Method and system for cooperative-competitive game decision of water regulation plan under incomplete information

Abstract

The application discloses a kind of non-complete information under water plan cooperation-competition game decision method and system, the method includes: cooperation-competition bayesian game model of water-receiving area water plan is constructed, water source area diversion and water-receiving area inflow sharing information, water-receiving area private information and water-receiving area water plan strategy set are used as model input, respectively using lumped decision optimization, bayesian learning evolutionary game etc. Method is solved game model cooperation part group benefit optimal value and the optimal value of individual relative advantage income of water-receiving area competition part, the cooperation-competition solution of each water-receiving area water plan game is obtained by comprehensive calculation, and the water plan decision of water-receiving area is determined by counteracting calculation.The application obtains the stable strategy of annual water diversion plan under non-complete information condition for water diversion project.

Description

Technical Field

[0001] This invention belongs to the field of water resource allocation and relates to game theory decision-making for annual water transfer project allocation plans. In particular, it relates to a cooperative-competitive game theory decision-making method and system for water transfer plans under incomplete information. Background Technology

[0002] The prerequisite for the orderly operation of inter-basin water transfer projects is the formulation of an annual water allocation plan. The formulation of the water transfer plan mainly follows two principles: first, it must meet the overall water resource allocation conditions and the operational safety requirements of the water transfer project; second, it must fully refer to the inflow forecast data of the water source area and the water receiving area, as well as the annual water use plan suggestions reported by each water receiving area. On this basis, by leveraging the regulation and storage functions of the reservoir group along the project route, a time-by-time water storage and supply plan is formulated to ensure the timely and accurate supply of water resources throughout the year. Furthermore, under different water inflow conditions, the water use allocation of various regions and industries can be coordinated to achieve efficient allocation of water resources and ultimately ensure the successful achievement of the water supply target designed for the water transfer project.

[0003] The water supply benefits obtained by the water-receiving areas of the water diversion project are influenced by their own proposed water use plans, as well as the proposed water use plans of other water-receiving areas. Therefore, the process of each water-receiving area proposing its water use plan can be viewed as a game between decision-making entities representing different regional interests. It is worth noting that the water use plan game among the water-receiving areas has both competitive and cooperative attributes. On the one hand, each water-receiving area pursues the maximization of water supply benefits and water supply security, resulting in competition for water diversion. On the other hand, the water-receiving areas share common demands in improving water diversion utilization and ensuring water supply fairness, thus providing a basis for cooperation. At the same time, this game process is characterized by incomplete information. Some information, such as meteorological and hydrological data, is generally observable or easily accessible in the water-receiving areas and is therefore shareable. However, other information, such as the status of water resource development and utilization in the water-receiving areas and the actual water demand determined based on this, is subject to certain barriers between regions, resulting in information asymmetry. In addition, the randomness brought about by uncertain factors in the forecasting of water inflow and water demand for water diversion projects is information that is difficult for all decision-making bodies to fully grasp. However, existing methods involving water resource allocation and scheduling in water diversion projects have not yet provided technical solutions for the dual nature of cooperation and competition in the water-receiving areas and the characteristics of incomplete information.

[0004] In summary, to address the aforementioned problems, this invention proposes a cooperative-competitive game-based decision-making method and system for water diversion planning under incomplete information conditions. This provides technical support for optimizing and determining annual water use plans under incomplete information conditions and contributes to further research on collaborative management and benefit balance in water network projects. Summary of the Invention

[0005] Objective: To provide a cooperative-competitive game-based decision-making method for water diversion projects with incomplete information, thereby addressing the aforementioned problems in existing technologies. Furthermore, to provide a cooperative-competitive game-based decision-making system for water diversion projects with incomplete information.

[0006] Technical Solution: A cooperative-competitive game theory decision-making method for water diversion projects with incomplete information, including the following steps:

[0007] Step S1: Collect data on the water source area and water receiving area of ​​the water transfer project, calculate the water diversion volume available in the water source area and the predicted water inflow in the water receiving area, and use it as the water inflow sharing information in the process of formulating the annual water transfer plan. Calculate the water demand process and its probability of occurrence in the water receiving area under different water inflow scenarios and use it as the water demand private information of each water receiving area.

[0008] Step S2: Treat the different water-receiving areas of the water diversion project as decision-making entities, collect annual water use plan suggestions for each time period of the water-receiving areas and treat them as strategies, and generate a water use plan strategy set for each water-receiving area.

[0009] Step S3: Use the shared water information, private water demand information, and water use planning strategy sets of each water-receiving area as inputs to the pre-constructed cooperative-competitive Bayesian game model of water use planning in the water-receiving area. Solve the model to obtain the optimal value of the group payoff in the cooperative part of the model and the optimal value of the individual relative advantage payoff of each water-receiving area.

[0010] Step S4: Based on the optimal value of group benefits and the optimal value of individual relative advantage benefits, calculate the CO-CO (cooperation-competition) value of the water use planning game for each water-receiving area, and solve the problem by back-calculating based on the CO-CO value of the water use planning game for each water-receiving area, and finally determine the water use planning decision for each water-receiving area.

[0011] According to one aspect of this application, step S1 is further comprising:

[0012] Step S11: Collect data on the water source area and water receiving area of ​​the water transfer project, including: meteorological and hydrological forecast data, regional population, economic and social data, water use efficiency data, water supply capacity data of the water transfer project, and regional water use management policies, and unify the time dimension of the collected data and synchronize them.

[0013] Step S12: Based on regional population, economic and social data, water supply capacity data of water diversion projects and regional water management policies, determine the ecological base flow guarantee threshold and water demand of the water source area, define the calculation boundary of the divertable water volume, and calculate the water demand, total available water resources and divertable water volume of the water source area respectively.

[0014] Step S13: Collect historical rainfall data and combine it with meteorological and hydrological forecast data to divide several water inflow scenarios, and calculate the local water inflow and water distribution of the water-receiving area under each water inflow scenario. Use the water diversion capacity of the water source area and the local water inflow and water distribution of the water-receiving area under each water inflow scenario as water inflow sharing information.

[0015] Step S14: Each water-receiving area calculates the water demand process and its corresponding probability of occurrence based on its own historical water use and water supply plan data, population, economic and social data, water use efficiency data, regional water use management policy data, and in conjunction with the water inflow scenario, and uses this as the water demand private information of each water-receiving area.

[0016] According to one aspect of this application, step S14 further comprises:

[0017] Step S14a: Calculate the total water demand of the water-receiving area and the water demand for domestic, industrial, agricultural and ecological purposes under each water inflow scenario according to the four major water use categories of domestic, industrial and agricultural and ecological purposes. Plot the water demand curves under different water inflow scenarios with time as the horizontal axis and water demand as the vertical axis to obtain the water demand process.

[0018] Step S14b: Collect historical rainfall data of the water-receiving area, count the number of years corresponding to different rainfall frequencies, and calculate the initial probability of each water inflow scenario;

[0019] Step S14c: Based on the reliability of the weather forecast, the initial probabilities of each water inflow scenario are corrected and normalized to ensure that the sum of the probabilities of all water inflow scenarios is 1, and the probability of occurrence of each water inflow scenario is obtained.

[0020] Step S14d: Determine the water demand process and its probability of occurrence data under different water inflow scenarios in the water-receiving area as private water demand information.

[0021] According to one aspect of this application, step S2 further comprises:

[0022] Step S21: Identify each water-receiving area covered by the water diversion project as an independent decision-making entity, and define the geographical scope, water use control authority, and responsibility boundaries of each entity;

[0023] Step S22: Based on the private water demand information and the shared water inflow information of each water-receiving area, generate a finite set of water use planning alternative strategies, which includes annual water consumption at different levels and monthly water use allocation, i.e., the water use planning strategy set.

[0024] According to one aspect of this application, step S3 further comprises:

[0025] Step S31: Select the most important online regulating reservoir along the entire water diversion project. Using this reservoir as a node, the water diversion project system is generalized into a water supply network system consisting of the regulating reservoir, the water receiving area above the reservoir, and the water receiving area below the reservoir.

[0026] Step S32: Construct a cooperative-competitive Bayesian game model of water use planning for water-receiving areas in the generalized system.

[0027] Step S33: Perform CO-CO decomposition on the Bayesian game of water use plan for the water-receiving areas, which includes two parts: cooperation and competition. The rules for the cooperation part are that the water-receiving areas cooperate completely, share information, and share benefits equally, with each water-receiving area receiving an average group benefit. The rules for the competition part are that the water-receiving areas do not share information and engage in a zero-sum game, that is, they pursue the maximum difference between the benefits of their own area and those of other areas, with each water-receiving area receiving a relative advantage benefit.

[0028] Step S34: Convert the water supply guarantee rate, water use efficiency, water diversion utilization rate and water supply fairness game objective of the water receiving area into transferable utility, and set the comprehensive benefit function of the water receiving area based on this.

[0029] According to one aspect of this application, step S33 further comprises:

[0030] Step S33a: Each water-receiving area independently and randomly selects any water use planning strategy from its own water use planning strategy set to form a combination of water use planning strategies for the water-receiving area.

[0031] Step S33b: Input the current water use planning strategy combination, annual available water volume, water inflow data of the water receiving area, predicted water demand process and its probability distribution information into the annual water volume scheduling plan compilation model of the water diversion project. The annual water volume scheduling plan compilation model of the water diversion project simulates and calculates the water storage and supply volume of each time period under different water demand processes in each water receiving area, and then obtains the expected value of the total comprehensive benefit of the water receiving area.

[0032] Step S33c: Repeat the process of randomly selecting water use planning strategies for the water-receiving area. After traversing all combinations of water use planning strategies, compare the expected value of the total comprehensive benefits and select the maximum value to obtain the optimal value of the group benefits for the cooperative part of the model.

[0033] According to one aspect of this application, step S33 further comprises:

[0034] Step S33d: Set the initial selection probability of different water use planning strategies for the water-receiving areas to be equal, calculate the location potential energy factor of each water-receiving area, carry out the first round of game, traverse all combinations of water use planning strategies, and calculate the expected value of the individual relative advantage of the water-receiving areas under different water demand processes based on the annual water volume scheduling plan compilation model of the water diversion project.

[0035] Step S33e: Based on the information obtained during the game, each water-receiving area uses a spatiotemporally aware five-dimensional adaptive heterogeneity coupled replication dynamic equation to update its beliefs. The five dimensions include information receptivity, strategy response strength, risk preference coefficient, location potential factor, and temporal memory factor. The information receptivity and strategy response strength are dynamically adjusted with each round of the game, and the temporal memory factor is updated cumulatively based on the changes in the payoffs of previous games.

[0036] Step S33f: Iterate continuously. After each iteration, the probability of strategy selection is subject to boundary constraints and normalization, and the update magnitude is controlled by adaptive learning rate decay. When all water-receiving areas meet the multiple convergence criteria of strategy probability stability, dominant strategy significance, and continuous stable rounds, it is considered that the Nash equilibrium of the Bayesian game has been achieved. At this time, the optimal value of the individual relative advantage payoff of the water-receiving areas is obtained.

[0037] According to one aspect of this application, step S4 further comprises:

[0038] S41. Based on the optimal value of group payoff and the optimal value of individual relative advantage payoff, calculate the CO-CO (cooperation-competition) value of the Bayesian game of cooperation-competition for water use planning in the water-receiving areas above and below the reservoir.

[0039] S42. Using the CO-CO value as the target value for optimizing the water transfer revenue of the water-receiving area, and using the planned water consumption of each month of the year in the water-receiving area as the decision variable, construct a back-calculation solution model for the water consumption plan of the water-receiving area.

[0040] S43. Input the available water volume, water inflow data of the water receiving area, and different predicted water demand processes under the current water inflow forecast level into the back calculation solution model of the water use plan of the water receiving area. Use a multi-group particle swarm intelligent optimization algorithm to solve the model and obtain the water use plan decision of the water receiving area corresponding to various water demand processes.

[0041] According to another aspect of this application, a cooperative-competitive game decision-making system for water diversion planning with incomplete information is provided, comprising:

[0042] At least one processor; and

[0043] A memory communicatively connected to at least one of the processors; wherein,

[0044] The memory stores instructions that can be executed by the processor to implement the incomplete information water diversion plan cooperative-competitive game decision-making method described in any of the above technical solutions.

[0045] Beneficial effects: The cooperative-competitive game decision-making method and system for water transfer plans under incomplete information can reflect the dual nature of water use planning games in the water-receiving areas of water transfer projects, which involves both competition and cooperation. This aligns with the actual management practice of water transfer involving consultation and decision-making among different regions or river basins. Combining the calculation of cooperative-competitive payoffs with evolutionary game theory based on Bayesian ideas helps the water-receiving areas obtain the optimal equilibrium strategy for annual water use plans under incomplete information conditions, ensuring the realization of water transfer benefits. Attached Figure Description

[0046] Figure 1 This is a flowchart of the present invention.

[0047] Figure 2 This is a flowchart of step S1 of the present invention.

[0048] Figure 3 This is a flowchart of step S2 of the present invention.

[0049] Figure 4 This is a flowchart of step S3 of the present invention.

[0050] Figure 5 This is a flowchart of step S4 of the present invention. Detailed Implementation

[0051] like Figure 1 As shown, the following technical solution is proposed. According to one aspect of this application, a cooperative-competitive game decision-making method for water diversion planning with incomplete information is provided, characterized by comprising the following steps:

[0052] Step S1: Collect data on the water source area and water receiving area of ​​the water transfer project, calculate the water diversion volume available in the water source area and the predicted water inflow in the water receiving area, and use it as the water inflow sharing information in the process of formulating the annual water transfer plan. Calculate the water demand process and its probability of occurrence in the water receiving area under different water inflow scenarios and use it as the water demand private information of each water receiving area.

[0053] Step S2: Treat the different water-receiving areas of the water diversion project as decision-making entities, collect annual water use plan suggestions for each time period of the water-receiving areas and treat them as strategies, and generate a water use plan strategy set for each water-receiving area.

[0054] Step S3: Use the shared water information, private water demand information, and water use planning strategy sets of each water-receiving area as inputs to the pre-constructed cooperative-competitive Bayesian game model of water use planning in the water-receiving area. Solve the model to obtain the optimal value of the group payoff in the cooperative part of the model and the optimal value of the individual relative advantage payoff of each water-receiving area.

[0055] Step S4: Based on the optimal value of group benefits and the optimal value of individual relative advantage benefits, calculate the CO-CO (cooperation-competition) value of the water use planning game for each water-receiving area, and solve the problem by back-calculating based on the CO-CO value of the water use planning game for each water-receiving area, and finally determine the water use planning decision for each water-receiving area.

[0056] like Figure 2 As shown, according to one aspect of this application, step S1 further comprises:

[0057] Step S11: Collect data on the water source area and water receiving area of ​​the water transfer project, including: meteorological and hydrological forecast data, regional population, economic and social data, water use efficiency data, water supply capacity data of the water transfer project, and regional water use management policies, and unify the time dimension of the collected data and synchronize them.

[0058] Step S12: Based on regional population, economic and social data, water supply capacity data of water diversion projects and regional water management policies, determine the ecological base flow guarantee threshold and water demand of the water source area, define the calculation boundary of the divertable water volume, and calculate the water demand, total available water resources and divertable water volume of the water source area respectively.

[0059] Step S13: Collect historical rainfall data and combine it with meteorological and hydrological forecast data to divide several water inflow scenarios, and calculate the local water inflow and water distribution of the water-receiving area under each water inflow scenario. Use the water diversion capacity of the water source area and the local water inflow and water distribution of the water-receiving area under each water inflow scenario as water inflow sharing information.

[0060] Step S14: Each water-receiving area calculates the water demand process and its corresponding probability of occurrence based on its own historical water use and water supply plan data, population, economic and social data, water use efficiency data, regional water use management policy data, and in conjunction with the water inflow scenario, and uses this as the water demand private information of each water-receiving area.

[0061] In this embodiment, the data collection and preparation in step S1 mainly includes: meteorological and hydrological forecast data, including rainfall, temperature, flow rate, etc.; historical data on domestic, industrial, agricultural and ecological water use in the water-receiving area and planned water supply data for each year; regional population, economic and social data, including resident population, GDP per capita, industrial added value, irrigation system and planting structure; water use efficiency data, including per capita domestic water consumption, industrial added value water consumption, irrigation water utilization coefficient, etc.; regional water use management policies, including total water consumption control indicators, water price, water conservation indicators and policies; and data related to the water supply capacity of water diversion projects.

[0062] Among them, the analysis based on meteorological and hydrological forecasts, total water consumption control indicators, and engineering water supply capacity yields the predicted water volume of the source area and the water inflow of the receiving area. This data is regarded as the water inflow shared information in the annual water transfer plan formulation process, meaning that all receiving areas can obtain information on water diversion from the source area, water inflow from their own area, and water inflow from other areas. The analysis based on regional historical water consumption and water supply plan data, population, economic and social data, water use efficiency, and management policies yields the water demand process and its probability of occurrence under different water inflow scenarios in the receiving area. This data is regarded as the water demand private information in the water transfer plan formulation process, meaning that each receiving area only has access to its own water demand information.

[0063] According to one aspect of this application, step S14 further comprises:

[0064] Step S14a: Calculate the total water demand of the water-receiving area and the water demand for domestic, industrial, agricultural and ecological purposes under each water inflow scenario according to the four major water use categories of domestic, industrial and agricultural and ecological purposes. Plot the water demand curves under different water inflow scenarios with time as the horizontal axis and water demand as the vertical axis to obtain the water demand process.

[0065] Step S14b: Collect historical rainfall data of the water-receiving area, count the number of years corresponding to different rainfall frequencies, and calculate the initial probability of each water inflow scenario;

[0066] Step S14c: Based on the reliability of the weather forecast, the initial probabilities of each water inflow scenario are corrected and normalized to ensure that the sum of the probabilities of all water inflow scenarios is 1, and the probability of occurrence of each water inflow scenario is obtained.

[0067] Step S14d: Determine the water demand process and its probability of occurrence data under different water inflow scenarios in the water-receiving area as private water demand information.

[0068] like Figure 3 As shown, according to one aspect of this application, step S2 further comprises:

[0069] Step S21: Identify each water-receiving area covered by the water diversion project as an independent decision-making entity, and define the geographical scope, water use control authority, and responsibility boundaries of each entity;

[0070] Step S22: Based on the private water demand information and the shared water inflow information of each water-receiving area, generate a finite set of water use planning alternative strategies, which includes annual water consumption at different levels and monthly water use allocation, i.e., the water use planning strategy set.

[0071] In this embodiment, the water use planning strategy set for the water-receiving area in step S2, that is, a finite set of water use planning candidate strategies for the water-receiving area that includes annual water consumption at different levels and monthly water allocation, can be represented as:

[0072] ;

[0073] In the formula, S A S B These are water use planning strategy sets for water-receiving area A and water-receiving area B, respectively, containing M and L water use planning strategies; taking water-receiving area A as an example. For S A The i-th water usage plan strategy, },in This is the planned water consumption vector for the h-th month in the i-th water use planning strategy for the water-receiving area, including domestic water consumption, industrial water consumption, agricultural irrigation water consumption, and ecological water consumption (unit: 10). 4 m 3 The four dimensions can be represented as ;

[0074] Water use planning in water-receiving areas often requires reference to historical water consumption records. Currently, some newly constructed inter-basin water transfer projects have been recently completed and put into operation, resulting in a lack of historical water use and water supply plan data. On the other hand, water use data from multiple sectors, including domestic, industrial, agricultural, and ecological sectors, are complex and multi-dimensional, leading to high computational complexity in model calculations. To address these issues, this invention proposes a lightweight, small-sample augmented deep learning-driven improved scenario tree method for generating water use planning strategy sets for water-receiving areas. This method effectively solves the problems of limited and complex historical water use data in water transfer project receiving areas. Through data augmentation, it can still generate high-quality data and models with strong generalization capabilities under small-sample conditions, avoiding scheduling strategy deviations due to insufficient data. At the same time, the lightweight model reduces the demand for high-performance computing power, facilitating on-site deployment in engineering projects.

[0075] In one embodiment, the specific steps include:

[0076] Small sample data enhancement preprocessing: First, collect 3-5 year monthly historical data of the water diversion project's receiving area, including domestic, industrial, agricultural, and ecological water use data, as well as driving factor data (rainfall, temperature, resident population, industrial output, monthly water supply ceiling of the water diversion project, etc.); second, clean the data, i.e., fill missing values ​​with linear interpolation and annual average values ​​of the same climate, and correct outliers using the isolated forest algorithm; third, enhance the data, i.e., perturb the original monthly data by ±5% based on the water use data fluctuation threshold to generate perturbed samples with 3 times the original amount, and use a Generative Lightweight Adversarial Network (CGAN) model (simplifying the network layer to 3 layers) to generate virtual samples with 2 times the original amount that conform to seasonal patterns, based on the month and meteorological factors; finally, select time series / seasonal / driving features with Pearson correlation coefficient ≥0.2, and output standardized training data after Z-score standardization;

[0077] Lightweight Temporal Model Construction and Training: For the basic model, a hybrid prediction model was constructed using a MobileNetV3 backbone combined with temporal fully connected layers. MobileNetV3 was used to extract high-dimensional features of driving factors, and a 0.75x width coefficient was used to simplify network parameters. The temporal fully connected layers consisted of only two hidden layers (with the number of neurons decreasing gradually from 64 to 32) to handle the monthly correlations in water usage data. The total number of parameters in the entire model was kept below 500,000, only 1 / 10 of the traditional LSTM model. For the training strategy, MAML meta-learning was adopted. Task sets were constructed by month, with each task containing three support sets for similar months and one query set for the target month. Meta-training was first performed on all tasks (learning rate 0). The model was fine-tuned using small sample data from the target water-receiving area (learning rate 0.0001, iterations 50 times) to understand the general patterns of water use in different months. The MobileNetV3 backbone layers were frozen, and only the temporal fully connected layers were trained to accurately adapt to the water use characteristics of the target area. For model quantization optimization, QAT quantization-aware training technology was used. In the PyTorch framework, the torch.quantization tool was used to convert the model weights from 32-bit floating-point numbers to 8-bit integers. A precision compensation mechanism was added during the quantization process to ensure that the model's prediction accuracy loss after quantization was ≤3%. After optimization, the model's inference speed was increased by 3 times, and the storage volume was reduced by 75%, meeting the deployment requirements of engineering projects.

[0078] Deep learning-driven hierarchical construction of scene trees: First, the scene tree hierarchy and prediction dimensions are defined. The scene tree is set with months as the core hierarchy, with a total of 12 layers. The root node is the initial water use status of the water-receiving area at the beginning of the year. Each layer node is input with the water use data of the previous month and the meteorological prediction factor of the current month, and outputs the conditional probability distribution of four types of water use, including key parameters such as mean and variance, to provide a quantitative basis for branch division. Second, scene branches are divided. For each type of water use conditional probability distribution output by the model, it is divided into three scene levels: low, medium and high according to the 25th, 50th and 75th percentiles. Combining the combination relationship of the four types of water use levels, only one type of water use is allowed to change in low / medium / high at each time, while the other types are fixed at the median. Each layer generates 3×4=12 basic branches (such as low for residential use, medium for industrial use, high for agricultural use, and medium for ecological use), covering multi-dimensional water use scenarios. Third, branch probability is assigned. The joint probability of the four types of water use output by the model is directly used as the branch weight to ensure that the sum of all branch weights is 1 and completely conforms to the actual water use pattern of the water-receiving area, avoiding the bias risk of subjective assignment in traditional scene trees.

[0079] Engineering constraint embedding and scenario tree pruning: First, the hard constraints of the water diversion project are quantitatively defined, including annual total water consumption control targets, water balance constraints, project water supply capacity constraints, ecological release constraints, water supply priority constraints, and project operation stability constraints, and are transformed into mathematical expressions to ensure that scenario branches conform to the actual project scheduling. Second, branch selection and pruning: First, the 12 basic branches of the initialized water use probability are constrained and verified one by one, and invalid branches that violate any constraint are eliminated. After pruning, if the number of effective branches at a certain level is less than 3, the trained lightweight model is called to supplement and generate suboptimal probability branches, that is, reasonable combination scenarios ranked 13th to 18th in the model's predicted probability are selected to ensure that at least 3 effective branches are retained at each level, taking into account the feasibility and diversity of the scenario tree. Finally, a scenario tree containing 12 months of water use plans is formed, namely the monthly water use plan strategy set for the water-receiving area.

[0080] like Figure 4 As shown, according to one aspect of this application, step S3 further comprises:

[0081] Step S31: Select the most important online regulating reservoir along the entire water diversion project. Using this reservoir as a node, the water diversion project system is generalized into a water supply network system consisting of the regulating reservoir, the water receiving area above the reservoir, and the water receiving area below the reservoir (including the direct supply area of ​​the reservoir).

[0082] All water-receiving areas have needs for domestic, industrial, irrigation, and ecological water use;

[0083] Step S32: Construct a cooperative-competitive Bayesian game model of water use planning for water-receiving areas in the generalized system. ;

[0084] Cooperative-competitive Bayesian game model for water use planning in water-receiving areas The definition is as follows:

[0085] ;

[0086] Where N is the set of water-receiving areas participating in the decision-making process (N = {1, 2}); for water-receiving area i, S i The representative decision set, in this case, is the water use planning strategy set for the water-receiving area; Y i The representative signal set is the water demand forecast information obtained by the water-receiving area for this region; U i The set of revenue functions represents the comprehensive water transfer revenue obtained by the water-receiving area after adopting a certain water use planning strategy; μ is the prior probability distribution on Y×U.

[0087] Step S33: Perform CO-CO decomposition on the Bayesian game of water use plans for the water-receiving areas, including two parts: cooperation and competition. The cooperation part stipulates that the water-receiving areas cooperate completely, share information, and share benefits equally. Each water-receiving area receives an average group benefit, expressed as u_{A}^{eq}(s)=u_{B}^{eq}(s)≡(u_{A}(s)+u_{B}(s)) / 2, where A and B represent the water-receiving areas above and below the reservoir, respectively, and s represents a combination of water use plans. The competition part stipulates that the water-receiving areas do not share information, engaging in a zero-sum game, i.e., maximizing the difference between the area's benefit and that of other areas. Each water-receiving area receives a relatively advantageous individual benefit, expressed as:

[0088] u_{A}^{ad}(s)≡(u_{A}(s)-u_{B}(s)) / 2;

[0089] u_{B}^{ad}(s)≡(u_{B}(s)-u_{A}(s)) / 2;

[0090] Step S34: Convert the objective benefits of the water supply security, water use efficiency, water diversion utilization rate and water supply fairness game of the water receiving area into transferable utility, and set the comprehensive benefit function of the water receiving area based on this.

[0091] The annual water allocation plan for a water diversion project is formulated based on meeting the overall water resource allocation requirements of the basin, the water resource allocation conditions of the project, and ensuring the safe operation of the project. It takes into account the water storage status of reservoirs along the route, the predicted inflow of water in the source and receiving areas, and the water use plans proposed by each receiving area. Utilizing the regulation and storage capacity of reservoirs along the project route, a time-period (monthly, ten-day) water supply plan is developed. The water supply benefits obtained by each receiving area are affected by its own proposed water use plans, and are also related to the water use plans of other receiving areas. If we consider the receiving areas as decision-making entities representing different regional interests, then the process of each receiving area proposing its water use plan involves a game theory relationship. Due to the uncertainty of incoming water and certain information barriers between regions, this game takes place under conditions of incomplete information. Therefore, the water use plan game of the receiving areas should be described as a Bayesian game involving different decision-making entities.

[0092] The water use planning game in the water-receiving areas has both competitive and cooperative attributes. On the one hand, each water-receiving area pursues the maximization of water supply efficiency and water supply security rate, resulting in competition for water diversion volume. On the other hand, the water-receiving areas share common demands in terms of improving water diversion utilization rate and ensuring water supply fairness, thus providing a basis for cooperation. Based on this, the Bayesian game of water use planning in the water-receiving areas is decomposed into a CO-CO (cooperation-competition) decomposition, dividing it into two units: a cooperative part involving group benefits and a competitive part involving individual relative advantage benefits.

[0093] Taking the water-receiving area above the reservoir as an example, the game payoff of the water-receiving area mainly consists of the following objectives:

[0094] The water supply security target, represented by the water supply / irrigation guarantee rate, can be maximized as follows:

[0095] ;

[0096] In the formula, T s T represents the total number of time periods during which the water supply for domestic, industrial, or agricultural irrigation can meet the water demand within the statistical period.

[0097] The water use efficiency target, expressed as the benefits of power generation and water supply minus the costs of diverting water from the main canal of the project, can be maximized as: maxf A，2 =Σw k E k -GS A θ A ;

[0098] In the formula, E k For the k economic benefits obtained by the water-receiving area through water diversion projects, w k Σw is the weight of this benefit. k =1; GS A The actual water supply to the receiving area (10 4 m 3 ), θ A Cost coefficient for water diversion to the water receiving area (yuan / m³) 3 );

[0099] The target for water diversion utilization rate, i.e., the proportion of actual water use in the water-receiving area to the actual water diversion, is to be maximized and can be expressed as: maxf A，3 =XS A / C A ;

[0100] In the formula, XS A The actual water demand of the water-receiving area (10 4 m 3 ), C A Water diversion volume for the receiving area (10 4 m 3 ).

[0101] The equity objective, represented by the difference in water supply guarantee levels between the districts and other districts, can be minimized as follows:

[0102] minf A，4 =∣f A，1 -f B，1 |;

[0103] In the formula, fA，1 f B，1 These are the water supply / irrigation guarantee rates for the upstream and downstream water-receiving areas, respectively.

[0104] According to one aspect of this application, step S33 further comprises:

[0105] Step S33a: Each water-receiving area independently and randomly selects any water use planning strategy from its own water use planning strategy set to form a combination of water use planning strategies for the water-receiving area.

[0106] Step S33b: Input the current water use planning strategy combination, annual available water volume, water inflow data of the water receiving area, predicted water demand process and its probability distribution information into the annual water volume scheduling plan compilation model of the water diversion project. The annual water volume scheduling plan compilation model of the water diversion project simulates and calculates the water storage and supply volume of each time period under different water demand processes in each water receiving area, and then obtains the expected value of the total comprehensive benefit of the water receiving area.

[0107] Step S33c: Repeat the process of randomly selecting water use planning strategies for the water-receiving area. After traversing all combinations of water use planning strategies, compare the expected values ​​of the total comprehensive benefits, and select the maximum value to obtain the optimal group benefit value for the cooperative part of the model. The formula is as follows:

[0108] ;

[0109] In the formula, c:Y→S represents the coordinated water use plan strategy for all water-receiving areas in this cooperative game.

[0110] In this embodiment, the annual water allocation plan model for the water transfer project refers to taking the predicted water inflow for each period of the year, the water use plan suggestions for each water-receiving area, and the initial water storage of the project at each period as inputs, and the water allocation scheme and total water use control indicators as constraints. After setting the water diversion rules for the reservoirs in the water source area and the scheduling rules for the online regulating reservoirs involved in the project, the water supply plan for each period of the year and its allocation among the water-receiving areas are determined based on the reservoir scheduling rules. Inter-basin water transfer projects often involve multiple reservoirs or lakes along the route. Depending on the different regulation and storage functions of the reservoirs, they can be divided into online regulating reservoirs. The water diversion system includes compensation regulation and storage regulation. Among these, the online regulating reservoir is an important node for water allocation in water diversion projects. It can coordinate the water supply of water-using units upstream and downstream of the reservoir. The scheduling rule for this type of reservoir is that when the available water volume is abundant and the water level is lower than the limiting water diversion line, water is drawn from the main canal to fill the reservoir. When the water diversion is insufficient, water is supplied to the main canal. The reservoir scheduling map sets 6 limiting water supply lines, namely the limiting water supply line for agriculture, domestic and industrial use upstream of the reservoir, the limiting water supply line for agriculture, domestic and industrial use downstream of the reservoir, and the limiting water supply line for agriculture, domestic and industrial use in the direct supply area of ​​the reservoir. At the same time, a limiting water diversion line is set to indicate the water diversion rules of the reservoir.

[0111] According to one aspect of this application, step S33 further comprises:

[0112] Step S33d: Set the initial selection probability of different water use planning strategies for the water-receiving areas to be equal, calculate the location potential energy factor of each water-receiving area, carry out the first round of game, traverse all combinations of water use planning strategies, and calculate the expected value of the individual relative advantage of the water-receiving areas under different water demand processes based on the annual water volume scheduling plan compilation model of the water diversion project.

[0113] Step S33e: Based on the information obtained during the game, each water-receiving area uses a spatiotemporally aware five-dimensional adaptive heterogeneity coupled replication dynamic equation to update its beliefs. The five dimensions include information receptivity, strategy response strength, risk preference coefficient, location potential factor, and temporal memory factor. The information receptivity and strategy response strength are dynamically adjusted with each round of the game, and the temporal memory factor is updated cumulatively based on the changes in the payoffs of previous games.

[0114] The spatiotemporal-aware five-dimensional adaptive heterogeneity coupling replication dynamic equation is:

[0115] x_(i,k)^(t+1)=x_(i,k)^(t)+λ_i^(t)·η_i^(t)·φ_i·x_(i,k)^(t)·(1-x_(i,k)^(t))·(û_(i,k)^(t)-ū_i^(t)+μ*·ψ_i^(t));

[0116] In the formula, x_(i,k)^(t) is the probability that the water-receiving area i chooses the k-th water use plan strategy in the t-th round of the game, x_(i,k)^(t+1) is the updated strategy selection probability, λ_i^(t) is the dynamically adjusted information receptivity, η_i^(t) is the dynamically adjusted strategy response strength, φ_i is the position potential factor, û_(i,k)^(t) is the expected value of the payoff of choosing strategy k after risk preference correction, ū_i^(t) is the average expected value of the payoff of the water-receiving area i under the current strategy probability distribution, μ* is the temporal memory influence coefficient, ψ_i^(t) is the temporal memory factor, and factor x_(i,k)^(t)·(1-x_(i,k)^(t) is the Logistic damping term;

[0117] The formula for calculating the position potential energy factor φᵢ is as follows:

[0118] φ_i=ω1(C_i^cap / (C_i^cap+C_j^cap))+ω2·α_i^pos;

[0119] In the formula, C_i^cap is the engineering design water supply capacity of water receiving area i, C_j^cap is the engineering design water supply capacity of another water receiving area j, α_i^pos is the location correction coefficient, with values ​​of 1.1 and 0.9 for the reservoir area, and ω1 and ω2 are weighting coefficients that satisfy ω1+ω2=1.

[0120] The recursive formula for the temporal memory factor ψ_i^(t) is:

[0121] ψ_i^(t)=β·ψ_i^(t−1)+Δu_i^(t−1);

[0122] In the formula, β is the memory decay coefficient, with a value range of (0, 1), Δu_i^(t−1) is the change in payoff in the (t-1)th round relative to the (t-2)th round, and the initial condition is ψ_i^(1)=0;

[0123] The dynamic adjustment formula for the information acceptability λ_i^(t) is:

[0124] λ_i^(t)=λ_i^(0)·(1+γ·ln(1+t))^(-1);

[0125] In the formula, λ_i^(0) is the initial information acceptance level, γ is the attenuation adjustment coefficient, and t is the current game round;

[0126] The dynamic adjustment formula for the strategy response intensity η_i^(t) is:

[0127] η_i^(t)=η_i^(0)·exp(-δ·σ_i^(t));

[0128] In the formula, η_i^(0) is the initial response strength, δ is the volatility sensitivity coefficient, and σ_i^(t) is the return volatility calculated based on the sliding window.

[0129] The formula for calculating the risk preference-adjusted return û_(i,k)^(t) is as follows:

[0130] û_(i,k)^(t)=E[u_(i,k)]-ρ_i·√(Var[u_(i,k)]);

[0131] In the formula, E[u_(i,k)] is the expected return of strategy k under all possible water demand scenarios, Var[u_(i,k)] is the variance of the return of strategy k, and ρ_i is the risk aversion coefficient of water-receiving area i.

[0132] Step S33f: Iterate continuously. After each iteration, the probability of strategy selection is subject to boundary constraints and normalization, and the update magnitude is controlled by adaptive learning rate decay. When all water-receiving areas meet the multiple convergence criteria of strategy probability stability, dominant strategy significance, and continuous stable rounds, it is considered that the Nash equilibrium of the Bayesian game has been achieved. At this time, the optimal value of the individual relative advantage payoff of the water-receiving areas is obtained.

[0133] The formula for handling the boundary constraints is as follows:

[0134] x_(i,k)^(t+1)=max(ε,x_(i,k)^(t+1));

[0135] In the formula, ε is the lower bound of the probability, and its value ranges from 0.001 to 0.01;

[0136] The normalization formula is as follows:

[0137] x_(i,k)^(t+1)←x_(i,k)^(t+1) / ∑(k'=1)^(K_i)x_(i,k)^(t+1);

[0138] In the formula, K_i is the total number of water use planning strategies for water-receiving area i;

[0139] The formula for adaptive learning rate decay is:

[0140] k^(t)=k0·(1+t / T_half)^(-1);

[0141] In the formula, k0 is the initial learning rate, and T_half is the half-life round;

[0142] The multiple convergence criteria include: Condition 1 is the stability of the strategy probability, i.e., max|x_(i,k)^(t+1)-x_(i,k)^(t)|<θ_x holds for all regions, where θ_x is the probability change threshold; Condition 2 is the significance of the dominant strategy, i.e., maxx_(i,k)^(t)>θ_dom holds for all regions, where θ_dom is the dominant strategy threshold; Condition 3 is the number of consecutive stable rounds, i.e., the above conditions must be satisfied continuously for N rounds. staβ1e wheel.

[0143] In the competitive game part of the model, apart from shared information, each water-receiving area possesses private information about its own water demand process and its probability of occurrence. They do not have relevant information about other areas, meaning they only hold initial prior beliefs. Therefore, at this point, the water-receiving areas set the initial selection probability of different water use planning strategies to be equal and participate in the game. After the first round of the game, by observing the strategies and payoffs of other water-receiving areas, the water-receiving areas update their beliefs based on the newly acquired information, thereby adjusting the probability of selecting water use planning strategies and participating in the next round of the game.

[0144] The aforementioned process of updating beliefs and adjusting water use plans in the water-receiving areas is achieved through the replication dynamic equation of evolutionary game theory. In this embodiment, considering the heterogeneity of different water-receiving areas in terms of new information acceptance, strategy update rules, and risk preferences, an improved three-dimensional heterogeneity coupled replication dynamic equation is proposed, as follows:

[0145] ;

[0146] In the formula, , λ represents the probabilities of the upstream and downstream water-receiving areas choosing the i-th and j-th water use planning strategies after belief updates, respectively; A , λ B The receptiveness of the upstream and downstream water-receiving areas to new information, respectively; η A η B The response strength to the difference in returns when different strategies are used to update the rules for the upper and lower water-receiving areas of the reservoir, respectively. For example, the response is strong (η=1) when the best imitation rule is used, medium (η=0.5) when the near-optimal rule is used, and weak (η=0.2) when the random imitation rule is used. , Let P be the risk-preference adjusted returns for selecting the i-th and j-th water use planning strategies for the upstream and downstream water-receiving areas, respectively. A The relevant P-th water use planning strategy, Q is B The relevant Q-th water use planning strategy, taking the reservoir's water-receiving area as an example, includes:

[0147] ;

[0148] Where ρ A This represents the risk preference coefficient for the water-receiving area.

[0149] ;

[0150] .

[0151] like Figure 5 As shown, according to one aspect of this application, step S4 further comprises:

[0152] S41. Based on the optimal value of group payoff and the optimal value of individual relative advantage payoff, calculate the CO-CO (cooperation-competition) value of the cooperative-competitive Bayesian game for water use planning in the water-receiving areas above and below the reservoir, as shown in the following formula:

[0153] ;

[0154] ;

[0155] In the formula, G ad = (S, Y, V, μ) ad ( ) represents a zero-sum game played across the water-receiving intervals, minmax A (G) ad ), minmax B (G) ad The numbers () and () represent the optimal relative advantage payoffs for individuals in the upper and lower water-receiving areas of the reservoir, respectively, in a competitive game. V is the core set describing the achievable payoffs under all participants' strategy combinations, and μ is the core set describing the achievable payoffs under all participants' strategy combinations. ad It is a function defined on the strategy space S and the strategy combination, and is the core tool for mapping the strategy combination of participants to specific payout values;

[0156] S42. Using the CO-CO value as the target value for optimizing the water transfer revenue of the water-receiving area, and using the planned water consumption of each month of the year in the water-receiving area as the decision variable, establish a back-calculation solution model for the water consumption plan of the water-receiving area.

[0157] S43. Input the available water volume and the water inflow data of the receiving area, and input different predicted water demand processes under the current water inflow forecast level. Use a multi-swarm particle swarm intelligent optimization algorithm to finally solve for the water use plan decision of the receiving area corresponding to various water demand processes.

[0158] According to another aspect of this application, a cooperative-competitive game decision-making system for water diversion planning with incomplete information is provided, characterized by comprising:

[0159] At least one processor; and

[0160] A memory communicatively connected to at least one of the processors; wherein,

[0161] The memory stores instructions that can be executed by the processor to implement the cooperative-competitive game decision-making method for incomplete information water diversion projects as described above.

[0162] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A cooperative-competitive game decision-making method for water diversion projects with incomplete information, characterized in that: Includes the following steps: Step S1: Collect data on the water source area and water receiving area of ​​the water transfer project, calculate the water diversion volume available in the water source area and the predicted water inflow in the water receiving area, and use it as the water inflow sharing information in the process of formulating the annual water transfer plan. Calculate the water demand process and its probability of occurrence in the water receiving area under different water inflow scenarios and use it as the water demand private information of each water receiving area. Step S2: Treat the different water-receiving areas of the water diversion project as decision-making entities, collect annual water use plan suggestions for each time period of the water-receiving areas and treat them as strategies, and generate a water use plan strategy set for each water-receiving area. Step S3: Use the shared water information, private water demand information, and water use planning strategy sets of each water-receiving area as inputs to the pre-constructed cooperative-competitive Bayesian game model of water use planning in the water-receiving area. Solve the model to obtain the optimal value of the group payoff in the cooperative part of the model and the optimal value of the individual relative advantage payoff of each water-receiving area. Step S4: Based on the optimal value of group benefits and the optimal value of individual relative advantage benefits, calculate the CO-CO (cooperation-competition) value of the water use planning game for each water-receiving area, and solve the problem by back-calculating based on the CO-CO value of the water use planning game for each water-receiving area, and finally determine the water use planning decision for each water-receiving area.

2. The cooperative-competitive game decision-making method for water diversion planning with incomplete information as described in claim 1, characterized in that, Step S1 further comprises: Step S11: Collect data on the water source area and water receiving area of ​​the water transfer project, including: meteorological and hydrological forecast data, regional population, economic and social data, water use efficiency data, water supply capacity data of the water transfer project, and regional water use management policies, and unify the time dimension of the collected data and synchronize them. Step S12: Based on regional population, economic and social data, water supply capacity data of water diversion projects and regional water management policies, determine the ecological base flow guarantee threshold and water demand of the water source area, define the calculation boundary of the divertable water volume, and calculate the water demand, total available water resources and divertable water volume of the water source area respectively. Step S13: Collect historical rainfall data and combine it with meteorological and hydrological forecast data to divide several water inflow scenarios, and calculate the local water inflow and water distribution of the water-receiving area under each water inflow scenario. Use the water diversion capacity of the water source area and the local water inflow and water distribution of the water-receiving area under each water inflow scenario as water inflow sharing information. Step S14: Each water-receiving area calculates the water demand process and its corresponding probability of occurrence based on its own historical water use and water supply plan data, population, economic and social data, water use efficiency data, regional water use management policy data, and in conjunction with the water inflow scenario, and uses this as the water demand private information of each water-receiving area.

3. The cooperative-competitive game decision-making method for water diversion planning with incomplete information as described in claim 2, characterized in that, Step S14 further comprises: Step S14a: Calculate the total water demand of the water-receiving area and the water demand for domestic, industrial, agricultural and ecological purposes under each water inflow scenario according to the four major water use categories of domestic, industrial and agricultural and ecological purposes. Plot the water demand curves under different water inflow scenarios with time as the horizontal axis and water demand as the vertical axis to obtain the water demand process. Step S14b: Collect historical rainfall data of the water-receiving area, count the number of years corresponding to different rainfall frequencies, and calculate the initial probability of each water inflow scenario; Step S14c: Based on the reliability of the weather forecast, the initial probabilities of each water inflow scenario are corrected and normalized to ensure that the sum of the probabilities of all water inflow scenarios is 1, and the probability of occurrence of each water inflow scenario is obtained. Step S14d: Determine the water demand process and its probability of occurrence data under different water inflow scenarios in the water-receiving area as private water demand information.

4. The cooperative-competitive game decision-making method for water diversion planning with incomplete information as described in claim 1, characterized in that, Step S2 further comprises: Step S21: Identify each water-receiving area covered by the water diversion project as an independent decision-making entity, and define the geographical scope, water use control authority, and responsibility boundaries of each entity; Step S22: Based on the private water demand information and the shared water inflow information of each water-receiving area, generate a finite set of water use planning alternative strategies, which includes annual water consumption at different levels and monthly water use allocation, i.e., the water use planning strategy set.

5. The cooperative-competitive game decision-making method for water diversion planning with incomplete information as described in claim 1, characterized in that, Step S3 further comprises: Step S31: Select the most important online regulating reservoir along the entire water diversion project. Using this reservoir as a node, the water diversion project system is generalized into a water supply network system consisting of the regulating reservoir, the water receiving area above the reservoir, and the water receiving area below the reservoir. Step S32: Construct a cooperative-competitive Bayesian game model of water use planning for water-receiving areas in the generalized system. Step S33: Perform CO-CO decomposition on the Bayesian game of water use plan for the water-receiving areas, which includes two parts: cooperation and competition. The rules for the cooperation part are that the water-receiving areas cooperate completely, share information, and share benefits equally, with each water-receiving area receiving an average group benefit. The rules for the competition part are that the water-receiving areas do not share information and engage in a zero-sum game, that is, they pursue the maximum difference between the benefits of their own area and those of other areas, with each water-receiving area receiving a relative advantage benefit. Step S34: Convert the water supply guarantee rate, water use efficiency, water diversion utilization rate and water supply fairness game objective of the water receiving area into transferable utility, and set the comprehensive benefit function of the water receiving area based on this.

6. The cooperative-competitive game decision-making method for incomplete information water diversion plans as described in claim 5, characterized in that, Step S33 further comprises: Step S33a: Each water-receiving area independently and randomly selects any water use planning strategy from its own water use planning strategy set to form a combination of water use planning strategies for the water-receiving area. Step S33b: Input the current water use planning strategy combination, annual available water volume, water inflow data of the water receiving area, predicted water demand process and its probability distribution information into the annual water volume scheduling plan compilation model of the water diversion project. The annual water volume scheduling plan compilation model of the water diversion project simulates and calculates the water storage and supply volume of each time period under different water demand processes in each water receiving area, and then obtains the expected value of the total comprehensive benefit of the water receiving area. Step S33c: Repeat the process of randomly selecting water use planning strategies for the water-receiving area. After traversing all combinations of water use planning strategies, compare the expected value of the total comprehensive benefits and select the maximum value to obtain the optimal value of the group benefits for the cooperative part of the model.

7. The cooperative-competitive game decision-making method for water diversion planning with incomplete information as described in claim 5, characterized in that, Step S33 further comprises: Step S33d: Set the initial selection probability of different water use planning strategies for the water-receiving areas to be equal, calculate the location potential energy factor of each water-receiving area, carry out the first round of game, traverse all combinations of water use planning strategies, and calculate the expected value of the individual relative advantage of the water-receiving areas under different water demand processes based on the annual water volume scheduling plan compilation model of the water diversion project. Step S33e: Based on the information obtained during the game, each water-receiving area uses a spatiotemporally aware five-dimensional adaptive heterogeneity coupled replication dynamic equation to update its beliefs. The five dimensions include information receptivity, strategy response strength, risk preference coefficient, location potential factor, and temporal memory factor. The information receptivity and strategy response strength are dynamically adjusted with each round of the game, and the temporal memory factor is updated cumulatively based on the changes in the payoffs of previous games. Step S33f: Iterate continuously. After each iteration, the probability of strategy selection is subject to boundary constraints and normalization, and the update magnitude is controlled by adaptive learning rate decay. When all water-receiving areas meet the multiple convergence criteria of strategy probability stability, dominant strategy significance, and continuous stable rounds, it is considered that the Nash equilibrium of the Bayesian game has been achieved. At this time, the optimal value of the individual relative advantage payoff of the water-receiving areas is obtained.

8. The cooperative-competitive game decision-making method for water diversion planning with incomplete information as described in claim 1, characterized in that, Step S4 further comprises: S41. Based on the optimal value of group payoff and the optimal value of individual relative advantage payoff, calculate the CO-CO (cooperation-competition) value of the Bayesian game of cooperation-competition for water use planning in the water-receiving areas above and below the reservoir. S42. Using the CO-CO value as the target value for optimizing the water transfer revenue of the water-receiving area, and using the planned water consumption of each month of the year in the water-receiving area as the decision variable, construct a back-calculation solution model for the water consumption plan of the water-receiving area. S43. Input the available water volume, water inflow data of the water receiving area, and different predicted water demand processes under the current water inflow forecast level into the back calculation solution model of the water use plan of the water receiving area. Use a multi-group particle swarm intelligent optimization algorithm to solve the model and obtain the water use plan decision of the water receiving area corresponding to various water demand processes.

9. A cooperative-competitive game decision-making system for water diversion planning with incomplete information, characterized in that: include: At least one processor; as well as A memory communicatively connected to at least one of the processors; wherein, The memory stores instructions that can be executed by the processor to implement the incomplete information down-water diversion plan cooperative-competitive game decision-making method according to any one of claims 1 to 8.

Images