Energy network dynamic game optimization method under carbon quota constraint

Abstract

The invention relates to the technical field of industrial data management, in particular to an energy network dynamic game optimization method under a carbon quota constraint, and the method is suitable for strategy linkage and response relationships generated by different management subjects due to carbon resource allocation and scheduling policy adjustment under a carbon quota system. According to the method, various local policy adjustment events are simulated, the response influence of the events on other management subjects is analyzed, and the possible management conflict risk is recognized; on the basis, a strategy buffer area and a time sequence transition mechanism are introduced, and transition path planning from a local strategy to global coordination is achieved; and the management strategy is dynamically adjusted in combination with feedback information in the execution process. According to the method, the potential instability region can be dynamically identified in the operation process, local reconstruction and iterative optimization are implemented in real time, and the compliance and stability of an energy system in a complex carbon quota management environment are guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial data management, in particular to an energy network dynamic game optimization method under carbon quota constraint. BACKGROUND

[0002] The carbon quota constraint mechanism is becoming a core institutional element in energy management. The regulatory department sets the carbon emission quota and makes compliance requirements for the operation activities of energy-using units. Various energy-using subjects need to meet energy efficiency targets while strictly controlling emission levels to keep the overall operation within the carbon quota permission range. This poses new challenges to existing energy dispatching, industrial data management and operation management methods.

[0003] In a multi-agent collaborative energy system, common operation modes include regional energy collaborative dispatching and multi-energy complementary resource sharing. Under the carbon constraint, each operation agent not only needs to meet its own business objectives, but also needs to comply with the upper-level carbon quota boundary, forming a complex game relationship at the strategy level. Due to the dynamic adjustment of carbon policy, such as sudden tightening of quota, correction of emission factor, and temporary effect of violation punishment, it often has an unstable impact on the original dispatching strategy, and in severe cases, it can even cause continuous failure of the operation scheme.

[0004] In particular, in the process of multi-agent collaborative game, once part of the strategy combination responds slowly or violates the policy boundary to the carbon constraint, it may cause the solution space of the local region in the whole system dispatching plan to be unstable, which is manifested as the inexecutable dispatching strategy between key nodes, frequent task conflicts or multiple boundary violations of emission boundary. Such "local strategy instability" phenomenon is highly concealed and has the risk of spreading. The existing dispatching and management system generally relies on global re-optimization to respond, resulting in strategy repair lag, too large intervention range and high management cost.

[0005] Therefore, the existing energy operation and dispatching system under carbon constraint generally lacks an optimization method with the ability to identify local unstable regions and quickly adjust strategies, and cannot efficiently solve the structural strategy risk caused by carbon policy disturbance without interfering with the global dispatching. At the same time, due to the lack of industrial data management capability, the system's fine perception and dynamic response efficiency to the execution state of carbon quota are also limited. Therefore, it is urgent to establish a strategy optimization method for operation management system based on dynamic change perception of carbon quota boundary, which can dynamically identify potential unstable regions in the operation process, implement local reconstruction and iterative optimization in real time, and thus ensure the compliance and stability of the energy system in the complex carbon quota management environment. SUMMARY

[0006] Invention purposes: In order to solve the problems mentioned in the background art, the application discloses an energy network dynamic game optimization method under carbon quota constraint, so as to construct a "local solution instability detection mechanism" for carbon quota regulation fluctuation response, and realize rapid elimination and reconstruction of high-risk strategy combination. The application simulates a plurality of local policy adjustment events, analyzes the response influence on other management subjects, and identifies the possible management conflict risk. On this basis, the strategy buffer zone and the time sequence transition mechanism are introduced to realize the transition path planning from the local strategy to the global coordination. And the management strategy is dynamically adjusted in combination with the feedback information during the execution process.

[0007] Technical scheme:

[0008] The application discloses an energy network dynamic game optimization method under carbon quota constraint, comprising:

[0009] S1, based on the current carbon quota implementation state, simulating different local scheduling strategy adjustment events, generating a corresponding disturbance event set, and constructing a potential game response prediction model for the disturbance event set;

[0010] S2, based on the historical response data and the current carbon quota execution deviation, using the potential game response prediction model to predict the response propagation path, and judging whether there is a game conflict risk;

[0011] S3, if the game conflict risk exists, based on the state gap between the current local strategy and the target strategy, constructing a strategy boundary elastic buffer zone, and generating an intermediate state buffer strategy set;

[0012] S4, according to the carbon quota limit and the resource allocation state, screening the initial time sequence transition path from the intermediate state buffer strategy set which satisfies the preset global constraint condition, and obtaining the optimal time sequence transition path by calculating the global execution feasibility index of the initial time sequence transition path;

[0013] S5, gradually executing the optimal time sequence transition path in the scheduling period, and collecting feedback data in real time; if the feedback data does not satisfy the preset global coordination index, the deviation amount is taken as a new disturbance event, and S2-S4 are executed again.

[0014] Preferably, the local scheduling strategy adjustment event includes: resource allocation proportion change event of energy management node, carbon emission weight adjustment event, scheduling response event caused by energy price fluctuation, renewable energy access proportion change event, load demand mode change event and carbon quota policy adjustment event.

[0015] Preferably, the process of generating a corresponding disturbance event set is:

[0016] extract a feature mode library of each local dispatching strategy adjustment event, and calculate a triggering probability and an influence range of each local dispatching strategy adjustment event based on the feature mode library;

[0017] parameterize sampling of each local dispatching strategy adjustment event, and construct a disturbance vector in combination with a current carbon quota implementation state; the disturbance vector includes an intensity and a direction of a disturbance event;

[0018] perform cluster analysis on the disturbance vector, and identify a representative disturbance mode;

[0019] calculate an influence degree of each disturbance mode on a key index of an energy network, establish an influence score matrix, and screen out a key disturbance event;

[0020] processing the key disturbance event according to a current carbon quota constraint condition, and generating a disturbance event set.

[0021] Preferably, the potential game response prediction model comprises a disturbance propagation structure construction unit, a response propagation path prediction unit, a round game simulation unit, and a game conflict risk judgment unit.

[0022] The disturbance propagation structure construction unit constructs a strategy response relationship graph between multiple management subjects according to the disturbance event set.

[0023] The response propagation path prediction unit analyzes the strategy response relationship graph, identifies a response propagation path of a disturbance event among multiple management subjects, and predicts an evolution trend and an influence intensity of the disturbance in multiple rounds.

[0024] The round game simulation unit simulates a strategy adjustment behavior of the multiple management subjects in a dynamic game process based on the response propagation path and historical response data, and generates a disturbance-induced response sequence.

[0025] The game conflict risk judgment unit determines a game conflict risk of the disturbance event according to a result of the round game simulation and a carbon quota implementation deviation.

[0026] Preferably, the process of constructing the strategy boundary elastic buffer zone comprises:

[0027] calculating a strategy offset vector in a multi-dimensional state space according to differences between a current local dispatching strategy and a target strategy in a carbon resource allocation state, an implementation frequency, and a strategy adaptability;

[0028] constructing a continuous strategy boundary elastic buffer zone outside a strategy boundary based on the strategy offset vector in combination with a dispatching flexibility of the multiple management subjects, a carbon quota implementation elasticity, and a response tolerance threshold.

[0029] Within the policy boundary elastic buffer zone, a multi-stage interpolation and rolling planning mechanism is adopted to generate a set of intermediate state buffer strategies from the current local strategy to the target strategy.

[0030] Preferably, an initial timing transition path meeting the preset global constraint condition is screened out from the intermediate state buffer strategy set, and an optimal timing transition path is obtained by calculating the global execution feasibility index of the initial timing transition path, and the specific process is as follows:

[0031] By constructing the mapping relationship between the local strategy change and the global carbon quota execution state, the contribution vector of each intermediate state buffer strategy in the intermediate state buffer strategy set to the overall scheduling balance, resource coordination and carbon emission index is extracted.

[0032] A plurality of feasible initial timing transition paths are constructed in the intermediate state buffer strategy set, and the contribution vector is combined for pre-screening to eliminate paths in the initial timing transition path that cannot meet the preset global constraint condition.

[0033] Each of the initial timing transition paths is simulated in the scheduling period, the global execution feasibility index is calculated, and the optimal path is screened out as the optimal timing transition path by using a weight-based multi-objective optimization algorithm.

[0034] Preferably, the specific process of S5 is as follows:

[0035] Based on the real-time collected feedback data, the deviation between the current timing transition path and the preset global coordination index is calculated.

[0036] The deviation is input into the potential game response prediction model as a new disturbance event, an updated disturbance event set is regenerated, and a new response propagation path and a game conflict risk level are predicted.

[0037] The optimal timing transition path is optimized based on the new response propagation path.

[0038] Compared with the prior art, the present application has the following advantages:

[0039] 1、The present application constructs a potential game response prediction model, effectively identifies the response chain and conflict propagation path that may be caused by local strategy adjustment events among multiple management subjects. The model combines the current carbon quota execution deviation and historical data to improve the predictability of disturbance consequences and avoid the strategy friction risk caused by "scheduling first and post-correction". By embedding this model into the entire dynamic scheduling process, the disturbance prediction and path planning links are cooperatively integrated, thereby improving the sensitivity and response foresight of the system to sudden local events.

[0040] 2、The application proposes a "strategy boundary elastic buffer zone" construction mechanism, supports multi-dimensional offset calculation between the current local strategy and the target strategy, and constructs a continuous strategy transition space based on scheduling flexibility, carbon quota elasticity and other factors. The intermediate state buffer strategy generated by interpolation and rolling planning not only avoids the impact of carbon resources caused by strategy mutation, but also gives the scheduling process the ability to transition flexibly. The mechanism provides a strategy envelope for the entire dynamic path planning, ensures that path selection will not trigger critical state violations, and at the same time, cooperates with the game prediction model to optimize the path again when the feedback does not meet the preset indicators, forming a flexible closed loop of strategy-path-feedback, greatly enhancing the execution stability and anti-disturbance ability under the carbon quota system.

[0041] 3、After constructing the intermediate state buffer strategy set, the application introduces a path feasibility modeling mechanism, combines the carbon quota execution state and resource coordination indicators, filters out the initial time sequence transition path, and further obtains the final execution path based on a multi-objective optimization algorithm. This method can dynamically simulate the execution effect of each path, and through global execution feasibility index evaluation, ensures the effectiveness of the selected path under the multi-objective of carbon constraint, scheduling balance and resource utilization. According to the disturbance intensity and carbon quota variation, the application flexibly adjusts the path strategy, realizes the real-time linkage of path planning and strategy buffer, and significantly improves the execution success rate and dynamic adaptability of carbon resource scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 A flowchart of an energy network dynamic game optimization method under carbon quota constraints provided by an embodiment of the application is shown.

[0043] Figure 2 A structure diagram of a potential game response prediction model provided by an embodiment of the application is shown.

[0044] Figure 3 A flowchart of constructing a strategy boundary elastic buffer zone provided by an embodiment of the application is shown. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0046] Carbon quota constraint mechanism is becoming a core institutional element in energy management. Under the background of carbon constraint, each operation subject not only needs to meet its own business objectives, but also needs to comply with the carbon quota boundary, forming a complex game relationship at the strategy level. Because carbon policy has the attribute of dynamic adjustment, such as: quota sudden tightening, emission factor correction, temporary effect of violation punishment, etc., it often has an unstable impact on the original dispatch strategy, and in severe cases it can even cause continuous failure of the operation scheme.

[0047] The present application proposes a kind of dynamic game optimization method of energy network under carbon quota constraint, can build " local solution instability detection mechanism " for carbon quota regulation fluctuation response, realize the quick elimination and reconstruction of high-risk strategy combination, to ensure the compliance and stability of energy system in complex carbon quota management environment. In order to explain the method of the present application can play the role of energy operation scheduling optimization under carbon constraint through local instability area identification and strategy rapid adjustment, the following will explain the effectiveness of the present application from two embodiments.

[0048] Example one

[0049] In the embodiments of the present application, the method proposed in the present application is used to build " local solution instability detection mechanism " for carbon quota regulation fluctuation response, and the process of realizing the quick elimination and reconstruction of high-risk strategy combination is described in detail. Figure 1 The specific flow chart of the method proposed in the present application includes: based on the current carbon quota implementation state, simulate different local dispatch strategy adjustment events, generate corresponding disturbance event set;For disturbance event set, construct potential game response prediction model, predict response propagation path based on historical response data and current carbon quota execution deviation, judge whether there is game conflict risk;If there is game conflict risk, then based on the state gap between current local strategy and target strategy, construct strategy boundary elastic buffer zone, generate intermediate state buffer strategy set;According to the current carbon quota limit and resource allocation state, select the initial time sequence transition path that meets the preset global constraint condition from the intermediate state buffer strategy set, and obtain the optimal time sequence transition path by calculating the global execution feasibility index of the initial time sequence transition path;In the dispatching period, the optimal time sequence transition path is executed gradually, and feedback data is collected in real time;If feedback data does not meet the preset global coordination index, re-predict and optimize. The following is described according to the content of Figure 1

[0050] Based on the current carbon quota implementation state, simulate different local dispatch strategy adjustment events, generate corresponding disturbance event set;

[0051] ​The local dispatch strategy adjustment events include: resource allocation proportion change event of the energy management node, carbon emission weight adjustment event, dispatch response event triggered by energy price fluctuation, renewable energy access proportion change event, load demand mode transition event, and carbon quota policy adjustment event.

[0052] By simulating different local dispatch strategy adjustment events, the composition dimension of the disturbance event set is enriched, the authenticity and diversity of the disturbance modeling in the overall scheme are improved, and the actual fluctuation situation in the carbon quota execution process can be simulated more accurately. Not only does it enhance the adaptability of the subsequent potential game response prediction model to the behavior reaction of different management subjects, but it also provides clear intervention basis for the subsequent strategy buffer zone construction and timing transition path screening, thereby realizing more targeted and globally coordinated optimization decisions in the carbon quota dynamic regulation process, embodying the deep synergy effect between disturbance modeling and dynamic game mechanism.

[0053] Further, the process of generating the corresponding disturbance event set is:

[0054] Extract the feature mode library of each local dispatch strategy adjustment event, and calculate the trigger probability and influence range of each local dispatch strategy adjustment event based on the feature mode library;

[0055] Parameterize sampling of each local dispatch strategy adjustment event, and construct a disturbance vector combining the current carbon quota implementation state; the disturbance vector includes the intensity and direction of the disturbance event;

[0056] Cluster analysis is performed on the disturbance vector to identify representative disturbance patterns;

[0057] The influence degree of each disturbance pattern on the key indicators of the energy network is calculated, an influence score matrix is established, and key disturbance events are selected;

[0058] According to the current carbon quota constraint condition, the key disturbance events are processed to generate a disturbance event set.

[0059] Specifically, first, collect the current carbon quota implementation state data of the energy network, including the basic information of each management subject such as carbon emissions and resource allocation proportion.

[0060] For possible local dispatch strategy adjustment events, use historical data and machine learning models for event identification and trigger probability evaluation; historical data such as 1-3 years of hourly operation data, machine learning models such as Bayesian networks or logistic regression models; the influence range is determined by analyzing the changes in carbon emissions and energy consumption of associated energy nodes when historical events occur, and modeling the intensity boundary; energy nodes such as specific regional power grids and specific industry factories.

[0061] Monte Carlo simulation is used to generate a series of parameter combinations within a preset reasonable range for key parameters of each local dispatch strategy adjustment event; the key parameters include price fluctuation amplitude, proportion change size, and policy adjustment intensity; the preset reasonable range is, for example, a price fluctuation amplitude of ±5% to ±20% with a step size of 1%; in the Monte Carlo simulation, a Gaussian process regression is used to predict the influence of the disturbance parameters, and high-risk areas are preferentially sampled.

[0062] The parameter combinations obtained by sampling are converted into disturbance vectors in combination with the deviation of the actual execution value from the target value of the current carbon quota; the deviation is the percentage of the current used quota exceeding the stage target; the disturbance vector is represented as: [event type ID, intensity (such as a 15% price increase), direction (such as a positive impact on carbon emissions), duration, current quota deviation].

[0063] Cluster analysis is performed on the generated disturbance vectors to identify representative disturbance patterns, such as high-intensity positive disturbance or low-intensity mixed disturbance.

[0064] A pre-trained energy system simulation model, such as an extended model based on the power system analysis software PSASP, is used to input the parameters of each representative disturbance pattern, simulate the impact of the disturbance pattern on key indicators of the energy network, and quantify the impact degree; the key indicators of the energy network include regional total carbon emissions, grid frequency stability, key section power flow, and renewable energy consumption rate.

[0065] An impact score matrix is constructed, with rows representing disturbance patterns and columns representing key indicators of the energy network, and the matrix elements being normalized scores of the impact degree. A screening threshold for key disturbance events is set to screen out key disturbance events that pose a significant threat to the system; the screening threshold is, for example, a disturbance pattern that causes a carbon emission exceeding probability of greater than 70% or causes a deterioration of more than 20% in the stability indicators of the power grid; scenario combination and filtering are performed according to the current carbon quota constraint, and finally a set of N high-risk disturbance events that need to be focused on is formed, which is used for subsequent game analysis.

[0066] Through multi-dimensional and parameterized disturbance simulation and system clustering screening, key disturbance scenarios that pose the most significant threat to the carbon quota control target can be accurately identified from complex and variable external and internal factors, potential disturbance events in the energy network can be fully captured, and the comprehensiveness and pertinence of risk identification are significantly improved, laying a solid foundation for subsequent game analysis and strategy formulation.

[0067] Further, for the set of disturbance events, a potential game response prediction model is constructed to predict the response propagation path based on historical response data and the current carbon quota execution deviation, and to determine whether there is a game conflict risk.

[0068] Referring to Figure 2The potential game response prediction model comprises a disturbance propagation structure construction unit, a response propagation path prediction unit, a round game simulation unit, and a game conflict risk judgment unit.

[0069] The disturbance propagation structure construction unit constructs a strategy response relationship graph among the multiple management subjects according to the set of disturbance events.

[0070] The response propagation path prediction unit analyzes the strategy response relationship graph, identifies the response propagation path of the disturbance event among the multiple management subjects, and predicts the evolution trend and influence intensity of the disturbance in multiple rounds.

[0071] The round game simulation unit simulates the strategy adjustment behavior of the multiple management subjects in the dynamic game process based on the response propagation path and historical response data, and generates a disturbance-induced response sequence.

[0072] The game conflict risk judgment unit determines the game conflict risk of the disturbance event according to the results of the round game simulation and the carbon quota execution deviation.

[0073] Specifically, the disturbance propagation structure construction unit models the influence path of the disturbance event and constructs a strategy response relationship graph among the multiple management subjects.

[0074] The response propagation path prediction unit, based on the strategy response relationship graph, combines historical disturbance-response samples, and uses a pre-trained graph neural network and a time series regression model to predict how the disturbance spreads in multiple rounds of games, and identify the path nodes and propagation strength that may trigger non-cooperative behavior.

[0075] The round game simulation unit simulates the strategy game process of the multiple management subjects under a given disturbance and current carbon quota execution deviation by establishing a reinforcement learning framework, generates a multi-round response sequence, such as whether the scheduling compression behavior of subject A causes the emission weighting of subject B to increase, etc.

[0076] The game conflict risk judgment unit calculates the non-coordination index in the disturbance-induced response chain based on the simulation results and the current carbon quota execution deviation, including the emission over-standard rate, response time delay, etc., and sets a conflict risk threshold to identify whether the current disturbance has the risk of triggering multi-party strategy conflict; the conflict risk threshold is dynamically updated by a machine learning model such as support vector machine. Table 1 shows the performance table of response path prediction and game conflict identification of the potential game response prediction model.

[0077] Table 1 Performance table of potential game response prediction model

[0078] Experimental scenario Number of perturbation events Response path prediction error (%) Game conflict identification accuracy (%) A 25 6.3 92.1 B 30 5.7 93.7 C 28 4.9 94.6 D 33 4.1 95.3 E 35 3.8 96.0

[0079] In Table 1, scenario A is a single energy node for scheduling ratio adjustment, the carbon quota pressure is slight, and the response chain is unidirectional propagation; scenario B is a double energy node for scheduling conflict adjustment, which triggers limited cascade response, and there is a double coupling of scheduling-carbon emission; scenario C is a sudden change in the proportion of renewable energy, which causes sudden adjustment of carbon quota and causes multi-node strategy synchronous deviation; scenario D is the superposition of energy price disturbance and carbon quota execution deviation, the game chain propagation length is lengthened, and the round response intensity is enhanced; scenario E is a comprehensive large disturbance event, containing carbon emission policy mutation + high intensity scheduling heterogeneous disturbance, and the multi-agent collaborative game risk is significant;

[0080] The accuracy of the system to determine whether a disturbance can cause a "strategy conflict" is represented; wherein TP is the number of events correctly predicted as having a conflict; TN is the number of events correctly predicted as having no conflict; FP is the number of events incorrectly predicted as having a conflict; and FN is the number of events incorrectly predicted as having no conflict;

[0081] The game conflict recognition accuracy is obtained by measuring the deviation between the predicted disturbance response path and the actual response path.

[0082] Through the construction and integration of the game response prediction model, the system can predict the conflicts and non-cooperative behaviors that may occur between multiple management subjects before the disturbance actually occurs, significantly improving the ability of the scheduling mechanism to understand the strategy evolution trend. At the same time, the combination of the strategy response relationship diagram and the round game simulation can quantify the disturbance path and risk level, providing a basis for the reasonable construction of the subsequent elastic buffer zone. The game response prediction model gives the system the ability of "predictive regulation", effectively avoiding resource mismatch, carbon right waste and other problems caused by game misjudgment under the constraints of complex energy networks and carbon quota, improving the stability of the overall strategy transition and the efficiency of global carbon regulation and execution, and embodying the deep integration between disturbance modeling and dynamic game intelligent identification.

[0083] Further, if there is a game conflict risk, based on the state gap between the current local strategy and the target strategy, a strategy boundary elastic buffer zone is constructed to generate an intermediate state buffer strategy set;

[0084] Referring to Figure 3 , the process of constructing the strategy boundary elastic buffer zone is as follows:

[0085] According to the differences between the current local scheduling strategy and the target strategy in carbon resource allocation state, execution frequency and strategy adaptability, a strategy offset vector in a multi-dimensional state space is calculated;

[0086] Based on the strategy offset vector, combined with the scheduling flexibility of the multi-management subject, the carbon quota execution flexibility and the response tolerance threshold, a continuous strategy boundary elastic buffer zone is constructed outside the strategy boundary to envelope the strategy transition state;

[0087] Within the strategy boundary elastic buffer zone, a set of intermediate state buffer strategies that transit from the current local strategy to the target strategy is generated by using multi-level interpolation and rolling planning mechanism.

[0088] Specifically, after identifying the risk of game conflict, the state gap between the current local strategy and the target strategy is represented by calculating the displacement vector of the strategy in the multi-dimensional state space according to the multi-dimensional indicators such as the carbon resource allocation ratio, the scheduling execution frequency and the historical response adaptability between the current local strategy and the target strategy.

[0089] According to the scheduling flexibility parameters of each management subject, such as response delay tolerance and resource switching cost, as well as the carbon quota execution elasticity and response tolerance threshold, a continuous strategy boundary elastic buffer zone is dynamically constructed near the boundary of the strategy displacement direction in the state space. The strategy boundary elastic buffer zone not only envelops the current strategy state and the target state, but also covers the possible intermediate transition state space, thereby forming a feasible region of the transition strategy.

[0090] Within the strategy boundary elastic buffer zone, a set of intermediate state buffer strategies that transit from the current local strategy to the target strategy is generated by using multi-level interpolation and rolling planning mechanism.

[0091] By constructing the strategy boundary elastic buffer zone, the transition state idea is introduced, which converts the original rigid strategy adjustment into a continuous controllable elastic strategy evolution path, effectively avoiding system instability or management subject rebound behavior caused by strategy mutation. The introduction of the elastic buffer zone realizes the expansion envelope of the strategy space, so that the system can flexibly coordinate the tension between resources and response capacity when facing carbon constraints and scheduling inconsistency of multiple management subjects. The construction of the intermediate state strategy set provides a rich selection of scheduling strategies for subsequent selection of the optimal transition path, significantly improving the robustness of carbon scheduling and the smoothness of system execution. In addition, the use of rolling optimization and multi-level interpolation technology ensures the gradual evolution of the strategy in the time dimension, providing high response and sustainability support for the "conflict risk-strategy generation-path evaluation" chain, greatly enhancing the execution stability and anti-disturbance ability under the carbon quota system. Table 2 shows the effect of constructing the strategy boundary elastic buffer zone.

[0092] Table 2 Effect of constructing the strategy boundary elastic buffer zone

[0093]

[0094] The envelope coverage rate is obtained by calculating the proportion of the potential strategy space enveloped by the buffer strategy set to all feasible strategy spaces;

[0095] where S iis the strategy vector of the ith step, ||·|| is the vector difference, a is the volatility penalty factor, which is used to measure the continuity and low volatility of the strategy sequence in the transition process, and the higher the strategy stability score indicates the smaller the volatility;

[0096] The number of buffer start times represents the number of times the strategy backtracking and entering the elastic buffer zone for re-planning is triggered in the scheduling process;

[0097] The target achievement efficiency is obtained by calculating the ratio of the time required for the strategy to achieve the preset scheduling target to the standard expected time.

[0098] Further, according to the carbon quota limit and the resource allocation state, an initial timing transition path that satisfies the preset global constraint condition is selected from the intermediate state buffer strategy set, and the optimal timing transition path is obtained by calculating the global execution feasibility index of the initial timing transition path; the specific process is:

[0099] By constructing the mapping relationship between local strategy change and global carbon quota execution state, the contribution vector of each intermediate state buffer strategy in the intermediate state buffer strategy set to the overall scheduling balance, resource coordination and carbon emission indicators is extracted;

[0100] A plurality of feasible initial timing transition paths are constructed in the intermediate state buffer strategy set, and pre-screening is performed in combination with the contribution vector to eliminate paths in the initial timing transition path that cannot satisfy the preset global constraint condition.

[0101] Each initial timing transition path is simulated in the scheduling period, the global execution feasibility index is calculated, and the optimal path is selected as the optimal timing transition path by using a weight-based multi-objective optimization algorithm.

[0102] Specifically, a mapping model between local strategy change and global carbon quota execution state is constructed, historical data fitting and graph neural network structure are used to extract the scheduling balance influence, resource coordination influence and carbon emission change trend of each intermediate state strategy in the global system, and the corresponding contribution vector is constructed; the scheduling balance influence, for example, the peak-valley mismatch caused by local load transfer; the resource coordination influence, for example, the complementary efficiency between energy sources.

[0103] Taking the intermediate state buffer strategy as a node, a plurality of possible initial timing transition paths are constructed, and pre-screening is performed according to the contribution vector of each path at each key stage to eliminate paths that do not satisfy the preset global constraint condition; the global constraint conditions include carbon quota constraint, resource constraint and timing coordination requirement.

[0104] Simulate each initial timing transition path in the entire scheduling period, analyze the global influence of each initial timing transition path on system operation, and calculate the global execution feasibility score including carbon emission error rate, resource utilization efficiency, system response delay and other indicators.

[0105] Based on the weight multi-objective optimization algorithm, the global indicator weight is comprehensively evaluated, the path with the best comprehensive performance is screened out, and the optimal timing transition path is determined as the reference sequence for subsequent scheduling execution. The multi-objective optimization algorithm can use, for example, NSGA-II algorithm based on Pareto frontier or entropy weight TOPSIS method.

[0106] By introducing the global contribution vector mechanism and path simulation evaluation, a high-dimensional screening and simulation verification linkage process from the strategy set to the execution path is realized, ensuring that the selected path is not only locally reasonable, but also has optimal execution capability within the carbon quota constraint at the global level. Through path-level simulation and multi-objective optimization, the present application avoids the problems of "late imbalance" or "resource bottleneck" in the actual execution of the scheduling path, and improves the sustainability, consistency and stability of system strategy evolution. Especially the four-layer process of "building influence mapping relationship + contribution vector + path feasibility simulation + optimal path optimization" embeds response foresight judgment and decision credibility evaluation in system strategy selection, which reflects the systematic innovation and engineering applicability advantage of the present application in carbon scheduling path generation and optimization.

[0107] Further, the optimal timing transition path is gradually executed within the scheduling period, and feedback data is collected in real time; if the feedback data does not meet the preset global coordination index, reforecasting and optimization are performed; the specific process is as follows:

[0108] Based on the real-time collected feedback data, the deviation amount between the current timing transition path and the preset global coordination index is calculated;

[0109] The deviation amount is input as a new disturbance event into the potential game response prediction model to regenerate an updated disturbance event set, and predict a new response propagation path and game conflict risk level;

[0110] Optimize the optimal timing transition path based on the new response propagation path.

[0111] By introducing the real-time feedback closed-loop mechanism, dynamic monitoring and adaptive adjustment of the execution effect of the optimal timing transition path are realized, effectively improving the response agility and strategy robustness of the present application to dynamic changes in the environment, ensuring that the scheduling process always evolves towards the global optimal direction, and significantly enhancing the adaptability and steady-state controllability of the energy network under carbon quota regulation.

[0112] Through the whole-process collaborative mechanism of disturbance event simulation-response prediction-conflict identification-strategy buffer-path screening-closed-loop feedback, the early perception and dynamic relief of strategy conflict among multiple management subjects under the condition of carbon quota uncertainty are realized. First, based on the current carbon quota implementation state, local disturbance is simulated, and a game response prediction model is constructed by using historical response data and carbon quota execution deviation, so that potential strategy conflicts are effectively identified. Then, for the local strategies with conflicts, a transition strategy set is generated by constructing a strategy boundary elastic buffer zone to guide the smooth transition of strategies. Subsequently, combined with the global carbon quota and resource constraints, the multi-path is screened and the optimal execution scheme is optimized by comprehensively considering the feasibility index. Finally, through the real-time correction mechanism driven by feedback data, the system stability and coordination of the scheduling path are ensured. The present application not only has the advantages of high risk identification accuracy, strategy adjustment flexibility and scheduling path adaptability, but also can dynamically identify the potential instability area in the complex and changeable carbon quota constraint scene, and realize local reconstruction and iterative optimization in real time, which significantly improves the response efficiency, game stability and global coordination level of energy network regulation, and has strong technical innovation and practical value.

[0113] Embodiment two

[0114] In embodiment one, the method proposed by the present application successfully realizes the rapid elimination and reconstruction of high-risk strategy combination by constructing a "local solution instability detection mechanism" for carbon quota regulation fluctuation response. To further verify the effectiveness of the present application, another carbon-constrained energy operation scheduling system is optimized in the embodiment of the present application.

[0115] Simulate different local scheduling strategy adjustment events based on the current carbon quota implementation state to generate a corresponding disturbance event set;

[0116] The local scheduling strategy adjustment events include but are not limited to: energy management node resource allocation ratio change event, carbon emission weight adjustment event, scheduling response event triggered by energy price fluctuation, renewable energy access ratio change event, load demand mode change event and carbon quota policy adjustment event.

[0117] Further, the process of generating a corresponding disturbance event set is:

[0118] Extract the feature mode library of each local scheduling strategy adjustment event, and calculate the trigger probability and influence range of each local scheduling strategy adjustment event based on the feature mode library;

[0119] Parameterize sampling of each local scheduling strategy adjustment event, and construct a disturbance vector combining the current carbon quota implementation state; the disturbance vector includes the intensity and direction of the disturbance event;

[0120] Cluster analysis is performed on the disturbance vector to identify representative disturbance modes;

[0121] calculating the degree of influence of each disturbance mode on the key indicators of the energy network, establishing an influence score matrix and screening out key disturbance events;

[0122] processing the key disturbance events according to the current carbon quota constraint condition to generate a disturbance event set.

[0123] Further, for the disturbance event set, a potential game response prediction model is constructed, and based on historical response data and current carbon quota execution deviation, a response propagation path is predicted to determine whether there is a game conflict risk;

[0124] The potential game response prediction model includes: a disturbance propagation structure construction unit, a response propagation path prediction unit, a round game simulation unit and a game conflict risk judgment unit;

[0125] The disturbance propagation structure construction unit constructs a strategy response relationship graph between multiple management subjects according to the disturbance event set;

[0126] The response propagation path prediction unit analyzes the strategy response relationship graph, identifies the response propagation path of the disturbance event among the multiple management subjects, and predicts the evolution trend and influence intensity of the disturbance in multiple rounds;

[0127] The round game simulation unit simulates the strategy adjustment behavior of the multiple management subjects in the dynamic game process based on the response propagation path and the historical response data, and generates a disturbance triggered response sequence;

[0128] The game conflict risk judgment unit determines the game conflict risk of the disturbance event according to the results of the round game simulation and the carbon quota execution deviation.

[0129] Further, if there is a game conflict risk, a strategy boundary elastic buffer zone is constructed based on the state gap between the current local strategy and the target strategy, and a set of intermediate state buffer strategies is generated;

[0130] The process of constructing the strategy boundary elastic buffer zone is:

[0131] According to the differences between the current local scheduling strategy and the target strategy in carbon resource allocation state, execution frequency and strategy adaptability, a strategy offset vector in a multi-dimensional state space is calculated;

[0132] Based on the strategy offset vector, combined with the scheduling flexibility of the multiple management subjects, the carbon quota execution elasticity and the response tolerance threshold, a continuous strategy boundary elastic buffer zone is constructed outside the strategy boundary to envelope the strategy transition state;

[0133] In the strategy boundary elastic buffer zone, a set of intermediate state buffer strategies from the current local strategy to the target strategy is generated by using a multi-level interpolation and rolling planning mechanism.

[0134] Table 3 gives the policy flexibility scheduling ability and multi-objective achievement degree comparison table of introducing policy boundary elastic buffer.

[0135] Table 3 introduces policy boundary elastic buffer for policy flexibility scheduling ability and multi-objective achievement degree comparison table

[0136]

[0137] In the above table 3, the policy jump frequency is the number of times of drastic change in the policy sequence; the elastic buffer usage rate is the proportion of actually used buffer strategy; the multi-objective average achievement rate is obtained by calculating the ratio of the number of achieved targets to the number of target achievements of the scheduling target (such as load balancing, carbon emission control, etc.); the carbon emission standard reaching rate is obtained by calculating the frequency of carbon emission control below the target threshold in the scheduling process.

[0138] Further, according to the carbon quota limit and the resource allocation state, an initial time sequence transition path that satisfies the preset global constraint condition is selected from the intermediate state buffer strategy set, and an optimal time sequence transition path is obtained by calculating the global execution feasibility index of the initial time sequence transition path; the specific process is:

[0139] By constructing the mapping relationship between local policy change and global carbon quota execution state, the contribution vector of each intermediate state buffer strategy in the intermediate state buffer strategy set to the overall scheduling balance, resource coordination and carbon emission index is extracted;

[0140] A plurality of feasible initial time sequence transition paths are constructed in the intermediate state buffer strategy set, and pre-screening is performed in combination with the contribution vector to eliminate paths in the initial time sequence transition path that cannot satisfy the preset global constraint condition;

[0141] Each initial time sequence transition path is simulated in the scheduling period, the global execution feasibility index is calculated, the optimal path is selected by using a weight-based multi-objective optimization algorithm, and the optimal time sequence transition path is obtained.

[0142] Further, the optimal time sequence transition path is executed step by step in the scheduling period, and feedback data is collected in real time; if the feedback data does not satisfy the preset global coordination index, re-prediction and optimization are performed; the specific process is:

[0143] Based on the feedback data collected in real time, the deviation amount between the current time sequence transition path and the preset global coordination index is calculated;

[0144] The deviation amount is input as a new disturbance event into the potential game response prediction model, an updated disturbance event set is regenerated, and a new response propagation path and game conflict risk level are predicted;

[0145] Optimizing optimal timing transition paths based on new response propagation paths.

[0146] The above description of the embodiments allows those skilled in the art to realize or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application should not be limited to the embodiments shown herein but should cover the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A dynamic game-theoretic optimization method for energy networks under carbon quota constraints, characterized in that, The method includes the following steps: S1 simulates different local scheduling strategy adjustment events based on the current carbon quota implementation status, generates a corresponding set of disturbance events, and constructs a potential game response prediction model for the set of disturbance events. S2 uses a potential game response prediction model to predict the response propagation path based on historical response data and current carbon quota execution deviations, and determines whether there is a risk of game conflict. S3 If the aforementioned game conflict risk exists, then based on the state gap between the current local strategy and the target strategy, construct a policy boundary elastic buffer and generate an intermediate buffer strategy set. Based on carbon quota restrictions and resource allocation status, S4 selects an initial time-series transition path that meets preset global constraints from the intermediate buffer strategy set, and obtains the optimal time-series transition path by calculating the global execution feasibility index of the initial time-series transition path. S5 executes the optimal timing transition path step by step within the scheduling cycle and collects feedback data in real time; if the feedback data does not meet the preset global coordination index, the deviation is treated as a new disturbance event, and S2-S4 are re-executed.

2. The dynamic game optimization method for energy networks under carbon quota constraints according to claim 1, characterized in that, The local scheduling strategy adjustment events include: resource allocation ratio changes at energy management nodes, carbon emission weight adjustments, scheduling response events triggered by energy price fluctuations, renewable energy integration ratio changes, load demand pattern shifts, and carbon quota policy adjustments.

3. The dynamic game optimization method for energy networks under carbon quota constraints according to claim 2, characterized in that, The process of generating the corresponding set of disturbance events is as follows: Extract the feature pattern library of each local scheduling strategy adjustment event, and calculate the trigger probability and impact range of each local scheduling strategy adjustment event based on the feature pattern library; The local scheduling strategy adjustment events are parameterized and sampled, and a disturbance vector is constructed in conjunction with the current carbon quota implementation status; the disturbance vector includes the intensity and direction of the disturbance event; Cluster analysis is performed on the perturbation vectors to identify representative perturbation patterns; Calculate the impact of each disturbance mode on key indicators of the energy network, establish an impact scoring matrix, and screen out key disturbance events; The key disturbance events are processed according to the current carbon quota constraints to generate a set of disturbance events.

4. The dynamic game optimization method for energy networks under carbon quota constraints according to claim 3, characterized in that, The potential game response prediction model includes: a disturbance propagation structure construction unit, a response propagation path prediction unit, a round game simulation unit, and a game conflict risk judgment unit; The disturbance propagation structure construction unit constructs a strategy response relationship diagram among multiple management entities based on the set of disturbance events; The response propagation path prediction unit analyzes the strategy response relationship diagram, identifies the response propagation path of the disturbance event among multiple management entities, and predicts the evolution trend and impact intensity of the disturbance in multiple rounds. The round-based game simulation unit simulates the strategy adjustment behavior of multiple management entities in a dynamic game process based on the response propagation path and historical response data, and generates a disturbance-triggered response sequence. The game conflict risk assessment unit determines the game conflict risk of the disturbance event based on the results of the round game simulation and the deviation in carbon quota implementation.

5. The dynamic game optimization method for energy networks under carbon quota constraints according to claim 1, characterized in that, The process of constructing the elastic buffer at the policy boundary is as follows: Based on the differences between the current local scheduling strategy and the target strategy in terms of carbon resource allocation status, execution frequency, and strategy adaptability, calculate the strategy offset vector in the multidimensional state space; Based on the policy offset vector, and combined with the scheduling flexibility of multiple management entities, the carbon quota execution elasticity, and the response tolerance threshold, a continuous policy boundary elastic buffer is constructed outside the policy boundary. Within the policy boundary elastic buffer, a multi-level interpolation and rolling planning mechanism is used to generate a set of intermediate buffer policies that transition from the current local policy to the target policy.

6. The dynamic game optimization method for energy networks under carbon quota constraints according to claim 1, characterized in that, The initial timing transition path that meets the preset global constraints is selected from the intermediate buffer strategy set, and the optimal timing transition path is obtained by calculating the global execution feasibility index of the initial timing transition path. The specific process is as follows: By constructing a mapping relationship between the impact of local policy changes on the global carbon quota execution status, the contribution vector of each intermediate buffer policy in the intermediate buffer policy set to the overall scheduling balance, resource coordination and carbon emission indicators is extracted. Multiple feasible initial temporal transition paths are constructed from the intermediate buffer strategy set, and pre-screened using the contribution vector to eliminate paths that cannot meet the preset global constraints. For each of the initial time-series transition paths, simulation is performed within the scheduling period to calculate the global execution feasibility index. The optimal path is selected as the optimal time-series transition path using a weighted multi-objective optimization algorithm.

7. The dynamic game optimization method for energy networks under carbon quota constraints according to claim 1, characterized in that, The specific process of S5 is as follows: Based on real-time collected feedback data, the deviation between the current time-series transition path and the preset global coordination index is calculated; The deviation is used as a new perturbation event to input into the potential game response prediction model, and an updated set of perturbation events is regenerated to predict the new response propagation path and the game conflict risk level. The optimal timing transition path is optimized based on the new response propagation path.

Images