Electric energy quotation decision-making method, system and equipment based on electric power prediction, and medium

Abstract

The invention relates to the technical field of carbon emission transactions, in particular to an electric energy quotation decision-making method, system and device based on electric power prediction and a medium. The method comprises the steps of obtaining power grid prediction data, historical operation data and self-contained power supply basic data, and performing preprocessing to obtain a standardized feature data set; an elastic network model is constructed based on feature vectors, and by balancing between Ridge regression and Lasso regression, feature correlation information is reserved, and feature sparse selection is realized; and finally formulating a quotation strategy according to a prediction result and historical transaction data. According to the invention, the key correlation between the price and the operation of the self-contained power supply can be effectively identified, and reliable decision support is provided for enterprises.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of carbon emission trading, and in particular to an electricity energy quotation decision method and system based on power prediction, a device and a medium. BACKGROUND

[0002] With the deepening of power market reform, the power market trading mechanism represented by the spot market is increasingly perfect. Electricity enterprises need to make electricity energy quotation and transaction in the day-ahead market and the real-time market, and reasonable quotation strategy is of great significance to reduce electricity cost and improve enterprise efficiency.

[0003] Because the power market price is affected by many factors, the current main methods for electricity energy price prediction are statistical analysis method based on historical data and deep learning method based on neural network, and the quotation strategy is made according to the prediction result.

[0004] However, the statistical analysis method of the prior art cannot effectively process multi-dimensional feature data, the deep learning method is prone to overfitting, and there is a lack of effective connection between the prediction result and the actual quotation strategy, resulting in poor accuracy and practicability of the quotation decision; this situation needs to be further improved. SUMMARY

[0005] In order to solve the problem of poor accuracy and practicability of the quotation decision of the existing statistical analysis method, the present application provides an electricity energy quotation decision method and system based on power prediction, a device and a medium, which adopts the following technical solution: In the first aspect, the present application provides an electricity energy quotation decision method based on power prediction, comprising the following steps: Obtain power grid prediction data, historical operation data and enterprise self-provided power supply basic data, preprocess the data to obtain a normalized feature data set, wherein the feature data set includes periodic characteristics of market clearing price, operating state characteristics of self-provided power supply and enterprise electricity load characteristics; According to the normalized feature data set, the periodic characteristics of market clearing price, the operating state characteristics of self-provided power supply and the enterprise electricity load characteristics are used to construct an elastic network model, and the elastic network model balances the model fitting degree and complexity by introducing a penalty term to determine the optimal weight coefficient; According to the elastic network model and the current power prediction data, the electricity energy price prediction result of the target period is calculated; According to the electricity energy price prediction result and the historical transaction data, the optimal quotation strategy of each period in the target period is determined; according to the optimal quotation strategy of each period and the prediction error analysis, a quotation decision support scheme is generated.

[0006] By adopting the technical scheme, since there is a complex interaction between the operation state of the self-provided power supply of the enterprise and the market price fluctuation in the power market, a large amount of redundant and noise information is contained in the historical data, and it is difficult to accurately grasp the market price trend by relying on the traditional regression method alone; and the correlation between the characteristics of the traditional statistical method is easy to cause the prediction model to be over-fitted; the power grid prediction data, the historical operation data and the enterprise self-provided power supply basic data are acquired first, and the normalized feature data set is obtained through preprocessing; then, the elastic network model is constructed based on the feature vector (including the historical same period value of the market price, the historical output value of the self-provided power supply and the historical power consumption value of the enterprise load), the correlation information between the features is retained, and the sparse selection of the features is realized by balancing between the Ridge regression and the Lasso regression; finally, the pricing strategy is formulated according to the model prediction result and the historical transaction data; the key correlation between the market price and the operation of the self-provided power supply of the enterprise is effectively identified and retained, and reliable decision support is provided for the enterprise to formulate a reasonable pricing strategy.

[0007] Optionally, the objective function of the elastic network model is specifically: wherein y i is the market clearing price of the i th sample, i = 1, 2,..., N, N is the total number of samples, β 0 is a bias term, β is a weight coefficient vector, is the feature vector of the i th sample, including the historical same period value of the market price, the historical output value of the self-provided power supply and the historical power consumption value of the enterprise load, P α (β) is a penalty term, and λ is a complexity parameter, α is a mixing parameter and 0 ≤ α ≤ 1, when α is equal to 0, it is Ridge regression, and when α is equal to 1, it is Lasso regression.

[0008] Optionally, according to the electric energy price prediction result and the historical transaction data, the optimal pricing strategy of each period in the target period is determined, specifically including the following steps: According to the electric energy price prediction result and the historical transaction data, a market participant risk preference portrait is established, the decision mode of different types of market subjects is identified, the market subject behavior characteristic data is obtained, and the types of the market subjects include power generation enterprises, power selling companies, power large users and self-provided power supply enterprises; According to the market subject behavior characteristic data, the correlation between the price fluctuation and the market subject behavior in the historical transaction data is analyzed, the correlation mechanism of the abnormal price fluctuation and the transaction behavior is quantified, and the abnormal behavior influence factor is obtained; According to the abnormal behavior influence factor, a market fluctuation prediction model considering the influence factor of the behavior mode is constructed in combination with the transaction behavior aggregation characteristics of the market participants, the irrational fluctuation deviation in the electric energy price prediction result is corrected, and the behavior mode corrected price prediction result is obtained. According to the price prediction result corrected according to the behavior mode, the strategy transaction probability and the bias cost in the historical transaction data, and the market state cycle output by the market fluctuation prediction model, a differential pricing strategy is designed for different market state stages, and an optimal pricing strategy combination in each period is determined.

[0009] By adopting the above technical solution, the traditional price prediction method only focuses on the statistical characteristics of historical data and cannot effectively depict the influence of the behavior mode factor of market subjects on the price. The application first establishes a market participant risk preference portrait, identifies the decision mode characteristics of different types of subjects such as power generation enterprises, power selling companies, power users and self-provided power enterprises; then analyzes the correlation between price fluctuations and market subject behaviors in historical transaction data, quantifies the correlation degree between abnormal price fluctuations and transaction behaviors, and obtains abnormal behavior influence factors; then a market fluctuation prediction model is constructed combined with the transaction behavior aggregation characteristics to correct the abnormal fluctuation bias in the price prediction result; finally, a differential pricing strategy is designed according to the corrected price prediction result, the strategy transaction probability and the bias cost, and the market state cycle. By introducing the behavior mode factor of market subjects, abnormal fluctuations in the market can be effectively identified and responded to, and the accuracy and reliability of the pricing decision can be improved.

[0010] Optionally, according to the price prediction result corrected according to the behavior mode, the strategy transaction probability and the bias cost in the historical transaction data, and the market state cycle output by the market fluctuation prediction model, a differential pricing strategy is designed for different market state stages, and an optimal pricing strategy combination in each period is determined, specifically including the following steps: According to the price prediction result corrected according to the behavior mode and the market subject behavior characteristic data, the historical transaction frequency, the transaction capacity proportion and the price correlation between market subjects are calculated, a market subject correlation matrix considering transaction intensity is established, direct and indirect transaction relationships are identified, and market subject correlation data is obtained; According to the market subject correlation data, the price transmission time lag and the transmission coefficient under different correlation strengths are statistically analyzed, a price transmission characteristic vector is established, the price elasticity coefficient between each correlation enterprise is calculated, and price influence transmission data is obtained; According to the market subject correlation data and the price influence transmission data, an enterprise scoring index system based on historical transaction credit is constructed, the transaction credit coefficient between enterprises is calculated, a price adjustment factor considering transaction credit is designed, and basic data for a synergistic optimization strategy is obtained; Based on the basic data of the collaborative optimization strategy, the price prediction results after behavioral pattern correction, and the market state cycle, combined with the strategy transaction probability and deviation cost in historical transaction data, the price adjustment coefficient and cost risk coefficient under different market state stages are calculated to optimize the quotation parameters for each time period and determine the optimal quotation strategy combination for each time period.

[0011] By adopting the above technical solution, traditional pricing decision-making methods ignore issues such as differences in transaction credit among market participants and inconsistent price transmission time lags, focusing only on the independent decisions of individual market participants and failing to effectively capture the impact mechanism of collaborative relationships among market participants on price formation. This application first constructs a market participant correlation matrix by analyzing historical transaction frequency, electricity share, and price correlation to identify direct and indirect transaction relationships. Then, it studies the price transmission characteristics under different correlation strengths, including transmission time lag, transmission coefficient, and price elasticity, characterizing the transmission path of price impact. Next, it establishes a corporate rating system based on historical transaction credit and introduces transaction credit coefficients to design price adjustment factors. Finally, it optimizes pricing parameters for each time period by combining market state cycles, strategy transaction probabilities, and deviation costs. By constructing a price transmission network that considers the collaborative relationships among market participants, it achieves precise design of differentiated pricing strategies, improving the adaptability and economic efficiency of pricing decisions.

[0012] Optionally, based on the basic data of the collaborative optimization strategy, the price prediction results after behavioral pattern correction, and the market state cycle, combined with the strategy transaction probability and deviation cost in historical transaction data, the price adjustment coefficient and cost risk coefficient under different market state stages are calculated to optimize the quotation parameters for each time period and determine the optimal quotation strategy combination for each time period. Specifically, this includes the following steps: Based on the price impact transmission data in the basic data of the collaborative optimization strategy, the changes in transaction frequency and price correlation among enterprises under each market state stage are statistically analyzed. Combined with the time series characteristics of the strategy transaction probability and the historical deviation cost distribution, the dynamic adjustment coefficient of the correlation strength is calculated to obtain the dynamic characteristic data of the network structure. Based on the dynamic characteristic data of the network structure, combined with the price prediction results after the behavioral pattern correction and the market state cycle, a risk assessment model based on transaction probability and deviation cost is constructed, the weight coefficients of enterprise rating indicators are updated, the price adjustment factor is reconstructed, and the transaction matching rules that take into account transaction credit are optimized. Based on the optimized transaction matching rules and the market characteristics of different market stages, the price adjustment parameters and risk aversion coefficients for each time period are calculated to generate the collaborative pricing coefficients of each market participant in the optimal pricing strategy combination.

[0013] By adopting the above technical solution, this application first analyzes the changing characteristics of inter-enterprise transaction relationships under different market state stages, and constructs a dynamic adjustment mechanism for association strength by combining the time-series characteristics of strategy transaction probability and historical deviation cost distribution; then, based on the dynamic characteristics of network structure, it establishes a risk assessment model that includes transaction probability and deviation cost, dynamically updates enterprise scoring weights, and optimizes transaction matching rules; finally, based on the characteristics of different market state stages, it generates a collaborative bidding coefficient that considers risk aversion; thus achieving adaptive optimization of bidding strategies and improving the stability and economy of electricity market transactions.

[0014] Optionally, based on the optimal pricing strategy for each time period and prediction error analysis, a pricing decision support solution is generated, which specifically includes the following steps: Obtain historical pricing strategy execution deviation data and calculate the prediction error level for each time period; Based on the prediction error level, price risk and transaction risk are identified, wherein the price risk includes the risk of non-execution due to overpricing and the risk of loss of profit due to underpricing. Based on the price risk and the transaction risk, determine the adjustment threshold and adjustment range of the pricing strategy; Based on the adjustment threshold, adjustment range, and optimal pricing strategy for each time period, a pricing decision support scheme is generated.

[0015] By adopting the above technical solution, this application first collects deviation data during the execution of historical pricing strategies, calculates the prediction error level in different time periods, and constructs a prediction reliability assessment system. Then, based on the prediction error analysis results, it identifies the price risks that may be faced in different time periods, including the risk of non-execution due to overpricing and the risk of loss of profit due to underpricing. Next, according to the identified risk type and degree, it sets the adjustment threshold and adjustment range of the pricing strategy to form a risk response mechanism. Finally, it combines the risk adjustment factors with the optimal pricing strategy to generate a defensive pricing decision support scheme, effectively balancing the execution rate and the rate of return. Especially in the case of drastic market price fluctuations and large prediction errors, it can maintain stable decision-making performance.

[0016] Optionally, based on the price risk and the transaction risk, the adjustment threshold and adjustment range of the pricing strategy are determined, specifically including the following steps: Based on historical price fluctuation patterns, a price risk assessment threshold is determined, which includes the deviation range between the intraday highest and lowest prices; Based on historical transaction data, a threshold for assessing transaction risk is determined, which includes the fluctuation range of the maximum and minimum transaction volume within a day. Calculate the adjustment range of the pricing strategy based on the price risk assessment threshold and the transaction risk assessment threshold; Based on the aforementioned adjustment range and the current market conditions, determine the adjustment threshold and adjustment range for the pricing strategy.

[0017] By adopting the above technical solution, this application first analyzes the fluctuation patterns of historical price data, extracts the deviation range of the highest and lowest prices within the day, and establishes a benchmark for price risk assessment; then, based on historical transaction data, it calculates the fluctuation range of the highest and lowest transaction volumes within the day, forming an assessment standard for transaction risk; next, it combines the assessment thresholds of price risk and transaction risk, and determines a reasonable adjustment range for the pricing strategy through multi-dimensional analysis; finally, based on the calculated adjustment range, and combined with factors such as the current market supply and demand status and competitive situation, it determines the specific adjustment threshold and adjustment range; this allows for flexible adjustment of the pricing strategy according to market changes, reducing the problems of insufficient or excessive adjustment.

[0018] Secondly, this application provides an electricity price quotation decision system based on electricity forecasting, comprising: The feature dataset acquisition module is used to acquire power grid forecast data, historical operation data and basic data of enterprise self-owned power sources, and preprocess the data to obtain a standardized feature dataset, wherein the feature dataset includes the periodic characteristics of market clearing prices, the operating status characteristics of self-owned power sources and the electricity load characteristics of enterprises. The elastic network model construction module is used to construct an elastic network model based on the normalized feature dataset, utilizing the periodic characteristics of market clearing prices, the operating status characteristics of self-provided power sources, and the electricity load characteristics of enterprises. The elastic network model determines the optimal weight coefficients by introducing a penalty term to balance the model fit goodness and complexity. The electricity price forecasting module is used to calculate the electricity price forecast result for the target period based on the elastic network model and the current electricity forecast data. The optimal pricing strategy acquisition module is used to determine the optimal pricing strategy for each period within the target period based on the electricity price forecast results and historical transaction data. The decision support solution acquisition module is used to generate a pricing decision support solution based on the optimal pricing strategy for each time period and the prediction error analysis.

[0019] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for making electricity price decisions based on electricity forecasting.

[0020] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method for making electricity price bidding based on electricity forecasting.

[0021] In summary, this application includes at least one of the following beneficial technical effects: 1. This application obtains power grid forecast data, historical operation data, and basic data of self-owned power sources, and preprocesses them to obtain a standardized feature dataset; it constructs an elastic network model based on feature vectors, and balances between Ridge regression and Lasso regression to retain feature correlation information while achieving feature sparsity selection; finally, it formulates a pricing strategy based on forecast results and historical transaction data, effectively identifying the key correlation between price and self-owned power source operation, and providing reliable decision support for enterprises. 2. This application first establishes a risk preference profile of market participants, identifying the decision-making patterns of different types of entities such as power generation companies, electricity sales companies, large power users, and self-owned power generation companies; then, it analyzes the correlation between price fluctuations and market participant behavior in historical transaction data, quantifies the degree of correlation between abnormal price fluctuations and trading behavior, and obtains the abnormal behavior influencing factor; next, it constructs a market volatility prediction model based on the clustering characteristics of trading behavior, correcting the abnormal fluctuation deviation in the price prediction results; finally, based on the corrected price prediction results, strategy execution probability and deviation cost, and market state cycle, it designs a differentiated pricing strategy; by introducing the market participant behavior pattern factor, it can effectively identify and respond to abnormal fluctuations in the market, improving the accuracy and reliability of pricing decisions; 3. This application first constructs a market participant correlation matrix by analyzing historical transaction frequency, electricity volume ratio, and price correlation to identify direct and indirect transaction relationships. Then, it studies the price transmission characteristics under different correlation strengths, including transmission lag, transmission coefficient, and price elasticity, characterizing the transmission path of price impact. Next, it establishes a corporate rating system based on historical transaction credit and introduces a transaction credit coefficient to design a price adjustment factor. Finally, it optimizes the pricing parameters for each time period by combining market state cycles, strategy transaction probability, and deviation costs. By constructing a price transmission network that considers the collaborative relationships of market participants, it achieves precise design of differentiated pricing strategies, improving the adaptability and economic efficiency of pricing decisions. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating an electricity price quotation decision method based on electricity forecasting, according to an embodiment of this application. Figure 2 This is a flowchart illustrating step S400 in an embodiment of the present application of an electricity price bidding decision method based on electricity forecasting. Figure 3 This is a flowchart illustrating step S440 in an embodiment of an electricity price bidding decision method based on electricity forecasting in this application. Figure 4 This is a flowchart illustrating step S444 in an embodiment of the present application of an electricity price quotation decision method based on electricity forecasting. Figure 5 This is a flowchart illustrating step S500 in an embodiment of an electricity price bidding decision method based on electricity forecasting in this application. Figure 6 This is a flowchart illustrating step S530 in an embodiment of the present application of an electricity price bidding decision method based on electricity forecasting. Figure 7 This is a schematic diagram of a power price decision system based on power forecasting according to an embodiment of this application; Figure 8 This is an internal structural diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0023] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0024] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0025] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0026] Firstly, this application provides a method for electricity pricing decisions based on electricity forecasting, referring to... Figure 1 It includes the following steps: S100: Obtain power grid forecast data, historical operation data, and basic data of enterprise self-owned power sources; preprocess the data to obtain a standardized feature dataset.

[0027] The feature dataset includes the periodicity of market clearing prices, the operating status of self-provided power sources, and the electricity load characteristics of enterprises.

[0028] In this embodiment, the power grid forecast data refers to the short-term load forecast data released by the power dispatching agency, which includes the time-of-use power load forecast values ​​for the next 24 hours; historical operating data includes market clearing prices, power generation output of various types of generating units, and historical records of enterprise power load; basic data of enterprise self-provided power sources includes the rated capacity of generating units, power generation costs, maintenance plans, and environmental protection restrictions.

[0029] Specifically, a data acquisition interface was first established to obtain daily forecast data from the power trading platform, one year's worth of historical operating records from the enterprise's production management system, and parameters of self-contained power sources from the equipment management system. A sliding time window method was used to preprocess the raw data, with a window length of 7 days and a step size of 1 day. Missing values ​​were filled using forward interpolation, and outliers were replaced using the median. Data features were categorized into three groups: historical price series (24×7 dimensions), self-contained power output curves (24×7 dimensions), and enterprise load curves (24×7 dimensions), with each group containing hourly data for the most recent 7 days. Finally, all data were standardized to the [0,1] interval to form a feature dataset.

[0030] S200. Based on the standardized feature dataset, a resilient network model is constructed using the periodic characteristics of market clearing prices, the operating status characteristics of self-provided power sources, and the electricity load characteristics of enterprises.

[0031] The objective function of the elastic network model is: y i Let be the market clearing price for the i-th sample, i = 1, 2, ..., N, where N is the total number of samples, β0 is the bias term, and β is the weight coefficient vector. Let P be the feature vector of the i-th sample, including the historical value of the market price for the same period, the historical output value of the self-provided power source, and the historical electricity consumption value of the enterprise load. α (β) is the penalty term, and λ is the complexity parameter, α is the mixture parameter and 0≤α≤1. When α equals 0, it is Ridge regression, and when α equals 1, it is Lasso regression.

[0032] In this embodiment, the elastic network model is a machine learning method that combines Ridge regression and Lasso regression. The objective function consists of a loss term and a penalty term, where the loss term measures the deviation between the predicted and actual values, and the penalty term controls the model complexity through a weighted sum of the L1 and L2 norms. The mixing parameter α is used to adjust the weight of the two penalties.

[0033] Specifically, a feature vector is first constructed, containing the historical price series for the past 7 days, the output curve of self-contained power sources, and the enterprise load curve. Mean squared error is chosen as the loss function, and an objective function is established by combining L1 and L2 regularization terms. Cross-validation is used to select the optimal mixture parameter α and regularization coefficient λ. The range of α is determined based on the data dimensionality ratio and feature correlation analysis: when the data dimensionality ratio is greater than 5 and there is a moderate degree of correlation between features, the search range of α is set to [0.2, 0.8]. The range of λ is determined based on the standard deviation of the response variable, with its maximum value set to 0.1 times the standard deviation and its minimum value set to 0.01 times the standard deviation, i.e., the search range is [0.01, 0.1]. The coordinate descent method is used to solve for the model parameters, and the optimal parameter combination is selected through grid search combined with cross-validation to ensure the model's predictive performance and generalization ability.

[0034] S300. Based on the elastic network model and current electricity forecast data, calculate the electricity price forecast results for the target period.

[0035] In this embodiment, the target time period refers to the next 24 hours in which electricity pricing needs to be quoted, divided into hourly price periods. The electricity price forecast result includes the predicted price value and prediction confidence interval for each period.

[0036] Specifically, a time-based data classification mechanism is established, categorizing historical data by weekdays, weekends, and holidays, and creating reference price range tables for different date types. Current power grid forecast data is input into a trained elastic network model to obtain preliminary forecast results. These are then corrected using the reference price ranges; if the forecast value exceeds a reasonable range, upper and lower limits of the reference price are used as constraints, and the final forecast result is output.

[0037] S400: Based on the electricity price forecast and historical transaction data, determine the optimal pricing strategy for each period within the target period.

[0038] In this embodiment, the optimal pricing strategy refers to a time-segmented pricing scheme that comprehensively considers the probability of execution and risk control while maximizing expected returns. Historical transaction data includes past pricing records and their execution results.

[0039] Specifically, firstly, a mapping table between price ranges and transaction probabilities is established. Historical quotes are divided into five levels based on their deviation from market prices, and the transaction rate for each level is calculated. Then, combining the predicted price and transaction probability, a pricing decision matrix based on risk preference is constructed. For periods with high prediction confidence, an aggressive pricing strategy is adopted, with quotes set in the 90%-100% range of the predicted price; for periods with low prediction confidence, a conservative pricing strategy is adopted, with quotes set in the 100%-110% range of the predicted price.

[0040] S500 generates a pricing decision support scheme based on the optimal pricing strategy for each time period and prediction error analysis.

[0041] In this embodiment, the pricing decision support scheme refers to a complete decision-making scheme that includes specific pricing recommendations and risk warnings. Prediction error analysis refers to the accuracy assessment of historical prediction results.

[0042] Specifically, a time-segmented forecasting error assessment index system is constructed, including three dimensions: mean absolute error, root mean square error, and maximum error. Early warning thresholds are set; when the forecasting error index for a certain time period exceeds the threshold, the pricing strategy for that period is automatically adjusted. A risk warning list is established to generate targeted pricing adjustment suggestions for situations such as changes in maintenance plans, stricter environmental restrictions, and drastic market price fluctuations. The final output is a complete decision support solution including suggested pricing values, risk warnings, and adjustment suggestions.

[0043] In one embodiment, refer to Figure 2 In step S400, based on the electricity price forecast results and historical transaction data, the optimal pricing strategy for each time period within the target period is determined, specifically including the following steps: S410. Based on the electricity price forecast results and historical transaction data, establish a risk preference profile of market participants, identify the decision-making patterns of different types of market entities, and obtain data on the behavioral characteristics of market entities.

[0044] The types of market entities include power generation companies, electricity sales companies, large electricity users, and self-owned power generation companies.

[0045] In this embodiment, the market participant risk preference profile refers to a characteristic description reflecting the risk tolerance and investment decision-making tendencies of market participants, established by analyzing their historical pricing behavior. The decision-making models include three types: conservative, moderate, and aggressive, used to characterize the pricing strategy selection tendencies of different market participants in the face of market volatility.

[0046] Specifically, the first step is to establish a database of risk appetite characteristics indicators, including indicators across three dimensions: historical price deviation, transaction price volatility, and price adjustment frequency. By setting threshold ranges for each indicator, market participants are categorized into corresponding decision-making models. For example, price deviations below the lower limit of the market average price fluctuation range are classified as conservative, those within the fluctuation range as moderate, and those above the upper limit as aggressive.

[0047] S420. Based on the behavioral characteristics data of market participants, analyze the correlation between price fluctuations and market participant behavior in historical transaction data, quantify the correlation mechanism between abnormal price fluctuations and trading behavior, and obtain the abnormal behavior impact factor.

[0048] In this embodiment, the abnormal behavior impact factor refers to an indicator used to quantify the impact of abnormal trading behaviors such as abnormal price fluctuations on electricity prices, which is obtained by analyzing the bidding behavior characteristics of market participants during periods of sharp price fluctuations.

[0049] Specifically, a table is established to correlate price fluctuations with pricing behavior, recording the direction and magnitude of price adjustments by market participants under different fluctuation ranges. When there is a continuous rise, the percentage of participants following the price increase is used as the "follow-the-rise" indicator; when there is a continuous fall, the percentage of participants following the price decrease is used as the "follow-the-fall" indicator. The abnormal behavior impact factor is calculated by combining these two indicators.

[0050] S430. Based on the abnormal behavior influencing factors and combined with the clustering characteristics of market participants' trading behavior, construct a market fluctuation prediction model that considers the influencing factors of behavioral patterns, correct the abnormal fluctuation deviation in the electricity price prediction results, and obtain the price prediction results after behavioral pattern correction.

[0051] In this embodiment, the market volatility prediction model is a machine learning model trained on historical data, used to predict price trends after considering the influence factors of market participants' behavioral patterns. The inputs include basic price prediction results, abnormal behavior influence factors, and trading behavior clustering characteristic indicators.

[0052] Specifically, a support vector regression model is used to construct a price corrector, with the basic prediction results and various behavioral characteristic indicators as input variables. The model is trained on actual transaction data during periods of significant historical price volatility, and the model parameters are optimized by minimizing the prediction error. For example, when abnormal market fluctuations are detected, the model will correct the basic prediction results based on price trends in similar historical scenarios.

[0053] S440. Based on the price prediction results after behavioral pattern correction, the strategy transaction probability and deviation cost in historical transaction data, and the market state cycle output by the market fluctuation prediction model, design differentiated pricing strategies for different market state stages and determine the optimal pricing strategy combination for each time period.

[0054] In this embodiment, the differentiated pricing strategy refers to formulating corresponding price adjustment plans based on the different characteristics of different market phases. The market phase cycle includes four phases: stable period, rising period, falling period, and period of violent fluctuation.

[0055] Specifically, a rule library mapping market state stages to pricing strategies is established. During stable periods, a robust pricing strategy based on historical average prices is employed; during rising periods, a follow-the-rise strategy is designed in conjunction with transaction probability; during falling periods, a defensive strategy is designed by controlling deviation costs; and during periods of high volatility, a risk-averse strategy is designed by comprehensively considering transaction probability and deviation costs. These strategies are combined and applied to different time periods to form a complete pricing decision-making scheme.

[0056] In one embodiment, refer to Figure 3 In step S440, based on the price prediction results after behavioral pattern correction, the strategy transaction probability and deviation cost in historical transaction data, and the market state cycle output by the market volatility prediction model, differentiated pricing strategies are designed for different market state stages to determine the optimal pricing strategy combination for each time period. Specifically, this includes the following steps: S441. Based on the price forecast results after behavioral pattern correction and the behavioral characteristic data of market entities, calculate the historical transaction frequency, transaction volume ratio and price correlation among market entities, establish a market entity association matrix considering transaction intensity, identify direct and indirect transaction relationships, and obtain data on the association relationships among market entities.

[0057] In this embodiment, the market participant association matrix is ​​a two-dimensional data structure that reflects the strength of transaction relationships between market participants. Direct transaction relationships refer to direct electricity transaction records between two market participants, while indirect transaction relationships refer to transaction connections established through intermediary entities.

[0058] Specifically, a database of indicators for assessing the strength of transaction relationships is constructed, including three dimensions: monthly transaction frequency, the proportion of monthly transaction volume in their respective total transaction volume, and price correlation coefficient. Weighting coefficients are set at 0.3, 0.4, and 0.3, respectively, and the weighted result of the three indicators is used as the correlation strength value and filled into the matrix. A correlation strength value exceeding 0.6 is considered a direct transaction relationship, while a correlation strength value between 0.3 and 0.6 with a common counterparty is considered an indirect transaction relationship.

[0059] S442. Based on the data on the relationships between market entities, statistically analyze the price transmission lag and transmission coefficient under different correlation strengths, establish a price transmission characteristic vector, calculate the price elasticity coefficient among related enterprises, and obtain price impact transmission data.

[0060] In this embodiment, the price transmission feature vector includes two key indicators: price transmission lag and transmission coefficient. Price transmission lag refers to the time difference in response to price changes among market participants, while the transmission coefficient represents the degree of impact of price changes. The price elasticity coefficient reflects the sensitivity of related enterprises to price changes.

[0061] Specifically, a price transmission characteristic analysis database is established to record price change sequences under different correlation strength intervals. Transmission lags are determined by calculating the time difference of price changes, and transmission coefficients are obtained using regression analysis. The price elasticity coefficient is obtained by calculating the ratio of price change rates among related enterprises, forming a complete dataset of price impact transmission.

[0062] S443. Based on the data on the relationship between market entities and the data on the transmission of price impacts, construct a corporate scoring index system based on historical transaction credit, calculate the transaction credit coefficient between enterprises, design a price adjustment factor that takes into account transaction credit, and obtain basic data for collaborative optimization strategies.

[0063] In this embodiment, the enterprise rating index system is a multi-dimensional evaluation system built based on historical transaction credit. The transaction credit coefficient is used to quantify the reliability of cooperation between enterprises, and the price adjustment factor adjusts the base price according to the transaction credit level.

[0064] Specifically, a scoring system is designed that includes transaction fulfillment rate, price stability, and settlement timeliness, and a scoring rule base is established. A transaction credit coefficient is determined by calculating a comprehensive score of historical transactions between enterprises, and a price adjustment range is set based on the transaction credit level to generate a price adjustment factor that takes transaction credit into account.

[0065] S444. Based on the basic data of the collaborative optimization strategy, the price prediction results after the behavior pattern correction, and the market state cycle, combined with the strategy transaction probability and deviation cost in the historical transaction data, calculate the price adjustment coefficient and cost risk coefficient under different market state stages, optimize the quotation parameters for each time period, and determine the optimal quotation strategy combination for each time period.

[0066] In this embodiment, the price adjustment coefficient is used to adjust the base price under different market conditions, and the cost risk coefficient reflects the potential economic loss risk caused by price adjustments. The optimal pricing strategy combination is a pricing scheme determined under the constraints of transaction probability and controllable deviation costs.

[0067] Specifically, a market state stage characteristic database is established to record price fluctuation characteristics and transaction patterns at different stages. Based on the foundational data of the collaborative optimization strategy, rules for calculating price adjustment coefficients are designed. A cost-risk assessment model is constructed through historical transaction data analysis to calculate the risk coefficients of each adjustment scheme. Finally, based on the transaction probability and risk control objectives, the optimal combination of pricing parameters is selected.

[0068] In one embodiment, refer to Figure 4In step S444, based on the basic data of the collaborative optimization strategy, the price prediction results after behavioral pattern correction, and the market state cycle, combined with the strategy transaction probability and deviation cost in historical transaction data, the price adjustment coefficient and cost risk coefficient under different market state stages are calculated to optimize the quotation parameters for each time period and determine the optimal quotation strategy combination for each time period. Specifically, the steps are as follows: S4441. Based on the price impact transmission data in the basic data of the collaborative optimization strategy, statistically analyze the changes in transaction frequency and price correlation among enterprises under each market state stage. Combined with the temporal characteristics of the strategy transaction probability and the historical deviation cost distribution, calculate the dynamic adjustment coefficient of the correlation strength to obtain the dynamic characteristic data of the network structure.

[0069] In this embodiment, the dynamic characteristic data of the network structure reflects the evolution of the market entity network as the market state changes cyclically. The dynamic adjustment coefficient of association strength is used to describe the time-varying characteristics of the strength of transaction relationships between enterprises, including two dimensions: the rate of change of transaction frequency and the rate of change of price correlation.

[0070] Specifically, a market state stage characteristic database is constructed to record the transaction network structure parameters under four stages: stable period, rising period, falling period, and period of violent fluctuation. The rate of change in inter-firm transaction frequency and the rate of change in price correlation are calculated using a sliding time window. Combined with the statistical distribution of current strategy execution probability and historical deviation cost, a dynamic adjustment function for correlation strength is established. The transaction network is reconstructed based on the adjusted correlation strength, forming a dynamic feature dataset of the network structure.

[0071] S4442. Based on the dynamic characteristics of the network structure, combined with the price prediction results after behavioral pattern correction and market state cycle, construct a risk assessment model based on transaction probability and deviation cost, update the weight coefficients of enterprise scoring indicators, reconstruct the price adjustment factor, and optimize the transaction matching rules that take into account transaction credit.

[0072] In this embodiment, the risk assessment model evaluates the risk and return of different pricing strategies by analyzing the historical distribution characteristics of the probability of a transaction and the cost of deviation. The weighting coefficients of the enterprise scoring indicators are dynamically adjusted according to market cycles to optimize the transaction matching rules.

[0073] Specifically, a risk assessment indicator system is established, including three dimensions: predicted transaction probability, historical average deviation cost, and cost volatility. The analytic hierarchy process (AHP) is used to determine the initial weights of each indicator, and weight adjustment rules are designed based on the dynamic characteristics of the network structure. Under different market conditions, the enterprise rating weights are dynamically updated, price adjustment factors are recalculated, and optimized transaction matching schemes are generated.

[0074] S4443. Based on the optimized transaction matching rules and the market characteristics of different market stages, calculate the price adjustment parameters and risk aversion coefficients for each time period, and generate the collaborative pricing coefficients of each market participant in the optimal pricing strategy combination.

[0075] In this embodiment, the coordinated bidding coefficient is an adjustment parameter used to coordinate bidding strategies among market participants, determined by comprehensively considering transaction matching rules and market condition characteristics. The price adjustment parameter and risk aversion coefficient control the offensive and defensive aspects of the bidding strategy, respectively.

[0076] Specifically, a rule base for mapping market state stages to pricing strategies is established, and a pricing parameter system including benchmark prices, adjustment ranges, and risk constraints is designed. Price adjustment coefficients for each time period are calculated based on transaction matching rules, and the risk aversion level is determined by considering the characteristics of the current market state stage. Finally, differentiated pricing schemes that consider synergistic effects are generated to achieve strategy coordination among market participants.

[0077] In one embodiment, refer to Figure 5 In step S500, a pricing decision support scheme is generated based on the optimal pricing strategy for each time period and the prediction error analysis. This specifically includes the following steps: S510. Obtain the execution deviation data of historical pricing strategies and calculate the prediction error level for each time period.

[0078] In this embodiment, execution deviation data refers to the record of the difference between the actual transaction results and the expected target during the historical price execution process; the prediction error level includes quantitative indicators in three dimensions: absolute error, relative error, and error direction. Time periods are divided into hourly price periods.

[0079] Specifically, a pricing execution monitoring database is established to record historical pricing strategies, predicted prices, and actual transaction prices. A prediction error calculation matrix is ​​created, with the horizontal axis representing time periods and the vertical axis representing error index types. A sliding time window method is used, with a window length of 7 days, to calculate the prediction error index for each time period. An error characteristic analysis table is established to identify the prediction accuracy characteristics of different time periods and price ranges.

[0080] S520. Based on the forecast error level, identify price risk and transaction risk. Price risk includes the risk of non-execution due to overpricing and the risk of loss of profit due to underpricing.

[0081] In this embodiment, price risk refers to the risk of economic loss due to prediction deviation; non-trading risk refers to the risk that the electricity volume cannot be traded due to the bid price being higher than the market clearing price; and revenue loss risk refers to the risk that the potential revenue will be damaged due to the bid price being lower than the market price.

[0082] Specifically, a risk identification and assessment framework is constructed, establishing a mapping relationship between prediction errors and risk types. A risk classification matrix is ​​designed, dividing risks into high, medium, and low levels based on the magnitude and direction of prediction errors. A risk quantification indicator system is established, including indicators for non-trading probability, profit / loss rate, and comprehensive risk score. Judgment thresholds for different risk levels are established through historical data statistics.

[0083] S530. Based on price risk and transaction risk, determine the adjustment threshold and adjustment range of the pricing strategy.

[0084] In this embodiment, the adjustment threshold refers to the critical condition that triggers an adjustment to the pricing strategy; the adjustment magnitude refers to the specific amount of correction to the pricing strategy. The adjustment plan needs to strike a balance between risk control and return optimization.

[0085] Specifically, a strategy adjustment decision table is established to establish a correspondence between risk levels and adjustment strategies. A tiered adjustment mechanism is designed: for high-risk periods, a large adjustment is adopted, with an adjustment range of 10%-15% of the benchmark price; for medium-risk periods, a medium adjustment is adopted, with an adjustment range of 5%-10% of the benchmark price; and for low-risk periods, a small adjustment is adopted, with an adjustment range of 0%-5% of the benchmark price.

[0086] S540. Generate a pricing decision support scheme based on the adjustment threshold, adjustment range, and optimal pricing strategy for each time period.

[0087] Specifically, a decision-making framework is constructed, integrating optimal pricing strategies with risk adjustment strategies. A time-segmented pricing suggestion table is established, recording four types of information: benchmark price, adjustment direction, reason for adjustment, and risk warning. A scheme presentation template is designed, using a combination of charts and graphs to visually present the decision-making suggestions. A scheme execution tracking mechanism is created to record the implementation effects, providing a basis for subsequent optimization. For special periods, such as high-risk periods and key control periods, supplementary special analysis suggestions are provided.

[0088] In one embodiment, refer to Figure 6 In step S530, based on price risk and transaction risk, the adjustment threshold and adjustment range of the pricing strategy are determined, specifically including the following steps: S531. Based on historical price fluctuation patterns, determine the assessment threshold for price risk, which includes the deviation range between the intraday highest and lowest prices.

[0089] In this embodiment, historical price fluctuation patterns refer to the characteristics of market price changes at different times and under different market conditions; deviation range refers to the upper and lower limits of price fluctuations; the statistical period for the intraday highest and lowest prices is 30 consecutive days to ensure the representativeness and timeliness of the data.

[0090] S532. Based on historical transaction data, determine the assessment threshold for transaction risk. The assessment threshold includes the fluctuation range of the maximum and minimum transaction volume within a day.

[0091] In this embodiment, historical transaction data includes three basic dimensions: transaction volume, transaction price, and transaction time; fluctuation range refers to the range of changes in transaction volume; the statistics of the maximum and minimum transaction volume are on a daily basis, recording the distribution characteristics of transaction volume in different time periods.

[0092] S533. Calculate the adjustment range of the pricing strategy based on the price risk assessment threshold and the transaction risk assessment threshold.

[0093] In this embodiment, the adjustment range refers to the range of price adjustments allowed by the pricing strategy; the calculation process needs to consider the constraints of both price risk and transaction risk to ensure the feasibility of the adjustment plan.

[0094] Specifically, a risk collaborative assessment framework is established to comprehensively quantify price risk and transaction risk. An adjustment range calculation matrix is ​​created, with the horizontal axis representing the price risk level and the vertical axis representing the transaction risk level. Matrix elements represent the suggested adjustment range for the corresponding risk combination. A time-segmented adjustment range table is designed to determine differentiated adjustment spaces based on the risk characteristics of different time periods.

[0095] S534. Based on the adjustment range and the current market conditions, determine the adjustment threshold and adjustment range of the pricing strategy.

[0096] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0097] Secondly, this application provides an electricity price quotation decision system based on electricity forecasting. The electricity price quotation decision system based on electricity forecasting of this application will be described below in conjunction with the above-mentioned electricity price quotation decision method based on electricity forecasting.

[0098] Reference Figure 7 A power price bidding decision system based on power forecasting, comprising: The feature dataset acquisition module is used to acquire power grid forecast data, historical operation data, and basic data of enterprise self-owned power sources, and to preprocess the data to obtain a standardized feature dataset.

[0099] The feature dataset includes the periodicity of market clearing prices, the operating status of self-provided power sources, and the electricity load characteristics of enterprises.

[0100] The elastic network model construction module is used to construct an elastic network model based on a normalized feature dataset, utilizing the periodic characteristics of market clearing prices, the operating status characteristics of self-provided power sources, and the electricity load characteristics of enterprises. The elastic network model balances the goodness of fit and complexity by introducing a penalty term to determine the optimal weight coefficients.

[0101] The objective function of the elastic network model is: y i Let be the market clearing price for the i-th sample, i = 1, 2, ..., N, where N is the total number of samples, β0 is the bias term, and β is the weight coefficient vector. Let P be the feature vector of the i-th sample, including the historical value of the market price for the same period, the historical output value of the self-provided power source, and the historical electricity consumption value of the enterprise load. α (β) is the penalty term, and λ is the complexity parameter, α is the mixture parameter and 0≤α≤1. When α equals 0, it is Ridge regression, and when α equals 1, it is Lasso regression.

[0102] The electricity price forecasting module is used to calculate the electricity price forecast for a target period based on the elastic network model and current electricity forecast data.

[0103] The optimal pricing strategy acquisition module is used to determine the optimal pricing strategy for each time period within the target period based on the electricity price forecast results and historical transaction data.

[0104] The decision support solution acquisition module is used to generate a pricing decision support solution based on the optimal pricing strategy for each time period and the prediction error analysis.

[0105] In one embodiment, this application provides an electronic device, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, this electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements an energy pricing decision-making method based on power forecasting.

[0106] Those skilled in the art will understand that Figure 8The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0107] In one embodiment, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0108] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0109] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for electricity price bidding based on electricity forecasting, characterized in that, Includes the following steps: Acquire power grid forecast data, historical operation data, and basic data of enterprise self-provided power sources. Preprocess the data to obtain a standardized feature dataset, wherein the feature dataset includes the periodic characteristics of market clearing prices, the operating status characteristics of self-provided power sources, and the electricity load characteristics of enterprises. Based on the standardized feature dataset, a resilient network model is constructed using the periodic characteristics of market clearing prices, the operating status characteristics of self-provided power sources, and the electricity load characteristics of enterprises. The resilient network model determines the optimal weight coefficients by balancing the goodness of fit and complexity of the model by introducing a penalty term. Based on the elastic network model and current electricity forecast data, calculate the electricity price forecast results for the target period; Based on the electricity price forecast results and historical transaction data, the optimal pricing strategy for each time period within the target period is determined; based on the optimal pricing strategy for each time period and the forecast error analysis, a pricing decision support scheme is generated.

2. The electricity price bidding decision method based on electricity forecasting according to claim 1, characterized in that, The objective function of the elastic network model is specifically: Among them, y i Let be the market clearing price for the i-th sample, i = 1, 2, ..., N, where N is the total number of samples, β0 is the bias term, and β is the weight coefficient vector. Let P be the feature vector of the i-th sample, including the historical value of the market price for the same period, the historical output value of the self-provided power source, and the historical electricity consumption value of the enterprise load. α (β) is the penalty term, and λ is the complexity parameter, α is the mixture parameter and 0≤α≤1. When α equals 0, it is Ridge regression, and when α equals 1, it is Lasso regression.

3. The electricity price bidding decision method based on electricity forecasting according to claim 1, characterized in that, Based on the electricity price forecast results and historical transaction data, the optimal pricing strategy for each time period within the target period is determined, specifically including the following steps: Based on the electricity price forecast results and historical transaction data, a risk preference profile of market participants is established to identify the decision-making patterns of different types of market entities and obtain behavioral characteristic data of market entities. The types of market entities include power generation companies, electricity sales companies, large electricity users and self-owned power generation companies. Based on the aforementioned market participant behavior characteristic data, the correlation between price fluctuations and market participant behavior in historical transaction data is analyzed, the correlation mechanism between abnormal price fluctuations and trading behavior is quantified, and the abnormal behavior influencing factor is obtained. Based on the abnormal behavior influencing factors and combined with the clustering characteristics of market participants' trading behavior, a market fluctuation prediction model considering the influencing factors of behavioral patterns is constructed to correct the irrational fluctuation deviation in the electricity price prediction results and obtain the price prediction results after behavioral pattern correction. Based on the price prediction results after the behavioral pattern correction, the strategy transaction probability and deviation cost in historical transaction data, and the market state cycle output by the market fluctuation prediction model, differentiated pricing strategies are designed for different market state stages, and the optimal pricing strategy combination for each time period is determined.

4. The electricity price bidding decision method based on electricity forecasting according to claim 3, characterized in that, Based on the price prediction results after the behavioral pattern correction, the strategy execution probability and deviation cost in historical transaction data, and the market state cycle output by the market volatility prediction model, differentiated pricing strategies are designed for different market state stages to determine the optimal pricing strategy combination for each time period. Specifically, this includes the following steps: Based on the price prediction results after the behavioral pattern correction and the behavioral characteristic data of market entities, the historical transaction frequency, transaction volume ratio and price correlation among market entities are calculated. A market entity association matrix considering transaction intensity is established to identify direct and indirect transaction relationships and obtain association data among market entities. Based on the data on the relationships between market entities, the price transmission time lag and transmission coefficient under different correlation strengths are statistically analyzed to establish a price transmission characteristic vector, calculate the price elasticity coefficient among each related enterprise, and obtain price impact transmission data. Based on the data on the relationships between market participants and the data on the transmission of price impacts, a corporate scoring index system based on historical transaction credit is constructed, the transaction credit coefficient between enterprises is calculated, a price adjustment factor that takes into account transaction credit is designed, and basic data for collaborative optimization strategies are obtained. Based on the basic data of the collaborative optimization strategy, the price prediction results after behavioral pattern correction, and the market state cycle, combined with the strategy transaction probability and deviation cost in historical transaction data, the price adjustment coefficient and cost risk coefficient under different market state stages are calculated to optimize the quotation parameters for each time period and determine the optimal quotation strategy combination for each time period.

5. The electricity price bidding decision method based on electricity forecasting according to claim 4, characterized in that, Based on the basic data of the collaborative optimization strategy, the price prediction results after behavioral pattern correction, and the market state cycle, combined with the strategy execution probability and deviation cost in historical transaction data, the price adjustment coefficient and cost risk coefficient under different market state stages are calculated to optimize the pricing parameters for each time period and determine the optimal pricing strategy combination for each time period. The specific steps include the following: Based on the price impact transmission data in the basic data of the collaborative optimization strategy, the changes in transaction frequency and price correlation among enterprises under each market state stage are statistically analyzed. Combined with the time series characteristics of the strategy transaction probability and the historical deviation cost distribution, the dynamic adjustment coefficient of the correlation strength is calculated to obtain the dynamic characteristic data of the network structure. Based on the dynamic characteristic data of the network structure, combined with the price prediction results after the behavioral pattern correction and the market state cycle, a risk assessment model based on transaction probability and deviation cost is constructed, the weight coefficients of enterprise rating indicators are updated, the price adjustment factor is reconstructed, and the transaction matching rules that take into account transaction credit are optimized. Based on the optimized transaction matching rules and the market characteristics of different market stages, the price adjustment parameters and risk aversion coefficients for each time period are calculated to generate the collaborative pricing coefficients of each market participant in the optimal pricing strategy combination.

6. The electricity price bidding decision method based on electricity forecasting according to claim 1, characterized in that, Based on the optimal pricing strategy and prediction error analysis for each time period, a pricing decision support solution is generated, which includes the following steps: obtaining historical pricing strategy execution deviation data and calculating the prediction error level for each time period; Based on the prediction error level, price risk and transaction risk are identified, wherein the price risk includes the risk of non-execution due to overpricing and the risk of loss of profit due to underpricing. Based on the price risk and the transaction risk, determine the adjustment threshold and adjustment range of the pricing strategy; Based on the adjustment threshold, adjustment range, and optimal pricing strategy for each time period, a pricing decision support scheme is generated.

7. The electricity price bidding decision method based on electricity forecasting according to claim 6, characterized in that, Based on the price risk and the transaction risk, the adjustment threshold and adjustment range of the pricing strategy are determined, specifically including the following steps: Based on historical price fluctuation patterns, a price risk assessment threshold is determined, which includes the deviation range between the intraday highest and lowest prices; Based on historical transaction data, a threshold for assessing transaction risk is determined, which includes the fluctuation range of the maximum and minimum transaction volume within a day. Calculate the adjustment range of the pricing strategy based on the price risk assessment threshold and the transaction risk assessment threshold; Based on the aforementioned adjustment range and the current market conditions, determine the adjustment threshold and adjustment range for the pricing strategy.

8. A power price quotation decision system based on power forecasting, characterized in that, include: The feature dataset acquisition module is used to acquire power grid forecast data, historical operation data and basic data of enterprise self-owned power sources, and preprocess the data to obtain a standardized feature dataset, wherein the feature dataset includes the periodic characteristics of market clearing prices, the operating status characteristics of self-owned power sources and the electricity load characteristics of enterprises. The elastic network model construction module is used to construct an elastic network model based on the normalized feature dataset, utilizing the periodic characteristics of market clearing prices, the operating status characteristics of self-provided power sources, and the electricity load characteristics of enterprises. The elastic network model determines the optimal weight coefficients by introducing a penalty term to balance the model fit goodness and complexity. The electricity price forecasting module is used to calculate the electricity price forecast result for the target period based on the elastic network model and the current electricity forecast data. The optimal pricing strategy acquisition module is used to determine the optimal pricing strategy for each period within the target period based on the electricity price forecast results and historical transaction data. The decision support solution acquisition module is used to generate a pricing decision support solution based on the optimal pricing strategy for each time period and prediction error analysis.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the electricity pricing decision method based on electricity forecasting 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 computer program is executed by the processor, it implements the steps of the power price bidding decision method based on power forecasting as described in any one of claims 1-7.

Images