Agricultural water resource intelligent scheduling method based on multi-target bionic algorithm

Abstract

The invention provides an agricultural water resource intelligent scheduling method based on a multi-target bionic algorithm, and the method comprises the following steps: S1, collecting farmland data, obtaining soil humidity, crop moisture demands, meteorological conditions, water source availability and remote sensing image information, and constructing a water resource demand data set; s2, preprocessing the water resource demand data set, including data cleaning, screening, normalization, remote sensing image correction, noise removal and registration, performing trend analysis in combination with historical water demand data, and extracting a water demand high-frequency period and a fluctuation mode; s3, based on the water resource demand data set and the trend analysis result, establishing a hierarchical optimization model including regions, crop types and time periods; and S4, initializing population individuals and path information of the multi-target bionic algorithm on three levels of areas, crop types and time periods, and setting a differentiation fitness evaluation standard based on a water resource demand data set and a demand fluctuation mode.

Description

Technical Field

[0001] The present invention belongs to the field of water resource scheduling, and particularly relates to an intelligent scheduling method for agricultural water resources based on a multi-objective bionic algorithm. Background Art

[0002] At present, with the rapid development of intelligent agricultural technology, the management and optimization of agricultural water resources have become an important research direction for the sustainable development of agriculture. As the core element of agricultural production, the reasonable scheduling of water resources is directly related to crop growth, resource conservation, and ecological environment protection. However, the scheduling of agricultural water resources involves complex multi-objective optimization problems, including the balanced allocation of regional water resources, precise irrigation of crop types, and dynamic regulation of water demand in different time periods, which poses a huge challenge to traditional irrigation management methods.

[0003] At present, common agricultural water resource scheduling methods mainly rely on empirical management or single-objective optimization models. The allocation of regional water resources is usually regulated by preset ratios or simple supply-demand balances, lacking refined management of the balance of resources between regions. Crop irrigation strategies are mostly formulated based on the average water demand of crop classifications and fail to adjust the irrigation volume in combination with real-time environmental data, making it difficult to meet the water demand characteristics of different crops. The irrigation scheduling in time periods often takes fixed time periods as units, ignoring the dynamic change rules of water use peaks and troughs. Traditional methods can meet agricultural water demand to a certain extent, but have significant limitations when facing complex and changing farmland environments.

[0004] Specifically, the existing technologies have the following problems:

[0005] Lack of balance in regional water resource scheduling: Traditional methods rely more on experience or simple ratio allocation in the allocation of regional water resources and cannot dynamically adjust according to the actual water demand of regions, easily leading to problems such as overuse of water in some regions and water shortage in other regions, thus affecting the overall water use efficiency and crop yields.

[0006] Insufficient precision in crop water demand: In the irrigation optimization at the crop level in existing technologies, most are based on a single average water demand index and lack a dynamic adjustment mechanism that combines the real-time health status and actual water demand characteristics of crops, which may lead to under-irrigation or over-irrigation, affecting crop health and water resource utilization efficiency.

[0007] Lack of flexibility in time period regulation: Existing technologies do not fully consider the peak and trough changes in water demand in the water use scheduling of time periods, resulting in the disconnection between water resource scheduling and actual demand, making it difficult to quickly respond during water use peaks or reasonably conserve water during water use troughs, further exacerbating the risk of water resource waste.

[0008] Lack of collaborative processing of multi-level optimization objectives: Most existing methods usually focus on the optimization of a single objective, lacking the comprehensive balance and collaborative optimization of multi-level objectives, and it is difficult to take into account the complex interactions among regions, crops, and time periods, resulting in one-sided optimization results.

[0009] In summary, the existing technologies have significant deficiencies in the balance of regional water resource allocation, precise irrigation of crops, flexibility of time period regulation, and multi-level collaborative optimization. These defects limit the efficiency and effectiveness of agricultural water resource scheduling in practical applications. There is an urgent need for an intelligent scheduling method that combines multi-objective optimization algorithms. By fully integrating multi-level information of regions, crops, and time periods, it dynamically adjusts water resource allocation to improve the comprehensive utilization efficiency of agricultural water resources and achieve the goal of water conservation and efficiency increase. Summary of the Invention

[0010] The present invention proposes an intelligent agricultural water resource scheduling method based on a multi-objective bionic algorithm, which significantly improves the utilization efficiency of agricultural water resources and the operation effect.

[0011] The technical solution of the present invention is realized as follows: An intelligent agricultural water resource scheduling method based on a multi-objective bionic algorithm includes the following steps:

[0012] S1. Collect farmland data, obtain soil humidity, crop water requirements, meteorological conditions, water source availability, and remote sensing image information, and construct a water resource demand dataset;

[0013] S2. Preprocess the water resource demand dataset, including data cleaning, screening, normalization, and correction, noise removal, and registration of remote sensing images. Combine historical water demand data for trend analysis, and extract high-frequency time periods and fluctuation patterns of water demand;

[0014] S3. Establish a hierarchical optimization model including regions, crop types, and time periods based on the water resource demand dataset and the trend analysis results;

[0015] S4. Initialize the population individuals and path information of the multi-objective bionic algorithm at three levels of regions, crop types, and time periods, and set a differential fitness evaluation criterion based on the water resource demand dataset and the demand fluctuation pattern;

[0016] S5. Use the multi-objective bionic algorithm to perform iterative optimization at three levels of regions, crop types, and time periods;

[0017] S6. After the multi-level optimization of the multi-objective bionic algorithm is completed, extract the generated optimal irrigation time sequence and water volume allocation plan, and use it as the scheduling instruction of the agricultural irrigation system to guide the irrigation execution at the levels of regions, crops, and time periods.

[0018] Optionally, the S1 includes the following steps:

[0019] S11. Collect farmland data through Internet of Things sensors, including soil humidity H s (x,y), crop water demand W c (x,y), meteorological conditions M t (t) and water source availability A w :

[0020] H s (x,y) represents the soil humidity at the position with coordinates (x,y) in the farmland;

[0021] W c (x,y) represents the water demand per unit time of the crop at the position with coordinates (x,y);

[0022] M t (t) represents the meteorological conditions at time t, including precipitation, temperature and evaporation;

[0023] w represents the total available amount of the current water source;

[0024] S12. Collect remote sensing images covering the target farmland area through multi-spectral and thermal infrared sensors carried by an unmanned aerial vehicle, and obtain soil humidity image spectral data H i (x,y), crop water status image data W i (x,y) and temperature spectral data T i (x,y):

[0025] H i (x,y) represents the observed value of the soil humidity spectral value obtained by remote sensing at the coordinate (x,y);

[0026] W i (x,y) represents the observed value of the crop water status spectral value obtained by remote sensing at the coordinate (x,y);

[0027] T i (x,y) represents the observed value of the temperature spectral value obtained by remote sensing at the coordinate (x,y);

[0028] S13. Calculate the comprehensive soil humidity H f (x,y):

[0029] H f (x,y) = α1H s (x,y) + α2H i (x,y);

[0030] Wherein, H f (x,y) is the comprehensive soil humidity, and α1 and α2 are weight coefficients;

[0031] S14. Calculate the comprehensive crop water requirement W by fusing crop water requirement data f (x,y):

[0032] W f (x,y) = β1W c (x,y) + β2W i (x,y);

[0033] Wherein, W f (x,y) is the comprehensive water requirement, and β1 and β2 are weight coefficients;

[0034] S15. Modify the water requirement D by using the comprehensive soil humidity, comprehensive crop water requirement, meteorological conditions and temperature spectrum data r (x,y,t):

[0035] D r (x,y,t) = W f (x,y)·G(H f (x,y))·F m (M t (t),T i (x,y));

[0036] Wherein, D r (x,y,t) represents the modified water requirement at time t at position (x,y), G(H f (x,y)) is the soil humidity correction function, which dynamically adjusts the water requirement according to H f (x,y), and F m (M t (t),T i (x,y)) is the meteorological and temperature correction function;

[0037] S16. Calculate the final water resource demand dataset D by incorporating water source availability w (x,y,t):

[0038]

[0039] Wherein, A t is the total area of the irrigation area, A w / A t represents the available water resource limit per unit area, and D w (x,y,t) is the final water resource demand at time t at coordinate (x,y).

[0040] Optionally, the S2 includes the following steps:

[0041] S21. Clean the farmland data in the water resource demand dataset, remove outliers and missing values, and use the interpolation method to complement the missing data to generate a cleaned water resource demand dataset;

[0042] S22. Normalize the cleaned water resource demand dataset;

[0043] S23. Perform geometric correction and registration processing on the remote sensing image;

[0044] S24. Combine historical water demand data with the normalized water resource demand dataset for trend analysis, and extract high-frequency water demand periods and fluctuation patterns:

[0045]

[0046] Among them, represents the average normalized water demand at time t, is the normalized water demand, and N is the total number of location points;

[0047]

[0048] Among them, is the normalized water demand fluctuation amplitude at time t.

[0049] Optionally, the S3 includes the following steps:

[0050] S31. Based on the cleaned and normalized water resource demand dataset and the trend analysis results, establish a regional-level optimization model according to the farmland area division, set the regional-level optimization goal as the balanced distribution and water conservation of water resources, and define the objective function as:

[0051]

[0052] Among them, R is the total number of regions, and A r is the coverage area of the rth region, |A r | is the number of pixel points in the region, D avg is the average water demand of the entire farmland area, W r is the total water consumption of the rth region, and λ region is the water conservation weight factor;

[0053] S32. Based on the water demand characteristics W of each crop in the water resource demand dataset c (x, y) and the crop health status, construct a crop-level optimization model, set the crop-level optimization goal as precise irrigation, and define the objective function as:

[0054]

[0055] Among them, K is the total number of crop types, C k is the area covered by the k-th type of crop, W f (x,y) is the comprehensive water demand, W c (x,y) is the water demand characteristics of the crop, H c (x,y) is the crop health index, with a value range of [0,1], λ crop is the health status weight factor;

[0056] S33. Based on the results of the water resource demand trend analysis and combined with the peak water demand in the time period, establish a time period hierarchical optimization model, and the objective function is defined as:

[0057]

[0058] Among them, T is the total number of time segments, is the normalized water resource demand, is the average normalized water demand at time t, is the water demand fluctuation range at time t, λ time is the time period optimization weight factor.

[0059] Optionally, the S4 includes the following content:

[0060] S41. Initialize the population individuals and path information of the multi-objective bionic algorithm at the three levels of region, crop type, and time period, and define the initial state of the population individuals as a set of decision variables, including the set of decision variables X at the region level region , the set of decision variables X at the crop level crop and the set of decision variables X at the time period level time ;

[0061] S42. Set the differential fitness evaluation criteria to evaluate the optimization effects of population individuals at different levels:

[0062] The fitness evaluation criteria at the region level are defined as F based on the balanced distribution of water resources and the water-saving goal region =-f region , the fitness evaluation criteria at the crop level are defined as F based on the water demand priority of the crop and the crop health index crop =-f crop , the fitness evaluation criteria at the time period level are defined as F based on the peak and trough of water demand time =-f time ;

[0063] S43. Use the multi-objective bionic algorithm to perform iterative optimization at the three levels of region, crop type, and time period:

[0064] At the regional level, according to the regional-level fitness evaluation criterion F region Optimize the balanced distribution and water-saving distribution of water resources;

[0065] At the crop level, according to the crop-level fitness evaluation criterion F crop Optimize the priority irrigation order and water volume distribution of crops;

[0066] At the time period level, according to the time period-level fitness evaluation criterion F time Dynamically adjust the irrigation time period and optimize the irrigation timing;

[0067] S44. Update the global and local fitness during each iteration, and achieve collaborative optimization among the three levels through the information sharing mechanism of the bionic algorithm, so that the final optimization result converges and meets the global goal.

[0068] Optionally, the S5 includes the following steps:

[0069] S51. Use the multi-objective bionic algorithm for the balanced scheduling and water-saving distribution of water resources at the regional level. Taking the decision variable set at the regional level as the input, combining with the regional-level fitness evaluation criterion, and optimizing the water resource distribution within the region through iterative update. In each iteration, apply the improved particle swarm algorithm to update the regional water resource distribution amount x r :

[0070]

[0071] Among them, is the water resource distribution amount of region r at the t-th iteration, is the velocity vector of region r at the t-th iteration, representing the water resource distribution adjustment amount, is the individual optimal water resource distribution amount of region r, g best is the global optimal water resource distribution amount, ω is the inertia weight, adjusting the influence of the particle velocity, c1 and c2 are acceleration factors, adjusting the influence degree of the individual and global optimal solutions, and rand() is a random number with a value range in [0,1];

[0072] The updated needs to meet the following constraint conditions:

[0073]

[0074] Among them, is the maximum allowable water resource distribution amount of region r, depending on the irrigation capacity of the region;

[0075] S52. Optimize the priority irrigation order and water volume allocation of crops using a multi-objective bionic algorithm at the crop level. Taking the set of decision variables at the crop level as the input, optimize through an improved genetic algorithm in combination with the fitness evaluation criteria at the crop level, and perform the following operations in each iteration:

[0076] Selection operation: Calculate the selection probability P of an individual according to the fitness evaluation criteria at the crop level k :

[0077]

[0078] where F crop,k is the fitness value of the k-th type of crop, reflecting the degree of water demand matching and the health status of the crop;

[0079] Crossover operation: Perform crossover on the selected parent individuals and to generate offspring individuals:

[0080]

[0081] where γ is the crossover coefficient, and its value range is [0, 1];

[0082] Mutation operation: Mutate the offspring individuals with the mutation probability P mutation :

[0083]

[0084] where δ k is the mutation amount, which follows a normal distribution σ k is the mutation intensity;

[0085] The updated x k needs to satisfy the following constraint conditions:

[0086]

[0087] where is the maximum allowable irrigation amount of crop k, is the total water resources allocated at the regional level;

[0088] S53. Optimize the irrigation timing using a multi-objective bionic algorithm at the time period level. Taking the set of decision variables at the time period level as the input, optimize through an improved ant colony algorithm in combination with the fitness evaluation criteria at the time period level, and perform the following operations in each iteration:

[0089] Path selection: The probability that ant m selects the next irrigation time period at time t

[0090]

[0091] Among them, is the pheromone concentration of ant m in time period t, representing the historical experience of choosing this time period, and η t is the heuristic factor, defined as:

[0092]

[0093] Among them, is the normalized water resource demand, is the average normalized water demand at time t, and α and β are parameters;

[0094] Pheromone update: After completing the path selection, update the pheromone concentration:

[0095]

[0096] Among them, ρ is the pheromone evaporation coefficient, which controls the degree of forgetting of pheromone, is the pheromone increment of ant m in time period t, defined as:

[0097]

[0098] Among them, Q is a constant, and L (m) is the total cost of the path of ant m;

[0099] The updated x t needs to satisfy the following constraints:

[0100]

[0101] Among them, is the maximum allowable irrigation amount in time period t, which depends on the capacity of the irrigation system and the peak water use limit, is the total irrigation amount allocated at the crop level, ensuring that the total water use at the time period level does not exceed the total water demand of the crop;

[0102] S54. In each iteration process, by combining the fitness evaluations at the regional, crop type, and time period levels, the global and local fitness are updated through the information sharing mechanism of the multi-objective bionic algorithm;

[0103] S55. After completing the predetermined number of iterations or meeting the convergence conditions, output the optimal solutions at each level. The optimal solutions are used as the scheduling instructions for the agricultural irrigation system to guide the irrigation execution at the regional, crop, and time period levels.

[0104] After adopting the above technical solution, the beneficial effects of the present invention are as follows: The hierarchical optimization model proposed by the present invention realizes the balanced scheduling of water resources and the water-saving goal at the regional level. At the crop level, it realizes precise irrigation by combining the crop health index and water demand characteristics. At the time period level, it conducts dynamic scheduling for the peak and trough of water demand. Through the multi-level design and optimization of the objective function, the present invention solves the problem of coordinated water resource scheduling among regions, crops, and time periods on the basis of traditional single-objective optimization, significantly improving the scientificity and accuracy of resource allocation. Experimental verification shows that the method of the present invention improves the water resource utilization efficiency in the region by about 18%, while reducing the risk of excessive concentration of regional water resources and effectively alleviating the problem of imbalance between water supply and demand in agricultural production.

[0105] The present invention respectively introduces the particle swarm optimization, genetic algorithm, and ant colony algorithm into the multi-objective bionic algorithm, and designs differentiated fitness evaluation criteria for different levels of optimization requirements: the fitness function based on the balanced allocation of water resources at the regional level, the fitness function based on the health status and water demand priority at the crop level, and the fitness function based on the dynamic water demand and fluctuation characteristics at the time period level. Each level realizes the balance of the global optimization goal through the cooperation mechanism. Compared with the limitations of a single algorithm in the prior art, the present invention effectively improves the flexibility and robustness of the optimization process. The simulation experiment results show that the optimization convergence speed of the multi-objective bionic algorithm is increased by about 23%, and the ability to adapt to the changes in the dynamic farmland environment is significantly enhanced.

[0106] The present invention combines Internet of Things sensors and unmanned aerial vehicle remote sensing data, constructs a high-precision water resource demand data set through data cleaning, normalization, and trend analysis, and introduces a real-time dynamic feedback mechanism in the algorithm optimization process, enabling water resource scheduling to respond to the changes in the farmland environment in a timely manner. On this basis, the present invention realizes the intelligent scheduling of water resources in a dynamic environment by dynamically adjusting the regional water resource allocation amount, crop irrigation order, and time period irrigation plan. Experiments show that the method of the present invention achieves a water use accuracy improvement of more than 25% in the complex farmland environment, and at the same time, the average increase rate of crop yield reaches 10%, fully reflecting its application value in smart agriculture. Brief Description of the Drawings

[0107] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0108] Figure 1 It is a flowchart of an intelligent agricultural water resource scheduling method based on a multi-objective bionic algorithm proposed by the present invention;

[0109] Figure 2 This is a schematic diagram of the optimization process of the multi-objective bionic algorithm in an intelligent agricultural water resource scheduling method proposed by the present invention. Specific implementation manner

[0110] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0111] Embodiment:

[0112] As Figures 1-2 shown, an intelligent agricultural water resource scheduling method based on a multi-objective bionic algorithm includes the following steps:

[0113] S1. Collect farmland data, obtain soil humidity, crop water requirements, meteorological conditions, water source availability, and remote sensing image information, and construct a water resource demand data set;

[0114] S2. Preprocess the water resource demand data set, including data cleaning, screening, normalization, and correction, noise removal, and registration of remote sensing images, conduct trend analysis in combination with historical water demand data, and extract high-frequency time periods and fluctuation patterns of water demand;

[0115] S3. Establish a hierarchical optimization model including regions, crop types, and time periods based on the water resource demand data set and the trend analysis results;

[0116] S4. Initialize the population individuals and path information of the multi-objective bionic algorithm at the three levels of region, crop type, and time period, and set differential fitness evaluation criteria based on the water resource demand data set and the demand fluctuation pattern;

[0117] S5. Use the multi-objective bionic algorithm to perform iterative optimization at the three levels of region, crop type, and time period;

[0118] S6. After the multi-level optimization of the multi-objective bionic algorithm is completed, extract the generated optimal irrigation time sequence and water volume allocation plan, and use it as the scheduling instruction of the agricultural irrigation system to guide the irrigation execution at the levels of region, crop, and time period.

[0119] Taking a specific application scenario as an example, it is applied to a cotton planting base in City A. The cotton planting base in City A covers an area of about 2,000 hectares. The planting distribution within the area is uneven and mainly consists of 3 functional areas: a concentrated contiguous planting area (covering 800 hectares), a small scattered planting area (covering 600 hectares), and an ecological protection area (covering 600 hectares). The irrigation method within the area is a drip irrigation system, and the main water source is a water supply channel. The total water resources are 15,000 cubic meters per day. The local precipitation is less, with an average annual precipitation of about 100 millimeters, and the evaporation is large, resulting in relatively high agricultural water use pressure. The main cotton varieties planted in the base include long-staple cotton and common cotton, with different water demand characteristics. The goal is to maximize water resource conservation while ensuring cotton production.

[0120] The base introduces the intelligent agricultural water resource scheduling method based on a multi-objective bionic algorithm of the present invention, and combines unmanned aerial vehicle (UAV) remote sensing data with Internet of Things sensors to achieve hierarchical optimization scheduling for regions, crop types, and time periods.

[0121] First, the Internet of Things sensor network deployed in the base is used to collect information on soil humidity, meteorological conditions, and water supply volume of the channel water source in real time. At the same time, a UAV equipped with multi-spectral and thermal infrared sensors is used to comprehensively monitor the farmland area to obtain data on crop water demand and health status. The sensor data shows that the average soil humidity in the concentrated contiguous planting area is 21%, in the small scattered planting area is 18%, and in the ecological protection area is 30%. The UAV remote sensing analysis finds that the average health index of common cotton is 0.82, and that of long-staple cotton is 0.75. The water demand of common cotton is about 1.2 times that of long-staple cotton.

[0122] Next, the collected data is cleaned and normalized to construct a water resource demand data set. Combining historical water use data, the water demand change trends of each region and crop are analyzed. The analysis results show that the water demand in the concentrated contiguous planting area is relatively high, and the peak water demand is concentrated at 6:00 - 8:00 in the morning and 4:00 - 6:00 in the afternoon every day. Due to the scattered distribution of the scattered planting area, the water demand fluctuates greatly. The ecological protection area has a relatively low water resource demand, but a certain degree of humidity needs to be maintained to promote ecological balance.

[0123] Based on the above analysis results, a hierarchical optimization model is established. At the regional level, with the goal of balanced water resource distribution and water conservation, the daily average water supply of each region is determined. The concentrated contiguous planting area is set at 8,000 cubic meters, the small scattered planting area is set at 5,000 cubic meters, and the ecological protection area is set at 2,000 cubic meters. At the crop level, considering the water demand characteristics and health status of common cotton and long-staple cotton, the irrigation amount and priority are dynamically adjusted, and the irrigation demand of common cotton is preferentially guaranteed. At the time period level, according to the daily water demand peak, the irrigation time period and water supply distribution are adjusted to ensure the water supply stability during peak hours, and at the same time, water is conserved during low peak hours.

[0124] Optimized by a multi-objective bionic algorithm, the particle swarm algorithm is applied to optimize water resource allocation at the regional level, the genetic algorithm is used to optimize the crop priority irrigation strategy, and the ant colony algorithm is used to optimize the irrigation timing. After the algorithm iteration converges, the final optimization results are output: 8000 cubic meters of water supply for the concentrated contiguous planting area, 4800 cubic meters of water supply for the small scattered planting area, and 2200 cubic meters of water supply for the ecological protection area; ordinary cotton is preferentially irrigated, with a daily irrigation volume of 1.8 cubic meters per mu, and long-staple cotton is 1.5 cubic meters per mu; the irrigation periods are 6:00-8:00 in the morning, 10:00-12:00 in the morning, and 4:00-6:00 in the afternoon every day.

[0125] To verify the feasibility and effectiveness of the present invention, the method of the present invention is compared with the traditional empirical scheduling method. Using the actual operation data of 2019 and 2020, the water resources are scheduled by the traditional method and the method of the present invention respectively. The results are shown in Table 1 below:

[0126] Table 1 Comparison data table of scheduling methods

[0127]

[0128] It can be seen from the data that the method of the present invention increases the yield per mu by 7.7% while reducing the total water consumption by 15%, and the water resource utilization efficiency is improved by 12 percentage points. At the same time, by dynamically adjusting the crop priority irrigation order, the health indexes of ordinary cotton and long-staple cotton are on average increased by 10%, further verifying the superiority of this method in intelligent scheduling.

[0129] It can be proved by this Example 1 that the method of the present invention can significantly improve the scheduling accuracy and resource utilization efficiency in actual agricultural water resource management, and provides an efficient and intelligent solution for the sustainable development of modern agriculture.

[0130] The hierarchical optimization model proposed by the present invention realizes the balanced scheduling of water resources and the water-saving goal at the regional level. At the crop level, precise irrigation is realized by combining the crop health index and water demand characteristics. At the time period level, dynamic scheduling is carried out for the peak and trough of water demand. Through the multi-level design and optimization of the objective function, the present invention solves the problem of water resource scheduling coordination among regions, crops and time periods on the basis of traditional single-objective optimization, and significantly improves the scientificity and accuracy of resource allocation. Experimental verification shows that the method of the present invention improves the water resource utilization efficiency in the region by about 18%, while reducing the risk of excessive concentration of regional water resources, and effectively alleviates the problem of water supply and demand imbalance in agricultural production.

[0131] In the present invention, the particle swarm optimization, genetic algorithm, and ant colony algorithm are respectively introduced into the multi-objective bionic algorithm, and differential fitness evaluation criteria are designed for different levels of optimization requirements: the fitness function based on the balanced distribution of water resources at the regional level, the fitness function based on the health status and water demand priority at the crop level, and the fitness function based on the dynamic water demand and fluctuation characteristics at the time period level. Each level achieves the balance of the global optimization goal through a cooperation mechanism. Compared with the limitations of a single algorithm in the prior art, the present invention effectively improves the flexibility and robustness of the optimization process. The simulation experiment results show that the optimization convergence speed of the multi-objective bionic algorithm is increased by about 23%, and the ability to adapt to the changes in the dynamic farmland environment is significantly enhanced.

[0132] The present invention combines Internet of Things sensors and drone remote sensing data, constructs a high-precision water resource demand data set through data cleaning, normalization, and trend analysis, and introduces a real-time dynamic feedback mechanism in the algorithm optimization process, enabling the water resource scheduling to respond to the changes in the farmland environment in a timely manner. On this basis, the present invention realizes the intelligent scheduling of water resources in a dynamic environment by dynamically adjusting the regional water resource allocation amount, crop irrigation order, and time period irrigation plan. Experiments show that the method of the present invention has achieved a water use accuracy improvement of more than 25% in the complex farmland environment, and at the same time, the average increase rate of crop yield reaches 10%, fully reflecting its application value in intelligent agriculture.

[0133] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent scheduling method for agricultural water resources based on a multi-objective bionic algorithm, characterized in that It includes the following steps: S1. Collect farmland data, obtain soil moisture, crop water requirements, meteorological conditions, water source availability and remote sensing image information, and construct a water resource demand dataset; S2. Preprocess the water resource demand dataset. The preprocessing includes data cleaning, screening, normalization, and correction, noise removal and registration of remote sensing images. Combine historical water demand data for trend analysis, and extract high-frequency water demand periods and fluctuation patterns; S3. Based on the water resource demand dataset and the results of trend analysis, establish a hierarchical optimization model including regions, crop types and time periods; S4. Initialize the population individuals and path information of the multi-objective bionic algorithm at the three levels of region, crop type and time period, and set differential fitness evaluation criteria based on the water resource demand dataset and demand fluctuation patterns; S5. Use the multi-objective bionic algorithm to perform iterative optimization at the three levels of region, crop type and time period; S6. After the multi-level optimization of the multi-objective bionic algorithm is completed, extract the generated optimal irrigation timing and water volume allocation plan, and use it as the scheduling instruction of the agricultural irrigation system to guide the irrigation execution at the levels of region, crop and time period.

2. The intelligent scheduling method for agricultural water resources based on a multi-objective bionic algorithm according to claim 1, characterized in that: The S1 includes the following steps: S11. Collect farmland data through Internet of Things sensors. The collected farmland data includes soil moisture, crop water requirements, meteorological conditions and water source availability; S12. Collect remote sensing images covering the target farmland area through multi-spectral and thermal infrared sensors carried by drones, and obtain soil moisture image spectral data, crop water status image data and temperature spectral data; S13. Calculate the comprehensive soil moisture based on the collected soil moisture data; S14. Calculate the comprehensive crop water requirement by fusing crop water requirement data; S15. Use the comprehensive soil moisture, comprehensive crop water requirement, meteorological conditions and temperature spectral data to correct the water requirement; S16. Incorporate water source availability to calculate the final water resource demand dataset.

3. The agricultural water resource intelligent scheduling method based on a multi-objective bionic algorithm according to claim 1, wherein: The S2 includes the following steps: S21. Clean the farmland data in the water resource demand dataset, remove outliers and missing values, and use the interpolation method to fill in the missing data to generate a cleaned water resource demand dataset; S22. Perform normalization processing on the cleaned water resource demand dataset; S23. Perform geometric correction and registration processing on the remote sensing images; S24. Combine historical water demand data to perform trend analysis on the normalized water resource demand dataset, and extract high-frequency water demand periods and fluctuation patterns.

4. The intelligent scheduling method for agricultural water resources based on a multi-objective bionic algorithm according to claim 1, characterized in that: The S3 includes the following steps: S31. Based on the water resource demand dataset after cleaning and normalization and the results of trend analysis, establish a regional-level optimization model according to farmland area division, and set the regional-level optimization goal as the balanced distribution and water saving of water resources; S32. Build a crop-level optimization model based on the water requirement characteristics and crop health status of each crop in the water resource demand dataset, and set the crop-level optimization goal as precise irrigation; S33. Based on the results of the water resource demand trend analysis, establish a time period-level optimization model in combination with the peak water demand in the time period.

5. The intelligent scheduling method of agricultural water resources based on a multi-objective bionic algorithm according to claim 1, characterized in that, The S4 includes the following contents: S41. Initialize the population individuals and path information of the multi-objective bionic algorithm at three levels: region, crop type, and time period. Define the initial state of the population individuals as a set of decision variables, including the set of decision variables X at the region level region , the set of decision variables X at the crop level crop , and the set of decision variables X at the time period level time ; S42. Set the differential fitness evaluation criteria to evaluate the optimization effects of population individuals at different levels: The fitness evaluation criteria at the regional level are defined as F based on the balanced allocation of water resources and water conservation goals region =-f region , and the fitness evaluation criteria at the crop level are defined as F based on the water demand priority of crops and the crop health index crop =-f crop , and the fitness evaluation criteria at the time period level are defined as F based on the peak and trough of water demand time =-f time ; S43. Use the multi-objective bionic algorithm to perform iterative optimization at three levels: region, crop type, and time period: At the regional level, according to the regional-level fitness evaluation criterion F region Optimize the balanced distribution and water-saving distribution of water resources; At the crop level, optimize the priority irrigation order and water volume allocation according to the crop-level fitness evaluation criterion F crop Optimize the priority irrigation order and water volume allocation of crops; At the time period level, according to the time period level fitness evaluation criterion F time Dynamically adjust the irrigation time period and optimize the irrigation timing sequence; S44. Update the global and local fitness during each iteration, and achieve collaborative optimization among the three levels through the information sharing mechanism of the bionic algorithm, so that the final optimization result converges and meets the global goal.

6. The intelligent scheduling method for agricultural water resources based on a multi-objective bionic algorithm according to claim 1, characterized in that The said S5 includes the following steps: S51. Use a multi-objective bionic algorithm to perform balanced scheduling and water-saving allocation of water resources at the regional level. Taking the decision variable set at the regional level as the input, combined with the fitness evaluation criteria at the regional level, optimize the water resource allocation within the region through iterative update. In each iteration, apply an improved particle swarm algorithm to update the regional water resource allocation amount x r ; S52. Use the multi-objective bionic algorithm to optimize the priority irrigation order and water volume allocation of crops at the crop level. Taking the decision variable set at the crop level as the input, optimize through the improved genetic algorithm in combination with the fitness evaluation criteria at the crop level; S53. Use the multi-objective bionic algorithm to optimize the irrigation timing at the time period level. Taking the decision variable set at the time period level as the input, optimize through the improved ant colony algorithm in combination with the fitness evaluation criteria at the time period level; S54. During each iteration, combine the fitness evaluations at the three levels of region, crop type, and time period, and update the global and local fitness through the information sharing mechanism of the multi-objective bionic algorithm; S55. After completing the predetermined number of iterations or meeting the convergence conditions, output the optimal solutions at each level. The optimal solutions are used as the scheduling instructions for the agricultural irrigation system to guide the irrigation execution at the levels of region, crop, and time period.

Images