An unmanned aerial vehicle water resource scheduling scheme data processing system and method

Abstract

The application relates to the technical field of data processing and discloses a kind of unmanned aerial vehicle water resource scheduling scheme data processing system and method. First, initial water resource scheduling parameters are obtained using unmanned aerial vehicles, and initial water resource scheduling parameters are subjected to abnormal parameter data deletion and missing parameter data filling to obtain processed water resource scheduling parameters;Then, water resource scheduling objective functions and water resource scheduling constraints are calculated, and a water resource scheduling model is established;The wolf optimization algorithm is used to adjust and optimize the water resource scheduling model, the fitness function is set, and the optimal fitness function value is obtained by searching to obtain the optimized value of each parameter;The key water resource scheduling parameters are extracted using the grey correlation analysis method, and finally the GRU neural network is trained to predict the flow and adjust the water resource scheduling scheme. The application realizes water resource scheduling scheme data processing by processing water resource scheduling parameters, and the method is accurate and objective.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a data processing system and method for an unmanned aerial vehicle (UAV) water resource scheduling solution. Background Art

[0002] Chinese patent CN113626923B discloses a method for calculating and improving the accuracy of water resource scheduling in a plain river network. The method specifically includes dividing an assessment area for water resource scheduling accuracy, wherein the assessment of water resource scheduling accuracy includes calculating the current water quantity scheduling accuracy, the current water quality scheduling accuracy, and the current average water resource scheduling accuracy in the assessment area; determining the current water resource scheduling plan for the assessment area, then constructing a river network hydrodynamic-water quality coupling model for the assessment area, and formulating a current optimized scheduling plan and a planned optimized scheduling plan for the assessment area; the current optimized scheduling plan and the planned optimized scheduling plan are formulated based on the water system pattern and gate control layout of the assessment area, while ensuring the accuracy of water resource scheduling; and comparing the accuracy improvement ratio before and after the plan based on the water resource scheduling accuracy, thereby achieving the calculation and improvement of water resource scheduling accuracy. However, the data of this invention may contain errors, the data calculation speed is slow, and the accuracy is low.

[0003] Traditional water resource scheduling schemes have brought troubles to water resource scheduling due to insufficient coordination of water resource scheduling and untimely processing of water resource scheduling-related data. During the water resource scheduling process, traditional water resource data processing often has errors, and high-tech technologies such as neural networks are not used, resulting in slow data processing speed. Summary of the Invention

[0004] In response to the problems in the related technology, the present invention provides a drone water resource scheduling plan data processing system and method to overcome the above-mentioned technical problems existing in the existing related technology.

[0005] To solve the above technical problems, the present invention is achieved through the following technical solutions:

[0006] The present invention provides a method for processing data of a UAV water resource scheduling scheme, comprising the following steps:

[0007] S1. Use drones to obtain initial water resource scheduling parameters to form an initial water resource scheduling parameter set. Based on the LOF abnormal data identification algorithm and the improved nearest neighbor filling algorithm, abnormal parameter data are deleted and missing parameter data are filled in the initial water resource scheduling parameter set to obtain a processed water resource scheduling parameter set.

[0008] S2. Obtaining an objective function of a water resource scheduling model based on the processed water resource scheduling parameter set, and establishing a water resource scheduling model in combination with water resource scheduling constraints;

[0009] S3. Obtaining a fitness function through the objective function of the water resource scheduling model, adjusting and optimizing the water resource scheduling model based on the coyote optimization algorithm while satisfying the water resource scheduling constraints, and obtaining an optimized water resource scheduling model by finding the optimal fitness function value;

[0010] S4. Based on the optimized water resource scheduling model, the grey correlation analysis method is used to extract key water resource scheduling parameters, and then the GRU neural network is trained to obtain a GRU neural network model to predict the flow rate and adjust the drone water resource scheduling plan.

[0011] The invention uses a drone to obtain initial water resource scheduling parameters, uses the LOF abnormal data identification algorithm to identify and delete abnormal parameter data in the initial water resource scheduling parameters, and uses the improved nearest neighbor filling algorithm to fill the missing parameter data of the initial water resource scheduling parameters to complete parameter preprocessing; deleting abnormal data and supplementing missing data effectively improves the accuracy of subsequent data processing and avoids interference; secondly, using the processed water resource scheduling parameters, a water resource scheduling model is established based on the objective function of the water resource scheduling model and the water resource scheduling constraints; according to the supply and demand balance of water resource scheduling, it is conducive to the coordinated scheduling of water resources; then the coyote optimization algorithm is used to adjust and optimize the water resource scheduling model, and the optimal fitness function value is found through continuous iteration to obtain the optimized value of each parameter; the algorithm imitates the environmental adaptation mechanism of the coyote population, continuously iterates to find the optimal solution to the problem, and has strong optimization performance and fast convergence speed; finally, the gray correlation analysis method is used to extract key water resource scheduling parameters, train the GRU neural network, and obtain the GRU neural network model. By predicting the flow, the water resource scheduling plan can be adjusted in advance; the neural network data prediction is accurate and the plan adjustment has high credibility.

[0012] Preferably, the S1 comprises the following steps:

[0013] S11. UAVs carrying laser scanning equipment, hyperspectral sensors and other equipment obtain water resource-related data of reservoirs and surrounding basins, including reservoir capacity, reservoir water level, flow, etc., which are recorded as the initial water resource scheduling parameter set. ,in Indicates the m A data set consisting of water resource scheduling parameters;

[0014] The LOF (local outlier factor) outlier data identification algorithm is used to identify and delete outlier parameter data in the initial water resource scheduling parameter set to obtain a processed water resource scheduling parameter set. The specific steps are as follows:

[0015] S111, select the first The data set consisting of water resource scheduling parameters is recorded as the first parameter data set; n dimensional space, mapping the first parameter data set to n In the k-dimensional space, any parameter data point is selected from the first parameter data set and recorded as the initial parameter data point. The Euclidean distance between the initial parameter data point and other parameter data points in the first parameter data set is calculated, and the Euclidean distances between the initial parameter data point and other parameter data points in the first parameter data set are sorted in ascending order to obtain a Euclidean distance sorted set. The k-th distance is set, and the parameter data points in the Euclidean distance sorted set whose Euclidean distances are less than or equal to the k-th distance are added to the k-th domain of the initial parameter data point. The local reachable density calculation formula of the initial parameter data point is as follows:

[0016] ;

[0017] in, represents the local reachable density of the initial parameter data points, b represents the number of parameter data points in the kth region of the initial parameter data point, a represents the initial parameter data point, Indicates the initial parameter data point in the kth field The Euclidean distance from the parameter data point to the initial parameter data point, ;

[0018] S112, calculate the local reachable density of the parameter data points in the kth domain of the initial parameter data points, obtain the local reachable density set, set Indicates the The local reachable density of parameter data points and , then the local outlier factor of the initial parameter data point is The calculation formula is as follows:

[0019] ;

[0020] If the local outlier factor of the initial parameter data point is less than or equal to 1, the initial parameter data point is normal parameter data; otherwise, the initial parameter data point is abnormal parameter data, the abnormal parameter data is deleted, and the initial water resource scheduling parameter set repeats S111 and S112 to obtain a processed water resource scheduling parameter set;

[0021] S12, select the first The data set consisting of water resource scheduling parameters is recorded as the second parameter data set. dimensional space, mapping the second parameter data set to dimensional space; select any parameter data point in the second parameter data set, record it as a sample parameter data point, select the missing parameter data point in the four quadrants The Euclidean distance between the parameter data points and the sample parameter data points in the associated parameter data point set is calculated. Indicates the first parameter in the set of sample parameter data points to the associated parameter data points. The Euclidean distance of each parameter data point is used to assign weights to missing parameter data points. The calculation formula is as follows:

[0022] ;

[0023] in, Indicates the first parameter in the set of sample parameter data points to the associated parameter data points. The weight coefficient of each parameter data point, ;

[0024] Calculate the first The attribute value corresponding to the parameter data point , combined with the weight coefficient, we can get the missing parameter data B , the calculation formula is as follows:

[0025] ;

[0026] The missing parameter data is filled into the second parameter data set, and S12 is repeated until all the missing parameter data of the processed water resource scheduling parameter set are filled, thereby obtaining a processed water resource scheduling parameter set.

[0027] This invention uses drones to obtain initial water resource scheduling parameters, uses the LOF abnormal data identification algorithm to delete abnormal parameter data in the initial water resource scheduling parameters, and uses an improved nearest neighbor filling algorithm to fill in missing parameter data in the initial water resource scheduling parameters. Deleting abnormal data and supplementing missing data effectively improves the accuracy of subsequent data processing and avoids interference.

[0028] Preferably, said S2 comprises the following steps:

[0029] S21. According to the processed water resource scheduling parameter set, the total number of water demand areas is obtained as follows: , the number of reservoirs is d , set the Water demand area The water demand for the time period is , No. Water demand area The actual water supply during the time period is , No. Reservoir No. The water level in the time period is ,exist t During the time period, the minimum water supply shortage is taken as the objective function , the maximum sum of reservoir water levels is used as the objective function , the calculation formula is as follows:

[0030] ,

[0031] ;

[0032] S22. Construct water resource scheduling constraints based on the processed water resource scheduling parameter set. The specific steps are as follows:

[0033] S221. According to the processed water resource scheduling parameter set, the first Reservoir No. The reservoir capacity at the beginning of the time period is , No. Reservoir No. The reservoir capacity at the end of the time period is , No. Reservoir No. The inflow rate during the time period is , No. Reservoir No. The outflow volume during the time period is , then in t During the time period, the reservoir water balance constraint is established, and the calculation formula is as follows:

[0034] ;

[0035] S222, set the Reservoir No. The maximum reservoir capacity at the end of the time period is , No. Reservoir No. The minimum reservoir capacity at the end of the time period is , establish reservoir capacity constraints ; Set the Reservoir No. The maximum outflow rate during the time period is , No. Reservoir No. The minimum outflow flow in the time period is , establish reservoir flow constraints ; Set the Reservoir No. The maximum reservoir water level during the time period is , No. Reservoir No. The minimum reservoir water level during the time period is , establish reservoir water level constraints ;

[0036] S23. Taking the minimum water supply shortage and the maximum reservoir water level as the objective function of water resource scheduling, a water resource scheduling model is established in combination with reservoir capacity constraints, reservoir flow constraints and reservoir water level constraints.

[0037] Preferably, the step S3 includes the following steps:

[0038] S31, according to the water resources scheduling model, assign the objective function and the objective function Weight, get the water resource scheduling objective function, use the water resource scheduling objective function as the fitness function, and satisfy the water resource scheduling constraint condition; set the coyote population size to , the individual dimensions of coyotes in the coyote subpopulation are E， The individual dimensions of the coyotes in the coyote subpopulation are equal to the number of water resource scheduling parameters in the water resource scheduling model, generating the initial population matrix as follows:

[0039] ;

[0040] in, Indicates the first D The dimensions are E of coyote individuals;

[0041] In the search space, set the first e Coyote individuals in g The upper limit of the dimension is , the first among the coyote subpopulations e Coyote individuals in g The lower limit of the dimension is , represents a random number and , then the first e Initial positions of individual coyotes The calculation formula is as follows:

[0042] ;

[0043] The ability of individual coyotes in a coyote subpopulation to adapt to the environment is called fitness, which is expressed using a fitness function;

[0044] S32. The coyote subpopulation is in the growth stage. The number of coyote subpopulation is set to G , a leader is selected from the coyote subpopulation, and the social ranking of individual coyotes in the coyote subpopulation is recorded asH , and coyotes are influenced by population culture during their growth stage, the formula for calculating the cultural trend of coyote subpopulations is as follows:

[0045] ;

[0046] in, Indicates the cultural trends of coyote subpopulations, Indicates the coyote subpopulation The social ranking of individual coyotes, Indicates the coyote subpopulation The social ranking of individual coyotes;

[0047] Random coyote individuals are selected from the coyote subpopulation and random coyote individuals , set a random coyote individual The impact factor is , random coyote individuals The impact factor is , random coyote individuals The location is , random coyote individuals The location is , represents a random number and , represents a random number and , then the first e Coyote individual locations The calculation formula is as follows:

[0048] , , ;

[0049] Set the current number of iterations to m , the first e Coyote individuals m The iteration position is recorded as , the first among the coyote subpopulations e Coyote individuals m The fitness function value corresponding to the iteration position is , the first among the coyote subpopulations e Coyote individuals m The fitness function value corresponding to the +1 iteration position is , then the first e Coyote individuals m +1 iteration position The calculation formula is as follows:

[0050] ;

[0051] S33, in the elimination stage of the coyote sub-population, random coyote individuals are selected from the coyote sub-population. and random coyote individuals As the first subpopulation of coyotes e The parent coyote of the coyote individual, the random coyote individual The dimension is , random coyote individuals The dimension is , express Random numbers in dimension, random coyote individuals The location is , random coyote individuals The location is , represents the association probability, represents discrete probability, represents a random number and , then the elimination probability The calculation formula is as follows:

[0052] ;

[0053] Under the elimination probability, the coyote individuals are eliminated according to the fitness function values ​​corresponding to the coyote individuals in the coyote sub-population, and when the fitness function value corresponding to the newborn coyote individual is the smallest, the newborn coyote individual is eliminated; when in each iteration, the fitness function value corresponding to the coyote individual in the coyote sub-population is less than the fitness function value corresponding to the newborn coyote individual, the coyote individual in the coyote sub-population is eliminated;

[0054] S34, as the number of iterations increases, the position of the coyote individual in the coyote subpopulation is updated according to the best fitness function value of each iteration, and the maximum number of iterations is set to M. When the current number of iterations reaches the maximum number of iterations, the iteration is stopped to obtain the final position of the coyote individual. E The coordinate values ​​of the dimensional positions correspond to the water resources scheduling parameters in the water resources scheduling model, and the optimized water resources scheduling model is obtained.

[0055] This invention uses the coyote optimization algorithm to adjust and optimize the water resource scheduling model, and obtains the optimized values ​​of each parameter and the optimization purpose through continuous iteration to find the best fitness function value; this algorithm imitates the living environment and elimination mechanism of the coyote population, and searches for the optimal solution to the problem in the search space according to the location of the coyotes, with fast convergence speed and strong optimization performance.

[0056] Preferably, the S4 comprises the following steps:

[0057] S41. Use the grey correlation analysis method to calculate the grey correlation degree of the water resource scheduling parameters in the optimized water resource scheduling model and extract the key water resource scheduling parameters. The specific steps are as follows:

[0058] S411. Generate a reference set based on the water resource scheduling parameters in the optimized water resource scheduling model, set a comparison set, perform dimensionless processing on the reference set using the mean method to obtain a dimensionless reference set; calculate a correlation coefficient between the comparison set and the dimensionless reference set to obtain a correlation coefficient corresponding to the water resource scheduling parameters in the reference set;

[0059] S412. Average the correlation coefficients corresponding to the water resource scheduling parameters in the reference set to obtain the grey correlation degrees of the comparison set and the reference set, sort the grey correlation degrees in descending order, select the top five grey correlation degrees, and obtain the water resource scheduling parameters corresponding to the top five grey correlation degrees, which are recorded as the key water resource scheduling parameter set;

[0060] S42. Obtain water resource scheduling parameters from previous years, build a new water resource scheduling model, obtain a sample water resource scheduling parameter set, normalize the sample water resource scheduling parameter set to obtain a normalized sample set, and divide the normalized sample set into a sample training set and a sample test set; set the number of neurons in the input layer of the GRU (Gated Recurrent Unit) neural network to 5, the number of neurons in the output layer to 1, select random gradient descent for optimization, input the sample training set into the GRU neural network, and iterate continuously until the GRU neural network converges to obtain a trained GRU neural network; then input the sample test set into the trained GRU neural network, and set the error threshold to , when the output result error is less than the error threshold, the GRU neural network model is obtained, otherwise the weight is adjusted until the output result error is less than the error threshold;

[0061] S43. Input the key water resource scheduling parameter set into the GRU neural network model, output the flow prediction value, and use the flow prediction value to adjust the optimized water resource scheduling model in real time to adjust the drone water resource scheduling plan.

[0062] This invention extracts key water resource scheduling parameters, trains the GRU neural network, and obtains a GRU neural network model. By predicting the flow rate, it can achieve early adjustment of the water resource scheduling plan; the gray correlation analysis method can effectively extract more relevant parameters, reduce the amount of data processing, and increase the speed of neural network training; the GRU neural network data prediction is accurate, and the plan adjustment is highly credible.

[0063] The present invention also discloses a system for processing data of a UAV water resource scheduling scheme, which specifically includes: a water resource scheduling parameter processing module, a water resource scheduling model establishment module, a water resource scheduling model optimization module and a water resource scheduling parameter prediction module;

[0064] The water resource scheduling parameter processing module is used to delete and fill in abnormal parameter data and missing parameter data in the water resource scheduling parameters respectively;

[0065] The water resource scheduling model establishment module is used to establish a water resource scheduling model based on the objective function and constraint conditions of the water resource scheduling model;

[0066] The water resource scheduling model optimization module is used to adjust and optimize the water resource scheduling model using the coyote optimization algorithm;

[0067] The water resource scheduling parameter prediction module is used to train the GRU neural network and predict the flow rate to adjust the UAV water resource scheduling plan.

[0068] The present invention has the following beneficial effects:

[0069] 1. This invention effectively improves the accuracy of subsequent data processing and avoids interference by deleting abnormal parameter data in the initial water resource scheduling parameters and filling the missing parameter data in the initial water resource scheduling parameters.

[0070] 2. The invention establishes a water resource scheduling model based on the objective function and water resource scheduling constraints of the water resource scheduling model, establishes a supply and demand balance, and is conducive to the coordinated scheduling of water resources.

[0071] 3. This invention adjusts and optimizes the water resource scheduling model by using the coyote optimization algorithm. By imitating the living environment and elimination mechanism of the coyote population, it searches for the optimal solution to the problem in the search space according to the location of the coyotes. It has a fast convergence speed and strong optimization performance.

[0072] 4. This invention uses the grey correlation analysis method to extract key water resource scheduling parameters according to the degree of correlation between parameters, which greatly reduces the amount of data and speeds up data processing.

[0073] 5. This invention uses the GRU neural network model to predict flow and achieve early adjustment of water resource scheduling plans. The GRU neural network data prediction is accurate and the plan adjustment is highly credible.

[0074] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying any creative work.

[0076] Figure 1 This is a schematic diagram of the flow of water resource scheduling plan data processing by a UAV water resource scheduling plan data processing system provided by the present invention. DETAILED DESCRIPTION

[0077] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0078] In the description of the present invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0079] Example 1

[0080] The present invention provides a method for processing data of a UAV water resource scheduling scheme, comprising the following steps:

[0081] S1. Use drones to obtain initial water resource scheduling parameters to form an initial water resource scheduling parameter set. Based on the LOF abnormal data identification algorithm and the improved nearest neighbor filling algorithm, abnormal parameter data are deleted and missing parameter data are filled in the initial water resource scheduling parameter set to obtain a processed water resource scheduling parameter set.

[0082] Said S1 comprises the following steps:

[0083] S11. UAVs carrying laser scanning equipment, hyperspectral sensors and other equipment obtain water resource-related data of reservoirs and surrounding basins, including reservoir capacity, reservoir water level, flow, etc., which are recorded as the initial water resource scheduling parameter set. ,in Indicates the m A data set consisting of water resource scheduling parameters;

[0084] The LOF abnormal data identification algorithm is used to identify and delete abnormal parameter data in the initial water resource scheduling parameter set to obtain a processed water resource scheduling parameter set. The specific steps are as follows:

[0085] S111, select the first The data set consisting of water resource scheduling parameters is recorded as the first parameter data set; n dimensional space, mapping the first parameter data set to n In the k-dimensional space, any parameter data point is selected from the first parameter data set and recorded as the initial parameter data point. The Euclidean distance between the initial parameter data point and other parameter data points in the first parameter data set is calculated, and the Euclidean distances between the initial parameter data point and other parameter data points in the first parameter data set are sorted in ascending order to obtain a Euclidean distance sorted set. The k-th distance is set, and the parameter data points in the Euclidean distance sorted set whose Euclidean distances are less than or equal to the k-th distance are added to the k-th domain of the initial parameter data point. The local reachable density calculation formula of the initial parameter data point is as follows:

[0086] ;

[0087] in, represents the local reachable density of the initial parameter data points, b represents the number of parameter data points in the kth region of the initial parameter data point, a represents the initial parameter data point, Indicates the initial parameter data point in the kth field The Euclidean distance from the parameter data point to the initial parameter data point, ;

[0088] S112, calculate the local reachable density of the parameter data points in the kth domain of the initial parameter data points, obtain the local reachable density set, set Indicates the The local reachable density of parameter data points and , then the local outlier factor of the initial parameter data point is The calculation formula is as follows:

[0089] ;

[0090] If the local outlier factor of the initial parameter data point is less than or equal to 1, the initial parameter data point is normal parameter data; otherwise, the initial parameter data point is abnormal parameter data, the abnormal parameter data is deleted, and the initial water resource scheduling parameter set repeats S111 and S112 to obtain a processed water resource scheduling parameter set;

[0091] S12, select the first The data set consisting of water resource scheduling parameters is recorded as the second parameter data set. dimensional space, mapping the second parameter data set to dimensional space; select any parameter data point in the second parameter data set, record it as a sample parameter data point, select the missing parameter data point in the four quadrants The Euclidean distance between the parameter data points and the sample parameter data points in the associated parameter data point set is calculated. Indicates the first parameter in the set of sample parameter data points to the associated parameter data points. The Euclidean distance of each parameter data point is used to assign weights to missing parameter data points. The calculation formula is as follows:

[0092] ;

[0093] in, Indicates the first parameter in the set of sample parameter data points to the associated parameter data points. The weight coefficient of each parameter data point, ;

[0094] Calculate the first The attribute value corresponding to the parameter data point , combined with the weight coefficient, we can get the missing parameter data B , the calculation formula is as follows:

[0095] ;

[0096] Filling the missing parameter data into the second parameter data set, repeating S12 until all missing parameter data of the processed water resource scheduling parameter set are filled, thereby obtaining a processed water resource scheduling parameter set;

[0097] S2. Obtaining an objective function of a water resource scheduling model based on the processed water resource scheduling parameter set, and establishing a water resource scheduling model in combination with water resource scheduling constraints;

[0098] The S2 comprises the following steps:

[0099] S21. According to the processed water resource scheduling parameter set, the total number of water demand areas is obtained as follows: , the number of reservoirs is d , set the Water demand area The water demand for the time period is , No. Water demand area The actual water supply during the time period is , No. Reservoir No. The water level in the time period is ,exist t During the time period, the minimum water supply shortage is taken as the objective function , the maximum sum of reservoir water levels is used as the objective function , the calculation formula is as follows:

[0100] ,

[0101] ;

[0102] S22. Construct water resource scheduling constraints based on the processed water resource scheduling parameter set. The specific steps are as follows:

[0103] S221. According to the processed water resource scheduling parameter set, the first Reservoir No. The reservoir capacity at the beginning of the time period is , No. Reservoir No. The reservoir capacity at the end of the time period is , No. Reservoir No. The inflow rate during the time period is , No. Reservoir No. The outflow volume during the time period is , then in t During the time period, the reservoir water balance constraint is established, and the calculation formula is as follows:

[0104] ;

[0105] S222, set the Reservoir No. The maximum reservoir capacity at the end of the time period is , No. Reservoir No. The minimum reservoir capacity at the end of the time period is , establish reservoir capacity constraints ; Set the Reservoir No. The maximum outflow rate during the time period is , No. Reservoir No. The minimum outflow flow in the time period is , establish reservoir flow constraints ; Set the Reservoir No. The maximum reservoir water level during the time period is , No. Reservoir No. The minimum reservoir water level during the time period is , establish reservoir water level constraints ;

[0106] S23, taking the minimum water supply shortage and the maximum reservoir water level as the objective function of water resource scheduling, and establishing a water resource scheduling model in combination with reservoir capacity constraints, reservoir flow constraints, and reservoir water level constraints;

[0107] S3. Obtaining a fitness function through the objective function of the water resource scheduling model, adjusting and optimizing the water resource scheduling model based on the coyote optimization algorithm while satisfying the water resource scheduling constraints, and obtaining an optimized water resource scheduling model by finding the optimal fitness function value;

[0108] The S3 includes the following steps:

[0109] S31, according to the water resources scheduling model, assign the objective function and the objective function Weight, get the water resource scheduling objective function, use the water resource scheduling objective function as the fitness function, and satisfy the water resource scheduling constraint condition; set the coyote population size to , the individual dimensions of coyotes in the coyote subpopulation are E， The individual dimensions of the coyotes in the coyote subpopulation are equal to the number of water resource scheduling parameters in the water resource scheduling model, generating the initial population matrix as follows:

[0110] ;

[0111] in, Indicates the first D The dimensions are E of coyote individuals;

[0112] In the search space, set the first e Coyote individuals in g The upper limit of the dimension is , the first among the coyote subpopulations e Coyote individuals in g The lower limit of the dimension is , represents a random number and , then the first e Initial positions of individual coyotes The calculation formula is as follows:

[0113] ;

[0114] The ability of individual coyotes in a coyote subpopulation to adapt to the environment is called fitness, which is expressed using a fitness function;

[0115] S32. The coyote subpopulation is in the growth stage. The number of coyote subpopulation is set to G , a leader is selected from the coyote subpopulation, and the social ranking of individual coyotes in the coyote subpopulation is recorded as H , and coyotes are influenced by population culture during their growth stage, the formula for calculating the cultural trend of coyote subpopulations is as follows:

[0116] ;

[0117] in, Indicates the cultural trends of coyote subpopulations, Indicates the coyote subpopulation The social ranking of individual coyotes, Indicates the coyote subpopulation The social ranking of individual coyotes;

[0118] Random coyote individuals are selected from the coyote subpopulation and random coyote individuals , set a random coyote individual The impact factor is , random coyote individuals The impact factor is , random coyote individuals The location is , random coyote individuals The location is , represents a random number and , represents a random number and , then the first e Coyote individual locations The calculation formula is as follows:

[0119] , , ;

[0120] Set the current number of iterations to m , the first e Coyote individuals m The iteration position is recorded as , the first among the coyote subpopulations e Coyote individuals m The fitness function value corresponding to the iteration position is , the first among the coyote subpopulations e Coyote individualsm The fitness function value corresponding to the +1 iteration position is , then the first e Coyote individuals m +1 iteration position The calculation formula is as follows:

[0121] ;

[0122] S33, in the elimination stage of the coyote sub-population, random coyote individuals are selected from the coyote sub-population. and random coyote individuals As the first subpopulation of coyotes e The parent coyote of the coyote individual, the random coyote individual The dimension is , random coyote individuals The dimension is , express Random numbers in dimension, random coyote individuals The location is , random coyote individuals The location is , represents the association probability, represents discrete probability, represents a random number and , then the elimination probability The calculation formula is as follows:

[0123] ;

[0124] Under the elimination probability, the coyote individuals are eliminated according to the fitness function values ​​corresponding to the coyote individuals in the coyote sub-population, and when the fitness function value corresponding to the newborn coyote individual is the smallest, the newborn coyote individual is eliminated; when in each iteration, the fitness function value corresponding to the coyote individual in the coyote sub-population is less than the fitness function value corresponding to the newborn coyote individual, the coyote individual in the coyote sub-population is eliminated;

[0125] S34, as the number of iterations increases, the position of the coyote individual in the coyote subpopulation is updated according to the best fitness function value of each iteration, and the maximum number of iterations is set to M. When the current number of iterations reaches the maximum number of iterations, the iteration is stopped to obtain the final position of the coyote individual. E The coordinate values ​​of the dimensional positions correspond to the water resource scheduling parameters in the water resource scheduling model, and the optimized water resource scheduling model is obtained;

[0126] S4. Based on the optimized water resource scheduling model, use the grey correlation analysis method to extract key water resource scheduling parameters, then train the GRU neural network to obtain a GRU neural network model, predict the flow rate, and adjust the drone water resource scheduling plan;

[0127] The S4 comprises the following steps:

[0128] S41. Use the grey correlation analysis method to calculate the grey correlation degree of the water resource scheduling parameters in the optimized water resource scheduling model and extract the key water resource scheduling parameters. The specific steps are as follows:

[0129] S411. Generate a reference set based on the water resource scheduling parameters in the optimized water resource scheduling model, set a comparison set, perform dimensionless processing on the reference set using the mean method to obtain a dimensionless reference set; calculate a correlation coefficient between the comparison set and the dimensionless reference set to obtain a correlation coefficient corresponding to the water resource scheduling parameters in the reference set;

[0130] S412. Average the correlation coefficients corresponding to the water resource scheduling parameters in the reference set to obtain the grey correlation degrees of the comparison set and the reference set, sort the grey correlation degrees in descending order, select the top five grey correlation degrees, and obtain the water resource scheduling parameters corresponding to the top five grey correlation degrees, which are recorded as the key water resource scheduling parameter set;

[0131] S42. Obtain water resource scheduling parameters from previous years, build a new water resource scheduling model, obtain a sample water resource scheduling parameter set, normalize the sample water resource scheduling parameter set to obtain a normalized sample set, and divide the normalized sample set into a sample training set and a sample test set; set the number of neurons in the input layer of the GRU neural network to 5, the number of neurons in the output layer to 1, select random gradient descent for optimization, input the sample training set into the GRU neural network, and iterate continuously until the GRU neural network converges to obtain a trained GRU neural network; then input the sample test set into the trained GRU neural network, and set the error threshold to , when the output result error is less than the error threshold, the GRU neural network model is obtained, otherwise the weight is adjusted until the output result error is less than the error threshold;

[0132] S43. Input the key water resource scheduling parameter set into the GRU neural network model, output the flow prediction value, and use the flow prediction value to adjust the optimized water resource scheduling model in real time to adjust the drone water resource scheduling plan.

[0133] Example 2

[0134] The present invention also discloses a system for processing data of a UAV water resource scheduling scheme, which specifically includes: a water resource scheduling parameter processing module, a water resource scheduling model establishment module, a water resource scheduling model optimization module and a water resource scheduling parameter prediction module;

[0135] The water resource scheduling parameter processing module is used to delete and fill in abnormal parameter data and missing parameter data in the water resource scheduling parameters respectively;

[0136] The water resource scheduling model establishment module is used to establish a water resource scheduling model based on the objective function and constraint conditions of the water resource scheduling model;

[0137] The water resource scheduling model optimization module is used to adjust and optimize the water resource scheduling model using the coyote optimization algorithm;

[0138] The water resource scheduling parameter prediction module is used to train the GRU neural network and predict the flow rate to adjust the UAV water resource scheduling plan.

[0139] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0140] The preferred embodiments of the invention disclosed above are intended only to help illustrate the invention. These preferred embodiments do not exhaust all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for processing data of a UAV water resource scheduling scheme, characterized in that: The steps include: S1. Use drones to obtain initial water resource scheduling parameters to form an initial water resource scheduling parameter set. Based on the LOF abnormal data identification algorithm and the improved nearest neighbor filling algorithm, abnormal parameter data are deleted and missing parameter data are filled in the initial water resource scheduling parameter set to obtain a processed water resource scheduling parameter set. The S1 comprises the following steps: S11. The drone obtains initial water resource scheduling parameters to form an initial water resource scheduling parameter set; uses the LOF abnormal data identification algorithm to identify and delete abnormal parameter data in the initial water resource scheduling parameter set to obtain a processed water resource scheduling parameter set; S12. Based on the processed water resource scheduling parameter set, using an improved nearest neighbor filling algorithm to identify and fill in missing parameter data to obtain a processed water resource scheduling parameter set; S2. Obtaining an objective function of a water resource scheduling model based on the processed water resource scheduling parameter set, and establishing a water resource scheduling model in combination with water resource scheduling constraints; The S2 comprises the following steps: S21. Based on the processed water resource scheduling parameter set, the minimum water supply shortage is used as the objective function F1, and the maximum reservoir water level is used as the objective function F2; S22. Constructing water resource scheduling constraints based on the processed water resource scheduling parameter set; S23, taking the minimum water supply deficit and the maximum reservoir water level as the objective function of water resource scheduling, and establishing a water resource scheduling model in combination with water resource scheduling constraints; The construction of water resource scheduling constraints includes the following steps: Establish reservoir water balance constraints and reservoir capacity constraints based on reservoir capacity in different time periods; establish reservoir flow constraints based on outflow in different time periods; establish reservoir water level constraints based on reservoir water levels in different time periods; S3. Obtaining a fitness function through the objective function of the water resource scheduling model, adjusting and optimizing the water resource scheduling model while satisfying the water resource scheduling constraints, and obtaining an optimized water resource scheduling model by finding the optimal fitness function value; The S3 comprises the following steps: S31. According to the water resource scheduling model, a water resource scheduling objective function is obtained, and the water resource scheduling objective function is used as a fitness function, and the water resource scheduling constraint condition is satisfied; the number of coyotes in the coyote population is set to D, and the dimension of coyotes in the coyote subpopulation is set to E, where the dimension of coyotes in the coyote subpopulation is equal to the number of water resource scheduling parameters in the water resource scheduling model, and an initialization population matrix is ​​generated; S32. During the growth phase of a coyote subpopulation, the number of coyotes in the subpopulation is set to G, an alpha wolf is selected from the subpopulation, the social ranking of individual coyotes in the subpopulation is recorded as H, and the coyotes are influenced by the culture of the subpopulation during the growth phase, and the cultural trend of the subpopulation is calculated; Selecting a random coyote individual from the coyote subpopulation, setting an influence factor of the random coyote individual, and calculating a position of the coyote individual in the coyote subpopulation based on the influence factor of the random coyote individual and the cultural trend of the coyote subpopulation; S33, in the elimination phase, the coyote sub-population again selects random coyote individuals from the coyote sub-population, and then calculates the elimination probability of the coyote sub-population; under the elimination probability, the coyote individuals are eliminated according to the fitness function values ​​corresponding to the coyote individuals in the coyote sub-population, and when the fitness function value corresponding to the newborn coyote individual is the smallest, the newborn coyote individual is eliminated; when in each iteration, the fitness function value corresponding to the coyote individual in the coyote sub-population is less than the fitness function value corresponding to the newborn coyote individual, the coyote individual in the coyote sub-population is eliminated; As the number of iterations increases, the position of the coyote individual in the coyote subpopulation is updated according to the best fitness function value of each iteration. The maximum number of iterations is set to M. When the current number of iterations reaches the maximum number of iterations, the iteration is stopped to obtain the final coyote individual position. The coordinate values ​​of the final coyote individual E-dimensional position are respectively corresponding to the water resource scheduling parameters in the water resource scheduling model to obtain the optimized water resource scheduling model; S4. Based on the optimized water resource scheduling model, use the grey correlation analysis method to extract key water resource scheduling parameters, then train the GRU neural network to obtain a GRU neural network model, predict the flow rate, and adjust the drone water resource scheduling plan; The S4 comprises the following steps: S41. Calculate the grey correlation degree of the water resource scheduling parameters in the optimized water resource scheduling model using a grey correlation analysis method, extract key water resource scheduling parameters, and form a set of key water resource scheduling parameters; S42. Obtain water resource scheduling parameters from previous years, construct a new water resource scheduling model, obtain a sample water resource scheduling parameter set, normalize the sample water resource scheduling parameter set to obtain a normalized sample set, and divide the normalized sample set into a sample training set and a sample test set; set the GRU neural network to select stochastic gradient descent for optimization, input the sample training set into the GRU neural network, and iterate continuously until the GRU neural network converges to obtain a trained GRU neural network; then input the sample test set into the trained GRU neural network, set an error threshold to ω, and when the output result error is less than the error threshold, obtain the GRU neural network model; otherwise, adjust the weight until the output result error is less than the error threshold; S43. Input the key water resource scheduling parameter set into the GRU neural network model, output the flow prediction value, and use the flow prediction value to adjust the optimized water resource scheduling model in real time to adjust the drone water resource scheduling plan.

2. The method for processing data of a drone water resource scheduling scheme according to claim 1 is characterized in that: The method of using the LOF abnormal data identification algorithm to identify and delete abnormal parameter data in the initial water resource scheduling parameter set includes the following steps: Select the initial parameter data point in the initial water resource scheduling parameter set, and calculate the local reachable density of the initial parameter data point to obtain the local outlier factor of the initial parameter data point; if the local outlier factor of the initial parameter data point is less than or equal to 1, the initial parameter data point is normal parameter data; otherwise, the initial parameter data point is abnormal parameter data, and the abnormal parameter data is deleted until all the abnormal parameter data in the initial water resource scheduling parameter set are deleted.

3. A system for implementing the method for processing data of a drone water resource scheduling plan according to any one of claims 1-2, characterized in that: Specifically include: Water resources scheduling parameter processing module, water resources scheduling model establishment module, water resources scheduling model optimization module and water resources scheduling parameter prediction module; The water resource scheduling parameter processing module is used to delete and fill in abnormal parameter data and missing parameter data in the water resource scheduling parameters respectively; The water resource scheduling model establishment module is used to establish a water resource scheduling model based on the objective function and constraint conditions of the water resource scheduling model; The water resource scheduling model optimization module is used to adjust and optimize the water resource scheduling model using the coyote optimization algorithm; The water resource scheduling parameter prediction module is used to train the GRU neural network and predict the flow rate to adjust the UAV water resource scheduling plan.

Images