Game-based electricity selling company decision optimization method

Abstract

The invention provides a game-based electricity selling company decision optimization method, which belongs to the technical field of electric power decision optimization, and comprises the following steps: constructing an electricity selling market inverse demand function model, carefully analyzing user load shortage power supply quantity, dividing market users into four grades, and establishing an electricity price reliability coefficient. On the basis, a trader formulates a strategy combination scheme according to a user capacity market proportion and a reliability coefficient, and constructs an expected benefit function. And deducing a dealer reaction function by solving an optimal first-order condition, and solving a Bayesian equilibrium solution by connecting the dealer reaction function in parallel, so as to accurately determine the optimal electric energy transaction volume. The research is further expanded to a marginal electricity price model of an electricity purchasing side system, an electricity selling and electricity purchasing bilateral strategy combination is constructed, a Bayesian Nash equilibrium solution in a multi-strategy mode is solved through a complex mathematical model, and finally the optimal transaction electric quantity of a dealer under different strategies is obtained, so that the problem of lack of consideration of differentiated demands of different types of users at present is solved, and the optimal transaction electric quantity of the dealer under different strategies is solved. And a targeted electricity selling strategy is difficult to push out.

Description

Technical Field

[0001] The present invention belongs to the technical field of power decision optimization, and more specifically, relates to a decision optimization method for electricity selling companies based on game theory. Background Art

[0002] With the advancement of the electricity market reform, the competition in the electricity retail market has been intensifying continuously. Different electricity selling companies have launched differentiated electricity sales products to attract various users. However, in the complex and changeable market environment, how electricity selling companies make optimal pricing and trading decisions has become a key issue to be solved urgently.

[0003] There are mainly two categories of existing decision optimization methods for electricity selling companies: one is the optimization model based on static programming, such as linear programming, dynamic programming, etc. These methods solve the optimal electricity selling strategy by establishing constraint conditions such as supply-demand balance and cost minimization. However, these models cannot accurately describe the dynamic game characteristics of the market and are prone to obtaining solutions that deviate from reality. The other category is the decision-making method based on price prediction, that is, first predicting the price trend of the electricity market and then formulating the corresponding electricity selling strategy. This method relies on an accurate price prediction model. However, the price of the electricity market is affected by multiple factors, and there is a large uncertainty in prediction.

[0004] That is to say, there is a technical problem in the prior art that the differentiated demands of different types of users are not considered, and it is difficult to launch targeted electricity selling strategies. Summary of the Invention

[0005] In view of this, the present invention provides a decision optimization method for electricity selling companies based on game theory, which can solve the technical problem in the prior art that the differentiated demands of different types of users are not considered, and it is difficult to launch targeted electricity selling strategies.

[0006] The present invention is implemented as follows: The present invention provides a decision optimization method for electricity selling companies based on game theory, including the following steps: S01. Establish an inverse demand function model for the electricity selling market, and the inverse demand function model includes the relationship between the electricity selling price in the market and the transaction electricity volume; S02. Determine the system average power shortage supply quantity index according to the power shortage supply quantity of the user load, divide the market users into 4 levels, and establish the electricity price reliability coefficients for each level of users; S03. Based on the market share of each level of capacity users and the electricity price reliability coefficients, construct a strategy combination plan for traders, and obtain the strategy set on the electricity selling side; S04. Establish an expected benefit function for traders, and the expected benefit function includes the trader strategy plan and the probability combination of the opponent choosing each strategy plan; S05. Solve the optimization first-order condition according to the expected benefit function to obtain the reaction function of each trader; S06. Establish an optimization equation system for the trader reaction function to optimize the reaction function; S07. Simultaneously solve the reaction functions to obtain the Bayesian equilibrium solution and get the optimal electricity trading volume of each trader under each strategy; S08. Establish a model of the system marginal price on the power purchase side and the system load function, and construct a bilateral strategy combination plan for the power selling side and the power purchase side; S09. Based on the bilateral strategy combination plan, establish an expected benefit function for the trader, solve the optimization first-order condition to obtain the reaction function; simultaneously solve the reaction functions to obtain the Bayesian Nash equilibrium solution under the bilateral multi-strategy mode, and get the optimal trading volume of each trader under each strategy.

[0007] Among them, the optimization equation system for the trader reaction function includes a market equilibrium equation, a profit maximization equation, a constraint condition equation, a convergence equation, and a risk assessment equation; The market equilibrium equation is used to solve the market supply and demand balance point. The inputs include the trader strategy plan, the trading volume, the system average power supply shortage index, and the market power selling price, and the outputs are the market clearing price and the optimal electricity trading volume of the trader; The profit maximization equation is used to calculate the optimal strategy of the trader. The inputs include the market clearing price, the optimal electricity trading volume of the trader, the electricity price reliability coefficient, and the market share of each level of capacity users, and the output is the maximum expected revenue of the trader; The constraint condition equation is used to ensure the feasibility of the solution. The inputs include grid operation constraints, trader quotation limits, system capacity requirements, and grid security constraints, and the output is the Bayesian equilibrium solution space; The convergence equation is used to ensure the stability of the iterative solution process. The inputs include iterative parameters, convergence parameters, iterative times limit, and convergence error limit, and the output is a stable solution; The risk assessment equation is used to evaluate the decision-making risk. The inputs include market volatility indicators, competitive strategy distributions, trading default records, and price prediction errors, and the output is the optimal decision-making plan for risk optimization.

[0008] Specifically, the relevant equations or calculation processes are described in detail as follows: 1. The inverse demand function model in the power selling market in S01 is specifically expressed as follows: ; In the formula, is the market power selling price; is the trading volume; and are the constant coefficients of the inverse demand function; is the demand disturbance term, which follows a normal distribution ; 2. The system average power supply shortage index in S02 is specifically expressed as follows: ; In the formula, is the system average power supply shortage index; is the power supply shortage of the th type of user; is the electricity load of the th type of user; is the total number of user types; is the system error term, with a range of 0.01 - 0.05; 3. The strategy combination plan in S03 is specifically expressed as follows: ; In the formula, is the strategy plan of dealer ; ; is the electricity price reliability coefficient corresponding to the th type of user; is the reliability level coefficient of the th level user; is the market share of the th level capacity users; 4. The expected profit function of the dealer in S04 is specifically expressed as follows: For the first strategy of dealer 1: ; For the second strategy of dealer 1: ; For the two strategies of dealer 2: ; ; In the formula, is the profit prediction error term, which follows a normal distribution ; 5. The reaction function in S05 is specifically expressed as follows: The reaction function of dealer 1: ; ; The reaction function of dealer 2: ; ; 6. The optimization equation system of the dealer reaction function in S06 is specifically expressed as follows: Market equilibrium equation: ; In the formula, is the optimal trading volume of the th dealer; is the market demand function; is the demand forecast error term, subject to a normal distribution ; is the demand elasticity adjustment coefficient; Profit maximization equation: ; In the formula, is the risk aversion coefficient; is the market power coefficient; Constraint condition equation: ; Convergence equation: ; In the formula, is the number of iterations; is the convergence threshold, and its value range is 0.001 - 0.01; Risk assessment equation: ; In the formula, is the value at risk; is the conditional value at risk; is the variance of returns; is the weight coefficient, and it satisfies ; 7. The system marginal electricity price and system load function model in S08 are specifically expressed as follows: ; In the formula, is the system marginal electricity price; is the system parameter; is the system disturbance term; 8. The bilateral multi-strategy game equilibrium solution in S09 is specifically expressed as follows: ; In the formula, is the strategy combination of other dealers except ; is the risk aversion coefficient; The theoretical basis and innovation points of these equations: 1. The introduction of a disturbance term in the demand function reflects the randomness of the market; 2. The profit maximization equation takes into account the factors of risk aversion and market power; 3. The constraint conditions add a change rate constraint, which is more in line with the actual trading characteristics; 4. The risk assessment adopts a multi-index comprehensive evaluation method; 5. The bilateral game model takes into account the risk aversion characteristics.

[0009] Compared with the prior art, a decision-making optimization method for electricity selling companies based on game theory provided by the present invention integrates theories such as demand modeling, profit maximization, and risk management, fully considers the dynamic game characteristics of the electricity market, and proposes differentiated electricity selling strategies for the differentiated demands of different types of users.

[0010] Specifically, the technical solution of the present invention has the following advantages: 1. An inverse demand function model including market price, trading volume, and random demand factors is established, which can more accurately reflect the price-volume relationship in the electricity market.

[0011] 2. The system average power shortage supply quantity index is determined according to the user load power shortage supply quantity, and the electricity price reliability coefficients of various types of users are established, so as to better describe the differentiated demands of different users for the reliability of electricity prices.

[0012] 3. A multi-strategy combination plan for traders is constructed, and differentiated electricity selling plans can be launched for different types of users.

[0013] 4. The game theory method is used to solve the optimal strategy response function of the trader, which better describes the dynamic game characteristics of the electricity market.

[0014] 5. A bilateral market model is established, comprehensively considering the market demand characteristics on the power purchase side, and an optimal decision-making plan for the coordination of the power purchase side and the electricity selling side is proposed.

[0015] 6. Risk assessment indicators are introduced, and the balance between revenue and risk is taken into account when solving the optimal decision.

[0016] Compared with the prior art, the solution of the present invention has obvious improvements and innovations in reflecting the dynamic characteristics of the electricity market, meeting the differentiated demands of users, and improving the accuracy and robustness of decision-making. This will help electricity selling companies maintain an advantageous position in the fierce market competition. Brief Description of the Drawings

[0017] Figure 1 It is a flowchart of the method provided by the present invention. Detailed Embodiment

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0019] like Figure 1 As shown, it is a flow chart of a decision optimization method for a power sales company based on game theory provided by the present invention. The method comprises the following steps: S01. Establish an inverse demand function model for the electricity sales market. The inverse demand function model includes the relationship between the market electricity sales price and the transaction volume; S02. Determine the average power shortage index of the system according to the power shortage of user loads, divide market users into 4 levels, and establish the power price reliability coefficient of users at each level; S03. Based on the market share of users at all levels of capacity and the reliability coefficient of electricity prices, a trader strategy combination scheme is constructed to obtain a set of strategies for the electricity sales side; S04. Establishing the expected profit function of the trader, which includes the trader's strategy plan and the probability combination of the opponent's choice of each strategy plan; S05. Solve the optimization first-order condition according to the expected profit function to obtain the reaction function of each trader; S06. Establish a trader reaction function optimization equation group to optimize the reaction function; S07. Simultaneously calculate the reaction function, solve the Bayesian equilibrium solution, and obtain the optimal electricity trading volume of the trader under each strategy; S08. Establish a system marginal electricity price and system load function model on the power purchasing side, and construct a bilateral strategy combination plan on the power selling side and the power purchasing side; S09. Establish the expected profit function of traders based on the bilateral strategy combination plan, solve the first-order optimization conditions to obtain the reaction function; solve the Bayesian Nash equilibrium solution under the bilateral multi-strategy model by combining the reaction functions to obtain the optimal trading volume of traders under each strategy.

[0020] The specific implementation methods of the above steps are described in detail below: The specific implementation of step S01 is to establish an inverse demand function model of the electricity sales market. The model can be expressed as .in The market electricity price, For trading electricity, and is the constant coefficient of the inverse demand function, is the demand disturbance term, which follows a normal distribution The model reflects the linear relationship between market price and trading volume, and takes into account the influence of random demand factors. Through the fitting analysis of historical data, it can be determined , and The specific value. This provides a theoretical basis for market pricing in subsequent steps.

[0021] The specific implementation of step S02 is to determine the system average power shortage index , and establish the electricity price reliability coefficients for users at all levels . The system average power shortage index can be expressed as . Among them is the power shortage of the th type of user, is the electricity load of the th type of user, is the total number of user types, is the system error term, and its value range is . After dividing users into 4 levels, the electricity price reliability coefficients for users at each level are determined respectively. This can better describe the different demand differences of users for electricity price reliability.

[0022] The specific implementation of step S03 is to construct the strategy combination plan of the dealer . It can be expressed as . Among them is the electricity price reliability coefficient of the th type of user, is the reliability level coefficient of the th level user, is the market share of the th level capacity user. The dealer can provide differentiated electricity sales plans for different types of users through different combined strategies .

[0023] The specific implementation of step S04 is to establish the expected profit function of the dealer. Taking the two strategies of dealer 1 as an example, its expected profit function can be expressed as: For the first strategy: ; For the second strategy: ; Among them and are the probabilities that dealer 2 selects the two strategies respectively, and are the two strategy plans of dealer 1, is the marginal cost of dealer 1, and are the profit prediction error terms, which follow the normal distribution Similarly, the expected profit functions of the two strategies of Dealer 2 can be derived. These functions describe the profit expectations of the dealer under uncertain opponent strategies.

[0024] The specific implementation of step S05 is to solve the reaction function of the dealer Taking Dealer 1 as an example, the reaction functions under its two strategies can be expressed as: For the first strategy: ; For the second strategy: ; The reaction functions of Dealer 2 under its two strategies can also be derived similarly. These reaction functions describe the optimal behavior of the dealer given the opponent's strategies.

[0025] The specific implementation of step S06 is to establish and optimize the reaction function of the dealer. First, construct an optimization system of equations including the following equations: Market equilibrium equation: ; Profit maximization equation: ; Constraint condition equation: ; Convergence equation: ; Risk assessment equation: ; Where is the optimal trading volume of the th dealer, is the market demand function, is the demand forecast error term, following a normal distribution , is the demand elasticity adjustment coefficient, is the risk aversion coefficient, is the market power coefficient, is the value at risk, is the conditional value at risk, is the variance of returns, , and are weight coefficients. By solving these equations, the optimal reaction functions of each dealer can be obtained.

[0026] The specific implementation of step S07 is to solve the Bayesian-Nash equilibrium solution. First, substitute the reaction functions of each dealer obtained in step S05 into the expected profit function in step S04 to form a system of simultaneous equations. Then, use an iterative method to solve this system of equations, and finally obtain the optimal electricity trading volume of the dealer under each strategy The purpose of this step is to determine the optimal decision for each trader given the opponent's strategy.

[0027] The specific implementation method of step S08 is to establish a function model of the system marginal electricity price on the power purchasing side and the system load, which can be expressed as .in is the system marginal electricity price, and is the system parameter, is the system disturbance term. This model reflects the market demand characteristics of the power purchasing side and provides a theoretical basis for the bilateral game in the subsequent steps.

[0028] The specific implementation method of step S09 is to establish the expected profit function of the trader based on the bilateral strategy combination scheme. And solve the optimization first-order condition to obtain the reaction function. This step is similar to step S04, except that the market characteristics of the power purchase side need to be considered. The reaction function obtained by solving It reflects the optimal decision-making behavior of traders in the power purchasing market environment. Next, we solve the Bayesian-Nash equilibrium solution under the bilateral multi-strategy model, which can be expressed as .in For Strategy portfolio of other traders, is the risk aversion coefficient. By combining the reaction functions of each trader obtained in step S09, the iterative algorithm is used to solve this set of equations, and finally the optimal transaction power of each trader under the bilateral multi-strategy game is obtained. The purpose of this step is to find the optimal decision-making plan for the coordination between the power purchase side and the power sales side.

[0029] Specifically, the principle of the present invention is to use the game theory method to build a dynamic game model between traders and solve the optimal decision-making strategy of each participant. Specifically, firstly, an inverse demand function model reflecting the price-volume relationship of the electricity market is established, and different levels of user groups are divided according to user load characteristics, and the electricity price reliability coefficient of each type of user is determined.

[0030] Next, the present invention constructs a multi-strategy combination scheme for traders, each of which provides differentiated electricity sales schemes for different types of users. Then, the expected profit function of the trader is established, which takes into account the strategy selection probability of the trader's opponents. By solving the reaction function of the trader, the optimal decision of each under a given opponent's strategy can be obtained.

[0031] In order to further optimize the decision, the present invention constructs a set of equations including market equilibrium, profit maximization, constraints, convergence and risk assessment, and solves and optimizes the reaction function of the dealer. This ensures the rationality and feasibility of the decision, and takes into account the balance between benefits and risks.

[0032] Finally, the present invention establishes a market model on the power purchase side. On this basis, a bilateral game model between the power selling side and the power purchase side is constructed, and the optimal decision-making scheme for each trader is obtained by solving its equilibrium solution. This fully considers both the supply and demand sides of the power market and better reflects the actual market game characteristics.

[0033] To better understand and implement the present invention, an embodiment of a specific application scenario of the present invention is provided below: Under the background of the power market reform, in order to improve its competitiveness, a power selling enterprise ABC decides to adopt the game-based power selling decision optimization method proposed by the present invention. The enterprise currently has a total load of about 1 million kilowatts for various types of users, mainly including three categories: industrial, commercial, and residential. Through market research and historical data analysis, the enterprise has mastered the following basic situations: 1. According to the statistics of the power supply shortage for user loads, users are divided into four levels: industrial users account for 35%, high-quality commercial users account for 25%, ordinary commercial users account for 20%, and residential users account for 20%. The system average power supply shortage index is about 0.03. Among them, the power supply shortage of industrial users accounts for a relatively high proportion, followed by commercial users, and residential users have a lower proportion.

[0034] 2. After measurement, the power price reliability coefficients of the four types of users are respectively: industrial users 0.85, high-quality commercial users 0.9, ordinary commercial users 0.8, and residential users 0.75. This reflects the different demands of different user groups for power price reliability.

[0035] 3. According to the fitting of historical power selling data, the market inverse demand function can be expressed as , where follows a normal distribution . This shows that there is a strong linear negative correlation between the market power selling price and the transaction power volume.

[0036] 4. The strategy combination options available to the power selling enterprise ABC include: for industrial users, there are two options: a basic power price of 170 yuan / MWh and a preferential power price of 150 yuan / MWh; for high-quality commercial users, there are two options: a basic power price of 200 yuan / MWh and a preferential power price of 180 yuan / MWh; for ordinary commercial users, there are two options: a basic power price of 220 yuan / MWh and a preferential power price of 200 yuan / MWh; for residential users, there are two options: a basic power price of 250 yuan / MWh and a preferential power price of 230 yuan / MWh. These differentiated sales strategies reflect the enterprise's targeting of different user groups.

[0037] Next, according to the steps of the present invention, it will be detailed how enterprise ABC uses this decision optimization method to formulate the best power selling strategy: Step S01: According to the above market inverse demand function, a price - quantity relationship model for the electricity sales market can be established as . This model can better describe the supply - demand characteristics of the electricity market and provide a theoretical basis for subsequent decision - making.

[0038] Step S02: Based on the statistical data of the power supply shortage for user loads, the system average power supply shortage index is approximately 0.03. After classifying users into four categories, the electricity price reliability coefficients of each type of user are respectively: 0.85 for industrial users, 0.9 for high - quality commercial users, 0.8 for ordinary commercial users, and 0.75 for residential users. This can better depict the differentiated demands of different user groups for electricity price reliability.

[0039] Step S03: According to the above - mentioned differentiated sales strategies, ABC Enterprises can construct their own trader strategy combination plans . Taking industrial users as an example, the basic electricity price plan can be expressed as , and the preferential electricity price plan can be expressed as . The strategy combination plans for other types of users can also be constructed similarly. These differentiated sales strategies reflect the enterprise's targeted consideration of the needs of different users.

[0040] Step S04: Assume that ABC Enterprises has a competitive relationship with another electricity seller, XYZ Enterprises. For industrial users, the expected benefit functions of the two strategy plans of ABC Enterprises can be expressed as: ; ; where and are respectively the probabilities that XYZ Enterprises choose the two strategies, is the marginal cost of ABC Enterprises. Similarly, the expected benefit function of XYZ Enterprises can be obtained. These functions describe the respective revenue expectations under the condition of uncertain opponent strategies.

[0041] Step S05: According to the above - mentioned expected benefit functions, the reaction functions of ABC Enterprises under the two strategies can be solved: For the basic electricity price plan: ; For the preferential electricity price plan: ; These reaction functions describe the optimal decisions of ABC Enterprises given the strategies of XYZ Enterprises. Similarly, the reaction function of XYZ Enterprises can be obtained.

[0042] Step S06: To further optimize the decision - making, the following optimization equations are constructed: Market equilibrium equation: ; Profit maximization equation: ; Constraint condition equation: ; Convergence equation: ; Risk assessment equation: ; By solving these equations, the optimal reaction functions of each trader can be obtained.

[0043] Step S07: Simultaneously solve the reaction functions of ABC Enterprise and XYZ Enterprise, and use the iterative algorithm to solve the Bayesian-Nash equilibrium solution, and the optimal transaction power of each party under different strategy combinations can be obtained. For example, when ABC Enterprise adopts the basic electricity price strategy and XYZ Enterprise adopts the preferential electricity price strategy, the optimal transaction power of ABC Enterprise is megawatt-hours, and the optimal transaction power of XYZ Enterprise is megawatt-hours.

[0044] Step S08: According to the system marginal electricity price and system load function on the power purchase side , a bilateral strategy combination plan for the power purchase side and the power sales side can be constructed.

[0045] Step S09: Based on this bilateral strategy combination, the expected benefit functions of ABC Enterprise and XYZ Enterprise can be established, and the reaction functions can be obtained by solving the first-order optimization conditions. For example, for the basic electricity price strategy of ABC Enterprise, its expected benefit function can be expressed as: ; where and are the probabilities of XYZ Enterprise adopting two strategies respectively.

[0046] Next, the obtained respective reaction functions are simultaneously solved, and the Bayesian-Nash equilibrium solution under the bilateral multi-strategy game can be obtained. For example, when ABC Enterprise adopts the basic electricity price strategy and XYZ Enterprise adopts the preferential electricity price strategy, the respective optimal transaction powers are megawatt-hours, megawatt-hours. This plan takes into account the interests of both the power sales side and the power purchase side and is a coordinated optimal decision.

[0047] By implementing the above steps, ABC Enterprise finally formulates the following power sales strategies: 1. For industrial users, provide two options: a basic electricity price of 170 yuan / megawatt-hour and a preferential electricity price of 150 yuan / megawatt-hour. The expected optimal transaction powers are 32,000 megawatt-hours and 28,000 megawatt-hours respectively.

[0048] 2. For high-quality commercial users, two options are provided: a base electricity price of 200 yuan per megawatt-hour and a preferential electricity price of 180 yuan per megawatt-hour. The expected optimal trading electricity volumes are 25,000 megawatt-hours and 22,000 megawatt-hours respectively.

[0049] 3. For ordinary commercial users, two options are provided: a base electricity price of 220 yuan per megawatt-hour and a preferential electricity price of 200 yuan per megawatt-hour. The expected optimal trading electricity volumes are 18,000 megawatt-hours and 16,000 megawatt-hours respectively.

[0050] 4. For residential users, two options are provided: a base electricity price of 250 yuan per megawatt-hour and a preferential electricity price of 230 yuan per megawatt-hour. The expected optimal trading electricity volumes are 15,000 megawatt-hours and 13,000 megawatt-hours respectively.

[0051] Through the differentiated sales strategy, ABC Enterprise can fully meet the needs of different types of users in the competition with XYZ Enterprise, improving its market share and profitability. At the same time, the enterprise has also taken corresponding risk management measures according to the risk preference levels of different users, such as strengthening the credit review of high-risk customers, further enhancing the robustness of decision-making.

[0052] The optimal trading electricity volume of the present invention can be used for the enterprise scheduling of the power generation of subsequent power plants.

[0053] It should be noted that the variable explanations involved in the description of the present invention are shown in Table 1 below.

[0054] Table 1 Variable Explanation Table

[0055] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention.

Claims

1. A game-based decision optimization method for a power sales company, characterized in that: The following steps are involved: S01. Establishing an inverse demand function model for the electricity sales market, wherein the inverse demand function model includes the relationship between the market electricity sales price and the transaction volume; S02. Determine the average power shortage index of the system according to the power shortage of user loads, divide market users into 4 levels, and establish the power price reliability coefficient of users at each level; S03. Based on the market share of users at all levels of capacity and the reliability coefficient of the electricity price, a trader strategy combination scheme is constructed to obtain a power sales side strategy set; S04. Establishing a trader's expected profit function, wherein the expected profit function includes the trader's strategy schemes and the probability combination of the opponent's selection of each strategy scheme; S05. Solving the optimization first-order condition according to the expected profit function to obtain the reaction function of each trader; S06. Establishing a trader reaction function optimization equation group to optimize the reaction function; S07, combining the reaction functions, solving the Bayesian equilibrium solution, and obtaining the optimal electricity trading volume of the trader under each strategy; S08. Establish a system marginal electricity price and system load function model on the power purchasing side, and construct a bilateral strategy combination plan on the power selling side and the power purchasing side; S09. Based on the bilateral strategy combination scheme, the expected profit function of the trader is established, the first-order optimization condition is solved to obtain the reaction function, and the Bayesian Nash equilibrium solution under the bilateral multi-strategy model is solved by combining the reaction functions to obtain the optimal trading volume of the trader under each strategy.

2. The method according to claim 1, characterized in that The step S01 specifically includes: Step 101: Use historical transaction data to establish a training set for the inverse demand function model of the electricity sales market, and collect market electricity prices and transaction volume data for no less than 12 months; Step 102: Perform outlier detection on the market electricity price and transaction volume data, and remove data points that are obviously deviated from the normal range; Step 103: Establish a parameter estimation equation for the inverse demand function model of the electricity sales market, and use the least squares method to obtain the constant coefficient of the inverse demand function model of the electricity sales market; Step 104: Perform a significance test on the inverse demand function model of the electricity sales market, and determine the coefficient of the inverse demand function model of the electricity sales market when the determination coefficient is greater than 0.8; Step 105: Apply the inverse demand function model of the electricity sales market to the market electricity sales price forecast, obtain the forecast error range and establish the demand disturbance term parameters.

3. The method according to claim 1, characterized in that The step S02 specifically includes: Step 201: Collect load data of various users in the system, establish user classification standards, and divide market users into four levels according to power supply reliability requirements; Step 202: Obtain the power load values ​​of users at all levels, and record the power load curves of users at all levels during the monitoring period; Step 203, counting the actual power shortage of users at all levels during the monitoring period, and establishing a power shortage database; Step 204, calculating the system average power shortage index according to the power shortage database, and determining the system error range; Step 205: Establish an electricity price reliability coefficient model for users at all levels, and associate the system average power shortage index with the electricity price reliability coefficient.

4. The method according to claim 1, characterized in that: The step S03 specifically includes: Step 301: Obtain market share data of users at each level of capacity and establish a user capacity distribution database; Step 302: construct a basic strategy set for traders according to the electricity price reliability coefficient and the market share data; Step 303: Set the reliability level coefficients of users at all levels and establish a strategy evaluation matrix; Step 304: Combine the basic strategy set with the strategy evaluation matrix to generate a trader strategy combination plan; Step 305: Conduct a feasibility assessment on the trader strategy combination scheme to determine a power sales side strategy set.

5. The method according to claim 1, characterized in that The step S04 specifically includes: Step 401: Establish a trader's expected benefit matrix, combining the benefits of the trader's strategy plan with the probability of opponent selection; Step 402: construct an opponent strategy selection probability distribution model and establish a probability selection matrix; Step 403: combining the market electricity price of the trader's strategy with the transaction volume to calculate the basic income; Step 404: introducing the trader's electricity purchase cost factor to calculate the net profit of the trader's strategy plan; Step 405: Establish an expected benefit prediction error correction model to obtain a corrected expected benefit function.

6. The method according to claim 1, characterized in that The step S05 specifically includes: Step 501: Calculate the first-order derivative of the expected benefit function to establish an optimization condition equation; Step 502: Considering market fluctuation factors, introducing market power coefficient, and revising the optimization condition equation; Step 503: establish a risk aversion correction term, and add the risk aversion correction term into the optimization condition equation; Step 504, solving the zero point of the optimization condition equation to obtain the reaction function of each trader; Step 505: Verify the stability of the reaction function to ensure the convergence of the solution.

7. The method according to claim 1, characterized in that The step S06 specifically includes: Step 601, construct a market equilibrium equation, combining the market demand function with the demand forecast error term; Step 602: Establish a profit maximization equation, introduce the risk aversion coefficient and the market power coefficient; Step 603: setting constraint equations, including electricity price restrictions, capacity restrictions, supply and demand balance restrictions, and change rate restrictions; Step 604: Establish a convergence equation and set iterative convergence conditions and thresholds; Step 605: construct a risk assessment equation, taking into account the value at risk, conditional value at risk and return variance.

8. The method according to claim 1, characterized in that The step S07 specifically includes: Step 701: Combining the reaction functions of various traders into an equation system and establishing a simultaneous equation solving model; Step 702: Solve the simultaneous equations using a numerical iteration method to obtain an initial Bayesian equilibrium solution; Step 703: Perform a stability test on the initial Bayesian equilibrium solution to ensure the robustness of the solution; Step 704: Calculate the optimal electric energy trading volume of the trader under each strategy and form a trading volume decision plan; Step 705: verify the feasibility of the transaction volume decision plan to ensure that the system operation constraints are met.

9. The method according to claim 1, characterized in that: The step S08 specifically includes: Step 801: Establish a system marginal electricity price model on the power purchase side and introduce a system load function; Step 802: Collect system disturbance data and establish a system disturbance term model; Step 803: construct a bilateral strategy space by combining the power selling side strategy set; Step 804: Evaluate the feasibility of the bilateral strategy combination scheme and select effective strategy combinations; Step 805: Establish a bilateral strategy evaluation index system to determine the optimal strategy combination solution.

10. The method according to claim 1, characterized in that The step S09 specifically includes: Step 1001: construct a new expected benefit function based on the bilateral strategy combination scheme; Step 1002: introducing a risk aversion coefficient to modify the expected benefit function; Step 1003, solving the optimization condition of the modified expected benefit function to obtain a new response function; Step 1004: Combine all reaction functions to build a bilateral multi-strategy equilibrium solution model; Step 1005: Solve the bilateral multi-strategy equilibrium solution model to obtain the final optimal trading electricity of the trader.

Images