Water conservancy information intelligent management and auxiliary decision-making method

Abstract

The invention relates to a water conservancy information intelligent management and auxiliary decision-making method, which comprises the following steps of: acquiring water resource dynamic information at water conservancy equipment and a node terminal, and performing synchronous calibration on sensing information by using a multi-source data fusion technology; processing the water resource dynamic information, and extracting key feature data based on a distributed processing architecture; analyzing the key feature data by utilizing an edge computing device and combining with the dynamic optimization model, and generating a prediction result of the water resource state; according to the characteristic indexes collected in real time and the stored analysis result, a local management decision strategy is generated by applying an adaptive algorithm in combination with a set rule base; verifying and adjusting the local management decision strategy by using a multi-objective optimization algorithm in the centralized management system; transmitting the optimized global management decision strategy to execution end equipment; the water conservancy information intelligent management and auxiliary decision-making method has the advantages of being high in real-time performance, high in data processing efficiency, high in decision-making support scientificity and high in environmental adaptability.

Description

Technical Field

[0001] This application relates to the field of water resources management technology, specifically to a method for intelligent management and decision support of water resources information. Background Technology

[0002] With the rapid development of information technology and intelligent management technology, the informatization level of the water conservancy industry is constantly improving, and water resource management and decision-making are gradually moving towards automation and intelligence. However, the dynamic changes of water resources are complex and widely distributed, and traditional water conservancy information management methods still face many challenges in processing massive amounts of data, dynamic sensing, and decision support. Among related technologies, traditional water conservancy information management and decision support methods mainly rely on centralized data collection and manual analysis. Although these methods can play a certain role in specific scenarios, they have significant shortcomings in the following aspects when facing the complexity of current water resource management and the needs of emergency response: 1. Lack of real-time and dynamic adaptability: Most existing water resources information management systems adopt a centralized data processing architecture, resulting in a long data collection and analysis cycle, failing to reflect the dynamic changes in water resources in real time. This delay makes it difficult for traditional methods to quickly respond to emergencies, such as flood warnings and real-time monitoring and decision-making regarding water pollution. 2. Low data processing efficiency: With the development of IoT technology, a large number of distributed sensors have generated massive amounts of data. However, traditional methods lack distributed and edge computing capabilities, resulting in huge bandwidth and computing resource consumption for data transmission and centralized processing, making it inefficient for handling large-scale dynamic data from multiple sources and dimensions. 3. Insufficient decision support: Traditional water resources management systems rely heavily on static rules or experience-based decision-making models. These methods cannot fully utilize real-time dynamic data and historical data for comprehensive analysis, lacking scientific decision support capabilities and proving inadequate in multi-objective optimization and complex scenario control. 4. Incomplete sensing: Existing sensor networks are mostly based on single-type sensors, failing to achieve accurate perception and dynamic fusion of multi-dimensional water resources information, leading to incomplete and inaccurate decision-making information. Summary of the Invention

[0003] This application provides a method for intelligent management and decision support of water conservancy information, which has the advantages of strong real-time performance, high data processing efficiency, strong scientific decision support, and strong environmental adaptability.

[0004] The intelligent management and decision support method for water conservancy information provided in this application includes the following steps: S1. Construct a distributed sensor network to collect dynamic water resource information at water conservancy equipment and node terminals, and use multi-source data fusion technology to synchronously calibrate the sensor information. S2. Configure edge computing devices at water conservancy node terminals, transmit the collected dynamic water resources information to the edge computing devices, perform data cleaning, outlier removal and format standardization on the dynamic water resources information, and extract key feature data based on a distributed processing architecture. S3. Using the edge computing device, combined with the dynamic optimization model, analyze the key feature data, generate prediction results of water resource status, extract feature indicators related to decision-making, and save the prediction results in the local storage unit. S4. Based on the real-time collected feature indicators and stored prediction results, an adaptive algorithm is applied in conjunction with a set rule base to generate local management decision strategies, and the locally generated local management decision strategies are transmitted to the centralized management system through a communication network. S5. In the centralized management system, a multi-objective optimization algorithm is used to verify and adjust the local management decision-making strategy to generate a global management decision-making strategy that meets the optimization requirements. S6. Transmit the optimized global management decision strategy to the execution terminal equipment, control the relevant water conservancy equipment to implement the management decision, and monitor the data changes in the implementation process in real time through the sensor network, and record the execution results and environmental response.

[0005] In one alternative approach, step S1 specifically includes the following steps: S11. Deploy intelligent sensing nodes that support distributed collaboration in water conservancy equipment and node terminals to collect dynamic information on water resources. The intelligent sensing nodes include water level sensors, flow velocity sensors, water quality sensors and meteorological sensors. S12. When collecting dynamic information on water resources, calculate the rate of change of dynamic parameters of water resources in the monitoring environment in real time. Based on the calculated rate of change Combined with adjustment coefficient and minimum sampling period Dynamically determine the sampling period : ; In the formula, For dynamic parameter functions, These are time-varying parameters, including water level, flow velocity, and water quality indicators. Indicates the minimum sampling period; Setting a threshold range for adjusting the sampling period. Limit the range of change between two consecutive sampling periods: ; In the formula, and They represent the first The sampling period and the first Each sampling period The preset range of variation threshold; S13. A spatiotemporal coordination mechanism based on sensor nodes is used to perform time calibration on the collected dynamic water resources information using a time synchronization protocol, and the geographic reference coordinates of the measurement data are unified through a spatial position correction model between nodes. S14. For multi-source observation data in water resource dynamic information, calculate the dynamic weight of each observation data. And based on dynamic weights The results are generated by merging and integrating observation data to produce collaborative processing results.

[0006] In one alternative embodiment, step S14 specifically includes the following steps: Multi-source observation data in water resource dynamic information Based on the prior state of the environment Establish a conditional probability model , indicating that given a prior state The reliability of each sensor data point is assessed, and the dynamic weight of each observation is calculated using Bayesian inference methods. : ; In the formula, For the first Dynamic weights of each observation data point This represents the confidence probability of the observed data under the prior state. The total number of sensors participating in the collaborative processing; Through a dynamic update mechanism, adjustments are made in real time based on the accuracy of historical sensor data, the current ambient noise level, and the sensor's operating status. Value: ; In the formula, Indicates the first The reliability score of each sensor This represents the value before the prior state was updated; Based on the calculated dynamic weights All observation data are weighted and fused to generate the fused collaborative processing result: ; In the formula, For the fused observation data, For the first Observational data from each sensor.

[0007] In one alternative approach, step S2 specifically includes the following steps: S21. The dynamic water resources information collected through the distributed sensor network is transmitted in the form of data frames to the edge computing device configured in the water conservancy node terminal. The data frames include timestamps, sensor identifiers and dynamic parameter values. S22. In the edge computing device, the received dynamic information is cleaned to remove incomplete data frames and abnormal data exceeding the threshold range. S23. Standardize the format of the cleaned data frame and convert the dynamic information into structured data containing multi-dimensional attribute fields based on a predefined unified data model. The multi-dimensional attribute fields include timestamps, spatial coordinates, dynamic parameter values ​​and sensor identifiers. S24. Based on a distributed processing architecture, the standardized structured data is divided into multiple data partitions. The data partitioning rules are set according to spatial range and time period, and the partition index is... for: ; In the formula, For partitioned indexes, Represents the spatial coordinates of the data. For timestamps, , and These represent the spatial and temporal partitioning intervals, respectively. and These represent the number of spatial partitions; S25. Within each data partition, based on the dynamic parameter set Key feature data are extracted using a higher-order difference method: ; In the formula, For the rate of dynamic change, It is the difference order. The sampling time interval, It is the number of combinations.

[0008] In one alternative approach, step S3 specifically includes the following steps: S31. Load a dynamic optimization model into the edge computing device and analyze the extracted key feature data, which includes a set of dynamic parameters and corresponding timestamps. S32. When constructing water resource state prediction equations based on dynamic optimization models, time-series prediction algorithms are used to analyze the dynamic parameter set. Modeling was performed, using an autoregressive integral moving average model to fit the time series data and predict values. for: ; In the formula, For predicted values, and These are the parameters for the autoregressive and moving average models, respectively. and The model order is... For prediction error, As a bias term, the model parameters are optimized by fitting historical time series data. , and During the forecasting process, a sliding window technique is introduced to update the time series data in real time, retaining only the window length at a time. Latest data ; S33, Regarding the prediction results By comparing the data with real-time data, the parameters of the dynamic optimization model are adjusted using error correction methods; S34. Extract decision-related feature indicators from the prediction results, based on the predicted value set. Construct a set of feature extraction functions using dynamically changing parameters. Each feature extraction function Specific indicators are calculated based on different decision-making needs. All calculated feature indicators undergo initial screening. Invalid and redundant features are filtered out based on constraints and priorities set in rule base R, resulting in the final set of key feature indicators. Each of the key metrics Satisfy the rules And store them in order of their decision priority; S35. Store the analyzed and screened set of feature indicators and the prediction results of water resource status to the local storage unit of the edge computing device.

[0009] In one alternative embodiment, step S33 specifically includes the following steps: Regarding prediction error To minimize the objective, a parameter optimization process based on gradient descent is constructed: ; By calculating the error function, the model parameters are... , and The gradient is used to optimize the parameters using an iterative update rule: ; ; ; In the formula, For learning rate, , , These are the partial derivatives of the prediction error with respect to the model parameters; An adaptive learning rate adjustment mechanism is introduced to dynamically adjust the learning rate based on the current error change rate. Optimize the efficiency and stability of error correction: ; In the formula, To adjust the coefficient, This represents the change in error between two iterations. Symbols representing changes in error.

[0010] In one alternative approach, step S4 specifically includes the following steps: S41. Feature index set based on real-time acquisition and the collection of stored analysis results Initialize the input parameters of the adaptive algorithm, including the indicator weight vector. and optimization objective function ; S42, Combining the rule base The decision-making rules in the model are used to construct local management decision objective functions through a multi-objective optimization model. ; S43. Using a genetic algorithm to determine the objective function of local management decisions. Optimize; S44. Optimize the set of local management decision parameters obtained by the genetic algorithm. Mapped to specific management decision-making strategies The mapping process is based on a rule base. The strategy generation rules in the middle will generate each optimal parameter value Converted into specific management actions or control instructions, the mapped set of management decision strategies is represented as follows: Each strategy Parameters corresponding to the objective function Related; S45. Dynamically adjust the rule base based on real-time environmental feedback data. Constraints in and indicator weights Through real-time feedback data Calculate the execution deviation of the current management strategy. The deviation is introduced into the rule base, and the constraint function is dynamically updated. and the weights of the objective function Based on the rate of change of the feedback data The learning rate used in the real-time adjustment and optimization process and constraint penalty coefficient ; S46. Generate local management decision strategies The policies are stored in the local storage unit of the edge computing device and transmitted to the centralized management system via a communication network.

[0011] In one alternative embodiment, step S42 specifically includes the following steps: Combined with rule base The decision rules in the system determine the set of characteristic indicators. and analysis result set Initialize indicator weights Weight The constraints are satisfied: ; Calculate characteristic indicators With analysis results Matching degree function : ; In the formula, The scaling parameter of the matching degree function is used to adjust the matching sensitivity of the feature indicators and analysis results; Based on the rule base Constraints set in Define constraint function for: ; In the formula, These are the weighting coefficients. For the first Constraints The deviation function measures the degree to which the current state satisfies the constraints. Matching function and constraint function Substitute into the local management decision objective function : ; In the formula, To constrain the penalty coefficient, the rule base is dynamically adjusted to balance the weight between the matching degree and the constraint conditions; The parameters in the objective function are updated based on real-time feedback, including dynamically adjusting the indicator weights. and constraint penalty coefficient , so that the objective function The calculations are consistent with real-time data on the current water resource status.

[0012] In one alternative embodiment, step S43 specifically includes the following steps: Initialize population Each individual Let represent a set of candidate decision parameters, with a population size of . Individual codes are represented using real numbers, and each code parameter corresponds to a local management decision objective function. Input variables; Based on the objective function For each individual in the population Fitness evaluation, fitness function This indicates an individual's performance in optimizing their goals: ; In the formula, For the objective function in individuals The value on, This is the current optimal target value; Selection is performed on the population based on fitness values, using a roulette wheel selection method to determine the parents of the next generation, with selection probabilities... for: ; In the formula, For individuals fitness value, The sum of the fitness values ​​of all individuals; A crossover operation is performed on the selected parent individuals using a single-point crossover method, randomly swapping the codes of the parent individuals to generate candidate solutions for the next generation. The crossover probability is... Control the occurrence rate of crossover operations; The individuals generated after crossover undergo a mutation operation, which randomly changes a parameter value in the individual's encoding to generate a new solution space. The mutation probability... Controlling the occurrence rate of mutation operations: ; In the formula, For the amplitude of variation, In the interval Randomly generated values; The fitness of the mutated population is evaluated, and the individual with the highest fitness value is selected as the optimal solution for the current population and recorded as follows: And update the objective function. Optimal value .

[0013] In one alternative embodiment, step S5 specifically includes the following steps: S51, Receive the set of local management decision strategies and real-time collected global water resource dynamic data For local strategies With global data Perform a consistency check; S52. Combining global optimization requirements, construct a global management decision objective function based on a multi-objective optimization algorithm. : ; In the formula, Representing local strategies Fitness under globally dynamic data For the weights of the local strategy, Representing global constraints The deviation function, This is the global constraint penalty coefficient; S53. Using a multi-objective optimization algorithm to optimize the global management decision objective function. Optimize; S54. Optimize the global management decision parameter set. Mapped to specific global management decision-making strategies Each strategy From optimization parameters It was converted from; S55, Generate the global management decision strategy Distributed to relevant water conservancy equipment and node terminals via communication networks.

[0014] The beneficial effects of this application are as follows: The intelligent water resources information management and decision support method in this application makes full use of edge computing technology, distributed intelligent sensing network and multi-objective optimization algorithm, which better realizes the management and decision support of dynamic water resources monitoring, real-time data processing and adaptive decision generation. It has the advantages of strong real-time performance, high data processing efficiency, strong scientific decision support and strong environmental adaptability.

[0015] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the logical framework of the intelligent management and decision support method for water conservancy information in this application.

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. Detailed Implementation

[0018] To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0019] It should be understood that the described embodiments are merely some embodiments of this application, and not all embodiments. All other technical solutions obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0021] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0022] like Figure 1 As shown in the figure, this application provides a method for intelligent management and decision support of water conservancy information, which mainly includes the following steps: S1. Construct a distributed sensor network to collect dynamic water resource information at water conservancy equipment and node terminals, and use multi-source data fusion technology to synchronously calibrate the sensor information. Specifically, step S1 includes the following steps: S11. Deploy intelligent sensing nodes that support distributed collaboration in water conservancy equipment and node terminals to collect dynamic information on water resources. The intelligent sensing nodes include water level sensors, flow velocity sensors, water quality sensors and meteorological sensors. S12. When collecting dynamic information on water resources, calculate the rate of change of dynamic parameters of water resources in the monitoring environment in real time. Based on the calculated rate of change Combined with adjustment coefficient and minimum sampling period Dynamically determine the sampling period : ; In the formula, For dynamic parameter functions, These are time-varying parameters, including water level, flow velocity, and water quality indicators. Indicates the minimum sampling period; Setting a threshold range for adjusting the sampling period. Limit the range of change between two consecutive sampling periods: ; In the formula, and They represent the first The sampling period and the first Each sampling period The preset range of variation threshold; S13. A spatiotemporal coordination mechanism based on sensor nodes is used to perform time calibration on the collected dynamic water resources information using a time synchronization protocol, and the geographic reference coordinates of the measurement data are unified through a spatial position correction model between nodes. S14. Regarding multi-source observation data in dynamic water resources information Based on the prior state of the environment Establish a conditional probability model , indicating that given a prior state The reliability of each sensor data point is assessed, and the dynamic weight of each observation is calculated using Bayesian inference methods. : ; In the formula, For the first Dynamic weights of each observation data point This represents the confidence probability of the observed data under the prior state. The total number of sensors participating in the collaborative processing; Through a dynamic update mechanism, adjustments are made in real time based on the accuracy of historical sensor data, the current ambient noise level, and the sensor's operating status. Value: ; In the formula, Indicates the first The reliability score of each sensor This represents the value before the prior state was updated; Based on the calculated dynamic weights All observation data are weighted and fused to generate the fused collaborative processing result: ; In the formula, For the fused observation data, For the first Observational data from each sensor.

[0023] S2. Configure edge computing devices at water conservancy node terminals to transmit the collected dynamic water resources information to the edge computing devices, perform data cleaning, outlier removal and format standardization on the dynamic water resources information, and extract key feature data based on the distributed processing architecture. Specifically, step S2 includes the following steps: S21. The dynamic water resources information collected through the distributed sensor network is transmitted in the form of data frames to the edge computing device configured in the water conservancy node terminal. The data frame includes timestamp, sensor identifier and dynamic parameter value. S22. In the edge computing device, the received dynamic information is cleaned to remove incomplete data frames and abnormal data exceeding the threshold range. S23. Standardize the format of the cleaned data frame and convert the dynamic information into structured data containing multi-dimensional attribute fields based on a predefined unified data model. The multi-dimensional attribute fields include timestamps, spatial coordinates, dynamic parameter values ​​and sensor identifiers. S24. Based on a distributed processing architecture, the standardized structured data is divided into multiple data partitions. The data partitioning rules are set according to spatial range and time period, and the partition index is... for: ; In the formula, For partitioned indexes, Represents the spatial coordinates of the data. For timestamps, , and These represent the spatial and temporal partitioning intervals, respectively. and These represent the number of spatial partitions; S25. Within each data partition, based on the dynamic parameter set Key feature data are extracted using a higher-order difference method: ; In the formula, For the rate of dynamic change, It is the difference order. The sampling time interval, It is the number of combinations.

[0024] S3. Utilize edge computing devices and combine them with dynamic optimization models to analyze key feature data, generate prediction results of water resource status, extract feature indicators related to decision-making, and save the prediction results in local storage units. Specifically, step S3 includes the following steps: S31. Load a dynamic optimization model into the edge computing device and analyze the extracted key feature data, which includes a set of dynamic parameters and corresponding timestamps. S32. When constructing water resource state prediction equations based on dynamic optimization models, time-series prediction algorithms are used to analyze the dynamic parameter set. Modeling was performed, using an autoregressive integral moving average model to fit the time series data and predict values. for: ; In the formula, For predicted values, and These are the parameters for the autoregressive and moving average models, respectively. and The model order is... For prediction error, As a bias term, the model parameters are optimized by fitting historical time series data. , and During the forecasting process, a sliding window technique is introduced to update the time series data in real time, retaining only the window length at a time. Latest data ; S33, Regarding the prediction results By comparing the data with real-time acquired data, error correction methods are used to adjust the parameters of the dynamically optimized model to address prediction errors. To minimize the objective, a parameter optimization process based on gradient descent is constructed: ; By calculating the error function, the model parameters are... , and The gradient is used to optimize the parameters using an iterative update rule: ; ; ; In the formula, For learning rate, , , These are the partial derivatives of the prediction error with respect to the model parameters; An adaptive learning rate adjustment mechanism is introduced to dynamically adjust the learning rate based on the current error change rate. Optimize the efficiency and stability of error correction: ; In the formula, To adjust the coefficient, This represents the change in error between two iterations. The sign representing the change in error; S34. Extract decision-related feature indicators from the prediction results, based on the predicted value set. Construct a set of feature extraction functions using dynamically changing parameters. Each feature extraction function Specific indicators are calculated based on different decision-making needs. All calculated feature indicators undergo initial screening. Invalid and redundant features are filtered out based on constraints and priorities set in rule base R, resulting in the final set of key feature indicators. Each of the key metrics Satisfy the rules And store them in order of their decision priority; S35. Store the analyzed and screened set of feature indicators and the prediction results of water resource status to the local storage unit of the edge computing device.

[0025] S4. Based on the real-time collected feature indicators and stored prediction results, an adaptive algorithm is applied in conjunction with a set rule base to generate local management decision strategies, and the locally generated local management decision strategies are transmitted to the centralized management system through a communication network. Specifically, step S4 includes the following steps: S41. Feature index set based on real-time acquisition and the collection of stored analysis results Initialize the input parameters of the adaptive algorithm, including the indicator weight vector. and optimization objective function ; S42, Combining the rule base The decision-making rules in the model are used to construct local management decision objective functions through a multi-objective optimization model. ; S421, Combining the rule base The decision rules in the system determine the set of characteristic indicators. and analysis result set Initialize indicator weights Weight The constraints are satisfied: ; S422, Calculate characteristic indicators With analysis results Matching degree function : ; In the formula, The scaling parameter of the matching degree function is used to adjust the matching sensitivity of the feature indicators and analysis results; S423, Based on the rule base Constraints set in Define constraint function for: ; In the formula, These are the weighting coefficients. For the first Constraints The deviation function measures the degree to which the current state satisfies the constraints. S424, Matching degree function and constraint function Substitute into the local management decision objective function : ; In the formula, To constrain the penalty coefficient, the rule base is dynamically adjusted to balance the weight between the matching degree and the constraint conditions; S425. Update the parameters in the objective function based on real-time feedback, including dynamically adjusting the indicator weights. and constraint penalty coefficient , so that the objective function The calculations are consistent with real-time data on the current water resource status; S43. Using a genetic algorithm to determine the objective function of local management decisions. Optimize; S431. Initialize the population Each individual Let represent a set of candidate decision parameters, with a population size of . Individual codes are represented using real numbers, and each code parameter corresponds to a local management decision objective function. Input variables; S432, Based on the objective function For each individual in the population Fitness evaluation, fitness function This indicates an individual's performance in optimizing their goals: ; In the formula, For the objective function in individuals The value on, This is the current optimal target value; S433. Based on fitness values, a selection operation is performed on the population. A roulette wheel selection method is used to determine the parent individuals of the next generation, with selection probabilities... for: ; In the formula, For individuals fitness value, The sum of the fitness values ​​of all individuals; S434. Perform a crossover operation on the selected parent individuals using a single-point crossover method, randomly swapping the codes of the parent individuals to generate candidate solutions for the next generation. The crossover probability is... Control the occurrence rate of crossover operations; S435. Perform a mutation operation on the individuals generated after crossover by randomly changing a parameter value in the individual's encoding to generate a new solution space. The mutation probability... Controlling the occurrence rate of mutation operations: ; In the formula, For the amplitude of variation, In the interval Randomly generated values; S436. Evaluate the fitness of the mutated population and select the individual with the highest fitness value as the optimal solution for the current population, recording it as... And update the objective function. Optimal value .

[0026] S44. Optimize the set of local management decision parameters obtained by the genetic algorithm. Mapped to specific management decision-making strategies The mapping process is based on a rule base. The strategy generation rules in the middle will generate each optimal parameter value Converted into specific management actions or control instructions, the mapped set of management decision strategies is represented as follows: Each strategy Parameters corresponding to the objective function Related; S45. Dynamically adjust the rule base based on real-time environmental feedback data. Constraints in and indicator weights Through real-time feedback data Calculate the execution deviation of the current management strategy. The deviation is introduced into the rule base, and the constraint function is dynamically updated. and the weights of the objective function Based on the rate of change of the feedback data The learning rate used in the real-time adjustment and optimization process and constraint penalty coefficient ; S46. Generate local management decision strategies The policies are stored in the local storage unit of the edge computing device and transmitted to the centralized management system via a communication network.

[0027] S5. In the centralized management system, use multi-objective optimization algorithms to verify and adjust local management decision strategies, and generate global management decision strategies that meet optimization requirements. Specifically, step S5 includes the following steps: S51, Receive the set of local management decision strategies and real-time collected global water resource dynamic data For local strategies With global data Perform a consistency check; S52. Combining global optimization requirements, construct a global management decision objective function based on a multi-objective optimization algorithm. : ; In the formula, Representing local strategies Fitness under globally dynamic data For the weights of the local strategy, Representing global constraints The deviation function, This is the global constraint penalty coefficient; S53. Using a multi-objective optimization algorithm to optimize the global management decision objective function. Optimize; S54. Optimize the global management decision parameter set. Mapped to specific global management decision-making strategies Each strategy From optimization parameters It was converted from; S55, Generate the global management decision strategy Distributed to relevant water conservancy equipment and node terminals via communication networks.

[0028] S6. Transmit the optimized global management decision strategy to the execution terminal equipment, control the relevant water conservancy equipment to implement the management decision, and monitor the data changes in the implementation process in real time through the sensor network, and record the execution results and environmental response.

[0029] The intelligent water resources information management and decision support method in this embodiment makes full use of edge computing technology, distributed intelligent sensing network and multi-objective optimization algorithm, and better realizes the management and decision support of dynamic water resources monitoring, real-time data processing and adaptive decision generation. It has the advantages of strong real-time performance, high data processing efficiency, strong scientific decision support and strong environmental adaptability.

[0030] More specifically, the beneficial effects that this intelligent water resources information management and decision support method can achieve are: (1) By combining edge computing technology, distributed intelligent sensing network and dynamic adaptive algorithm, the system realizes real-time collection and accurate analysis of dynamic water resources information, enabling it to adapt to complex water resources changes and emergencies in real time, and significantly improving the dynamic response capability of water resources management. Especially in scenarios such as flood warning and water pollution monitoring, it can quickly generate targeted management strategies, which helps to ensure the timeliness and effectiveness of water resources regulation.

[0031] (2) By combining lightweight algorithms with multi-objective optimization models, edge computing devices are fully utilized to efficiently process and extract features from distributed sensor data, and local strategies are globally optimized through a centralized management system. This method effectively reduces the resource consumption of data transmission and centralized processing, and significantly improves the efficiency and accuracy of large-scale dynamic data processing.

[0032] (3) By using dynamic optimization models and multi-objective optimization algorithms, combined with a rule base, a scientific decision support system is constructed, realizing the collaborative optimization of management strategies from local to global perspectives. The system can adjust optimization parameters and weights in real time, dynamically generating management decision strategies that meet global optimization needs, thereby enhancing the scientific nature and adaptability of water conservancy management.

[0033] (4) Through the fusion sensing of multiple types of sensors and high-order feature extraction technology, a comprehensive dynamic monitoring network for water resources has been constructed, which can realize accurate perception and comprehensive analysis of multi-dimensional information such as water level, flow velocity, water quality, and meteorology. The system can automatically identify and eliminate abnormal data, improving the reliability and comprehensiveness of decision support information.

[0034] Example 1: To verify the feasibility of this invention, it was applied to the intelligent management system of a key provincial water conservancy project. This project covers a large area of ​​rivers, reservoirs, and irrigation networks, involving complex dynamic water resource management needs. In recent years, the project's operation and maintenance department has found that existing water resource monitoring and control systems suffer from slow response, low data processing efficiency, and insufficient decision support. Especially in critical scenarios such as flood warnings and optimized water resource scheduling, traditional methods struggle to meet the requirements of real-time performance and intelligence. Therefore, the project team decided to deploy the edge computing-based intelligent management and decision support method for water conservancy information proposed in this invention.

[0035] During implementation, the system first deployed various types of distributed sensors within the monitored area, including water level sensors, flow velocity sensors, water quality monitors, and weather stations, constructing a distributed intelligent sensing network. These sensor nodes collect and preprocess dynamic data in real time through edge computing devices, and perform data cleaning, anomaly removal, and format standardization using preset lightweight algorithms. After data processing, the system uses a dynamic optimization model to extract key feature data and stores it in local storage units, providing high-quality foundational data for subsequent decision support.

[0036] After deployment, the system was validated in a flood control scenario triggered by a rainstorm. During the rainstorm, real-time data collected by the sensor network showed a rapid rise in water level and flow velocity. Edge computing devices preprocessed and dynamically analyzed this data, extracting a set of characteristic indicators. Combining this with historical data, the system employed a dynamic adaptive algorithm to generate local control strategies, including the order of flood diversion zone opening, adjustments to reservoir discharge, and priorities for downstream dike reinforcement.

[0037] In the centralized management system, local strategies are globally verified and adjusted using a multi-objective optimization algorithm, generating a global control scheme covering upstream, midstream, and downstream areas. During the optimization process, the objective function design fully considers multi-dimensional objectives such as flood control safety, water resource utilization efficiency, and environmental impact, while constraints are dynamically adjusted based on the actual carrying capacity and operational rules of the region. The final global decision strategy is distributed to relevant water conservancy equipment, including reservoir gates, pumping stations, and flood diversion gates, through a communication network to execute control commands. The system's performance was fully validated during this rainstorm and flood control operation. The detailed data report is shown in Table 1.

[0038] Table 1 Comparison Report on Flood Control Effects of a Key Provincial Water Conservancy Project Data categories describe Before deploying the system After deploying the system Regulation response time Time from data acquisition to the generation of control strategies 25 minutes 3 minutes Control scheme precision Matching degree between water resource allocation and flood control areas in the regulation plan 78.6% 96.8% Data processing efficiency The amount of dynamic data that can be processed per minute (in records) 10,000 100,000 Decision support accuracy The proportion of management strategies generated by the system that meet actual needs 81.3% 97.2% Water resource loss reduction rate Water loss reduced compared to traditional methods - 34.5% Flood risk reduction rate Reduced flood-affected area compared to traditional methods - 42.7% As shown in Table 1, after deploying the system, the control response time was reduced from 25 minutes to 3 minutes, significantly improving the system's real-time performance. Regarding the accuracy of water resource control schemes, the optimized global control scheme achieved a 96.8% match with actual needs, an improvement of 18.2 percentage points compared to traditional methods. Simultaneously, data processing efficiency was significantly improved, processing 100,000 dynamic data points per minute, ten times that of traditional methods. Furthermore, the decision support accuracy increased from 81.3% to 97.2%, and the area affected by flood disaster risk decreased by 42.7%, significantly enhancing the scientific rigor and safety of water resource management.

[0039] Through the deployment of this invention, the water conservancy project has seen significant improvements in its ability to respond to emergencies, resource utilization efficiency, and the level of intelligent decision-making. The practical application of the system has demonstrated the advantages of this invention in terms of real-time performance, efficiency, and decision support, successfully addressing many shortcomings of existing technologies and showcasing broad application prospects.

[0040] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for intelligent management and decision support of water conservancy information, characterized in that, Includes the following steps: S1. Construct a distributed sensor network to collect dynamic water resource information at water conservancy equipment and node terminals, and use multi-source data fusion technology to synchronously calibrate the sensor information. S2. Configure edge computing devices at water conservancy node terminals, transmit the collected dynamic water resources information to the edge computing devices, perform data cleaning, outlier removal and format standardization on the dynamic water resources information, and extract key feature data based on a distributed processing architecture. S3. Using the edge computing device, combined with the dynamic optimization model, analyze the key feature data, generate prediction results of water resource status, extract feature indicators related to decision-making, and save the prediction results in the local storage unit. S4. Based on the real-time collected feature indicators and stored prediction results, an adaptive algorithm is applied in conjunction with a set rule base to generate local management decision strategies, and the locally generated local management decision strategies are transmitted to the centralized management system through a communication network. S5. In the centralized management system, a multi-objective optimization algorithm is used to verify and adjust the local management decision-making strategy to generate a global management decision-making strategy that meets the optimization requirements. S6. Transmit the optimized global management decision strategy to the execution terminal equipment, control the relevant water conservancy equipment to implement the management decision, and monitor the data changes in the implementation process in real time through the sensor network, and record the execution results and environmental response.

2. The intelligent management and decision support method for water conservancy information according to claim 1, characterized in that, Step S1 specifically includes the following steps: S11. Deploy intelligent sensing nodes that support distributed collaboration in water conservancy equipment and node terminals to collect dynamic information on water resources. The intelligent sensing nodes include water level sensors, flow velocity sensors, water quality sensors and meteorological sensors. S12. When collecting dynamic information on water resources, calculate the rate of change of dynamic parameters of water resources in the monitoring environment in real time. Based on the calculated rate of change Combined with adjustment coefficient and minimum sampling period Dynamically determine the sampling period : ； In the formula, For dynamic parameter functions, These are time-varying parameters, including water level, flow velocity, and water quality indicators. Indicates the minimum sampling period; Setting a threshold range for adjusting the sampling period. Limit the range of change between two consecutive sampling periods: ； In the formula, and They represent the first The sampling period and the first Each sampling period The preset range of variation threshold; S13. A spatiotemporal coordination mechanism based on sensor nodes is used to perform time calibration on the collected dynamic water resources information using a time synchronization protocol, and the geographic reference coordinates of the measurement data are unified through a spatial position correction model between nodes. S14. For multi-source observation data in water resource dynamic information, calculate the dynamic weight of each observation data. And based on dynamic weights The results are generated by merging and integrating observation data to produce collaborative processing results.

3. The intelligent management and auxiliary decision-making method for water conservancy information according to claim 2, characterized in that, Step S14 specifically includes the following steps: Multi-source observation data in water resource dynamic information Based on the prior state of the environment Establish a conditional probability model , indicating that given a prior state The reliability of each sensor data point is assessed, and the dynamic weight of each observation is calculated using Bayesian inference methods. : ； In the formula, For the first Dynamic weights of each observation data point This represents the confidence probability of the observed data under the prior state. The total number of sensors participating in the collaborative processing; Through a dynamic update mechanism, adjustments are made in real time based on the accuracy of historical sensor data, the current ambient noise level, and the sensor's operating status. Value: ； In the formula, Indicates the first The reliability score of each sensor This represents the value before the prior state was updated; Based on the calculated dynamic weights All observation data are weighted and fused to generate the fused collaborative processing result: ； In the formula, For the fused observation data, For the first Observational data from each sensor.

4. The intelligent management and decision support method for water conservancy information according to any one of claims 1-3, characterized in that, Step S2 specifically includes the following steps: S21. The dynamic water resources information collected through the distributed sensor network is transmitted in the form of data frames to the edge computing device configured in the water conservancy node terminal. The data frames include timestamps, sensor identifiers and dynamic parameter values. S22. In the edge computing device, the received dynamic information is cleaned to remove incomplete data frames and abnormal data exceeding the threshold range. S23. Standardize the format of the cleaned data frame and convert the dynamic information into structured data containing multi-dimensional attribute fields based on a predefined unified data model. The multi-dimensional attribute fields include timestamps, spatial coordinates, dynamic parameter values ​​and sensor identifiers. S24. Based on a distributed processing architecture, the standardized structured data is divided into multiple data partitions. The data partitioning rules are set according to spatial range and time period, and the partition index is... for: ； In the formula, For partitioned indexes, Spatial coordinates representing data, For timestamps, , and These represent the spatial and temporal partitioning intervals, respectively. and These represent the number of spatial partitions; S25. Within each data partition, based on the dynamic parameter set Key feature data are extracted using a higher-order difference method: ； In the formula, For the rate of dynamic change, Let be the difference order. The sampling time interval, It is the combination number.

5. The intelligent management and decision support method for water conservancy information according to any one of claims 1-3, characterized in that, Step S3 specifically includes the following steps: S31. Load a dynamic optimization model into the edge computing device and analyze the extracted key feature data, which includes a set of dynamic parameters and corresponding timestamps. S32. When constructing water resource state prediction equations based on dynamic optimization models, time-series prediction algorithms are used to analyze the dynamic parameter set. Modeling was performed, using an autoregressive integral moving average model to fit the time series data and predict values. for: ； In the formula, For predicted values, and These are the parameters for the autoregressive and moving average models, respectively. and The model order is... For prediction error, As a bias term, the model parameters are optimized by fitting historical time series data. , and During the forecasting process, a sliding window technique is introduced to update the time series data in real time, retaining only the window length at a time. Latest data ; S33, Regarding the prediction results By comparing the data with real-time data, the parameters of the dynamic optimization model are adjusted using error correction methods; S34. Extract decision-related feature indicators from the prediction results, based on the predicted value set. Construct a set of feature extraction functions using dynamically changing parameters. Each feature extraction function Specific indicators are calculated based on different decision-making needs. All calculated feature indicators undergo initial screening. Invalid and redundant features are filtered out based on constraints and priorities set in rule base R, resulting in the final set of key feature indicators. Each of the key metrics Satisfy the rules And store them in order of their decision priority; S35. Store the analyzed and screened set of feature indicators and the prediction results of water resource status to the local storage unit of the edge computing device.

6. The intelligent management and auxiliary decision-making method for water conservancy information according to claim 5, characterized in that, Step S33 specifically includes the following steps: Regarding prediction error To minimize the objective, a parameter optimization process based on gradient descent is constructed: ； By calculating the error function, the model parameters are... , and The gradient is used to optimize the parameters using an iterative update rule: ； ； ； In the formula, For learning rate, , , These are the partial derivatives of the prediction error with respect to the model parameters; An adaptive learning rate adjustment mechanism is introduced to dynamically adjust the learning rate based on the current error change rate. Optimize the efficiency and stability of error correction: ； In the formula, To adjust the coefficient, This represents the change in error between two iterations. Symbols representing changes in error.

7. The intelligent management and auxiliary decision-making method for water conservancy information according to any one of claims 1-3 or 6, characterized in that, Step S4 specifically includes the following steps: S41. Feature index set based on real-time acquisition and the collection of stored analysis results Initialize the input parameters of the adaptive algorithm, including the indicator weight vector. and optimization objective function ; S42, Combining the rule base The decision-making rules in the model are used to construct local management decision objective functions through a multi-objective optimization model. ; S43. Using a genetic algorithm to determine the objective function of local management decisions. Optimize; S44. Optimize the set of local management decision parameters obtained by the genetic algorithm. Mapped to specific management decision-making strategies The mapping process is based on a rule base. The strategy generation rules in the middle will generate each optimal parameter value Converted into specific management actions or control instructions, the mapped set of management decision strategies is represented as follows: Each strategy corresponds to a parameter in the objective function. Related; S45. Dynamically adjust the rule base based on real-time environmental feedback data. Constraints in and indicator weights Through real-time feedback data Calculate the execution deviation of the current management strategy. The deviation is introduced into the rule base, and the constraint function is dynamically updated. and the weights of the objective function Based on the rate of change of the feedback data The learning rate used in the real-time adjustment and optimization process and constraint penalty coefficient ; S46. Generate local management decision strategies The policies are stored in the local storage unit of the edge computing device and transmitted to the centralized management system via a communication network.

8. The intelligent management and auxiliary decision-making method for water conservancy information according to claim 7, characterized in that, Step S42 specifically includes the following steps: Combined with rule base The decision rules in the system determine the set of characteristic indicators. and analysis results set Initialize indicator weights Weight The constraints are satisfied: ； Calculate characteristic indicators With analysis results Matching degree function : ； In the formula, The scaling parameter of the matching degree function is used to adjust the matching sensitivity of the feature indicators and analysis results; Based on the rule base Constraints set in Define constraint function for: ； In the formula, These are the weighting coefficients. For the first Constraints The deviation function measures the degree to which the current state satisfies the constraints. Matching function and constraint function Substitute into the local management decision objective function : ； In the formula, To constrain the penalty coefficient, the rule base is dynamically adjusted to balance the weight between the matching degree and the constraint conditions; The parameters in the objective function are updated based on real-time feedback, including dynamically adjusting the indicator weights. and constraint penalty coefficient , so that the objective function The calculations are consistent with real-time data on the current water resource status.

9. The intelligent management and auxiliary decision-making method for water conservancy information according to claim 7, characterized in that, Step S43 specifically includes the following steps: Initialize population Each individual Let represent a set of candidate decision parameters, with a population size of . Individual codes are represented using real numbers, and each code parameter corresponds to a local management decision objective function. Input variables; Based on the objective function For each individual in the population Fitness evaluation, fitness function This indicates an individual's performance in optimizing their goals: ； In the formula, For the objective function in individuals The value on, This is the current optimal target value; Selection is performed on the population based on fitness values, using a roulette wheel selection method to determine the parents of the next generation, with selection probabilities... for: ； In the formula, For individuals fitness value, The sum of the fitness values ​​of all individuals; A crossover operation is performed on the selected parent individuals using a single-point crossover method, randomly swapping the codes of the parent individuals to generate candidate solutions for the next generation. The crossover probability is... Control the occurrence rate of crossover operations; The individuals generated after crossover undergo a mutation operation, which randomly changes a parameter value in the individual's encoding to generate a new solution space. The mutation probability... Controlling the occurrence rate of mutation operations: ； In the formula, For the amplitude of variation, For in the interval Randomly generated values; The fitness of the mutated population is evaluated, and the individual with the highest fitness value is selected as the optimal solution for the current population and recorded as [the optimal solution]. And update the objective function. Optimal value .

10. The intelligent management and auxiliary decision-making method for water conservancy information according to claim 7, characterized in that, Step S5 specifically includes the following steps: S51. Receive local management decision-making strategy sets and real-time collected global water resource dynamic data. For local strategies With global data Perform a consistency check; S52. Combining global optimization requirements, construct a global management decision objective function based on a multi-objective optimization algorithm. : ； In the formula, Representing local strategies Fitness under globally dynamic data For the weights of the local strategy, Representing global constraints The deviation function, This is the global constraint penalty coefficient; S53. Using a multi-objective optimization algorithm to optimize the global management decision objective function. Optimize; S54. Optimize the global management decision parameter set. Mapped to specific global management decision-making strategies Each strategy From optimization parameters It was converted from; S55, Generate the global management decision strategy Distributed to relevant water conservancy equipment and node terminals via communication networks.

Images