Market price prediction method based on reverse game deduction and related equipment

Abstract

The invention discloses a market price prediction method based on reverse game deduction and related equipment. The method comprises the following steps: constructing an external factor data set and a historical power data set; based on market prices in the external factor data set and the historical power data set, key driving factors related to market price changes in the external factor data set are determined, and a causal relationship model is constructed; constructing a game model for the market participants according to the key driving factors, and deducing an optimal strategy combination of each market participant in different market scenes by taking a set target market price as a starting point; based on the causal relationship model and the optimal strategy combination, constructing a market price prediction model; generating scene prediction results under different market scenes according to the market price prediction model; and generating a regulation and control scheme matched with the target market price through reverse deduction based on the scene prediction result. For a complex market scene and a plurality of market participants, the adaptability of the regulation and control scheme of the scheme is obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power price regulation, and in particular to a market price prediction method based on reverse game deduction and related equipment. Background Art

[0002] In the field of market price forecasting and regulation, existing technical solutions mainly predict and regulate market prices based on methods such as time series analysis and simple causal analysis. The above-mentioned known technologies are helpful in achieving basic market price forecasting and rough regulation of market supply and demand in the early stage of power market development, mainly including the following contents:

[0003] Time series analysis: Most traditional market price forecasting systems are based on time series analysis and use historical price data for trend modeling. For example, the autoregressive integrated moving average model and the generalized autoregressive conditional heteroskedasticity model are used to infer future price trends by analyzing past price data and supply and demand changes. However, such methods usually assume that market changes have a relatively linear and stable trend, and it is difficult to provide accurate predictions in extreme environments or when market factors fluctuate violently.

[0004] Causal analysis and simple data regression: In some market price forecasts, causal analysis methods are combined with data regression to identify the key factors of market price fluctuations, but they are usually limited to analysis based on the direct impact of specific factors on price changes, ignoring the interaction of multiple market factors and the impact of external shocks, and unable to achieve causal relationship modeling in complex situations.

[0005] That is, with the rapid development and complexity of the electricity market, market price forecasting and regulation play an important role in ensuring stable electricity supply and reasonable prices. However, traditional market price forecasting and regulation technologies have many limitations when faced with complex market scenarios and diverse behaviors of market participants, and it is difficult to meet the growing market demand. Summary of the invention

[0006] The present invention provides a market price prediction method based on reverse game deduction to solve the problem of market price prediction based on reverse game deduction.

[0007] In a first aspect, the present invention provides a market price prediction method based on reverse game deduction, comprising:

[0008] S1. Construct an external factor data set based on the external factor data, and construct a historical power data set based on the historical transaction data of the power market, the user's power demand data and the power generation data of the power generator. The historical transaction data includes the market price of electricity.

[0009] S2. Based on the external factor data set and market prices, determine the key driving factors in the external factor data set that are related to market price changes and build a causal relationship model;

[0010] S3. Construct a game model for market participants based on key driving factors, take the set target market price as the starting point, and use the reverse deduction method to reversely deduce the optimal strategy combination of each market participant under different market scenarios; market participants include power generators, power sales companies and users;

[0011] S4, based on the causal relationship model and the optimal strategy combination, the market price prediction model is constructed using support vector regression, and the model parameters of the market price prediction model are optimized through the historical power data set;

[0012] S5. Generate scenario prediction results under different market scenarios according to the market price prediction model;

[0013] S6. Based on the scenario forecast results, a combination of control strategies that matches the target market price is generated through reverse reasoning to obtain a control plan, and the control plan is optimized in combination with the causal relationship model.

[0014] In a second aspect, the present invention provides a market price prediction device based on reverse game deduction, comprising:

[0015] A data set construction module is used to construct an external factor data set based on the external factor data, and to construct a historical power data set based on the historical transaction data of the power market, the user's power demand data, and the power generation data of the power generator, wherein the historical transaction data includes the market price of power;

[0016] A causal relationship model building module, which is used to determine the key driving factors in the external factor data set that are related to the market price changes and build a causal relationship model based on the external factor data set and the market price;

[0017] The module for determining the optimal strategy combination is used to build a game model for market participants based on key driving factors. It takes the set target market price as the starting point and uses the reverse deduction method to reversely deduce the optimal strategy combination of each market participant under different market scenarios. Market participants include power generators, power sales companies and users.

[0018] A market price forecasting model building module is used to build a market price forecasting model based on a causal relationship model and an optimal strategy combination using support vector regression, and to optimize the model parameters of the market price forecasting model using a historical power data set;

[0019] A scenario forecasting module is used to generate scenario forecasting results under different market scenarios based on the market price forecasting model;

[0020] The control scheme determination module is used to generate a control strategy combination that matches the target market price through reverse deduction based on scenario forecast results, obtain the control scheme, and optimize the control scheme in combination with the causal relationship model.

[0021] In a third aspect, the present invention provides an electronic device, the electronic device comprising:

[0022] at least one processor; and

[0023] a memory communicatively connected to the at least one processor; wherein,

[0024] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the market price prediction method based on reverse game deduction described in the first aspect of the present invention.

[0025] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a processor to implement the market price prediction method based on reverse game theory described in the first aspect of the present invention when executed.

[0026] The embodiment of the present invention provides a market price prediction method based on reverse game deduction, comprising: constructing an external factor data set according to external factor data, and constructing a historical power data set according to historical transaction data of the power market, user power demand data and power generation data of the power generator, wherein the historical transaction data includes the market price of power; determining key driving factors related to market price changes in the external factor data set based on the external factor data set and the market price and constructing a causal relationship model; constructing a game model for market participants according to the key driving factors, taking the set target market price as the starting point, and using a reverse deduction method to reversely deduce the optimal strategy combination of each of the market participants under different market scenarios; the market participants include power generators, power sales companies and users; constructing a market price prediction model based on the causal relationship model and the optimal strategy combination and using support vector regression, and optimizing the model parameters of the market price prediction model through the historical power data set; generating scenario prediction results under different market scenarios according to the market price prediction model; generating a control strategy combination matching the target market price through reverse deduction based on the scenario prediction results, obtaining a control scheme, and optimizing the control scheme in combination with the causal relationship model. When faced with complex market scenarios and diverse behaviors of market participants, such as external factors such as extreme weather events, large-scale access to new energy, and frequent policy adjustments, the adaptability of the regulation scheme of this plan is significantly improved, ensuring stable power supply and reasonable prices.

[0027] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0029] Figure 1 This is a flow chart of a market price prediction method based on reverse game deduction provided by Embodiment 1 of the present invention;

[0030] Figure 2 It is a structural schematic diagram of a market price prediction device based on reverse game deduction provided by the second embodiment of the present invention;

[0031] Figure 3 It is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0033] Embodiment 1

[0034] Figure 1 This is a flowchart of a market price prediction method based on reverse game deduction provided in the first embodiment of the present invention. This embodiment is applicable to the situation of market price prediction based on reverse game deduction. The method can be executed by a market price prediction device based on reverse game deduction. The market price prediction device based on reverse game deduction can be implemented in the form of hardware and / or software. The market price prediction device based on reverse game deduction can be configured in an electronic device. Figure 1 As shown, the market price prediction method based on reverse game deduction includes:

[0035] S1. Construct an external factor data set based on the external factor data, and construct a historical power data set based on the historical transaction data of the power market, user power demand data and power generation data of power generators. The historical transaction data includes the market price of electricity.

[0036] External factor data include weather data, carbon emission data and policy change data.

[0037] S1 specifically includes the following steps:

[0038] S11, obtaining historical transaction data of the power market, the historical transaction data including price data, transaction volume data and time stamp;

[0039] S12, obtaining user electricity demand data, where the user electricity demand data includes historical electricity demand of user groups;

[0040] S13, obtaining power generation data of the power generator, where the power generation data includes the total power generation of the power generator and the output power of the power generation equipment;

[0041] S14, obtaining weather data, the weather data including temperature, humidity and wind speed;

[0042] S15. Obtain carbon emission data, including the total carbon emission and carbon emission intensity of different power generation sources;

[0043] S16. Obtain policy change data, including electricity price adjustment policy, subsidy policy and market access policy;

[0044] S17. Combine S11-S16 to construct the historical power data set D of the power market 历史 and external factor dataset D 外部 :

[0045] D 历史 ={(P t ,Q t ,D t ,G t ,E t,i ,t)∣t=1,2,…,T; i=1,2,…,N};

[0046] Where T represents the total number of time points, N represents the number of power generation equipment, and P t is the price data, Q t is the transaction volume data, Q t is the electricity demand data, G t is the total power generation, E t,i is the output power of the i-th power generation equipment;

[0047] D 外部 ={(W t ,Ht ,V t ,C t ,θ t ,R t ,S t ,M t ,t)|t=1,2,…,T};

[0048] Among them, W t is the air temperature, H t is humidity, V t is the wind speed, C t is the total carbon emissions, θ t is the carbon emission intensity, R t For electricity price adjustment policy, S t For subsidy policy, M t For market access policy.

[0049] S2. Based on the external factor data set and market prices, determine the key driving factors in the external factor data set that are related to market price changes and build a causal model.

[0050] Specifically, S2 includes:

[0051] S21, constructing an input matrix X and a target vector Y with causal relationships based on the external factor data set and the corresponding market prices;

[0052] n is the number of candidate drivers, T is the total number of time points, X i,t is the value of the i-th candidate driving factor at time point t; R (.) represents a real number, and n×T represents the rows and columns of a matrix; Represents a transpose operation.

[0053] P t is the market price at the tth time point; T represents the row of the matrix;

[0054] S22, using a structural learning algorithm to determine a causal relationship graph structure of the input matrix X and the target vector Y, where each input matrix X and target vector Y is a node in the causal relationship graph structure;

[0055] Using Bayesian network to build initial causal relationship model Use structural learning algorithms (greedy algorithms, PT algorithms) to determine the causal relationship graph structure G between variables:

[0056] G = (V, E);

[0057] S23, selecting the optimal causal relationship graph structure by maximizing the Bayesian information criterion scoring function;

[0058] By maximizing the Bayesian Information Criterion scoring function:

[0059] BIC(G)=-2lnL(G)+klnT;

[0060] Where L(G) = P(Y, X|G) is the log-likelihood function of the data under a given causal graph structure G, k is the number of initial causal model parameters, reflecting the complexity of the initial causal model, and T is the sample size, i.e., the total number of time points;

[0061] The optimal causal graph structure G is selected by maximizing the Bayesian information criterion to determine the direct causal relationship between each candidate driving factor and the electricity market price.

[0062] S24. In the optimal causal relationship graph structure, for the market price P t Directly connected candidate driver X i,t , calculate the Bayesian causal strength coefficient δ i :

[0063]

[0064] Among them, Pa(P t ) is P t The parent node set of Pa(P t )\X i,t Indicates removing X from the parent node set i,t , P(P t |X i ,Pa(P t )\X i,t )、P(P t |Pa(P t )\X i,t ) respectively represent the i,t Under the condition of t The conditional probability distribution of excluding X i,t Under the condition of t The conditional probability distribution of ;

[0065] δ i The direct causal impact intensity of the candidate driving factor Xi on the electricity market price Pt is quantified, and the larger the value, the more significant the impact.

[0066] S25, the Bayesian causal strength coefficient δ i Candidate driving factors X that are greater than the preset intensity coefficient and pass the Granger test i,t , and constitute the key driving factor set K;

[0067] S26. For each key driver X in the key driver set K,i,t , using Gaussian process regression method to establish the market price P t and key drivers X j,t The nonlinear relationship model between them gives the causal relationship model:

[0068] P t =∑ i∈K f i (X i,t )+ε t ;

[0069] Among them, f i (X i,t ) is the key driving factor X i,t For market price P t The nonlinear influence function is obtained by Gaussian process regression, ε t is an independent and identically distributed Gaussian noise term, σ 2 is the variance of the noise term, which is estimated through nonlinear relationship model training.

[0070] f i (X i,t ) is used to capture complex nonlinear relationships, ε t Represents random errors that are not explained by the nonlinear relationship model.

[0071] By training the causal relationship model, the kernel function parameters and model hyperparameters of the Gaussian process are learned, so that the causal relationship model can reflect the nonlinear impact of key driving factors on electricity market prices.

[0072] S3. Construct a game model for market participants based on key driving factors. Taking the set target market price as the starting point, use the reverse deduction method to reversely deduce the optimal strategy combination of each market participant under different market scenarios; market participants include power generators, power sales companies and users.

[0073] Specifically, S3 includes:

[0074] S31. Construct a game model for market participants based on the key driving factor set K, and set the market participant set A:

[0075] A={a1,a2,…a i …,a m}, where m represents the number of market participants, a i For market participants;

[0076] S32. Construct the profit function for each market participant

[0077]

[0078] in, Indicates that the market price P t and key drivers X i,t Under the joint action, market participants i The profit function at time t is, Indicates participant a i Cost function, cost function and transaction volume Market Price t and key drivers X i,t Related;

[0079] S33. Define the target market price P d , by reverse deduction from the target market price P d Starting from, reversely deduce the optimal strategy combination of market participants So that:

[0080]

[0081] Among them, S i Indicates participant a i The strategy combination, λ i is the penalty coefficient, the control strategy deviates from the target market price P d The degree of i (X i,t ) is the key driving factor X i,t For market price P t The nonlinear influence function of , K is the set of key driving factors;

[0082] S34. Simulate the strategic interactions of market participants in different market scenarios, and set the set of market scenarios M to be:

[0083] M={M1,M2,…M k …,M q};

[0084] Among them, M k is the market scenario, q represents the number of market scenarios;

[0085] For each market scenario M k , defining the optimal strategy combination of market participants as And the following Nash equilibrium conditions are met:

[0086]

[0087] in, Indicates that except for participant a i Other market participants in the market scenario M kThe optimal strategy combination under the condition that all market participants’ strategies are optimal, and no single market participant can increase its profit by changing its strategy.

[0088] You can also choose the M according to each market scenario k The optimal strategy combination under different scenarios is used to generate the expected trend of electricity market prices and strategy interaction results in various scenarios.

[0089] The optimal strategy combination is obtained by deducing the behavior of each market participant in order to maximize profits under certain market conditions. The strategy interaction result is equivalent to the dynamic simulation of the market results produced after all market participants execute the strategy.

[0090] S4. Based on the causal relationship model and the optimal strategy combination, a market price prediction model is constructed using support vector regression, and the model parameters of the market price prediction model are optimized through the historical electricity data set.

[0091] Specifically, S4 includes:

[0092] S41. Construct a market price prediction model based on the causal relationship model and the optimal strategy combination for market price P t For prediction, let the input features of the market price prediction model be the key driving factor set K, then the input feature matrix X of the market price prediction model SVR It is expressed as:

[0093] X SVR ={f i (X i,t )|i∈K,t=1,2,…,T};

[0094] Among them, f i (X i,t ) is the key driving factor X i,t For market price P t The nonlinear influence function of , T represents the total number of time points, and K is the set of key driving factors;

[0095] S42. Set the loss function L(w,b) and regularization parameter C of the market price prediction model. The loss function is defined as follows:

[0096]

[0097] Among them, P t is the real market price at time t, w is the weight vector of the market price prediction model, b is the bias term of the market price prediction model, ∈ is the tolerance error parameter, and regularization Used to avoid model overfitting;

[0098] S43, iterating the training of the market price prediction model by minimizing the loss function L(w,b), updating the model parameters w and b to optimize the market price prediction accuracy; the optimization goal of the market price prediction model training is:

[0099]

[0100] Among them, w * and b * Represents the optimal weights and biases obtained through iterative optimization;

[0101] S44. Based on the optimized market price forecasting model, combined with the input feature matrix X of the historical power data set SVR , generate market price prediction sequence Get the market price prediction model:

[0102]

[0103] in, is the predicted market price at time t.

[0104] S5. Generate scenario forecast results under different market scenarios based on the market price forecast model.

[0105] Market scenarios include specific scenarios such as extreme weather events, large-scale new energy access, and policy changes.

[0106] Specifically, S5 includes:

[0107] S51, construct a market scenario set M = {M1, M2, ... M k …,M q}; where M k is the market scenario, q represents the number of market scenarios;

[0108] Define each market scenario M k The eigenvector of

[0109] For example, the feature vector of each market scenario M is defined as It is expressed as:

[0110]

[0111] in, For market scenario M k The temperature influence factors are: For market scenario M k The wind speed factors under For market scenario M k The policy subsidy factor under For market scenario Mk Large-scale new energy access under For market scenario M k The electricity price adjustment factor.

[0112] S52. In each market scenario M k The eigenvector As input, combined with the market price prediction model, the market price series under different market scenarios are predicted

[0113]

[0114] in, In the market scenario M k and the market price forecast value at time t, w * and b * are the weight and bias of the market price prediction model respectively;

[0115] S53, according to each market scenario M k The eigenvector of Simulate the strategic interaction behavior of market participants in different market scenarios and define participants a i In scenario M k The following policy set And optimize based on the game model to determine the optimal strategy combination

[0116] in, Represents market scenario M k Participant A i The optimal strategy set is Represents market scenario M k Except for participant a i The strategy portfolio of other market participants other than In the scenario M k Next, market participants a i The profit function of

[0117] S54. In each market scenario M k Next, combine the best strategy combination of each market participant Generate scenario forecast results for analyzing price fluctuation characteristics under different market scenarios:

[0118] T represents the total number of time points.

[0119] The scenario forecast results in this step can take into account the price fluctuation characteristics under different market scenarios, especially in market scenarios where extreme weather, large-scale new energy access and policy changes have a greater impact on market prices.

[0120] S6. Based on the scenario forecast results, a combination of control strategies that matches the target market price is generated through reverse reasoning to obtain a control plan, and the control plan is optimized in combination with the causal relationship model.

[0121] Specifically, S6 includes:

[0122] S61. Based on scenario prediction results and target market price P d , set the power market regulation strategy combination set Where p is the number of control strategies.

[0123] Each control strategy S j Includes the following elements:

[0124]

[0125] in, For the control strategy S j The power generation plan adjustment at time t is: is the reserve power released at time t, is the upper limit adjustment value of the electricity price at time t, is the temporary subsidy policy value at time t;

[0126] S62, according to the market target market price P d and scenario forecast results Use reverse deduction to generate a combination of regulatory strategies that match the target market price

[0127]

[0128] in, Represents market scenario M k The predicted market price value at time t, α, β, γ, δ are the weight coefficients of each control measure, which are used to balance the control costs among power generation adjustment, reserve power release, electricity price cap adjustment and temporary subsidies;

[0129] S63. Combination of control strategies further optimized based on causal relationship model By adjusting the influence weights of each key driving factor, the optimized control strategy combination is expressed as

[0130]

[0131] Among them, δ i is the key driving factor X in the causal relationship model i,t The causal strength coefficient, f i (X i,t ) is the key driving factor X at time t i,t The impact function on market price, Represents the key driving factor value under the target state, making the price prediction model tend to the target market price P d ;

[0132] Exemplary, the control strategy is:

[0133] For power generators, it can increase high-efficiency motors and reduce inefficient energy;

[0134] For power sales companies, they can adopt a tiered pricing system, increasing the electricity price during peak hours to increase the total electricity sales price, and reducing the electricity price during low-peak hours to promote electricity consumption.

[0135] For users, they can respond to the electricity sales mechanism, reduce electricity consumption during peak hours, and increase electricity consumption during off-peak hours.

[0136] S64. Combination of control strategies generated by reverse deduction In scenario M k Generate a control plan.

[0137] In scenario M k Generate regulation plans and predict the adjustment effects of various strategy combinations on market prices, providing optimized strategy references for actual market regulation.

[0138] The regulatory plans include adjusting power generation plans, releasing reserve electricity, modifying the electricity price cap or formulating temporary subsidy policies.

[0139] The control plan obtained in this step can generate corresponding control plans in market scenarios such as extreme weather, large-scale new energy access and policy changes that have a greater impact on market prices, so as to avoid market price chaos.

[0140] The embodiment of the present invention provides a market price prediction method based on reverse game deduction, comprising: constructing an external factor data set according to external factor data, and constructing a historical power data set according to historical transaction data of the power market, user power demand data and power generation data of the power generator, wherein the historical transaction data includes the market price of power; determining key driving factors related to market price changes in the external factor data set based on the external factor data set and the market price and constructing a causal relationship model; constructing a game model for market participants according to the key driving factors, taking the set target market price as the starting point, and using a reverse deduction method to reversely deduce the optimal strategy combination of each of the market participants under different market scenarios; the market participants include power generators, power sales companies and users; constructing a market price prediction model based on the causal relationship model and the optimal strategy combination and using support vector regression, and optimizing the model parameters of the market price prediction model through the historical power data set; generating scenario prediction results under different market scenarios according to the market price prediction model; generating a control strategy combination matching the target market price through reverse deduction based on the scenario prediction results, obtaining a control scheme, and optimizing the control scheme in combination with the causal relationship model. When faced with complex market scenarios and diverse behaviors of market participants, such as external factors such as extreme weather events, large-scale access to new energy, and frequent policy adjustments, the adaptability of the regulation scheme of this plan is significantly improved, ensuring stable power supply and reasonable prices.

[0141] In an optional embodiment, the market price prediction method based on reverse game deduction further includes:

[0142] S7. When the market price meets the abnormal fluctuation conditions, the market early warning mechanism is triggered, and an emergency regulation strategy is generated based on the causal analysis model and the game deduction model;

[0143] Emergency control strategies include releasing emergency reserve power, adjusting market supply and demand balance plans, or launching temporary price control measures;

[0144] S8. After the implementation of the regulation plan, the market price forecasting model, causal analysis model and reverse game deduction model are adaptively optimized based on the implementation results of the regulation plan and real-time market data.

[0145] Specifically, the execution results of market regulation are compared and analyzed with real-time market data, and based on the analysis results, the market price prediction model, causal analysis model and reverse game deduction model are adaptively optimized. Through adaptive optimization, the prediction accuracy of the model and the execution effect of the regulation plan can be optimized.

[0146] Embodiment 2

[0147] Figure 2This is a schematic diagram of the structure of a market price prediction device based on reverse game deduction provided by the second embodiment of the present invention. Figure 2 As shown, the market price prediction device based on reverse game deduction includes:

[0148] The data set construction module 201 is used to construct an external factor data set according to the external factor data, and to construct a historical power data set according to the historical transaction data of the power market, the user's power demand data and the power generation data of the power generator, wherein the historical transaction data includes the market price of power;

[0149] A causal relationship model building module 202, for determining key driving factors related to market price changes in the external factor data set based on the external factor data set and the market price, and building a causal relationship model;

[0150] The optimal strategy combination determination module 203 is used to construct a game model for market participants based on key driving factors, and to reversely deduce the optimal strategy combination of each market participant under different market scenarios using a reverse deduction method, taking the set target market price as the starting point; market participants include power generators, power sales companies and users;

[0151] A market price prediction model building module 204 is used to build a market price prediction model based on a causal relationship model and an optimal strategy combination using support vector regression, and optimize the model parameters of the market price prediction model through a historical power data set;

[0152] A scenario prediction module 205 is used to generate scenario prediction results under different market scenarios according to the market price prediction model;

[0153] The control scheme determination module 206 is used to generate a control strategy combination that matches the target market price through reverse deduction based on the scenario prediction results, obtain the control scheme, and optimize the control scheme in combination with the causal relationship model.

[0154] The market price prediction device based on reverse game deduction provided by the embodiment of the present invention can execute the market price prediction method based on reverse game deduction provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0155] Embodiment 3

[0156] Figure 3A schematic diagram of an electronic device 40 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0157] like Figure 3 As shown, the electronic device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In the RAM 43, various programs and data required for the operation of the electronic device 40 can also be stored. The processor 41, the ROM 42, and the RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0158] A number of components in the electronic device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0159] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 41 executes the various methods and processes described above, such as a market price prediction method based on reverse game deduction.

[0160] In some embodiments, the market price prediction method based on reverse game deduction can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the market price prediction method based on reverse game deduction described above can be performed. Alternatively, in other embodiments, the processor 41 can be configured to execute the market price prediction method based on reverse game deduction in any other appropriate manner (for example, by means of firmware).

[0161] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0162] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0163] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0164] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0165] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0166] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0167] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0168] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A market price prediction method based on reverse game deduction, characterized in that: include: S1. Construct an external factor data set based on the external factor data, and construct a historical power data set based on the historical transaction data of the power market, the user's power demand data and the power generation data of the power generator. The historical transaction data includes the market price of electricity. S2. Based on the external factor data set and market prices, determine the key driving factors in the external factor data set that are related to market price changes and build a causal relationship model; S3. Construct a game model for market participants based on key driving factors, take the set target market price as the starting point, and use the reverse deduction method to reversely deduce the optimal strategy combination of each market participant under different market scenarios; market participants include power generators, power sales companies and users; S4, based on the causal relationship model and the optimal strategy combination, the market price prediction model is constructed using support vector regression, and the model parameters of the market price prediction model are optimized through the historical power data set; S5. Generate scenario prediction results under different market scenarios according to the market price prediction model; S6. Based on the scenario forecast results, a combination of control strategies that matches the target market price is generated through reverse reasoning to obtain a control plan, and the control plan is optimized in combination with the causal relationship model.

2. The method according to claim 1, characterized in that S2, including: S21, constructing an input matrix X and a target vector Y with causal relationships based on the external factor data set and the corresponding market prices; n is the number of candidate drivers, T is the total number of time points, X i,t is the value of the i-th candidate driving factor at time point t; P t is the market price at the tth time point; S22, using a structural learning algorithm to determine a causal relationship graph structure of the input matrix X and the target vector Y, where each input matrix X and target vector Y is a node in the causal relationship graph structure; S23, selecting the optimal causal relationship graph structure by maximizing the Bayesian information criterion scoring function; S24. In the optimal causal relationship graph structure, for the market price P t Directly connected candidate driver X i,t , calculate the Bayesian causal strength coefficient δ i : Among them, Pa(P t ) is P t The parent node set (that directly affects P t variable set), Pa(P t )\X i,t Indicates removing X from the parent node set i,t , P(P t |X i ,Pa(P t )\X i,t )、P(P t |Pa(P t )\X i,t ) respectively represent the i,t Under the condition of t The conditional probability distribution of excluding X i,t Under the condition of t The conditional probability distribution of ; S25, the Bayesian causal strength coefficient δ i Candidate driving factors X that are greater than the preset intensity coefficient and pass the Granger test i,t , and constitute the key driving factor set K; S26. For each key driver X in the key driver set K, i,t , using Gaussian process regression method to establish the market price P t and key drivers X j,t The nonlinear relationship model between them gives the causal relationship model: P t =∑ i∈K f i (X i,t )+ε t ; Among them, f i (X i,t ) is the key driving factor X i,t For market price P t The nonlinear influence function is obtained by Gaussian process regression, ε t is an independent and identically distributed Gaussian noise term, σ 2 is the variance of the noise term, which is estimated through nonlinear relationship model training.

3. The method according to claim 1, characterized in that S3 include: S31. Construct a game model for market participants based on the key driving factor set K, and set the market participant set A: A={a1,a2,…a i …,a m }, where m represents the number of market participants, a i For market participants; S32. Construct the profit function for each market participant in, Indicates that the market price P t and key drivers X i,t Under the joint action, market participants i The profit function at time t is, Indicates participant a i Cost function, cost function and transaction volume Market Price t and key drivers X i,t Related; S33. Define the target market price P d , by reverse deduction from the target market price P d Starting from, reversely deduce the optimal strategy combination of market participants So that: Among them, S i Indicates participant a i The strategy combination, λ i is the penalty coefficient, the control strategy deviates from the target market price P d The degree of i (X i,t ) is the key driving factor X i,t For market price P t The nonlinear influence function of , K is the set of key driving factors; S34. Simulate the strategic interactions of market participants in different market scenarios, and set the set of market scenarios M to be: M={M1,M2,…M k …,M q }; Among them, M k is the market scenario, q represents the number of market scenarios; For each market scenario M k , defining the optimal strategy combination of market participants as And the following Nash equilibrium conditions are met: in, Indicates that except for participant a i Other market participants in the market scenario M k The optimal strategy combination.

4. The method according to claim 1, characterized in that S4 include: S41. Construct a market price prediction model based on the causal relationship model and the optimal strategy combination for market price P t For prediction, let the input features of the market price prediction model be the key driving factor set K, then the input feature matrix X of the market price prediction model SVR It is expressed as: X SVR ={f i (X i,t )∣i∈K,t=1,2,…,T}; Among them, f i (X i,t ) is the key driving factor X i,t For market price P t The nonlinear influence function of , T represents the total number of time points, and K is the set of key driving factors; S42. Set the loss function L(w,b) and regularization parameter C of the market price prediction model. The loss function is defined as follows: Among them, P t is the real market price at time t, w is the weight vector of the market price prediction model, b is the bias term of the market price prediction model, ∈ is the tolerance error parameter, and regularization Used to avoid model overfitting; S43, iterating the training of the market price prediction model by minimizing the loss function L(w,b), updating the model parameters w and b to optimize the market price prediction accuracy; the optimization goal of the market price prediction model training is: Among them, w * and b * Represents the optimal weights and biases obtained through iterative optimization; S44. Based on the optimized market price forecasting model, combined with the input feature matrix X of the historical power data set SVR , generate market price prediction sequence Get the market price prediction model: in, is the predicted market price at time t.

5. The method according to claim 1, characterized in that S5 include: S51, construct a market scenario set M = {M1, M2, ... M k …,M q }; where M k is the market scenario, q represents the number of market scenarios; Define each market scenario M k The eigenvector of S52. In each market scenario M k The eigenvector As input, combined with the market price prediction model, the market price series under different market scenarios are predicted in, In the market scenario M k and the market price forecast value at time t, w * and b * are the weight and bias of the market price prediction model respectively; S53, according to each market scenario M k The eigenvector of Simulate the strategic interaction behavior of market participants in different market scenarios and define participants a i In scenario M k The following policy set And optimize based on the game model to determine the optimal strategy combination in, Represents market scenario M k Participant A i The optimal strategy set is Represents market scenario M k Except for participant a i The strategy portfolio of other market participants other than In the scenario M k Next, market participants a i The profit function of S54. In each market scenario M k Next, combine the best strategy combination of each market participant Generate scenario forecast results for analyzing price fluctuation characteristics under different market scenarios: T represents the total number of time points.

6. The method according to claim 1, characterized in that S6 include: S61. Set the power market regulation strategy combination set Where p is the number of control strategies, and each control strategy S j Includes the following elements: in, For the control strategy S j The power generation plan adjustment at time t is: is the reserve power released at time t, is the upper limit adjustment value of the electricity price at time t, is the temporary subsidy policy value at time t; S62, according to the market target market price P d and scenario forecast results Use reverse deduction to generate a combination of regulatory strategies that match the target market price in, Represents market scenario M k The predicted market price value at time t, α, β, γ, δ are the weight coefficients of each control measure; S63. Combination of control strategies further optimized based on causal relationship model By adjusting the influence weights of each key driving factor, the optimized control strategy combination is expressed as Among them, δ i is the key driving factor X in the causal relationship model i,t The causal strength coefficient, f i (X i,t ) is the key driving factor X at time t i,t The impact function on market price, Represents the key driving factor value under the target state, making the price prediction model tend to the target market price P d ; S64. Combination of control strategies generated by reverse deduction In scenario M k Generate a control plan.

7. The method according to any one of claims 1 to 6, characterized in that: Also includes: S7. When the market price meets the abnormal fluctuation conditions, the market early warning mechanism is triggered, and an emergency regulation strategy is generated based on the causal analysis model and the game deduction model; S8. After the emergency control strategy is executed, the market price forecasting model, causal analysis model and reverse game deduction model are adaptively optimized based on the execution results of the control plan and real-time market data.

8. A market price prediction device based on reverse game deduction, characterized in that: The method for predicting market prices based on reverse game deduction according to any one of claims 1 to 7 comprises: A data set construction module is used to construct an external factor data set based on the external factor data, and to construct a historical power data set based on the historical transaction data of the power market, the user's power demand data, and the power generation data of the power generator, wherein the historical transaction data includes the market price of power; A causal relationship model building module, which is used to determine the key driving factors in the external factor data set that are related to the market price changes and build a causal relationship model based on the external factor data set and the market price; The module for determining the optimal strategy combination is used to build a game model for market participants based on key driving factors. It takes the set target market price as the starting point and uses the reverse deduction method to reversely deduce the optimal strategy combination of each market participant under different market scenarios. Market participants include power generators, power sales companies and users. A market price forecasting model building module is used to build a market price forecasting model based on a causal relationship model and an optimal strategy combination using support vector regression, and to optimize the model parameters of the market price forecasting model using a historical power data set; A scenario forecasting module is used to generate scenario forecasting results under different market scenarios based on the market price forecasting model; The control scheme determination module is used to generate a control strategy combination that matches the target market price through reverse deduction based on scenario forecast results, obtain the control scheme, and optimize the control scheme in combination with the causal relationship model.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the market price prediction method based on reverse game deduction as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the market price prediction method based on reverse game deduction according to any one of claims 1 to 7 when executed.

Images