation of the Gaussian function, which determines the weighting range. is the scheduling error for the current period, which measures the difference between the predicted and actual loads. is the maximum scheduling error in the current period, used for normalization.

[0041] This mechanism adjusts the weighting coefficient according to the scheduling error and historical load changes, allowing the model to flexibly respond to different load changes and ensure scheduling accuracy and stability.

[0042] As a preferred embodiment, step S14, i.e., generating a resource scheduling strategy of the virtual power plant for the target user based on the processed load forecast value and the load demand of the target user, includes: According to the processed load forecast value and the load demand of the target user, the resource scheduling strategy of the virtual power plant for the user is generated according to the following scheduling formula: in, is the resource scheduling value of user i at time step t; is the load forecast value after smoothing; is the actual load demand; are adjustment coefficients, which control the correction of forecast error and the impact of historical scheduling respectively.

[0043] In this embodiment of the present invention, the dispatch system performs optimized dispatch based on the smoothed load forecast. During the dispatch process, a preliminary load forecast is first obtained through linear interpolation, which is then further optimized through local weighted regression interpolation. Finally, the dispatch strategy is adjusted based on real-time system feedback.

[0044] Through multi-stage scheduling optimization, the system can flexibly adjust the scheduling strategy based on real-time load demand and resource conditions, thereby avoiding scheduling errors caused by sudden changes in load demand. Through this scheduling optimization, the system can make flexible adjustments based on predicted load and real-time resource scheduling requirements, thereby avoiding scheduling errors caused by sudden changes in load demand.

[0045] As a preferred embodiment, the present invention is further implemented on the basis of any of the above embodiments. Figure 2 , is a flow chart of a resource scheduling method for a virtual power plant preferably according to an embodiment of the present invention, wherein the method further comprises steps S21 to S23: S21. Calculate the scheduling selection probability of the target user using a roulette algorithm; S22. Determine the policy scheduling priority of the target user based on the scheduling selection probability; wherein the scheduling selection probability is positively correlated with the policy scheduling priority; S23: Execute the resource scheduling policy for the target user according to the policy scheduling priority.

[0046] To ensure efficient operation of virtual power plants and their ability to respond to changing load demands and emergencies, dispatch optimization relies not only on accurate load forecasting but also on real-time adaptability. By optimizing the accuracy and stability of load forecasts to provide a reliable basis for dispatch decisions, the system further optimizes dispatch strategies through a roulette wheel algorithm, intelligently and dynamically allocating resources based on real-time feedback and optimized load forecasts. This approach, through fitness weighting and sequential dispatch control, ensures efficient and balanced resource allocation in the virtual power plant while avoiding over-concentration or over-dispersion of resources.

[0047] In the embodiment of the present invention, a roulette algorithm is used to calculate the scheduling selection probability of the target user. The scheduling selection probability is used to represent the policy scheduling priority for the target user. Generally, the higher the scheduling selection probability, the higher the policy scheduling priority. Then, the resource scheduling policy is executed for the target user according to the policy scheduling priority.

[0048] Preferably, step S21, i.e., using the roulette algorithm to calculate the scheduling selection probability of the target user, includes steps S211 to S213: S211. Calculate the utility value of the target user; wherein the utility value is used to evaluate the behavior value of the user; S212. Calculate the scheduling selection probability of the target user according to the utility value; S213. Introduce a timing scheduling control method to adjust the scheduling selection probability according to the historical behavior weight of the target user to obtain a final scheduling selection probability; wherein the historical behavior weight is used to characterize the responsiveness and reliability of the target user to the resource scheduling strategy within a historical period.

[0049] In this embodiment of the present invention, to better optimize the scheduling strategy of a virtual power plant, a spatiotemporal coupled utility function was first designed. This utility function takes into account the spatiotemporal distribution characteristics of electric vehicle (EV) clusters. This utility function design helps the system consider the distance between EVs and charging stations when scheduling resources, thereby achieving geographical optimization of resources. This utility function can be expressed as: in, For users and charging piles At time step Utility value. Represents a user and charging piles At time step distance and optimize geographical distribution. Represents the time step The overall electricity demand reflects market demand. and is the adjustment coefficient, which is used to balance the influence of different utility factors.

[0050] In the roulette algorithm, each user's scheduling strategy is based on the utility value to dynamically assign selection probabilities. The user's scheduling selection probability can be calculated using the following formula: in, For users At time step The scheduling selection probability. For users At time step Utility value.

[0051] In order to avoid the imbalance caused by centralized scheduling, the present invention innovatively introduces time-series scheduling control, and the user's scheduling selection probability is adjusted according to historical load performance and market demand fluctuations. The adjusted selection probability is: in, For users At time step The historical behavior weight is dynamically adjusted according to the user's historical load performance and response capability. For users At time step The current utility value of .

[0052] By adjusting the dynamic weighted selection probability, the system can respond to load changes more flexibly, ensure a more balanced allocation of resources among different users, and effectively prevent users with lower loads from occupying too many resources, thereby improving the fairness of overall scheduling.

[0053] Preferably, the historical behavior weight is dynamically adjusted based on the user's historical performance and current load demand. , ensuring that users with larger load demands and better responses are dispatched first. The historical behavior weight is dynamically adjusted in each dispatch cycle according to the following formula: in, For users The weight of historical behavior in the previous cycle. It is the maximum value of all user utility values ​​in the current period and is used to normalize user utility. The scheduling error of the current cycle measures the difference between the actual schedule and the predicted schedule. It is the maximum scheduling error in the current cycle and is used to normalize the scheduling error. and is the adjustment coefficient, which is used to balance the impact of utility and scheduling error on the adjustment of user behavior weights.

[0054] Through this dynamic timing adjustment mechanism, the system can flexibly adjust users' scheduling priorities based on historical behavior and real-time feedback, ensuring that the system responds to load fluctuations more accurately and promptly.

[0055] As a preferred embodiment, the embodiment of the present invention is further implemented on the basis of any of the above embodiments. In step S14, that is, after generating the resource scheduling strategy of the virtual power plant for the user based on the processed load forecast value and the load demand of the target user, the method further includes steps S24 to S27: S24. Calculate a scheduling error; wherein the scheduling error is used to represent the difference between the load prediction value and the actual load of all users; S25. Calculate the utility function values ​​of the global game and the local game, and calculate the global adjustment factor and the local adjustment factor through a global-local two-level game strategy; wherein the goal of the global game layer is to optimize the resource allocation of the entire virtual power plant to balance the grid load and user demand, and the goal of the local game layer is to optimize the resource scheduling strategy of each user; S26. Calculating a game theory adjustment term based on the utility function value of the global game, the utility function value of the local game, the global adjustment factor, and the local adjustment factor; S27. Modify the resource scheduling value according to the game theory adjustment item to form a final resource scheduling strategy.

[0056] In the embodiment of the present invention, the scheduling error The system will redistribute resources according to the scheduling error in each period to optimize energy usage. Assume that the scheduling error is , represents the difference between the predicted load and the actual load of all users, reflecting the accuracy of the entire virtual power plant system scheduling. The calculation formula is: in, and For users At time step The predicted load and actual load; is the total number of users in the virtual power plant.

[0057] The system dynamically adjusts the parameters of the regression model and resource scheduling based on the scheduling error in each period to ensure the accuracy and stability of scheduling and avoid resource waste due to load fluctuations.

[0058] Furthermore, the virtual power plant's dispatch strategy is optimized using a roulette wheel algorithm. Each user's dispatch strategy is dynamically adjusted based on the calculation of a utility function, which is the core of the optimized dispatch strategy. Within a game theory framework, the global and local game layers further utilize these utility functions to coordinate resource allocation and optimize dispatch strategies. By dynamically adjusting utility functions and optimizing game strategies, the system ensures more balanced and efficient dispatch across users and resources, thereby enabling adaptive dispatch and maximizing resource utilization in complex power environments.

[0059] Specifically, the goal of the global game layer is to optimize the resource allocation of the entire virtual power plant to balance the grid load and user demand. In this process, the utility function is used to measure each user at time step The efficiency and load demand of the game can be calculated to provide a basis for resource scheduling. The utility function optimization formula of the global game can be written as follows: At the local game layer, each user chooses a scheduling strategy based on their load demand and response capability. The utility function plays a key role in local games. By reflecting user preferences and response delays, it helps the system allocate appropriate resources and scheduling strategies to users. Each user's utility function can be expressed as: in, To predict load, is the actual load; is the scheduling error; and are adjustment coefficients, which respectively control the impact of prediction error and scheduling error on utility.

[0060] Through this utility function, users can adjust their scheduling strategies according to the difference between their load forecast error and actual load demand, ensuring efficient scheduling of virtual power plants under variable load environments.

[0061] To further enhance the system's adaptability and scheduling accuracy, this paper introduces a dynamic penalty factor and dual adjustment factors to optimize the strategy update process in the game. The penalty factor and adjustment factor help the system cope with fluctuations in user behavior, load changes, and market fluctuations.

[0062] Each participant (i.e., user or dispatch resource) chooses a strategy based on their own utility function. During the game, the system optimizes the participants' strategy choices to maximize overall utility. The introduction of penalty factors and adjustment factors can optimize game strategies and prevent excessive deviations.

[0063] Dynamic penalty factor Used to penalize policies when they deviate from the system's objectives, thereby guiding the system toward convergence to the optimal solution. The penalty factor can be dynamically adjusted based on user feedback and changes in global scheduling to avoid excessive or unreasonable scheduling strategies.

[0064] in, is the dynamic penalty factor; are the current global utility and optimal utility of the system respectively; is the adjustment coefficient used to control the intensity of the penalty factor. When the system utility deviates from the optimal value, the penalty factor increases, prompting the system to adjust its strategy to reduce the deviation.

[0065] To precisely balance global and local scheduling strategies, the system introduces two adjustment factors: a global adjustment factor and a local adjustment factor. The global and local adjustment factors dynamically adjust the update rate and accuracy of the scheduling strategy, ensuring that the scheduling process can both respond to changes and maintain decision stability. The update formula for the global adjustment factor is as follows: The formula for the local adjustment factor is: in: and are global and local regulatory factors; and is the sensitivity coefficient, which controls the relationship between the adjustment factor and the policy deviation; and are the scheduling error and maximum scheduling error of the current period respectively.

[0066] By dynamically adjusting global and local regulation factors, the system can flexibly adjust scheduling strategies based on changes in load demand, actual resource utilization, and market fluctuations to ensure efficient scheduling of virtual power plant resources.

[0067] Finally, through the above multi-level game adjustment process, the system calculates the game theory adjustment items based on the optimization results of global and local strategies, utility functions, dynamic penalty factors and adjustment factors, and then dynamically adjusts the resource scheduling value of the target user.

[0068] The final resource scheduling value is the initial calculated resource scheduling value and the game theory adjustment item The sum, that is: In this process, the final scheduling output can be expressed by the following formula: in, For users At time step The resource scheduling value. For users At time step load forecast value. For users At time step The actual load value. is the scheduling error, which represents the difference between the predicted load and the actual load. The target utility expected by the system. is the global utility function at time step The current value of . is the scheduling error of the current period. It is the maximum scheduling error in the current period and is used to normalize the scheduling error. , , , , are adjustment coefficients, which are used to control the impact of prediction error, historical scheduling, scheduling error, global adjustment factor and local adjustment factor on scheduling optimization.

[0069] By adopting the technical means of the embodiment of the present invention, a dynamic adaptive collaborative dispatching system architecture of a virtual power plant based on chain feedback of user behavior is constructed, and the intelligent evolution of the dispatching strategy is realized through multi-module collaborative innovation. Figure 2 As shown in the figure, this solution adopts a four-layer progressive architecture of "data-driven modeling-prediction optimization-strategy selection-game adjustment." First, a chain feedback model of user behavior is constructed through deep reinforcement learning to capture the implicit characteristics of user electricity consumption behavior in real time. Second, a local weighted regression interpolation and adaptive adjustment mechanism are innovatively integrated to improve load forecasting accuracy. A roulette wheel algorithm with time-series optimization is then used to dynamically optimize resource allocation strategies. Finally, a global-local two-level game framework is used to achieve intelligent coordination between supply and demand. Each module is organically linked through a distributed control architecture, with user behavior feedback data running throughout the entire process, forming a dynamic closed-loop optimization system of "perception-prediction-decision-execution." This significantly improves the system's adaptability to complex user behavior and its ability to balance multi-objective optimization.

[0070] See also Figure 3, is a structural diagram of a resource scheduling device for a virtual power plant provided by an embodiment of the present invention. The embodiment of the present invention provides a resource scheduling device 10 for a virtual power plant, comprising: A user scheduling strategy generating module 11 is configured to generate a user scheduling strategy based on the collected user behavior data; wherein the user scheduling strategy is used to characterize the resource allocation tendency for the target user; A load prediction value calculation module 12 is configured to calculate the load prediction value of the target user based on the user behavior data and the user scheduling policy; The load forecast value processing module 13 is used to perform curve smoothing processing on the load forecast value using a preset interpolation algorithm to obtain a processed load forecast value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; The resource scheduling strategy generating module 14 is used to generate the resource scheduling strategy of the virtual power plant for the user according to the processed load forecast value and the load demand of the target user; wherein the resource scheduling strategy includes a resource scheduling value.

[0071] It should be noted that the resource scheduling device for a virtual power plant provided in an embodiment of the present invention is used to execute all the process steps of the resource scheduling method for a virtual power plant in the above embodiment. The working principles and beneficial effects of the two correspond one to one, and therefore will not be repeated here.

[0072] See also Figure 4 , is a structural diagram of a resource scheduling device for a virtual power plant provided in an embodiment of the present invention. An embodiment of the present invention also provides a resource scheduling device 20 for a virtual power plant, comprising a processor 21, a memory 22, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the resource scheduling method for a virtual power plant as described in any one of the above embodiments.

[0073] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the resource scheduling method of the virtual power plant as described in any one of the above embodiments.

[0074] An embodiment of the present invention also provides a computer program product, which includes a computer program or computer instructions. When the computer program or the computer instructions are executed by a processor, the resource scheduling method of the virtual power plant as described in any one of the above embodiments is implemented.

[0075] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0076] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A resource scheduling method for a virtual power plant, characterized in that: include: Calculating a user scheduling strategy based on the collected user behavior data; wherein the user scheduling strategy is used to characterize the resource allocation tendency for the target user; Calculating a load forecast value of the target user based on the user behavior data and the user scheduling policy; A preset interpolation algorithm is used to perform curve smoothing processing on the load forecast value to obtain a processed load forecast value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; Based on the processed load forecast value and the load demand of the target user, a resource scheduling strategy of the virtual power plant for the user is generated; wherein the resource scheduling strategy includes a resource scheduling value.

2. The resource scheduling method for a virtual power plant according to claim 1, characterized in that: Generating a user scheduling strategy based on the collected user behavior data includes: Collecting user behavior data; wherein the user behavior data includes the load demand, response delay and preference window of the target user; Normalizing the load demand, the response delay, and the preference window and converting them into feature vectors; The feature vector, the feature vectors of the target user's neighbor users, and the adjacency matrix are input into a preset user behavior chain feedback model for processing, and a user scheduling policy for the target user is output; wherein the adjacency matrix is ​​used to characterize the degree of mutual influence between the target user and the neighbor users, and the neighbor users are other users with behavioral associations with the target user; the user behavior chain feedback model is constructed and trained based on deep reinforcement learning.

3. The resource scheduling method for a virtual power plant according to claim 1, wherein: The method of using a preset interpolation algorithm to perform curve smoothing processing on the load forecast value to obtain the processed load forecast value includes: According to the actual load value of the target user at time (t-1) and the load forecast value at time (t+1), a preset linear interpolation algorithm is used to calculate the load forecast value at time t, which is recorded as a first load forecast value; According to the actual load values ​​of neighboring users at time t, the load prediction value of the target user at time t is calculated using the local weighted regression interpolation algorithm as the second load prediction value; wherein the neighboring users are other users with behavioral associations with the target user; A load prediction value is calculated according to the first load prediction value, the second load prediction value and a preset adaptive weight coefficient.

4. The resource scheduling method for a virtual power plant according to claim 1, wherein: The method further comprises: Using a roulette algorithm, calculating the scheduling selection probability of the target user; Determining the policy scheduling priority of the target user according to the scheduling selection probability; wherein the scheduling selection probability is positively correlated with the policy scheduling priority; The resource scheduling policy is executed for the target user according to the policy scheduling priority.

5. The resource scheduling method for a virtual power plant according to claim 4, characterized in that: The roulette algorithm is used to calculate the scheduling selection probability of the target user, including: Calculating the utility value of the target user; wherein the utility value is used to evaluate the value of the user's behavior; Calculating a scheduling selection probability of the target user according to the utility value; A timing scheduling control method is introduced to adjust the scheduling selection probability according to the historical behavior weight of the target user to obtain the final scheduling selection probability; wherein the historical behavior weight is used to characterize the responsiveness and reliability of the target user to the resource scheduling strategy within the historical period.

6. The resource scheduling method for a virtual power plant according to claim 1, wherein: After generating the resource scheduling strategy of the virtual power plant for the user according to the processed load forecast value and the load demand of the target user, the method further includes: Calculating a scheduling error; wherein the scheduling error is used to represent the difference between the load forecast value and the actual load of all users; Through a global-local dual-level game strategy, the utility function values ​​of the global game and the local game are calculated, as well as the global and local adjustment factors. The goal of the global game layer is to optimize the resource allocation of the entire virtual power plant to balance the grid load and user demand, while the goal of the local game layer is to optimize the resource scheduling strategy for each user. Calculating a game theory adjustment term according to the utility function value of the global game, the utility function value of the local game, the global adjustment factor, and the local adjustment factor; The resource scheduling value is modified according to the game theory adjustment item to form a final resource scheduling strategy.

7. A resource scheduling device for a virtual power plant, characterized in that: include: A user scheduling strategy generation module is used to generate a user scheduling strategy based on the collected user behavior data; wherein the user scheduling strategy is used to represent the resource allocation tendency for the target user; A load forecast value calculation module, configured to calculate the load forecast value of the target user based on the user behavior data and the user scheduling policy; A load forecast value processing module, configured to perform curve smoothing processing on the load forecast value using a preset interpolation algorithm to obtain a processed load forecast value; wherein the interpolation algorithm is obtained by combining a linear interpolation algorithm and a local weighted regression interpolation algorithm; A resource scheduling strategy generation module is used to generate a resource scheduling strategy of the virtual power plant for the user based on the processed load forecast value and the load demand of the target user; wherein the resource scheduling strategy includes a resource scheduling value.

8. A resource scheduling device for a virtual power plant, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements the resource scheduling method of the virtual power plant as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein, when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the resource scheduling method of the virtual power plant according to any one of claims 1 to 6.

10. A computer program product, characterized in that The computer program product includes a computer program or computer instructions, and when the computer program or the computer instructions are executed by a processor, the resource scheduling method of the virtual power plant as described in any one of claims 1 to 6 is implemented.

Images