Pricing method, device and equipment for leader type power generator, and storage medium
By setting up a simulated environment in the power market and optimizing pricing strategies using reinforcement learning algorithms, the strict restrictions in the Markov decision-making process in the application of the power market are solved, and more flexible and efficient market response and profit optimization are achieved.
Patent Information
- Application Number
- CN202411632267.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-14
- Publication Date
- 2025-02-25
AI Technical Summary
The Markov decision-making process has strict restrictions in the application of the power market, and it is difficult to capture complex market dynamics and changing market conditions, which limits its application effect in the actual power market.
By setting up a simulation environment for the multi-pricing model of leadership-based power generation, target pricing action parameters are determined, and bidding simulation evolution is used using the strategy-based reinforcement learning algorithm PRLA algorithm to optimize pricing strategies to cope with complex and changeable market conditions.
Overcoming the strict restrictions on action selection and state transition of the traditional Markov decision-making process, we can more flexibly respond to complex and changeable market conditions, effectively optimize decision-making strategies, and thus improve the ultimate profit of leading power generation companies.
Smart Images
Figure CN119515448A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of electricity price pricing, and in particular to a pricing method, device, equipment and storage medium for a leading power producer. Background Art
[0002] At present, many studies apply reinforcement learning to the study of electricity markets. In related technologies, electricity market research based on reinforcement learning usually models the market environment as a Markov decision process (MDP). This modeling method provides a framework that enables the system to take a series of actions in a given state and update the state based on the results of these actions. Specifically, by converting the dynamic changes of the electricity market and the behavior of participants into states, actions, and rewards, reinforcement learning algorithms can be used to optimize decision-making strategies. However, the Markov decision process has strict restrictions on the selection of actions and the transition of states, which leads to relatively limited application scenarios. For example, complex market dynamics and changing market conditions are often difficult to be accurately captured by traditional MDP models, which limits its application effect in actual electricity markets.
[0003] Therefore, how to overcome the strict limitations of the Markov decision process in the application of the power market is a technical problem that those skilled in the art urgently need to solve. Summary of the invention
[0004] Based on the above problems, the present application provides a pricing method, device, equipment and storage medium for a leading power producer, which can overcome the strict limitations of the Markov decision process in the application of the power market.
[0005] The embodiments of the present application disclose the following technical solutions:
[0006] A pricing method for a leading power producer, the method comprising:
[0007] Set up the simulation environment of the multiple pricing model of the leading power generator;
[0008] Determine the target pricing action parameter of the leading power generator; the target pricing action parameter is a variable;
[0009] Utilizing the policy-based reinforcement learning algorithm PRLA algorithm, based on the target pricing action parameters, the multiple pricing model of the leading power producer in the simulation environment is subjected to bidding simulation evolution to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
[0010] In a possible implementation, the setting of a simulation environment for a multiple pricing model of a leading power generator includes:
[0011] Setting a market structure of the power market and power generation demand on the demand side of the power market; the market structure includes a day-ahead market and a real-time market;
[0012] Determine pricing parameters; the pricing parameters include the constant power generation of new energy, the fixed marginal cost of the leading power generator, the elasticity of the market supply of the following power generator to the price, the elasticity of the market supply of the following power generator to the price in real time, and the fixed marginal cost of the following power generator;
[0013] Under the market structure of the power market, a pricing strategy formula for the power sales price of a leading power producer is constructed based on the power generation demand on the demand side of the power market and the pricing parameters; the pricing strategy formula for the power sales price of a leading power producer is: in, is the target clearing price for the leading generator in the day-ahead market, is the target clearing price of the leading generator in the real-time market, p1 is the clearing price of the leading generator in the day-ahead market; p2 is the clearing price of the leading generator in the real-time market; c is the fixed marginal cost of the leading generator; q1 is the residual demand faced by the leading generator in the day-ahead market; q2 is the residual demand faced by the leading generator in the real-time market; p1 and p2 are variables.
[0014] In one possible implementation,
[0015] q1=AQ-b1*p1-c1, A is the market structure of the power market and the power generation demand on the demand side of the power market, Q is the constant power generation of the new energy, b1 is the day-ahead market supply elasticity of the follower power producer to the price, and c1 is the fixed marginal cost of the follower power producer;
[0016] q2=b1*p1-b2*p2, where b2 is the elasticity of the real-time market supply of the follower power producer to the price.
[0017] In a possible implementation, the determining of the target pricing action parameters of the leading power producer includes:
[0018] The clearing price of the leading power producer in the day-ahead market and the clearing price of the leading power producer in the real-time market in the pricing parameters are determined as variables to obtain the target pricing action parameters.
[0019] In a possible implementation, the strategy-based reinforcement learning algorithm PRLA algorithm is used to perform bidding simulation evolution on the multiple pricing model of the leading power generator in the simulation environment based on the target pricing action parameter to obtain the target value of the target pricing action parameter of the leading power generator and the target profit under the target value of the target pricing action parameter, including:
[0020] Initialize the learning rate and discount factor of the PRLA algorithm;
[0021] Utilizing the policy-based reinforcement learning algorithm PRLA algorithm, based on the target pricing action parameters, the multiple pricing model of the leading power producer in the simulation environment is subjected to bidding simulation evolution to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
[0022] A pricing device for a leading power producer, comprising:
[0023] The simulation environment unit is used to set up the simulation environment of the multiple pricing model of the leading power producer;
[0024] A first determining unit is used to determine a target pricing action parameter of a leading power generator; the target pricing action parameter is a variable;
[0025] The bidding simulation evolution unit is used to utilize the policy-based reinforcement learning algorithm PRLA algorithm to perform bidding simulation evolution on the multiple pricing model of the leading power producer in the simulation environment based on the target pricing action parameters to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
[0026] In a possible implementation, the simulation environment unit specifically includes:
[0027] A power generation demand setting unit, used to set the market structure of the power market and the power generation demand on the demand side in the power market; the market structure includes a day-ahead market and a real-time market;
[0028] The second determination unit is used to determine pricing parameters; the pricing parameters include the constant power generation of new energy, the fixed marginal cost of the leading power generator, the elasticity of the market supply of the following power generator to the price, the elasticity of the real-time market supply of the following power generator to the price and the fixed marginal cost of the following power generator;
[0029] A construction unit is used to construct a pricing strategy formula for the electricity sales price of a leading power producer based on the pricing parameters; the pricing strategy formula for the electricity sales price of the leading power producer is: in, is the target clearing price for the leading generator in the day-ahead market, is the target clearing price of the leading generator in the real-time market, p1 is the clearing price of the leading generator in the day-ahead market; p2 is the clearing price of the leading generator in the real-time market; c is the fixed marginal cost of the leading generator; q1 is the residual demand faced by the leading generator in the day-ahead market; q2 is the residual demand faced by the leading generator in the real-time market; p1 and p2 are variables.
[0030] In one possible implementation,
[0031] q1=AQ-b1*p1-c1, A is the market structure of the power market and the power generation demand on the demand side of the power market, Q is the constant power generation of the new energy, b1 is the day-ahead market supply elasticity of the follower power producer to the price, and c1 is the fixed marginal cost of the follower power producer;
[0032] q2=b1*p1-b2*p2, where b2 is the elasticity of the real-time market supply of the follower power producer to the price.
[0033] A pricing device for a leading power producer comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the pricing method for the leading power producer as described above is implemented.
[0034] A computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the pricing method for the leading power producer as described above.
[0035] Compared with the prior art, this application has the following beneficial effects:
[0036] The present application provides a pricing method, device, equipment and storage medium for a leading power producer. Specifically, when executing the pricing method for a leading power producer provided in an embodiment of the present application, a simulation environment for a multiple pricing model of a leading power producer can be set first. Then the target pricing action parameters of the leading power producer are determined, and the target pricing action parameters are variables. Finally, a policy-based reinforcement learning algorithm (PRLA algorithm) is used to perform a bidding simulation evolution on the multiple pricing model of a leading power producer with a simulation environment based on the target pricing action parameters to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter. By setting a simulation environment and determining the target pricing action parameters, the present application can more flexibly respond to complex and changeable market conditions, and overcome the strict limitations of the traditional Markov decision process in action selection and state transition. At the same time, by using the PRLA algorithm, the decision-making strategy can be more effectively optimized, thereby increasing the final profit of the leading power producer. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the technical solutions in this embodiment or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0038] Figure 1 A method flow chart of a pricing method for a leading power producer provided in an embodiment of the present application;
[0039] Figure 2 A schematic diagram of the structure of a pricing device for a leading power producer provided in an embodiment of the present application. DETAILED DESCRIPTION
[0040] To facilitate understanding of the technical solutions provided by the embodiments of the present application, the background technology involved in the embodiments of the present application will be described below.
[0041] Currently, many studies are devoted to exploring the application of reinforcement learning technology in the field of power market. Among these studies, the power market research based on reinforcement learning usually adopts Markov decision process (MDP) to model the market environment. This modeling method provides a framework for the system to take a series of actions in a specific state and update the state according to the results of these actions. Specifically, by converting the dynamic changes of the power market and the behavior of the participants into states, actions and rewards, reinforcement learning algorithms can be used to optimize decision strategies. However, Markov decision processes have strict restrictions on action selection and state transitions, which leads to a relatively limited scope of application. For example, complex market dynamics and changing market conditions are often difficult to be accurately captured by traditional MDP models, which in turn limits their effective application in actual power markets.
[0042] In order to solve this problem, a pricing method, device, equipment and storage medium for a leading power producer are provided in an embodiment of the present application. First, a simulation environment of a multiple pricing model of a leading power producer is constructed, and its target pricing action parameters (these parameters are used as variables) are determined. A bidding simulation evolution is performed using a policy-based reinforcement learning algorithm (PRLA algorithm), and finally the pricing strategy and profit of the leading power producer are obtained. The advantage of this method is that by setting up a simulation environment and determining the target pricing action parameters, it is possible to respond more flexibly to complex and changeable market conditions, thereby overcoming the strict limitations of the traditional Markov decision process in action selection and state transition. At the same time, using a reinforcement learning method based on the PRLA algorithm, the decision-making strategy can be more effectively optimized, thereby significantly improving the ultimate target profit of the leading power producer.
[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0044] See also Figure 1 , which is a flowchart of a pricing method for a leading power producer provided in an embodiment of the present application, such as Figure 1 As shown, the pricing method of the leading power producer may include steps S101-S103:
[0045] S101: Set up a simulation environment for the multiple pricing model of the leading power generator.
[0046] Setting up a simulation environment for the multiple pricing model of a leading power producer refers to creating a virtual market environment for simulating and testing the pricing strategies of leading power producers under different conditions.
[0047] Specifically, the setting of the simulation environment includes the following aspects:
[0048] Setting of market structure and demand-side generation requirements: This includes the definition of the day-ahead market and the real-time market. The day-ahead market is a market that determines the supply and price one day before the actual supply, while the real-time market is a market that adjusts the price and supply according to the real-time supply and demand situation. It is also necessary to set the generation requirements A on the demand side of the power market.
[0049] Determination of pricing parameters: These parameters include the constant power generation Q of new energy, the fixed marginal cost c of the leading generator, the elasticity b1 of the day-ahead market supply of the follower generator to the price, the elasticity b2 of the real-time market supply of the follower generator to the price and the fixed marginal cost c1 of the follower generator.
[0050] Construct pricing strategy formula: Based on the above-mentioned market structure and power generation demand, as well as the determined pricing parameters, construct the power sales price pricing strategy formula of the leading power generator. The formula is in, is the target clearing price for the leading generator in the day-ahead market, is the target clearing price of the leading generator in the real-time market, p1 is the clearing price of the leading generator in the day-ahead market; p2 is the clearing price of the leading generator in the real-time market; c is the fixed marginal cost of the leading generator; q1 is the residual demand faced by the leading generator in the day-ahead market; q2 is the residual demand faced by the leading generator in the real-time market; p1 and p2 are variables.
[0051] In one possible implementation, q1=AQ-b1*p1-c1, A is the market structure of the electricity market and the power generation demand on the demand side of the electricity market, Q is the constant power generation of the new energy, b1 is the price elasticity of the market supply of the follower power producer on the day before, and c1 is the fixed marginal cost of the follower power producer.
[0052] q2=b1*p1-b2*p2, where b2 is the elasticity of the real-time market supply of the follower power producer to the price.
[0053] It should be noted that A, Q, b1, b2, c, c1 are all fixed parameters, and p1 and p2 are variable parameters, which are variables.
[0054] Generation demandA: Generation demand in the electricity market represents the total amount of electricity that needs to be met, that is, the total demand for electricity by consumers or the market.
[0055] Constant power generation Q from renewable energy: This represents a constant output from renewable energy (such as solar, wind, etc.) that is not affected by fluctuations in market demand.
[0056] Fixed marginal cost c of leading power producers: This is the fixed cost that leading power producers need to pay for producing unit electricity. It is independent of output and usually includes fixed expenses such as equipment maintenance and employee salaries.
[0057] The day-ahead market supply elasticity of the follower generator to the price b1: This indicates the market price’s sensitivity to changes in supply when the follower generator adjusts its supply based on the market conditions of the previous day. This elasticity value can measure the market’s responsiveness to changes in supply.
[0058] The elasticity of the real-time market supply of the follower generator to the price b2: This indicates the sensitivity of the market price to the change in its supply when the follower generator adjusts its supply according to the real-time market situation. Similar to b1, this elasticity value is used to evaluate the market's responsiveness to the change in real-time supply.
[0059] Fixed marginal cost c1 of the follower generator: This represents the marginal cost that the follower generator needs to pay in addition to the fixed cost when producing a unit of electricity. These marginal costs may include fuel costs, operating costs, etc., which are related to each unit of output.
[0060] The remaining demand q1 faced by the leading generator in the day-ahead market: This represents the difference between the market demand faced by the leading generator and its supply the day before or after the market supply is determined in advance. q1 can be regarded as a parameter calculated based on the day-ahead market conditions and used to evaluate the market pricing and production decisions of the leading generator in the day-ahead market.
[0061] The remaining demand q2 faced by the leading generator in the real-time market: This represents the difference between the market demand and the supply of the leading generator on the day when the actual transaction occurs and when the market price changes dynamically. q2 can be regarded as a parameter calculated based on the real-time market conditions to evaluate the market pricing and production decisions of the leading generator in the real-time market.
[0062] The day-ahead market clearing price p1 of the leading generator: This is the final transaction price after all supply and demand are balanced in the day-ahead market. p1 represents the clearing price of the leading generator in the day-ahead market, that is, the final price at which they sell electricity in this market.
[0063] The real-time market clearing price p2 of the leading generator: This is the final transaction price after all supply and demand are balanced in the real-time market. p2 represents the clearing price of the leading generator in the real-time market, that is, the final price at which they sell electricity in this market.
[0064] It should also be noted that the multiple pricing model of leading power generators means that in the electricity market, some power generators may influence market prices and their own profits by setting multiple different prices. In this model, leading power generators flexibly adjust their selling prices based on market demand, competitor pricing and other factors to maximize profits or other strategic goals. Through multiple pricing, leading power generators can better adapt to market changes and competitive situations, adjust their own supply, and influence market balance and price trends. This strategic pricing behavior helps leading power generators gain competitive advantages in the market and maximize their profits.
[0065] In this way, leading power generators can test different pricing strategies in a virtual environment to more accurately predict and optimize their pricing behavior to achieve the goal of maximizing profits. This simulation environment provides a risk-controlled platform that enables power generators to make more intelligent and effective decisions in the actual market.
[0066] S102: Determine the target pricing action parameters of the leading power producer.
[0067] Determining the target pricing action parameters for leading power producers means taking key factors into account when formulating pricing strategies and using these factors as variables to adjust and optimize pricing behavior. Such dynamic adjustments can help power producers better adapt to changes in market demand and price fluctuations, thereby achieving the goal of maximizing profits.
[0068] In one possible implementation, to determine the target pricing action parameters of the leading generator, it is necessary to consider the clearing price p1 of the leading generator in the day-ahead market and the clearing price p2 of the leading generator in the day-ahead real-time market as important variables. The clearing price p1 of the day-ahead market and the clearing price p2 of the real-time market are crucial to determining the pricing strategy of the generator, because they directly affect the balance between market demand and supply and the formation of prices.
[0069] By using these two prices as variables, the maximum profit can be obtained (i.e., the target profit under the target value of the target pricing action parameter). This means that leading power producers can flexibly adjust their pricing strategies according to different market conditions to maximize profits. The process of determining the target pricing action parameters should include the analysis and monitoring of day-ahead and real-time market prices so that corresponding adjustments and decisions can be made according to market conditions, thereby achieving an effective pricing strategy.
[0070] S103: Utilizing the policy-based reinforcement learning algorithm PRLA algorithm, based on the target pricing action parameters, the multiple pricing model of the leading power producer in the simulation environment is subjected to bidding simulation evolution to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
[0071] Using the PRLA algorithm, the multiple pricing models of leading power producers with a simulation environment are simulated and evolved based on the target pricing action parameters. The purpose is to obtain the specific value of the target pricing action parameters that maximizes the target profit through iterative training of the algorithm (i.e., calculate the target clearing price of the leading power producer in the day-ahead market). and the target clearing price of the leading generator in the real-time market ) and the maximum profit result (i.e. profit max ). This means using reinforcement learning techniques to optimize power generators’ pricing strategies to achieve the goal of maximizing profits.
[0072] It should be noted that the target clearing price and target clearing price It refers to the optimal clearing price that maximizes the profit of the leading power generator. The "profit" here specifically refers to the value obtained by deducting the relevant costs from the total income obtained by the leading power generator from the sale of electricity.
[0073] In one possible implementation, the target profit
[0074] In one possible implementation, to use the PRLA algorithm, based on the target pricing action parameters, to perform bidding simulation evolution on the multiple pricing model of the leading power generator in the simulation environment to obtain the target profit, the following steps are required:
[0075] Initialize the learning rate and discount factor of the PRLA algorithm: Before starting training, set the appropriate learning rate and discount factor. These parameters will determine how the algorithm adjusts and updates the target pricing action.
[0076] Iterative training is performed using the initialized PRLA algorithm: the pricing model of the leading power generator in the simulation environment is trained multiple times, and the training is repeated according to the set training cutoff conditions to obtain the maximum profit (i.e. the target profit profit max ) and the day-ahead market-clearing price and real-time market-clearing price of the leading generator corresponding to maximum profit.
[0077] It should be noted that in the field of reinforcement learning, the PRLA algorithm is a reinforcement learning algorithm based on the policy gradient method, which is used to train the agent to learn the best strategy by interacting with the environment. The key idea of this type of algorithm is to adjust the agent's behavior strategy by maximizing the long-term cumulative return. Specifically, the PRLA algorithm updates the agent's strategy by optimizing the policy residual so that the predicted behavior is closer to the optimal behavior. During the training process, the agent will continuously adjust its strategy based on the reward signal fed back by the environment, thereby gradually improving its performance on specific tasks.
[0078] Through the above steps, combined with the PRLA algorithm and target pricing action parameters, the pricing model can be simulated and evolved through multiple bidding processes to ultimately determine the optimal pricing strategy for leading power producers under specific market conditions in order to maximize profits.
[0079] Based on the contents of S101-S103, it can be known that a simulation environment is first established to set the multiple pricing model of the leading power producer, and on this basis, the variable target pricing action parameters of the power producer are determined. Subsequently, a policy-based reinforcement learning algorithm (PRLA algorithm) is used to perform bidding simulation evolution on the multiple pricing model of the leading power producer for which the simulation environment has been set according to these target pricing action parameters, so as to obtain the final profit result and pricing result of the leading power producer. This application helps to better adapt to complex and changeable market conditions by setting up a simulation environment and flexibly adjusting the target pricing action parameters, while breaking through the strict limitations of the traditional Markov decision process in action selection and state transition. At the same time, the use of the PRLA algorithm can effectively optimize the decision-making strategy, thereby improving the ultimate target profit of the leading power producer.
[0080] See also Figure 2 , Figure 2 The schematic diagram of the structure of a pricing device for a leading power producer provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the pricing device of the leading power producer includes:
[0081] A simulation environment unit 201, used to set a simulation environment of a multiple pricing model of a leading power producer;
[0082] The first determining unit 202 is used to determine the target pricing action parameter of the leading power producer; the target pricing action parameter is a variable;
[0083] The bidding simulation evolution unit 203 is used to utilize the policy-based reinforcement learning algorithm PRLA algorithm to perform bidding simulation evolution on the multiple pricing model of the leading power producer in the simulation environment based on the target pricing action parameters to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
[0084] In a possible implementation, the simulation environment unit 201 specifically includes:
[0085] A power generation demand setting unit, used to set the market structure of the power market and the power generation demand on the demand side in the power market; the market structure includes a day-ahead market and a real-time market;
[0086] The second determination unit is used to determine pricing parameters; the pricing parameters include the constant power generation of new energy, the fixed marginal cost of the leading power generator, the elasticity of the market supply of the following power generator to the price, the elasticity of the real-time market supply of the following power generator to the price and the fixed marginal cost of the following power generator;
[0087] A construction unit is used to construct a pricing strategy formula for the electricity sales price of a leading power producer based on the pricing parameters; the pricing strategy formula for the electricity sales price of the leading power producer is: in, is the target clearing price for the leading generator in the day-ahead market, is the target clearing price of the leading generator in the real-time market, p1 is the clearing price of the leading generator in the day-ahead market; p2 is the clearing price of the leading generator in the real-time market; c is the fixed marginal cost of the leading generator; q1 is the residual demand faced by the leading generator in the day-ahead market; q2 is the residual demand faced by the leading generator in the real-time market; p1 and p2 are variables.
[0088] In one possible implementation,
[0089] q1=AQ-b1*p1-c1, A is the market structure of the power market and the power generation demand on the demand side of the power market, Q is the constant power generation of the new energy, b1 is the day-ahead market supply elasticity of the follower power producer to the price, and c1 is the fixed marginal cost of the follower power producer;
[0090] q2=b1*p1-b2*p2, where b2 is the elasticity of the real-time market supply of the follower power producer to the price.
[0091] In a possible implementation manner, the first determining unit 202 is specifically configured to:
[0092] The clearing price of the leading power producer in the day-ahead market and the clearing price of the leading power producer in the real-time market in the pricing parameters are determined as variables to obtain the target pricing action parameters.
[0093] In a possible implementation, the bidding simulation evolution unit 203 specifically includes:
[0094] An initialization unit, used to initialize the learning rate and discount factor of the PRLA algorithm;
[0095] An iterative training unit is used to utilize a policy-based reinforcement learning algorithm PRLA algorithm to perform bidding simulation evolution on the multiple pricing model of the leading power producer in the simulation environment based on the target pricing action parameters to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
[0096] In addition, an embodiment of the present application also provides a pricing device for a leading power producer, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the pricing method for the leading power producer as described above is implemented.
[0097] In addition, an embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes the pricing method for the leading power producer as described above.
[0098] The embodiment of the present application provides a pricing device for a leading power producer, which first uses a simulation environment unit 201 to set a simulation environment for a multiple pricing model of a leading power producer. The first determination unit 202 determines the target pricing action parameters of the leading power producer, where the target pricing action parameters are variables. Then, the bidding simulation evolution unit 203 uses the PRLA algorithm to perform bidding simulation evolution on the multiple pricing model of the leading power producer with the simulation environment set based on the target pricing action parameters to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter. By setting a simulation environment and determining the target pricing action parameters, the present application can more flexibly respond to complex and changeable market conditions, and overcome the strict limitations of the traditional Markov decision process in action selection and state transition. At the same time, by using the PRLA algorithm, the decision-making strategy can be more effectively optimized, thereby obtaining the optimal target profit for the leading power producer.
[0099] The above is a detailed introduction to the pricing method, device, equipment and storage medium of a leading power generator provided by the present application. The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the embodiments can be referenced to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.
[0100] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.
[0101] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
Claims
1. A pricing method for a leading power producer, characterized in that: The method comprises: Set up the simulation environment of the multiple pricing model of the leading power generator; Determine the target pricing action parameter of the leading power generator; the target pricing action parameter is a variable; Utilizing the policy-based reinforcement learning algorithm PRLA algorithm, based on the target pricing action parameters, the multiple pricing model of the leading power producer in the simulation environment is subjected to bidding simulation evolution to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
2. The method according to claim 1, characterized in that The simulation environment for setting up the multiple pricing model of the leading power generator includes: Setting a market structure of the power market and power generation demand on the demand side of the power market; the market structure includes a day-ahead market and a real-time market; Determine pricing parameters; the pricing parameters include the constant power generation of new energy, the fixed marginal cost of the leading power generator, the elasticity of the market supply of the following power generator to the price, the elasticity of the market supply of the following power generator to the price in real time, and the fixed marginal cost of the following power generator; Under the market structure of the power market, a pricing strategy formula for the power sales price of a leading power producer is constructed based on the power generation demand on the demand side of the power market and the pricing parameters; the pricing strategy formula for the power sales price of a leading power producer is: in, is the target clearing price for the leading generator in the day-ahead market, is the target clearing price of the leading generator in the real-time market, p1 is the clearing price of the leading generator in the day-ahead market; p2 is the clearing price of the leading generator in the real-time market; c is the fixed marginal cost of the leading generator; q1 is the residual demand faced by the leading generator in the day-ahead market; q2 is the residual demand faced by the leading generator in the real-time market; p1 and p2 are variables.
3. The method according to claim 2, characterized in that q1=AQ-b1*p1-c1, A is the market structure of the power market and the power generation demand on the demand side of the power market, Q is the constant power generation of the new energy, b1 is the day-ahead market supply elasticity of the follower power producer to the price, and c1 is the fixed marginal cost of the follower power producer; q2=b1*p1-b2*p2, where b2 is the elasticity of the real-time market supply of the follower power producer to the price.
4. The method according to claim 2, characterized in that: The target pricing action parameters for determining the leading power generator include: The clearing price of the leading power producer in the day-ahead market and the clearing price of the leading power producer in the real-time market in the pricing parameters are determined as variables to obtain the target pricing action parameters.
5. The method according to claim 2, characterized in that: The strategy-based reinforcement learning algorithm PRLA algorithm is used to perform bidding simulation evolution on the multiple pricing model of the leading power generator with the simulation environment set based on the target pricing action parameter to obtain the target value of the target pricing action parameter of the leading power generator and the target profit under the target value of the target pricing action parameter, including: Initialize the learning rate and discount factor of the PRLA algorithm; Utilizing the policy-based reinforcement learning algorithm PRLA algorithm, based on the target pricing action parameters, the multiple pricing model of the leading power producer in the simulation environment is subjected to bidding simulation evolution to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
6. A pricing device for a leading power producer, characterized in that: The device comprises: The simulation environment unit is used to set up the simulation environment of the multiple pricing model of the leading power producer; A first determining unit is used to determine a target pricing action parameter of a leading power generator; the target pricing action parameter is a variable; The bidding simulation evolution unit is used to utilize the policy-based reinforcement learning algorithm PRLA algorithm to perform bidding simulation evolution on the multiple pricing model of the leading power producer in the simulation environment based on the target pricing action parameters to obtain the target value of the target pricing action parameter of the leading power producer and the target profit under the target value of the target pricing action parameter.
7. The device according to claim 6, characterized in that The simulation environment unit specifically includes: A power generation demand setting unit, used to set the market structure of the power market and the power generation demand on the demand side in the power market; the market structure includes a day-ahead market and a real-time market; The second determination unit is used to determine pricing parameters; the pricing parameters include the constant power generation of new energy, the fixed marginal cost of the leading power generator, the elasticity of the market supply of the following power generator to the price, the elasticity of the real-time market supply of the following power generator to the price and the fixed marginal cost of the following power generator; A construction unit is used to construct a pricing strategy formula for the electricity sales price of a leading power producer based on the pricing parameters; the pricing strategy formula for the electricity sales price of the leading power producer is: in, is the target clearing price for the leading generator in the day-ahead market, is the target clearing price of the leading generator in the real-time market, p1 is the clearing price of the leading generator in the day-ahead market; p2 is the clearing price of the leading generator in the real-time market; c is the fixed marginal cost of the leading generator; q1 is the residual demand faced by the leading generator in the day-ahead market; q2 is the residual demand faced by the leading generator in the real-time market; p1 and p2 are variables.
8. The device according to claim 7, characterized in that q1=AQ-b1*p1-c1, A is the market structure of the power market and the power generation demand on the demand side of the power market, Q is the constant power generation of the new energy, b1 is the day-ahead market supply elasticity of the follower power producer to the price, and c1 is the fixed marginal cost of the follower power producer; q2=b1*p1-b2*p2, where b2 is the elasticity of the real-time market supply of the follower power producer to the price.
9. A pricing device for a leading power producer, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the pricing method for the leading power producer as described in any one of claims 1 to 5 is implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on the terminal device, the terminal device executes the pricing method for the leading power producer according to any one of claims 1 to 5.