Water network system water resources-ecological environment-economic and social coordination method and system

Abstract

The present invention discloses a collaborative method and system for water resources - ecological environment - economic society in a water network system. The marginal distributions of rainfall and evapotranspiration are fitted and a joint distribution is constructed. Massive scenarios are randomly sampled and input into a pre - constructed machine learning model to obtain the water resource quantity. Historical population and historical GDP data of the study area are collected to construct an economic and social volume prediction model and an economic and social volume prediction sample. A multi - objective optimization model is constructed to calculate the collaborative indicators of water resources - ecological environment - economic society. A system dynamics model is constructed and its model parameters are calibrated to calculate the values of the collaborative indicators of water resources - ecological environment - economic society. The collaborative indicators of water resources - ecological environment - economic society are dimensionally reduced and the corresponding indicator values are calculated, and a regulation plan is constructed to screen out the optimal optimization plan. The present invention improves the collaborative operation efficiency of the water conservancy model, realizes the comprehensive simulation of the water conservancy system, enhances the intelligent level of water network scheduling, and realizes the deep interactive integration of the physical water network and the digital water network.

Description

Technical Field

[0001] The present invention relates to a collaborative method for water resources - ecological environment - economic society in a water network system. Background Art

[0002] A water network system is a comprehensive system that is based on natural rivers and lakes, uses water diversion, drainage and regulation projects as channels, regulation and storage projects as nodes, and intelligent regulation as a means, integrating functions such as optimal allocation of water resources, basin flood control and disaster reduction, and protection of water ecological systems. It is an effective measure to solve the uneven spatial distribution of water resources, improve the water resource guarantee rate in the water receiving area, alleviate the contradiction between water supply and demand in water - shortage areas, and achieve the rational allocation of water resources. It is an important way to promote the economic development and comprehensive development and utilization of water resources in water - shortage areas.

[0003] The collaboration of water resources - ecological environment - economic society is the primary goal that the construction, management and operation of the water network system need to achieve. However, the relationships among the elements restricting the collaboration of water resources - ecological environment - economic society in the water network system are relatively complex. At present, the research on water network subsystems and their pairwise relationships has been continuously deepened, but the causal chain of water resources - ecological environment - economic society in the water network system is not yet clear, the direct and indirect impacts are intertwined, it is difficult to conduct a comprehensive evaluation of water resources - ecological environment - economic society, and at the same time, there is no mature multi - element collaboration technology for the entire water network system. The mutual feedback process among the water balance elements has not been analyzed from the overall water network system, and a collaborative plan for water resources - ecological environment - economic society has not been proposed.

[0004] The present invention proposes a collaborative method and system for water resources - ecological environment - economic society in a water network system to solve the above - mentioned existing problems, improve the collaborative operation efficiency of water conservancy models, achieve the comprehensive simulation of water conservancy systems, enhance the intelligent level of water network dispatching, realize the deep interactive integration of physical water networks and digital water networks, promote the gradual development of various water conservancy projects from points to networks and from dispersion to systems, and comprehensively improve the systematicness, comprehensiveness and strong toughness of the water network. Summary of the Invention

[0005] Object of the Invention: To provide a collaborative method for water resources - ecological environment - economic society in a water network system to solve the above - mentioned existing problems in the prior art. On the other hand, to provide a collaborative system for water resources - ecological environment - economic society in a water network system.

[0006] Technical Solution: The collaborative method for water resources - ecological environment - economic society in a water network system includes the following steps:

[0007] Step S1: Collect historical rainfall, historical evapotranspiration and historical water resource volume data in the study area, fit the marginal distributions of rainfall and evapotranspiration, construct their joint distribution, randomly sample to obtain a large number of scenarios and input them into a pre - constructed machine learning model to obtain the water resource volume corresponding to the large number of scenarios;

[0008] Step S2: Collect historical population and historical GDP data of the study area, construct an economic and social volume prediction model, use the rolling prediction method to obtain population prediction values and GDP prediction values, and combine them to obtain economic and social volume prediction samples;

[0009] Step S3: Input the massive scenario samples into the pre-constructed multi-objective optimization model to obtain water resources-ecological environment-economic and social coordination indicators, construct a system dynamics model and calibrate the model parameters. Randomly combine the water resource amounts and economic and social volume prediction samples corresponding to the massive scenarios to obtain massive scenario samples and input them into the system dynamics model in turn to obtain the water resources-ecological environment-economic and social coordination indicator values corresponding to the massive scenario samples;

[0010] Step S4: Use the maximum entropy projection pursuit method to reduce the dimension of the water resources-ecological environment-economic and social coordination indicators to obtain comprehensive coordination indicators and calculate the corresponding comprehensive coordination indicator values, construct a regulation plan, calculate the transfer degree between the comprehensive coordination indicator values of each scenario sample under different regulation plans, and screen to obtain the water network system water resources-ecological environment-economic and social coordinated optimization plan.

[0011] According to one aspect of the present application, the step S1 is further as follows:

[0012] Step S11: Collect historical rainfall, historical evapotranspiration and historical water resource amount data of the study area, construct a machine learning model, and optimize the model hyperparameters based on the historical rainfall, historical evapotranspiration and historical water resource amount data;

[0013] Step S12: Fit the marginal distributions of rainfall and evapotranspiration, construct the joint distribution of rainfall and evapotranspiration, and randomly sample to generate massive scenarios;

[0014] Step S13: Input the massive scenarios into the optimized machine learning model to obtain the water resource amounts corresponding to the massive scenarios.

[0015] According to one aspect of the present application, the step S11 is further as follows:

[0016] Step S11a: Collect historical rainfall, historical evapotranspiration and historical water resource amount data of the study area, and divide the data into a training set, a validation set and a test set according to the ratio of 80 / 10 / 10;

[0017] Step S11b: Construct a gradient boosting tree machine learning model, and use the training set to train the machine learning model;

[0018] Step S11c: Optimize the hyperparameters of the machine learning model by using the Bayesian optimization method, evaluate the model with the hyperparameters optimized by using the test set and verify it with the validation set.

[0019] According to one aspect of the present application, the step S2 is further as follows:

[0020] Step S21: Construct an economic and social volume prediction model based on a machine learning model, collect historical population and historical GDP data of the study area, and calibrate the parameters of the economic and social volume prediction model based on the historical population and historical GDP data;

[0021] Step S22: Use the rolling prediction method to obtain m population prediction values and n GDP prediction values, and combine them to obtain m×n economic and social volume prediction samples, where m and n are positive integers.

[0022] According to one aspect of the present application, the step S22 is further as follows:

[0023] Step S22a: Use the first four-fifths of the historical population and historical GDP data of the study area as the prediction set, and the last one-fifth as the verification set;

[0024] Step S22b: Input the prediction set into the economic and social volume prediction model, and use the moving average method to shift it backward to obtain the prediction values of the last one-fifth;

[0025] Step S22c: Compare the prediction values of the last one-fifth with the verification set and optimize the economic and social volume prediction model;

[0026] Step S22d: Input the historical population and historical GDP data of the study area into the optimized economic and social volume prediction model to obtain m population prediction values and n GDP prediction values, and combine them to obtain m×n economic and social volume prediction samples.

[0027] According to one aspect of the present application, the step S3 is further as follows:

[0028] Step S31: Construct a multi-objective optimization model, and the objective function is: the highest coordination degree of water resources - ecological environment - economic and society;

[0029] Step S32: Input the massive scenario samples into the multi-objective optimization model to obtain the water network system water resources - ecological environment - economic and social coordination index corresponding to the scenario sample;

[0030] Step S33: Collect the water resources volume, economic and social, and ecological environment data of the study area, construct a system dynamics model based on the relationship between water resources volume, economic and social, and ecological environment, and calibrate the parameters of the system dynamics model based on the water resources volume, economic and social, and ecological environment data of the study area;

[0031] Step S34: Extract the water resources volume corresponding to the massive scenarios and m×n economic and social volume prediction samples, and randomly combine the two to obtain massive scenario samples;

[0032] Step S35: Input the massive scenario samples into the system dynamics model in sequence to obtain the water resource - ecological environment - economic society coordination index values corresponding to the massive scenario samples.

[0033] According to one aspect of the present application, the step S33 is further as follows:

[0034] Step S33a: Determine system elements based on the relationship among water resource quantity, economic society and ecological environment, which are divided into three categories, namely water system connectivity, natural function and social function;

[0035] Step S33b: Construct several system causal loops based on the three system elements of water system connectivity, natural function and social function, including m positive feedback loops and n negative feedback loops, and construct a system causal loop diagram, where m and n are positive integers;

[0036] Step S33c: Determine the levels and rates of flow based on the system causal loop diagram, obtain three subsystems of water system connectivity, natural function and social function, and merge the three subsystems to construct a system dynamics model;

[0037] Step S33d: Calibrate the parameters of the system dynamics model based on the water resource quantity, economic society and ecological environment data of the study area by using the sensitivity analysis method and historical test.

[0038] According to one aspect of the present application, it is characterized in that the step S4 is further as follows:

[0039] Step S41: Use the maximum entropy projection pursuit method to reduce the dimension of the water resource - ecological environment - economic society coordination index to obtain a comprehensive coordination index;

[0040] Step S42: Calculate the comprehensive coordination index values corresponding to the massive scenario samples respectively based on the water resource - ecological environment - economic society coordination index values corresponding to the massive scenario samples;

[0041] Step S43: Construct a regulation plan, calculate the transfer degree between the comprehensive coordination index values under different regulation plans for each scenario sample, and select the one with the largest transfer degree as the regulation plan with the largest improvement degree, that is, the water network system water resource - ecological environment - economic society coordination optimization plan corresponding to this scenario sample.

[0042] According to one aspect of the present application, it is characterized in that the step S41 is further as follows:

[0043] Step S41a: Construct a projection function;

[0044] Step S41b: Use the entropy of the projection value as the projection index function, and use the projection index function to measure the information quantity of the projection value;

[0045] Step S41c: Determine the optimal projection direction using the maximum entropy method and construct an optimization model;

[0046] Step S41d: Input the coordinated index data of water resources - ecological environment - economic society into the optimization model, and calculate to obtain a one - dimensional projection vector, that is, the coordinated index after dimensionality reduction.

[0047] According to one aspect of the present application, it is characterized in that the step S43 is further as follows:

[0048] Step S43a: Input massive scenario samples into the multi - objective optimization model in sequence, set A regulation measures, and set B different regulation degrees for each regulation measure, obtaining A×B regulation schemes, where A and B are positive integers greater than 2;

[0049] Step S43b: Modify the parameters and boundary conditions of the multi - objective optimization model based on the A×B regulation schemes in sequence, and solve the model to obtain the comprehensive coordinated index values corresponding to each regulation scheme under each scenario sample;

[0050] Step S43c: Calculate the transfer degree between the original comprehensive coordinated index value of each scenario sample and the comprehensive coordinated index value after regulation of all regulation schemes using the Markov one - step transition probability, and select the one with the largest transfer degree as the regulation scheme with the largest improvement degree, that is, the coordinated optimization scheme of water resources - ecological environment - economic society of the water network system corresponding to this scenario sample.

[0051] According to another aspect of the present application, a coordinated system of water resources - ecological environment - economic society for a water network system is provided, including:

[0052] At least one processor; and

[0053] A memory communicatively connected to at least one of the processors; wherein,

[0054] The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the coordinated method of water resources - ecological environment - economic society of the water network system described in any one of the above technical solutions.

[0055] Beneficial effects: By using the coordinated method of water resources - ecological environment - economic society of the water network system, the collaborative operation efficiency of the water conservancy model is improved, the comprehensive simulation of the water conservancy system is realized, the intelligent level of water network dispatching is enhanced, the deep interaction and integration of the physical water network and the digital water network are realized, the development of various water conservancy projects from point to network and from dispersion to system is promoted, and the systematicness, comprehensiveness and strong toughness of the water network are comprehensively improved. Brief Description of the Drawings

[0056] Figure 1 is the flowchart of the present invention.

[0057] Figure 2 It is the flowchart of step S1 of the present invention.

[0058] Figure 3 It is the flowchart of step S2 of the present invention.

[0059] Figure 4 It is the flowchart of step S3 of the present invention.

[0060] Figure 5 It is the flowchart of step S4 of the present invention. Specific implementation manners

[0061] As Figure 1 shown, the following technical solutions are proposed. According to one aspect of the present application, a method for coordinating water resources - ecological environment - economic society in a water network system is provided, which is characterized by including the following steps:

[0062] Step S1: Collect historical rainfall, historical evapotranspiration, and historical water resource volume data in the study area, fit the marginal distributions of rainfall and evapotranspiration, construct their joint distribution, randomly sample to obtain a large number of scenarios and input them into a pre - constructed machine learning model to obtain the water resource volume corresponding to the large number of scenarios;

[0063] Step S2: Collect historical population and historical GDP data in the study area, construct an economic and social volume prediction model, use the rolling prediction method to obtain population prediction values and GDP prediction values, and combine them to obtain economic and social volume prediction samples;

[0064] Step S3: Input the large number of scenario samples into a pre - constructed multi - objective optimization model to obtain water resources - ecological environment - economic society coordination indicators, construct a system dynamics model and calibrate the model parameters, randomly combine the water resource volume corresponding to the large number of scenarios and the economic and social volume prediction samples to obtain a large number of scenario samples and input them into the system dynamics model in sequence to obtain the water resources - ecological environment - economic society coordination indicator values corresponding to the large number of scenario samples;

[0065] Step S4: Use the maximum entropy projection pursuit method to reduce the dimension of the water resources - ecological environment - economic society coordination indicators to obtain comprehensive coordination indicators and calculate the corresponding comprehensive coordination indicator values, construct a regulation plan, calculate the transfer degree between the comprehensive coordination indicator values of each scenario sample under different regulation plans, and screen to obtain the water network system water resources - ecological environment - economic society coordination optimization plan.

[0066] The coordination of water resources - ecological environment - economic society is the primary goal that the construction, management, and operation of the water network system need to achieve. To improve the collaborative operation efficiency of the water conservancy model, realize the comprehensive simulation of the water conservancy system, and enhance the intelligent level of water network dispatching, the present invention constructs a causal chain of water resources - ecological environment - economic society in the water network system to clarify the relationships among elements. First, it is necessary to collect various data on water resources, economic society, and ecological environment;

[0067] The present invention first collects historical rainfall, historical evapotranspiration, and historical water resources data in the study area. By fitting the marginal distributions of rainfall and evapotranspiration, a joint distribution of the two is constructed, and a large number of simulated scenario data are randomly sampled for training and calculation of the machine learning model. After training the machine learning model, the model is used to calculate the water resources volume;

[0068] Secondly, historical population and historical GDP data in the study area are collected. By constructing a prediction model for the economic and social volume, the population and GDP are predicted. Based on the corresponding relationship between historical population data and GDP data, the relationship curve between the two is predicted, and the future population and GDP data are respectively predicted. Based on the predicted data, cross - combinations are made to obtain prediction samples of the economic and social volume, that is, economic and social data;

[0069] Then, based on the calculated water resources volume and economic and social data, together with the ecological environment data in the study area, a system dynamics model of the study area is constructed to obtain the structural functions and connections of each subsystem within the study area, and based on this, the coordination and optimization of water resources - ecological environment - economic society in the water network system of the study area are carried out.

[0070] According to one aspect of the present application, the step S1 is further as follows:

[0071] Step S11: Collect historical rainfall, historical evapotranspiration, and historical water resources data in the study area, construct a machine learning model, and optimize the model hyperparameters based on the historical rainfall, historical evapotranspiration, and historical water resources data;

[0072] Step S12: Fit the marginal distributions of rainfall and evapotranspiration, construct the joint distribution of rainfall and evapotranspiration, and randomly sample to generate a large number of scenarios;

[0073] Step S13: Input the large number of scenarios into the optimized machine learning model to obtain the water resources volume corresponding to the large number of scenarios.

[0074] According to one aspect of the present application, the step S11 is further as follows:

[0075] Step S11a: Collect historical rainfall, historical evapotranspiration, and historical water resources data in the study area, and divide the data into a training set, a validation set, and a test set according to the ratio of 80 / 10 / 10;

[0076] Step S11b: Construct a gradient boosting tree machine learning model and train the machine learning model using a training set;

[0077] Step S11c: Optimize the hyperparameters of the machine learning model using the Bayesian optimization method, evaluate the model with optimized hyperparameters using a test set, and verify it using a validation set.

[0078] According to one aspect of the present application, step S2 is further as follows:

[0079] Step S21: Based on the machine learning model, construct an economic and social volume prediction model, collect historical population and historical GDP data of the research area, and calibrate the parameters of the economic and social volume prediction model based on the historical population and historical GDP data;

[0080] Step S22: Use the rolling prediction method to obtain m population prediction values and n GDP prediction values, and combine them to obtain m×n economic and social volume prediction samples, where m and n are positive integers.

[0081] According to one aspect of the present application, step S22 is further as follows:

[0082] Step S22a: Use the first four-fifths of the historical population and historical GDP data of the research area as the prediction set, and the last one-fifth as the validation set;

[0083] Step S22b: Input the prediction set into the economic and social volume prediction model, and use the moving average method to shift it backward to obtain the prediction values of the last one-fifth;

[0084] Step S22c: Compare the prediction values of the last one-fifth with the validation set and optimize the economic and social volume prediction model;

[0085] Step S22d: Input the historical population and historical GDP data of the research area into the optimized economic and social volume prediction model to obtain m population prediction values and n GDP prediction values, and combine them to obtain m×n economic and social volume prediction samples.

[0086] The rolling prediction method is a method for predicting time series data and is usually used in the fields of sales forecasting or inventory management. This method is based on the rolling concept, that is, imagining the data sequence as a rolling wheel, and each time data is observed, it rolls forward by a time window, and the data within this time window is used for prediction. Specifically:

[0087] Select a suitable time window size, that is, the number of data points included each time the wheel rolls forward;

[0088] Decompose the time series data into multiple time windows, and each time roll forward by one time window for prediction. Use the average value, moving average, and exponential smoothing method to predict the values within the next time window;

[0089] Use historical data to evaluate the accuracy of the roller prediction method.

[0090] In this embodiment, since the economic society is different from the natural law and it grows steadily, its distribution cannot be obtained by fitting. In this embodiment, the distribution of the predicted values of the economic and social volume is obtained by the rolling prediction method. In a certain embodiment, specifically:

[0091] Collect the historical population and historical GDP data of the study area in the previous 30 years;

[0092] Take the historical population and historical GDP data in the previous 24 years as the test set, and the historical population and historical GDP data in the latter 6 years as the validation set;

[0093] Based on the relationship between the historical population and historical GDP in the test set, obtain the relationship curve between the two. Based on the development law of the economic society, translate the relationship curve backward to obtain the predicted data of the historical population and historical GDP in the latter 6 years;

[0094] Use the validation set to compare with the predicted data and modify the relationship curve to obtain an optimized prediction model for the economic and social volume;

[0095] Input the historical population and historical GDP data of the study area in the previous 30 years into the optimized prediction model for the economic and social volume to obtain a number of population predicted values and a number of GDP predicted values, and arrange and combine the two to obtain a prediction sample set for the economic and social volume.

[0096] According to one aspect of the present application, the step S3 is further as follows:

[0097] Step S31: Construct a multi-objective optimization model, and the objective function is: the highest synergy degree of water resources - ecological environment - economic society;

[0098] Step S32: Input a large number of scenario samples into the multi-objective optimization model to obtain the water network system water resources - ecological environment - economic society synergy index corresponding to the scenario sample;

[0099] Step S33: Collect the water resources volume, economic society and ecological environment data of the study area, construct a system dynamics model based on the relationship between water resources volume, economic society and ecological environment, and calibrate the parameters of the system dynamics model based on the water resources volume, economic society and ecological environment data of the study area;

[0100] System dynamics is a method that combines qualitative and quantitative analysis and systematic analysis, and is applicable to long-term dynamic trend research. It can comprehensively simulate and analyze the internal relationships of various complex systems and the long-term dynamics under different decisions. In this embodiment, since the connection of river and lake water systems is a complex feedback system involving society, ecology, and resources, the connection effects between water system connectivity, social functions, and natural functions can all rely on the quantitative analysis and simulation of system dynamics to obtain results, which is convenient for simulation and analysis of the evolution laws of the main driving factors. Therefore, in this embodiment, a system dynamics model is constructed to depict the internal water network system of the study area.

[0101] Step S34: Extract the water resource amounts corresponding to a large number of scenarios and m×n economic and social volume prediction samples, and randomly combine the two to obtain a large number of scenario samples.

[0102] Step S35: Input the large number of scenario samples into the system dynamics model in sequence to obtain the water resource-ecological environment-economic and social coordination index values corresponding to the large number of scenario samples.

[0103] According to one aspect of the present application, the step S33 is further as follows:

[0104] Step S33a: Determine system elements based on the relationships among water resource amounts, economic society, and ecological environment, which are divided into three categories, namely water system connectivity, natural functions, and social functions.

[0105] Step S33b: Based on the three types of system elements of water system connectivity, natural functions, and social functions, construct several system causal loops, including m positive feedback loops and n negative feedback loops, and construct a system causal loop diagram, where m and n are positive integers.

[0106] Step S33c: Determine the levels and rates based on the system causal loop diagram to obtain three subsystems of water system connectivity, natural functions, and social functions, and combine the three subsystems to construct a system dynamics model.

[0107] Step S33d: Calibrate the parameters of the system dynamics model based on the water resource amount, economic society, and ecological environment data of the study area using the sensitivity analysis method and historical test.

[0108] System dynamics describes the dependence relationship of the change rates of each state variable of the system on each state variable or specific input. The principles of system dynamics modeling are:

[0109] The system can completely describe the structure of the system and its behavior in different periods using state variables.

[0110] Each feedback loop in the model should contain at least one state variable.

[0111] The principle of material conservation.

[0112] Any state variable in the system indirectly affects another state variable.

[0113] The steps for modeling the system dynamics model are specifically as follows:

[0114] Determine the system simulation objective, delimit the system boundary, and identify the problems to be solved by the system;

[0115] Identify the relevant factors of the system and the relationships between them, construct a causal relationship diagram and a feedback relationship, observe the mutual restraint relationships of the feedback loops, and formulate policies for controlling the system;

[0116] Construct a system flow diagram and structural equations. The structural equations include: level equations, rate equations, and auxiliary variable equations, to obtain the system dynamics model;

[0117] Assign initial values to the equations in the model and perform simulation. Input the initial values of the parameters and the variable values into the structural equations for simulation to obtain the predicted values of each variable;

[0118] Modify the system model based on the simulation results, including: model operation parameters, system structure, and boundary.

[0119] According to one aspect of the present application, it is characterized in that the step S4 is further as follows:

[0120] Step S41: Use the maximum entropy projection pursuit method to reduce the dimension of the water resources - ecological environment - economic and social coordination index to obtain a comprehensive coordination index;

[0121] Step S42: Based on the water resources - ecological environment - economic and social coordination index values corresponding to a large number of scenario samples, calculate the comprehensive coordination index values corresponding to the large number of scenario samples respectively;

[0122] Step S43: Construct a regulation plan, calculate the transfer degree between the comprehensive coordination index values under different regulation plans for each scenario sample, and select the one with the largest transfer degree as the regulation plan with the largest improvement degree, that is, the water network system water resources - ecological environment - economic and social coordination optimization plan corresponding to this scenario sample.

[0123] According to one aspect of the present application, it is characterized in that the step S41 is further as follows:

[0124] Step S41a: Construct a projection function;

[0125] Step S41b: Use the entropy of the projection value as the projection index function, and use the projection index function to measure the information amount of the projection value;

[0126] Step S41c: Use the maximum entropy method to determine the optimal projection direction and construct an optimization model;

[0127] Step S41d: Input the coordinated index data of water resources - ecological environment - economic society into the optimization model, and calculate to obtain a one - dimensional projection vector, that is, the comprehensive coordinated index after dimensionality reduction.

[0128] The projection pursuit method projects high - dimensional data into a low - dimensional space. Through numerical calculation, it maximizes the index reflecting the data clustering degree, and then finds the optimal projection to reflect the data structure characteristics. In a certain embodiment, specifically:

[0129] Reduce the p - dimensional standardized data containing n samples to a one - dimensional projection value;

[0130] Construct a projection index function. In the projection pursuit method, the construction of the projection index function is the multiplication of the within - class density and the between - class distance of the projection values. Its purpose is to enable the projection values to reveal as much as possible the spatial distribution structure characteristics of the original high - dimensional data. In this embodiment, in order to extract as much information as possible from the high - dimensional risk factors, while solving the curse of dimensionality problem and retaining as much original information as possible, the entropy of the projection values is used as the projection index function to measure the information content of the projection values;

[0131] Use the maximum entropy method to select the solution, extract as much information as possible from the known data, and make the fewest assumptions about the unknown part. When the information content of the projection values is the largest, the optimal parameters are obtained;

[0132] Substitute the obtained best projection direction into the projection value calculation formula to obtain the projection vector.

[0133] By using the projection pursuit method to reduce the dimensionality of multiple coordinated equilibrium indicators into a single comprehensive coordinated equilibrium indicator, it can make the subsequent judgment of the equilibrium state and the screening of the regulation and control schemes more intuitive and clear, and at the same time reduce the amount of calculation. Therefore, in this embodiment, the maximum entropy projection pursuit method is used to reduce the dimensionality of the coordinated equilibrium indicators of water resources - ecological environment - economic society to obtain the comprehensive coordinated equilibrium indicator.

[0134] According to one aspect of the present application, the step S43 is further:

[0135] Step S43a: Input the massive scenario samples into the multi - objective optimization model in sequence, set A regulation and control measures, and set B different regulation and control degrees for each regulation and control measure, to obtain A×B regulation and control schemes, where A and B are positive integers greater than 2;

[0136] Step S43b: Based on the A×B regulation and control schemes, modify the parameters and boundary conditions of the multi - objective optimization model in sequence, and solve the model to obtain the comprehensive coordinated index values corresponding to each regulation and control scheme under each scenario sample;

[0137] Step S43c: Calculate the transfer degree between the original comprehensive coordination index value of each scenario sample and the comprehensive coordination index value after the regulation of all regulation schemes in turn using the Markov one-step transition probability, and select the one with the largest transfer degree as the regulation scheme with the largest improvement degree, that is, the water network system water resources-ecological environment-economic and social collaborative optimization scheme corresponding to this scenario sample.

[0138] According to another aspect of the present application, there is provided a water network system water resources-ecological environment-economic and social collaborative system, characterized by including:

[0139] At least one processor; and

[0140] A memory communicatively connected to at least one of the processors; wherein,

[0141] The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the water network system water resources-ecological environment-economic and social collaborative method described in any one of the above.

[0142] 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 belong to the protection scope of the present invention.

Claims

1. Water resources-ecological environment-economic and social coordination method of water network system, characterized by: The steps include: Step S1, collect historical rainfall, historical evapotranspiration and historical water resources data of the study area, fit the marginal distribution of rainfall and evapotranspiration, construct the joint distribution of the two, randomly sample massive scenarios and input them into the pre-built machine learning model to obtain the water resources corresponding to the massive scenarios; Step S2: Collect historical population and GDP data of the study area, build an economic and social volume prediction model, use the rolling prediction method to obtain population prediction values ​​and GDP prediction values, and combine them to obtain economic and social volume prediction samples; Step S3, inputting the massive scenario samples into the pre-built multi-objective optimization model to obtain the water resources-ecological environment-economic and social synergy index, constructing the system dynamics model and calibrating the model parameters, randomly combining the water resources and economic and social volume prediction samples corresponding to the massive scenarios to obtain massive scenario samples and inputting them into the system dynamics model in turn to obtain the water resources-ecological environment-economic and social synergy index values ​​corresponding to the massive scenario samples; Step S4, using the maximum entropy projection pursuit method to reduce the dimension of the water resources-ecological environment-economic and social synergy index to obtain a comprehensive synergy index and calculate the corresponding comprehensive synergy index value, construct a control plan, calculate the transfer degree between the comprehensive synergy index values ​​under different control plans for each scenario sample, and screen out a water network system water resources-ecological environment-economic and social synergy optimization plan; The step S3 is further as follows: Step S31, construct a multi-objective optimization model, the objective function is: the highest degree of coordination between water resources, ecological environment and economic society; Step S32, inputting a large number of scenario samples into a multi-objective optimization model to obtain the water resources-ecological environment-economic and social coordination index of the water network system corresponding to the scenario samples; Step S33, collecting data on water resources, economic society and ecological environment in the study area, building a system dynamics model based on the relationship between water resources, economic society and ecological environment, and calibrating system dynamics model parameters based on the data on water resources, economic society and ecological environment in the study area; Step S34, extracting the water resources corresponding to the massive scenario and m×n economic and social volume prediction samples, and randomly combining the two to obtain massive scenario samples; Step S35, inputting the massive scenario samples into the system dynamics model in sequence to obtain the water resources-ecological environment-economic and social synergy index values ​​corresponding to the massive scenario samples; The step S33 is further as follows: Step S33a, determining system elements based on the relationship between water resources, economic society and ecological environment, which are divided into three categories: water system connectivity, natural functions and social functions; Step S33b, constructing a number of system causal loops based on the three system elements of water system connectivity, natural function and social function, including m positive feedback loops and n negative feedback loops, and constructing a system causal loop diagram, where m and n are positive integers; Step S33c, determining the flow position and flow rate based on the system causal loop diagram, obtaining three subsystems of water system connectivity, natural function, and social function, and merging the three subsystems to construct a system dynamics model; Step S33d, based on the water resources, economic, social and ecological environment data of the study area, the system dynamics model parameters are calibrated using sensitivity analysis and historical verification; The step S4 is further as follows: Step S41, using the maximum entropy projection pursuit method to reduce the dimension of the water resources-ecological environment-economic and social synergy index to obtain a comprehensive synergy index; Step S42, based on the water resources-ecological environment-economic and social synergy index values ​​corresponding to the massive scenario samples, respectively calculate the comprehensive synergy index values ​​corresponding to the massive scenario samples; Step S43, constructing a control scheme, calculating the transfer degree between the comprehensive synergy index values ​​under different control schemes for each scenario sample, and selecting the control scheme with the largest transfer degree as the one with the largest improvement, that is, the water resources-ecological environment-economic and social synergistic optimization scheme of the water network system corresponding to the scenario sample; The step S41 is further as follows: Step S41a, constructing a projection function; Step S41b, using the entropy of the projection value as a projection index function, and using the projection index function to measure the amount of information of the projection value; Step S41c, using the maximum entropy method to determine the best projection direction and construct an optimization model; Step S41d, input the water resources-ecological environment-economic and social synergy index data into the optimization model, and calculate a one-dimensional projection vector, i.e., a comprehensive synergy index after dimensionality reduction; The step S43 is further as follows: Step S43a, sequentially inputting a large number of scenario samples into the multi-objective optimization model, setting A control measures, setting B different control degrees for each control measure, and obtaining A×B control schemes, where A and B are positive integers greater than 2; Step S43b: based on A×B control schemes, modify the multi-objective optimization model parameters and boundary conditions in turn, and solve the model to obtain the comprehensive coordination index value corresponding to each control scheme under each scenario sample; Step S43c, using the Markov one-step transition probability to calculate the degree of transfer between the original comprehensive synergy index value of each scenario sample and the comprehensive synergy index value after adjustment of all control schemes, and selecting the control scheme with the largest degree of transfer as the one with the greatest improvement, that is, the water resources-ecological environment-economic and social coordinated optimization scheme of the water network system corresponding to the scenario sample.

2. The water network system water resources-ecological environment-economic and social coordination method according to claim 1, characterized in that: The step S1 is further as follows: Step S11, collecting historical rainfall, historical evapotranspiration and historical water resources data of the study area, building a machine learning model, and optimizing model hyperparameters based on the historical rainfall, historical evapotranspiration and historical water resources data; Step S12, fitting the marginal distribution of rainfall and evapotranspiration, constructing the joint distribution of rainfall and evapotranspiration, and randomly sampling to generate a large number of scenarios; Step S13: input the massive scenarios into the optimized machine learning model to obtain the water resources corresponding to the massive scenarios.

3. The water resources-ecological environment-economic and social coordination method of the water network system according to claim 2, characterized in that: The step S11 is further as follows: Step S11a, collect historical rainfall, historical evapotranspiration and historical water resources data of the study area, and divide the data into training set, validation set and test set according to the ratio of 80 / 10 / 10; Step S11b, constructing a gradient boosting tree machine learning model, and using a training set to train the machine learning model; Step S11c: Use the Bayesian optimization method to optimize the hyperparameters of the machine learning model, use the test set to evaluate the model after hyperparameter optimization, and use the validation set to verify it.

4. The water resources-ecological environment-economic and social coordination method of the water network system according to claim 1, characterized in that: The step S2 is further as follows: Step S21: construct an economic and social volume prediction model based on the machine learning model, collect historical population and historical GDP data of the study area, and calibrate the parameters of the economic and social volume prediction model based on the historical population and historical GDP data; Step S22: Use the rolling forecast method to obtain m population forecast values ​​and n GDP forecast values, and combine them to obtain m×n economic and social volume forecast samples, where m and n are positive integers.

5. The water resources-ecological environment-economic and social coordination method of the water network system according to claim 4, characterized in that: The step S22 is further as follows: Step S22a, taking the first four fifths of the historical population and historical GDP data of the study area as the prediction set, and the last one fifth as the verification set; Step S22b, input the prediction set into the economic and social volume prediction model, and use the moving average method to shift backward to obtain the prediction value of the last fifth; Step S22c, comparing the predicted values ​​of the last fifth with the validation set and optimizing the economic and social volume prediction model; Step S22d, input the historical population and historical GDP data of the study area into the optimized economic and social volume prediction model, obtain m population prediction values ​​and n GDP prediction values, and combine them to obtain m×n economic and social volume prediction samples.

6. Water network system Water resources-ecological environment-economic and social coordination system, characterized by: include: at least one processor; as well as, a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the water resources-ecological environment-economic and social coordination method of the water network system as described in any one of claims 1 to 5.

Images