A market clearing and unit collaborative decision-making method based on pattern recognition

Abstract

The present application provides a market clearing and unit collaborative decision-making method based on pattern recognition, comprising: analyzing market clearing data for a preset time period to extract periodic and trend changes therefrom; predicting energy demand for the preset time period using a linear regression algorithm based on the periodic and trend information; adjusting a supply and demand balance model, predicting economic benefits, and adjusting market strategies based on the current market supply and demand status; using an SJF algorithm to adjust unit configuration and set optimization parameters to adjust unit output; using a collaborative decision-making mechanism to distribute energy by time period based on the results of the unit configuration adjustment; calculating transaction costs if the energy distribution by time period results in a supply and demand balance; visualizing transaction costs and the supply and demand balance model and marking key performance indicators; continuously monitoring network stability through a data visualization interface, calibrating any abnormal fluctuations, and performing automatic adjustments.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a market clearing and unit collaborative decision-making method based on pattern recognition. Background Art

[0002] With the continuous development of the energy market and the increasing complexity of supply and demand relationships, traditional energy management methods are no longer able to accurately predict and rationally allocate energy demand. Existing energy market monitoring systems often lack the ability to deeply analyze and predict trends in market clearing data, and are unable to provide accurate supply and demand balance models and time-based energy allocation plans. In terms of data management, traditional methods are unable to accurately filter collected data values. Data processing is often ineffective, and market cyclical and trend changes are not accurately predicted, which is detrimental to market efficiency. The key to energy management systems is data application. The lack of accurate data monitoring, the inability to balance comprehensive data analysis capabilities and the ability to tap the potential application value of energy management systems, prevents the maximization of market benefits. Summary of the Invention

[0003] The present invention provides a market clearing and unit collaborative decision-making method based on pattern recognition, which mainly includes:

[0004] Analyze the market clearing data of the preset time period and extract cyclical and trend changes from it; use the linear regression algorithm to predict the energy demand of the preset time period based on the cyclical and trend information; calculate the price elasticity if the correlation between the energy demand forecast and the market clearing indicator derived from the ARIMA model is higher than the preset threshold; adjust the supply and demand balance model based on the current market supply and demand status, predict economic benefits, and adjust the market strategy; use the SJF algorithm to adjust the unit configuration and set optimization parameters to adjust the unit output; use the collaborative decision-making mechanism to allocate energy by time period based on the results of the unit configuration adjustment; if the energy allocation by time period results in a supply and demand balance, calculate the transaction cost; visualize the transaction cost and the supply and demand balance model, and mark the key performance indicators; continuously monitor the network stability through the data visualization interface, calibrate any abnormal fluctuations, and make automatic adjustments.

[0005] In one embodiment, analyzing the market clearing data for a preset period of time to extract periodic and trend changes therefrom includes:

[0006] Based on the market clearing data, data preprocessing methods are adopted, including data cleaning and normalization; outliers in the data are found, data correction is performed, and purified time series data is obtained; the ARIMA model in time series analysis is used to perform periodic analysis on the purified data to identify periodic patterns in the data; if the evaluation indicators AIC and BIC of the ARIMA model meet the optimization standards, the periodic characteristics of the data are determined, and the model is adjusted to improve the accuracy of the prediction; linear regression analysis is applied to perform trend analysis on data points with significant periodic characteristics to determine the long-term trend changes of the data; the output results of the linear regression model are used to predict the future trend of the market clearing data and obtain a prediction model for future market behavior; based on the output of the prediction model, analysis is performed to identify key change points and potential risk areas in the market; through the analysis results, the prediction model is refined to obtain more accurate predictions of market periodicity and trend changes; the refined prediction results are used to adjust the market strategy to adapt to the periodicity and trend changes in the market and optimize the market decision-making process.

[0007] In one embodiment, the energy demand for a preset period is predicted using a linear regression algorithm based on the periodicity and trend information, including:

[0008] Extract key features from cyclical and trend information, including seasonal fluctuations and long-term trends; use the extracted features to build a model, initially selecting a linear regression algorithm for energy demand forecasting; train a linear regression model based on historical energy demand data and the extracted features; evaluate the prediction accuracy of the trained model, and if it does not meet the preset threshold, use random forest training; tune the parameters of the random forest model to improve prediction accuracy; use the tuned model to forecast energy demand for a preset time period; verify the reliability of the prediction results and analyze the deviation of the model by comparing it with actual data; once the model is accurate, deploy it in a production environment for real-time prediction.

[0009] In one embodiment, if the correlation between the energy demand forecast and the market clearing indicator derived from the ARIMA model is higher than a preset threshold, calculating the price elasticity includes:

[0010] Based on the cyclical and trend information obtained from the ARIMA model, energy demand for a preset time period is predicted; if the correlation between the energy demand forecast and the market clearing indicator obtained from the ARIMA model is higher than the preset threshold, price elasticity is calculated; through price elasticity data, the supply and demand in the market are balanced and a supply and demand balance model is constructed; unit configuration is adjusted and optimization parameters are set to adjust unit output; based on the results of the unit configuration adjustment, energy is allocated by time period using a collaborative decision-making mechanism; if the energy allocation by time period results in a supply and demand balance, transaction costs are calculated; transaction costs and the supply and demand balance model are used to visualize data and mark important indicators; network stability is continuously monitored through the data visualization interface, any abnormal fluctuations are calibrated, and automatic adjustments are made.

[0011] In one embodiment, adjusting the supply and demand balance model, predicting economic benefits, and adjusting market strategies based on the current market supply and demand status include:

[0012] Obtain supply attribute data, and obtain relevant data on demand attributes through research and market surveys, including demand volume, demand price, number of demanders, and purchasing power of demanders; obtain market attribute data through market research, statistical data, and industry reports, including market size, market competition level, market growth rate, and market structure; obtain external environment attribute data, including economic conditions, policies and regulations, and social and cultural data; based on the obtained supply and demand attribute data, draw supply and demand curves, representing the supply curve and demand curve respectively; the supply curve slopes upward to the right, and the demand curve slopes downward to the left; determine the supply and demand balance by analyzing the intersection of the supply and demand curves. Point, that is, the point where the supply quantity equals the demand quantity, indicating the state of supply and demand equilibrium in the market; based on the supply and demand equilibrium point, predict the market economic benefits; evaluate the economic benefits by calculating the price and quantity corresponding to the supply and demand equilibrium point, as well as the supplier's profit and the demander's consumption expenditure indicators; adjust the market strategy in a timely manner according to the predicted results of the supply and demand equilibrium point; if the supply and demand equilibrium point price exceeds the preset threshold, increase the supply quantity or improve the supply efficiency to meet market demand and increase profits; establish a monitoring mechanism to regularly obtain supply and demand attribute data and market environment data to adjust the supply and demand balance model in a timely manner; continuously monitor market changes and make timely adjustments to adapt to changes in market demand.

[0013] In one embodiment, the SJF algorithm is used to adjust the unit configuration and set the optimization parameters to adjust the unit output, including:

[0014] Use real-time data monitoring to capture the output data of the current unit; use the SJF algorithm to analyze the unit data to determine whether configuration adjustments are needed; if the data analysis shows that adjustment is needed, optimize the parameter settings to make the unit output match the recommended level; if the unit output does not meet the predetermined efficiency after adjustment, use the parameter adjustment logic to analyze and identify the specific factors that have caused the efficiency to be lower than the preset threshold; improve the efficiency of the unit configuration through targeted adjustment of optimization parameters; based on the adjusted output data, use the efficiency evaluation standard to evaluate the performance; if the performance evaluation results do not meet the established goals, refine the optimization parameter settings and continue to adjust until the efficiency requirements are met; once the unit output reaches the optimal configuration, use the current optimized parameter settings as a benchmark; monitor real-time data and predict future changes in supply and demand, and adjust the preset parameter range in a timely manner according to the predicted results; use the configuration efficiency optimization index to plan the optimization goals for the next stage; if there is a difference between the current output and the next stage goal, fine-tune the parameters based on the difference and gradually approach the goal; integrate the optimized configurations of all units to ensure overall supply and demand balance.

[0015] In one embodiment, the energy allocation by time period is performed using a collaborative decision-making mechanism based on the result of the unit configuration adjustment, including:

[0016] The current energy usage status is analyzed through a real-time data processing system, and the unit configuration data is recorded; the unit configuration data is used to adopt the ARIMA model to predict energy demand in each future period; a collaborative decision-making mechanism is used to match energy demand based on the predicted results with the actual energy supply status; based on the matching results, if the predicted demand exceeds the actual supply, the resource scheduling system is activated to balance supply and demand; based on the balanced data, the gradient descent algorithm is used to adjust the energy allocation plan of each unit to improve efficiency; after executing energy allocation, the deviation between actual consumption and plan is analyzed through the energy consumption monitoring system; based on the deviation analysis results, the parameters of the ARIMA model are adjusted to accurately predict subsequent energy demand; based on the adjusted prediction, the collaborative decision-making mechanism is again used to optimize the resource scheduling plan; through the energy flow optimization control system, energy is ensured to be allocated according to the updated scheduling plan to achieve optimal energy use.

[0017] In one embodiment, if the energy allocation by time period results in a supply and demand balance, calculating the transaction cost includes:

[0018] Based on energy supply data and demand forecast analysis, calculate the expected energy supply and demand balance; design a real-time monitoring system to continuously track energy usage data by time period; use the ARIMA model to identify energy usage patterns and trends from energy usage data; if the monitored energy usage rate is lower than the expected threshold, adjust the energy quota based on historical data; use regression analysis to calculate the relationship between energy usage and cost in different time periods; if the cost analysis shows that it is lower than the preset threshold, make adjustment suggestions through the automated decision support system; based on the adjustment suggestions, optimize cost-effectiveness and implement a new energy allocation plan; through target evaluation, evaluate the deviation between the actual effect of the new strategy and the expected target; combine the deviation results with the trading strategy to further optimize the management of energy allocation and transaction costs.

[0019] In one embodiment, the data visualization of the transaction cost and supply-demand balance model and the annotation of key performance indicators include:

[0020] After data cleaning is completed, the Bayesian probability algorithm is used to conduct a preliminary analysis of the data to determine key performance indicators, including transaction costs, transaction prices, payment fees, freight, and market demand; after the key performance indicators are determined, a preliminary chart of the indicators is generated using data visualization software; if the chart matches the data of the supply and demand balance model, the chart elements are further refined to improve the visual presentation effect; if the user interface feedback is positive, an interactive chart is constructed through the front-end framework; if the system interoperability test passes, the chart is integrated into the data dashboard; if the indicator interpretation tool verifies the accuracy of the chart data, user usage status collection is initiated; if the collected user usage data indicates high user satisfaction, the model is finally optimized; after the optimization is completed, if the analysis model shows that the data is consistent with the market trend, the data visualization project is completed and released to the production environment.

[0021] In one embodiment, the continuous monitoring of network stability through a data visualization interface, identifying any abnormal fluctuations, and performing automatic adjustments include:

[0022] Through real-time data analysis, the real-time status of network performance is determined, and specific parameters that deviate from normal values ​​are identified; based on the results of real-time analysis, the local outlier factor algorithm is used to analyze the influencing factors of abnormal fluctuations; an automatic adjustment system is used to calibrate parameters, and deviated network performance indicators are corrected based on the parameters identified by the local outlier factor algorithm; based on the results of automatic adjustment, the adjustment effect is monitored through the data visualization interface to ensure that the network performance indicators are restored to the predetermined range; if the network performance indicators are monitored to deviate again, the local outlier factor algorithm is re-run and the adjustment strategy of the automatic adjustment system is updated; an early warning mechanism is designed to immediately send a system performance alarm when an abnormal network performance indicator is detected, achieving a rapid response; based on the system feedback logic, the accuracy of the early warning mechanism and the local outlier factor algorithm is evaluated, and the monitoring and adjustment strategies are continuously optimized.

[0023] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0024] The present invention discloses a method for market clearing and unit collaborative decision-making based on pattern recognition. By analyzing market clearing data from the past year and extracting information on cyclical and trend changes, energy demand for the next three months can be predicted. Based on the high correlation between energy demand forecasts and market clearing indicators derived from pattern recognition, price elasticity is calculated to understand the sensitivity of energy demand to price changes. A supply and demand balance model is constructed to optimize unit configuration and adjust unit output. This method allocates energy by time period to achieve supply and demand balance, visualizes data, and labels important indicators to enable decision makers to better understand market conditions. Network stability is continuously monitored through a data visualization interface, any abnormal fluctuations are calibrated, and automatic adjustments are made to ensure the consistency and accuracy of the data set. This method can achieve the goals of market clearing and unit collaboration, providing a stable and reliable energy supply for the energy market. This method can realize real-time monitoring and adjustment of network performance to ensure that network performance indicators always remain within a predetermined range. The above effects or functions include improving the accuracy of energy demand forecasts, achieving supply and demand balance, optimizing unit configuration, ensuring the flexibility and adaptability of unit configuration, reducing energy transaction costs, and ensuring the stability of the energy network. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 The present invention is a flow chart of a method for market clearing and unit collaborative decision-making based on pattern recognition.

[0026] Figure 2 Schematic diagram of a market clearing and unit collaborative decision-making method based on pattern recognition according to the present invention. DETAILED DESCRIPTION

[0027] The following will describe the technical solutions in the embodiments of the present invention in detail with reference to the accompanying drawings. The described embodiments are only a part of the embodiments of the present invention.

[0028] In this embodiment, a method for market clearing and unit collaborative decision-making based on pattern recognition may specifically include:

[0029] S101. Analyze the market clearing data of a preset period to extract periodic and trend changes.

[0030] Based on market-clearing data, data preprocessing methods, including data cleaning and normalization, are employed. Outliers are identified and corrected to obtain cleaned time series data. The ARIMA model, a model used in time series analysis, is then used to perform periodic analysis on the cleaned data to identify cyclical patterns. If the ARIMA model's evaluation metrics, AIC and BIC, meet the optimization criteria, the data's cyclical characteristics are determined, and the model is adjusted to improve forecast accuracy. Linear regression analysis is then used to perform trend analysis on data points with significant cyclical characteristics to identify long-term trends in the data. The output of the linear regression model is used to predict future trends in the market-clearing data, generating a predictive model for future market behavior. Based on the output of the predictive model, key market change points and potential risk areas are analyzed. Based on these analytical results, the predictive model is refined to obtain more accurate forecasts of market cyclical and trend changes. Using these refined forecasts, market strategies are adjusted to adapt to market cyclical and trend changes, optimizing the market decision-making process.

[0031] For example, let's perform time series analysis on the monthly sales of an e-commerce platform to predict future sales trends. First, perform data cleaning and normalization. The raw data is as follows: monthly sales: January 1.2 million, February 1.5 million, March 1.8 million, April 2 million, May 2.2 million, June 2.5 million, July 2.8 million, August 3 million, September 3.3 million, October 3.6 million, November 4 million, and December 4.2 million. Perform data cleaning and check for missing values ​​or outliers. There are no missing values ​​or outliers in the dataset. Next, perform normalization to convert sales to a range between 0 and 1. The minimum-maximum normalization method can be used, using the following formula: Normalized sales = (sales - minimum sales) / (maximum sales - minimum sales). Assuming the minimum sales is 1.2 million yuan and the maximum sales is 4.2 million yuan, the normalized data is as follows: |Month| Normalized sales: January: 0, February: 0.71 million yuan, March: 1.43 million yuan, April: 1.79 million yuan, May: 2.14 million yuan, June: 2.86 million yuan, July: 3.57 million yuan, August: 3.93 million yuan, September: 4.64 million yuan, October: 5.36 million yuan, November: 6.43 million yuan, December: 6.79 million yuan. Time series analysis can be performed to analyze cyclical patterns. If the ARIMA model's evaluation metrics, AIC and BIC, meet the optimization criteria, the data demonstrates cyclical characteristics. If the ARIMA model determines that the data exhibits a quarterly cyclical pattern, the model can be adjusted to improve forecast accuracy. Linear regression analysis can be used to analyze trends in data points with significant cyclical characteristics to identify long-term trends. The data demonstrates a positive linear trend. The output of the linear regression model can be used to predict future trends in market clearing data. The linear regression model equation is: Sales = 50 + 20 * Months. Based on the output of the forecast model, analysis can be performed to identify key market changes and potential risk areas. The forecast model predicts continued sales growth in the coming months. The analysis results allow the forecast model to be refined to obtain more accurate predictions of market cyclical and trend changes. If the model is refined into a combined forecast model of the ARIMA model and the linear regression model, the refined forecast results can be used to adjust marketing strategies to adapt to market cyclical and trend changes and optimize market decision-making. Based on the forecast results, decisions can be made to increase advertising investment and inventory management to meet future sales growth.

[0032] S102. Based on the periodicity and trend information, use a linear regression algorithm to predict the energy demand for a preset time period.

[0033] Extract key features from cyclical and trend information, including seasonal fluctuations and long-term trends. Use the extracted features to build a model, initially selecting a linear regression algorithm for energy demand forecasting. Train the linear regression model based on historical energy demand data and the extracted features. Evaluate the trained model's prediction accuracy. If it falls short of a preset threshold, train the model using a random forest algorithm. Optimize the parameters of the random forest model to improve prediction accuracy. Use the optimized model to forecast energy demand for a preset time period. Verify the reliability of the prediction results by comparing them with actual data and analyzing any deviations. Once the model is confirmed to be accurate, deploy it in a production environment for real-time prediction.

[0034] For example, suppose we have energy demand data for the past 12 months and want to predict energy demand for the next three months. We extract cyclical information about energy demand. We observe significant seasonal fluctuations in energy demand in summer and winter, as well as a long-term upward trend. We use the extracted features to build an initial linear regression model for energy demand forecasting. We find that energy demand for the past 12 months is positively correlated with average temperature. We train the linear regression model using average temperature as a feature input. We evaluate the prediction accuracy of the trained linear regression model. We use the root mean square error (RMSE) as the evaluation metric, with a preset threshold of 10. If the RMSE exceeds the threshold, the model's predictions are inaccurate. In this case, we try a more complex model, the random forest model. We tune the parameters of the random forest model to improve prediction accuracy. We use a grid search method to adjust the number of trees and the maximum depth of the random forest. We use the tuned random forest model to forecast energy demand for the next three months and compare it with the actual data. If the predicted results differ from the actual data by no more than 5%, the model's predictions are reliable. After confirming the model's accuracy, we deploy it in a production environment for real-time predictions. By inputting the model into the latest temperature data and historical data every day, energy demand for the next three months can be predicted in real time.

[0035] S103. If the correlation between the energy demand forecast and the market clearing indicator derived from the ARIMA model is higher than a preset threshold, the price elasticity is calculated.

[0036] Based on the cyclical and trend information obtained from the ARIMA model, energy demand for a preset time period is predicted; if the correlation between the energy demand forecast and the market clearing indicator obtained from the ARIMA model is higher than the preset threshold, price elasticity is calculated; through price elasticity data, the supply and demand in the market are balanced and a supply and demand balance model is constructed; unit configuration is adjusted and optimization parameters are set to adjust unit output; based on the results of the unit configuration adjustment, energy is allocated by time period using a collaborative decision-making mechanism; if the energy allocation by time period results in a supply and demand balance, transaction costs are calculated; transaction costs and the supply and demand balance model are used to visualize data and mark important indicators; network stability is continuously monitored through the data visualization interface, any abnormal fluctuations are calibrated, and automatic adjustments are made.

[0037] For example, we are forecasting electricity demand, and the market clearing indicator is the monthly electricity supply and demand balance rate. The threshold is set at 80%, that is, when the supply and demand balance rate is lower than 80%, the market supply and demand relationship is considered to be tight. Based on historical data analysis, it is found that the correlation between the market clearing indicator and electricity demand is higher than the preset threshold. Linear regression analysis is used to calculate the price elasticity. The regression model is: demand quantity

[0038] = α + β * price, where α and β are regression coefficients. Regression analysis yielded the following results: α = 1000 β = -5. This means that a 1% increase in price per unit will result in a 5% decrease in electricity demand. The regression model was optimized using the ARIMA model. If the ARIMA model can more accurately identify market dynamics, the forecasting model will be updated to improve its accuracy. The ARIMA model can identify more market factors, such as weather and economic indicators, thereby improving the forecasting model. The updated model becomes: Demand = α + β1 * price + β2 * weather + β3 * economic indicators. The updated forecasting model better matches the market clearing indicators. To further improve model accuracy, the market clearing indicators were refined, for example, by analyzing them in more detail based on time and region. Subsequently, the model predictions were tested using real-time monitoring of market dynamics. The real-time monitoring showed that the forecasting model matched the market dynamics well, validating its accuracy. The forecasting data was used to adjust the price elasticity calculation parameters. Based on the real-time monitoring results, a 1% increase in price per unit actually resulted in a 7% decrease in electricity demand. Adjust the price elasticity calculation parameter to -7. Generate a trend analysis report on energy demand and supply, which includes the price elasticity calculation results and real-time monitoring data of market dynamics, as well as forecasts and recommendations for future trends.

[0039] S104. Adjust the supply and demand balance model, predict economic benefits, and adjust market strategies based on the current supply and demand status of the market.

[0040] Obtain supply attribute data. Through research and market surveys, obtain demand attribute data, including demand volume, demand price, number of demanders, and their purchasing power. Obtain market attribute data, including market size, market competition, market growth rate, and market structure, through market research, statistical data, and industry reports. Obtain external environmental attribute data, including economic conditions, policies and regulations, and social and cultural data. Based on this acquired supply and demand attribute data, draw supply and demand curves, representing the supply curve and demand curve, respectively. The supply curve slopes upward to the right, while the demand curve slopes downward to the left. By analyzing the intersection of the supply and demand curves, determine the supply-demand equilibrium point—the point where supply equals demand, indicating a state of supply-demand equilibrium in the market. Based on the supply-demand equilibrium point, predict the market's economic benefits. Economic benefits are assessed by calculating the price and quantity corresponding to the supply-demand equilibrium point, as well as supplier profits and demander spending indicators. Based on the predicted supply-demand equilibrium point, adjust market strategies in a timely manner. If the price at the supply-demand equilibrium point exceeds a preset threshold, increase supply or improve supply efficiency to meet market demand and increase profits. Establish a monitoring mechanism to regularly obtain supply and demand attribute data and market environment data to timely adjust the supply and demand balance model. Continuously monitor market changes and make timely adjustments to adapt to changes in market demand.

[0041] For example, given supply attribute data, we know that the supply of a certain product is 1,000 units, the supply price is 10 yuan per unit, there are 20 suppliers, and the supplier production capacity is 50 units per month. Simultaneously, given demand attribute data, we know that the demand for the product is 800 units, the demand price is 15 yuan per unit, there are 10 demanders, and the purchasing power of the demanders is 100 yuan per month. Based on market attribute data, we know that the market size is 10,000 units, the market competition level is moderate, the market growth rate is 5%, and the market structure is monopolistic competition. Based on this, we construct a supply and demand curve. The supply curve slopes from the lower left to the upper right, while the demand curve slopes from the upper right to the lower left. The intersection is the equilibrium point of supply and demand. Data analysis shows that the equilibrium price is 15 yuan per unit and the quantity is 800 units. Based on this equilibrium point, we can predict the market's economic benefits. The supplier's cost is 8 yuan per unit, and the demander's spending power is 120 yuan. Based on the price and quantity at the equilibrium point for supply and demand, the supplier's profit is 7 yuan per unit, and the demander's expenditure is 100 yuan. Therefore, the predicted market economic benefits are: 7 yuan per unit * 800 units = 5,600 yuan for the supplier, and 100 yuan per unit * 800 units = 80,000 yuan for the demander. Adjust market strategies based on the forecast results. If the predicted equilibrium price is higher, increase supply or improve supply efficiency to meet market demand and increase profits. Increase supply by increasing the number of suppliers or improving their production capacity, or improve supply efficiency by optimizing supply chain management. To promptly respond to changes in market demand, establish a monitoring mechanism to regularly collect data on supply and demand attributes and market conditions. Continuously monitor market changes and make timely adjustments to adapt to changes in market demand. If market competition increases, increase market share by lowering prices or implementing marketing promotions. This solution enables timely adjustments to the supply and demand equilibrium model, predicts economic benefits, and adjusts market strategies accordingly.

[0042] S105. Use the SJF algorithm to adjust the unit configuration and set optimization parameters to adjust the unit output.

[0043] Real-time data monitoring captures current unit output data. The SJF algorithm analyzes unit data to determine whether configuration adjustments are necessary. If data analysis indicates adjustments are necessary, parameter settings are optimized to bring unit output to the recommended level. If the adjusted unit output does not meet the predetermined efficiency, parameter adjustment logic is used to analyze and identify the specific factors that caused the efficiency to fall below the preset threshold. Targeted parameter adjustments are made to improve unit configuration efficiency. Based on the adjusted output data, performance is evaluated using efficiency evaluation criteria. If the performance evaluation results do not meet the established targets, the optimized parameter settings are refined and adjustments are continued until efficiency requirements are met. Once the unit output reaches the optimal configuration, the current optimized parameter settings are used as a baseline. Real-time data is monitored and future supply and demand fluctuations are predicted. Preset parameter ranges are adjusted accordingly. Configuration efficiency optimization indicators are used to plan the next optimization target. If the current output differs from the next target, parameters are fine-tuned to gradually approach the target. The optimized configurations of all units are integrated to ensure overall supply and demand balance.

[0044] For example, a power plant has three coal-fired generating units, each with an installed capacity of 100 MW. Real-time data monitoring reveals the actual output data for the three units: Unit 1 outputs 95 MW, Unit 2 outputs 100 MW, and Unit 3 outputs 98 MW. Based on this data, the SJF algorithm analyzes the current grid demand, which is 290 MW. The calculated supply-demand gap is 290 MW - (95 MW + 100 MW + 98 MW) = -3 MW, meaning the power supply exceeds demand by 3 MW. Because the power supply exceeds demand, parameter settings are optimized, and it is determined that no unit configuration adjustment is necessary. Based on the current unit output data, parameter settings are optimized to bring the unit outputs closer to the recommended level. The parameters of Unit 2 are fine-tuned to bring its output closer to the model-recommended 100 MW. After the optimized parameter settings, the unit output is adjusted to 99 MW. If the unit output still fails to meet the target efficiency after adjustment, the parameter adjustment logic needs to be analyzed to identify the specific factors causing the efficiency to fall below the preset threshold. Output data analysis reveals that the output of unit 2 does not meet the target efficiency. Further analysis reveals that the fuel supply of unit 2 is unstable, causing the efficiency to fall below the preset threshold. To address this issue, fuel supply parameter settings can be optimized, such as increasing fuel reserves or optimizing the supply path, to improve the efficiency of the unit configuration. After adjusting the parameters of unit 2, another analysis is conducted to determine whether the preset efficiency requirements are met. If the output of unit 2 reaches the optimal configuration, the current optimized parameter settings are used as a benchmark to simplify subsequent adjustments. Based on the adjusted output data, performance is evaluated using efficiency evaluation criteria. The evaluation criteria are the deviation between unit output and demand. If the deviation is less than or equal to 1 MW, performance is considered to meet the requirements. Real-time data is continuously monitored and future supply and demand changes are forecasted. Based on the forecast results, the preset parameter ranges are adjusted as appropriate to ensure the flexibility and adaptability of the unit configuration. The preset parameter ranges are adjusted based on real-time data and forecast results. Based on future increases or decreases in grid demand, timely adjust the unit's operating capacity or adjust the unit's start-up and shutdown strategies. Based on the configuration efficiency optimization indicator, plan the next phase of optimization targets. The optimization goal is to control the supply-demand gap within 1MW. Fine-tune parameters: If the current output differs from the next phase target, fine-tune the parameters based on the difference to gradually approach the target. Integrate the optimized configurations of all units to ensure overall supply and demand balance. Perform similar optimization adjustments on other units to maintain overall power supply stability and reliability.

[0045] S106. Based on the results of the unit configuration adjustment, a collaborative decision-making mechanism is used to distribute energy in different time periods.

[0046] A real-time data processing system analyzes current energy usage and records unit configuration data. Using this unit configuration data, an ARIMA model is used to predict energy demand for each future period. A collaborative decision-making mechanism is used to match energy demand with the predicted results and actual energy supply. If the predicted demand exceeds the actual supply, the resource scheduling system is activated to balance supply and demand. Based on the balanced data, a gradient descent algorithm is used to adjust the energy allocation plan for each unit to improve efficiency. After energy allocation is executed, the energy consumption monitoring system analyzes the deviation between actual consumption and the plan. Based on the deviation analysis results, the ARIMA model parameters are adjusted to accurately predict subsequent energy demand. Based on the adjusted forecast, the collaborative decision-making mechanism is again used to optimize the resource scheduling plan. The energy flow optimization control system ensures that energy is allocated according to the updated scheduling plan, optimizing energy use.

[0047] For example, a real-time data processing system analyzes current energy usage and records unit configuration data. The system analyzes current energy usage and determines a total energy demand of 1,000 units. Using this recorded data, an ARIMA model is used to predict energy demand for each future period. Based on data from the past week, the ARIMA model predicts an energy demand of 1,200 units for the next period. A collaborative decision-making mechanism is used to match energy demand with the predicted results and actual energy supply. The actual energy supply is 1,100 units, while the predicted demand is 1,200 units. Based on the collaborative decision-making mechanism, the system determines that demand exceeds supply and needs to activate the resource scheduling system to balance supply and demand. Based on the matching results, if the predicted demand exceeds the actual supply, the resource scheduling system is activated to balance supply and demand. The resource scheduling system adjusts energy allocation to other units, bringing the actual supply to 1,200 units to match the predicted demand. Based on this balanced data, a gradient descent algorithm is used to adjust the energy allocation plan for each unit to improve efficiency. Energy reallocation through the optimization algorithm results in a more balanced energy allocation for each unit, improving overall efficiency. After energy allocation is executed, the energy consumption monitoring system analyzes the deviation between actual consumption and the plan. Actual consumption is 1150 units, a deviation from the planned 1200 units. Based on the deviation analysis results, the ARIMA model parameters are adjusted to accurately predict future energy demand. Based on the deviation analysis results, the ARIMA model parameters need to be adjusted to more accurately predict future energy demand. Based on the adjusted forecast, the collaborative decision-making mechanism is again used to optimize resource scheduling. Based on the adjusted forecast demand of 1250 units, the system again uses the collaborative decision-making mechanism to optimize resource scheduling to ensure supply and demand balance. The energy flow optimization control system ensures that energy is allocated according to the updated scheduling plan, optimizing energy use. The energy flow optimization control system allocates energy according to the optimal scheduling plan, improving energy efficiency. The actual energy supply is 1100 units, and the predicted demand is 1200 units. The resource scheduling system adjusts the actual supply to 1200 units, achieving supply and demand balance.

[0048] S107. If the energy allocation by time period results in a supply-demand balance, calculate the transaction cost.

[0049] Calculate the expected energy supply and demand balance based on energy supply data and demand forecast analysis. Design a real-time monitoring system to continuously track energy usage data by time period. Use the ARIMA model to identify patterns and trends in energy usage from energy usage data. If energy usage is monitored to be below the expected threshold, adjust energy quotas based on historical data. Use regression analysis to calculate the relationship between energy usage and costs in different time periods. If the cost analysis shows that it is below the preset threshold, make adjustment recommendations through the automated decision support system. Based on the adjustment recommendations, optimize cost-effectiveness and implement a new energy allocation plan. Use target evaluation to assess the deviation between the actual effect of the new strategy and the expected target. Combine the deviation results with the trading strategy to further optimize energy allocation and transaction cost management.

[0050] For example, based on energy supply data and demand forecast analysis, the current energy supply is 1000 units and the demand forecast is 1200 units. Calculate the expected energy supply and demand balance: supply and demand balance = demand forecast - energy supply = 1200-1000 = 200 units. Design a real-time monitoring system to continuously track energy usage data by time period. At different times of the day, the energy usage data is as follows: 08:00-10:00 energy usage is 150 units, 10:00-12:00 energy usage is 180 units, 12:00-14:00 energy usage is 200 units, 14:00-16:00 energy usage is 220 units, and 15:00-16:00 energy usage is 220 units.

[0051] Energy usage was 190 units between 4:00 PM and 6:00 PM. An ARIMA model was used to identify patterns and trends in energy usage data. The ARIMA model identified a gradually increasing trend in energy usage. If energy usage is lower than expected, the energy quota is adjusted based on historical data. Historical data shows that the average energy usage for the same period is 180 units. The energy quota is adjusted to 180 units. Regression analysis is used to calculate the relationship between energy usage and costs for different time periods. The regression analysis yields the following equation: Cost = 5 * Energy Usage. If the cost analysis shows a decrease below the preset threshold, the automated decision support system will provide adjustment recommendations. Based on the cost analysis, which indicates a gradually increasing trend, the system recommends reducing energy usage. Based on the adjustment recommendations, cost-effectiveness is optimized and a new energy allocation plan is implemented. Based on the recommended energy reduction, the energy quota is adjusted to 150 units. Target evaluation assesses the deviation between the actual effectiveness of the new strategy and the expected target. Actual energy usage is 150 units. Deviation = Expected Energy Use - Actual Energy Use = 150 - 150 = 0 units. Combining the deviation results with the trading strategy further optimizes energy allocation and transaction cost management. Since the deviation result is 0 units, the current energy allocation and transaction cost management have reached an ideal state and no further optimization is required.

[0052] S108. Visualize the transaction cost and supply-demand balance model data and mark the key performance indicators.

[0053] After data cleaning is complete, a Bayesian probability algorithm is used to conduct a preliminary analysis of the data to determine key performance indicators, including transaction costs, transaction prices, payment fees, shipping costs, and market demand. Once the key performance indicators are determined, data visualization software is used to generate preliminary charts of the indicators. If the chart matches the data from the supply and demand balance model, the chart elements are further refined to enhance the visual presentation. If the user interface feedback is positive, an interactive chart is constructed using the front-end framework. If the system interoperability test passes, the chart is integrated into the data dashboard. If the indicator interpretation tool verifies the accuracy of the chart data, user usage data collection is initiated. If the collected user usage data indicates high user satisfaction, the model is finally optimized. After optimization is complete, if the analysis model indicates that the data is consistent with market trends, the data visualization project is completed and released to the production environment.

[0054] For example, consider collecting transaction data, including information such as product prices, shipping costs, and payment fees. Consider a dataset containing 1,000 transaction records. Data cleaning is required to ensure accuracy and completeness. Using a Python script for data cleaning can remove duplicate data and address missing values. After data cleaning, perform a preliminary analysis using a Bayesian probability algorithm. Understand the relationship between transaction costs and prices. Calculate the total cost of each transaction and the average cost. For example, if the product price in a transaction is 100 yuan, the shipping cost is 10 yuan, and the payment fee is 5 yuan, the total cost of the transaction is 115 yuan. By summarizing the costs of all transactions, the average cost is 110 yuan. Use data visualization software to generate a bar chart to display the average cost. Create a bar chart with the x-axis representing different time periods and the y-axis representing the average cost. By observing the bar chart, you can visualize the trend of average cost over different time periods. If the average cost shows an upward trend over a certain period, further analysis may be necessary to identify the cause of the increase and take appropriate measures. If the bar chart matches the preset supply and demand balance model data, further refine the chart elements to enhance the visual presentation. Add colors, labels, and more. If user feedback on the chart's interface is positive, use the front-end framework to build an interactive chart. By adding a mouse-over effect, users can view specific transaction cost data. Build the interactive chart and conduct system interoperability testing to ensure that the chart can interact correctly with other systems. If the test passes, integrate the chart into the data dashboard. After integration into the data dashboard, use the indicator interpretation tool to verify the accuracy of the chart data. Interpret the average cost for a specific time period to determine whether the cost meets expectations. If the data accuracy is verified, initiate user usage collection. By collecting user usage data, understand user satisfaction with the data visualization project. If user satisfaction is high, conduct final optimization of the model. Model accuracy can be improved by adding more transaction features. If the analysis model shows that the data is consistent with market trends, publish the data visualization project to the production environment for relevant personnel to use.

[0055] S109. Continuously monitor network stability through a data visualization interface, identify any abnormal fluctuations, and make automatic adjustments.

[0056] Through real-time data analysis, the real-time status of network performance is determined and specific parameters that deviate from normal values ​​are identified. Based on the results of real-time analysis, the local outlier factor algorithm is used to analyze the factors influencing abnormal fluctuations. An automatic adjustment system is used to calibrate parameters and correct deviated network performance indicators based on the parameters identified by the local outlier factor algorithm. Based on the results of automatic adjustment, the adjustment effect is monitored through a data visualization interface to ensure that network performance indicators return to the predetermined range. If network performance indicators are monitored to deviate again, the local outlier factor algorithm is rerun and the adjustment strategy of the automatic adjustment system is updated. An early warning mechanism is designed to immediately issue system performance alarms when abnormal network performance indicators are detected, enabling rapid response. Based on system feedback logic, the accuracy of the early warning mechanism and the local outlier factor algorithm is evaluated, and the monitoring and adjustment strategies are continuously optimized.

[0057] For example, real-time data analysis can be used to determine the real-time status of network performance and identify specific parameters that deviate from normal values. To monitor a company's network latency, the latency of each packet can be collected in real time and the average latency calculated. If network latency exceeds a preset threshold, a local outlier factor algorithm is activated to analyze the factors contributing to the abnormal fluctuation. Network bandwidth utilization is checked for excessive or network failures. Once the cause of the anomaly is determined, the automatic tuning system is used to calibrate parameters. If low network bandwidth causes increased latency, the router's bandwidth settings are automatically adjusted to improve network performance. After parameter calibration, the effectiveness of the adjustments is monitored through a data visualization interface. A real-time curve of network latency is plotted to ensure that network performance indicators return to the predetermined range. If network performance indicators deviate again, the local outlier factor algorithm is rerun to update the automatic tuning system's adjustment strategy. The parameter calibration algorithm is optimized, or network device configurations are adjusted. To ensure rapid response to anomalies, an early warning mechanism is designed. When an abnormal network performance indicator is detected, the system immediately sends a system performance alert. Alerts are sent via email or text message to network administrators, allowing them to take timely action. Based on system feedback logic, evaluate the accuracy of the early warning mechanism and local outlier factor algorithm, and continuously optimize the monitoring and adjustment strategies. Analyze the accuracy and false alarm rate of alarm triggering to determine the direction of improvement.

[0058] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A market clearing and unit collaborative decision-making method based on pattern recognition, characterized in that: The method comprises: Analyze market-clearing data for a preset time period to extract cyclical and trend changes; use a linear regression algorithm to predict energy demand for the preset time period based on the cyclical and trend information; calculate price elasticity if the correlation between the energy demand forecast and the market-clearing indicator derived from the ARIMA model is higher than a preset threshold; adjust the supply and demand balance model based on the current market supply and demand status, predict economic benefits, and adjust market strategies; use the SJF algorithm to adjust unit configuration and set optimization parameters to adjust unit output; use a collaborative decision-making mechanism to allocate energy by time period based on the results of the unit configuration adjustment; calculate transaction costs if the energy allocation by time period results in a supply and demand balance; visualize transaction costs and the supply and demand balance model, and annotate key performance indicators; continuously monitor network stability through the data visualization interface, identify any abnormal fluctuations, and make automatic adjustments; The energy demand for a preset period is predicted using a linear regression algorithm based on the periodicity and trend information, including: Extract key features from cyclical and trend information, including seasonal fluctuations and long-term trends; use the extracted features to build an initial linear regression model for energy demand forecasting; train the linear regression model based on historical energy demand data and the extracted features, using the average temperature as the feature input for linear regression model training, and evaluate the prediction accuracy of the trained linear regression model using the root mean square error as the evaluation metric; based on the prediction accuracy of the evaluated trained model, if it does not meet the preset threshold, use random forest training; tune the parameters of the random forest model, using the grid search method to adjust the number and maximum depth of the random forest trees; use the tuned random forest model to predict energy demand for the next three months and compare it with the actual data; after confirming that the random forest model meets the preset requirements, deploy it in a production environment for real-time prediction, input the latest temperature data and historical data into the random forest model every day, and predict energy demand for the next three months in real time; Based on the results of the unit configuration adjustment, a collaborative decision-making mechanism is used to distribute energy in different time periods, including: analyzing the current energy usage status and recording the unit configuration data, analyzing the current energy usage to obtain the total energy demand, and using the ARIMA model to predict the energy demand for the next time period based on the data of the past week. The collaborative decision-making mechanism is used to match the energy demand with the actual energy supply status according to the prediction results. Based on the matching results, if the predicted demand exceeds the actual supply, the resource scheduling system is started to balance supply and demand. The energy allocation of other units is adjusted through the resource scheduling system to match the predicted demand. Based on the balanced data, the gradient descent algorithm is used to adjust the energy allocation plan of each unit. Energy is reallocated through the optimization algorithm. After executing the energy allocation, the deviation between the actual consumption and the plan is analyzed. According to the deviation analysis results, it is concluded that the parameters of the ARIMA model need to be adjusted. Based on the adjusted prediction, the collaborative decision-making mechanism is used again to optimize the resource scheduling plan. The collaborative decision-making mechanism is used again to optimize resource scheduling to ensure supply and demand balance. The energy flow optimization control system ensures that energy is allocated according to the updated scheduling plan to achieve optimal energy use.

2. The method according to claim 1, wherein The analysis of the market clearing data for a preset period of time to extract periodic and trend changes includes: Based on market-clearing data, data preprocessing methods are used, including data cleaning and normalization. Outliers in the data are identified and corrected to obtain cleaned time series data. Using the ARIMA model in time series analysis, we conduct a periodic analysis on the cleansed data to identify periodic patterns in the data. If the ARIMA model's evaluation indicators, AIC and BIC, meet the optimization criteria, we determine the periodic characteristics of the data and adjust the model to improve the accuracy of the forecast. Apply linear regression analysis to trend data points with significant cyclical characteristics to determine long-term trend changes in the data. Utilize the output of the linear regression model to predict future trends in market clearing data and develop a predictive model for future market behavior. Analyze the output of the predictive model to identify key market change points and potential risk areas. By analyzing the results, we refine the forecast model to obtain more accurate forecasts of market cyclical and trend changes; Utilize the refined forecast results to adjust market strategies to adapt to market cyclical and trend changes and optimize market decision-making processes.

3. The method according to claim 1, wherein If the correlation between the energy demand forecast and the market clearing indicator derived from the ARIMA model is higher than a preset threshold, the price elasticity is calculated, including: Based on the cyclical and trend information obtained from the ARIMA model, energy demand for a preset time period is predicted; if the correlation between the energy demand forecast and the market clearing indicator obtained from the ARIMA model is higher than the preset threshold, price elasticity is calculated; through price elasticity data, the supply and demand in the market are balanced and a supply and demand balance model is constructed; unit configuration is adjusted and optimization parameters are set to adjust unit output; based on the results of the unit configuration adjustment, energy is allocated by time period using a collaborative decision-making mechanism; if the energy allocation by time period results in a supply and demand balance, transaction costs are calculated; transaction costs and the supply and demand balance model are used to visualize data and mark important indicators; network stability is continuously monitored through the data visualization interface, any abnormal fluctuations are calibrated, and automatic adjustments are made.

4. The method according to claim 1, wherein The above mentioned adjustment of the supply and demand balance model, prediction of economic benefits and adjustment of market strategies based on the current supply and demand status of the market include: Obtain supply attribute data, and obtain relevant data on demand attributes through research and market surveys, including demand volume, demand price, number of demanders, and purchasing power of demanders; obtain market attribute data through market research, statistical data, and industry reports, including data on market size, market competition level, market growth rate, and market structure; obtain external environment attribute data, including economic conditions, policies and regulations, and social and cultural data; based on the obtained supply and demand attribute data, draw supply and demand curves, representing the supply curve and demand curve respectively; the supply curve slopes upward to the right, and the demand curve slopes downward to the left; determine the supply and demand balance by analyzing the intersection of the supply and demand curves. Point, that is, the point where the supply quantity equals the demand quantity, indicating the state of supply and demand equilibrium in the market; based on the supply and demand equilibrium point, predict the market economic benefits; evaluate the economic benefits by calculating the price and quantity corresponding to the supply and demand equilibrium point, as well as the supplier's profit and the demander's consumption expenditure indicators; adjust the market strategy in a timely manner according to the predicted results of the supply and demand equilibrium point; if the supply and demand equilibrium point price exceeds the preset threshold, increase the supply quantity or improve the supply efficiency to meet market demand and increase profits; establish a monitoring mechanism to regularly obtain supply and demand attribute data and market environment data to adjust the supply and demand balance model in a timely manner; continuously monitor market changes and make timely adjustments to adapt to changes in market demand.

5. The method according to claim 1, wherein The SJF algorithm is used to adjust the unit configuration and set the optimization parameters to adjust the unit output, including: Use real-time data monitoring to capture the output data of the current unit; use the SJF algorithm to analyze the unit data to determine whether configuration adjustments are needed; if the data analysis shows that adjustment is needed, optimize the parameter settings to make the unit output match the recommended level; if the unit output does not meet the predetermined efficiency after adjustment, use the parameter adjustment logic to analyze and identify the specific factors that cause the efficiency to be lower than the preset threshold; improve the efficiency of the unit configuration through targeted adjustment of optimization parameters; based on the adjusted output data, use the efficiency evaluation standard to evaluate the performance; if the performance evaluation results do not meet the established goals, refine the optimization parameter settings and continue to adjust until the efficiency requirements are met; once the unit output reaches the optimal configuration, use the current optimized parameter settings as a benchmark; monitor real-time data and predict future supply and demand changes, and adjust the preset parameter range in a timely manner based on the predicted results; Use configuration efficiency optimization indicators to plan the optimization goals for the next stage; If the current output differs from the target for the next stage, fine-tune the parameters based on the difference and gradually approach the target; integrate the optimized configurations of all units to ensure overall supply and demand balance.

6. The method according to claim 1, wherein If the energy allocation by time period is balanced between supply and demand, the transaction cost is calculated, including: Calculate the expected energy supply and demand balance based on energy supply data and demand forecast analysis; design a real-time monitoring system to continuously track energy usage data by time period; use the ARIMA model to identify energy usage patterns and trends from energy usage data; and adjust energy quotas based on historical data if energy usage is below the expected threshold. Utilize regression analysis to calculate the relationship between energy use and costs over different time periods. If the cost analysis shows a value below a preset threshold, an automated decision support system will be used to make adjustment recommendations. Based on the recommended adjustments, cost-effectiveness will be optimized, and a new energy allocation plan will be implemented. Target evaluation will be used to assess the deviation between the actual effects of the new strategy and the expected goals. Combining the deviation results with trading strategies will further optimize energy allocation and transaction cost management.

7. The method according to claim 1, wherein The data visualization of the transaction cost and supply-demand balance model is performed, and key performance indicators are marked, including: After data cleaning is completed, the Bayesian probability algorithm is used to conduct a preliminary analysis of the data to determine key performance indicators, including transaction costs, transaction prices, payment fees, freight, and market demand; after the key performance indicators are determined, a preliminary chart of the indicators is generated using data visualization software; if the chart matches the data of the supply and demand balance model, the chart elements are further refined to improve the visual presentation effect; if the user interface feedback is positive, an interactive chart is constructed through the front-end framework; if the system interoperability test passes, the chart is integrated into the data dashboard; if the indicator interpretation tool verifies the accuracy of the chart data, user usage status collection is initiated; if the collected user usage data indicates high user satisfaction, the model is finally optimized; after the optimization is completed, if the analysis model shows that the data is consistent with the market trend, the data visualization project is completed and released to the production environment.

8. The method according to claim 1, wherein The data visualization interface continuously monitors network stability, identifies any abnormal fluctuations, and automatically adjusts the network, including: Through real-time data analysis, the real-time status of network performance is determined, and specific parameters that deviate from normal values ​​are identified. Based on the results of the real-time analysis, the local outlier factor algorithm is used to analyze the factors affecting abnormal fluctuations. An automatic adjustment system is used to calibrate parameters and correct deviated network performance indicators based on the parameters identified by the local outlier factor algorithm. Based on the results of the automatic adjustment, the adjustment effect is monitored through a data visualization interface to ensure that the network performance indicators return to the predetermined range. If the network performance indicators are monitored to deviate again, the local outlier factor algorithm is re-run to update the adjustment strategy of the automatic adjustment system. Design an early warning mechanism to immediately send system performance alarms when abnormal network performance indicators are detected, enabling rapid response. Based on system feedback logic, evaluate the accuracy of the early warning mechanism and local outlier factor algorithm, and continuously optimize monitoring and adjustment strategies.

Images