Multi-objective collaborative irrigation nitrogen application intelligent decision method, device and program product

Abstract

The application belongs to the technical field of efficient utilization of agricultural water resources, and discloses a multi-target coordinated irrigation and nitrogen application intelligent decision method, equipment and program product. The method obtains simulated decision variables by using an AHC model, converts the simulated decision variables into floating-point simulated decision variable data in a COZE intelligent agent framework, and respectively inputs the floating-point simulated decision variable data into a VIKOR decision model, an AHP decision model and a TOPSIS decision model for parallel decision analysis, so as to respectively obtain a first irrigation and nitrogen application strategy and first irrigation and nitrogen application decision variable data under the best scenario of each decision model. The first irrigation and nitrogen application strategy is optimized based on the first irrigation and nitrogen application decision variable data by using a large language model until a second irrigation and nitrogen application strategy is obtained, and the percentage change of second irrigation and nitrogen application decision variable data compared with local control irrigation and nitrogen application decision variable data is calculated. The irrigation amount, the nitrogen application amount and the percentage change corresponding to the second irrigation and nitrogen application strategy are output.

Description

Technical Field

[0001] This invention relates to the field of efficient utilization of agricultural water resources, and in particular to a multi-objective collaborative irrigation nitrogen application intelligent decision-making method, equipment and program products. Background Technology

[0002] Based on a balanced consideration of multiple objectives including water, food, ecology, and economy, determining a reasonable irrigation nitrogen application strategy is of significant theoretical and practical importance for ensuring water security, ecological security, sustainable economic growth, and achieving the goal of green and high-quality agricultural development. Traditional methods for determining irrigation nitrogen application strategies are time-consuming, labor-intensive, and highly uncertain, making it difficult to obtain optimal solutions. Furthermore, they lack user-friendly interfaces, resulting in high learning costs for users.

[0003] Determining reasonable decision indicators and selecting reliable decision-making methods / tools are crucial for formulating appropriate agricultural irrigation nitrogen application strategies. Regarding decision indicators, previous studies primarily used field trials to obtain water use efficiency, crop yield, and economic benefits as decision targets. This approach is time-consuming, labor-intensive, and struggles to obtain indicators related to the service value of agro-ecosystems. Regarding decision-making methods, previous studies relied heavily on human experience to determine suitable irrigation nitrogen application strategies. While this method can provide preliminary solutions, it suffers from low efficiency and high subjectivity. With further research, many scholars have begun to apply multiple methods (such as VIKOR, TOPSIS, and AHP) for systematic decision-making. Although these methods can objectively weigh multiple decision objectives and generate appropriate irrigation nitrogen application strategies, strategies obtained from single decision-making methods have high uncertainty, and their level of intelligence is relatively low. In recent years, AI-based decision / optimization methods have gradually been applied in agricultural water and nitrogen management. However, these methods have not fully utilized the advantages of AI technology, resulting in limited capabilities in autonomous decision-making and the expression of decision results. For example, they require manual input of numerous decision indicators, demand high levels of expertise in extracting decision results, and struggle to directly obtain key information from the decision outcomes. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention aims to provide a method, equipment, and program product for intelligent decision-making on multi-objective collaborative irrigation nitrogen application based on AI agents, considering water, food, ecology, and economy. Employing the COZE agent framework, this method combines agricultural hydrological models, multi-objective decision-making methods, and large language models to optimize irrigation nitrogen application strategies. This method does not rely on human judgment or the results of a single decision model. After simulating water, food, ecological, and economic indicators, it automatically determines a highly efficient and precise multi-objective collaborative irrigation nitrogen application strategy, and intelligently analyzes the decision results through structured data presentation.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] According to a first aspect of the present invention, a multi-objective collaborative irrigation nitrogen application intelligent decision-making method is provided, comprising the following steps:

[0007] The AHC model was used to simulate the decision variables under a preset irrigation nitrogen application scenario. The preset scenarios can be set according to actual needs, and there is no limit to the number of scenarios. Preset scenario 1 is set as the local control scenario, and there are no restrictions on the order of other scenarios.

[0008] The simulated decision variables are extracted using the COZE agent framework, and the data type of the simulated decision variables is converted into floating point type to obtain the simulated decision variable data;

[0009] Within the COZE agent framework, the simulated decision variable data are input into the VIKOR decision model, AHP decision model, and TOPSIS decision model respectively for parallel decision analysis, to obtain the first irrigation nitrogen application strategy under the best scenario and the corresponding first irrigation nitrogen application decision variable data.

[0010] Within the COZE agent framework, a large language model is used to judge the first irrigation nitrogen application strategy and the first irrigation nitrogen application decision variable data until the second irrigation nitrogen application strategy is obtained.

[0011] Extract the decision variable data for the second irrigation nitrogen application strategy;

[0012] Calculate the percentage change in the second irrigation nitrogen application decision variable data compared to the control irrigation nitrogen application decision variable data of the local control irrigation nitrogen application strategy;

[0013] Output the irrigation amount, nitrogen application amount, and percentage change corresponding to the second irrigation nitrogen application strategy.

[0014] The present invention provides an intelligent decision-making method for multi-objective collaborative irrigation nitrogen application based on AI agents, which integrates agricultural hydrological models, multi-objective decision-making methods and large language models within the COZE agent framework. It performs two-level decision-making judgments on multiple indicators of water, food, ecology and economy under different irrigation nitrogen application scenarios, and finally determines the appropriate irrigation nitrogen application strategy and its effect on the regulation of the "water, food, ecology and economy" system, so as to realize intelligent and efficient management of agricultural water and nitrogen.

[0015] This invention, within the COZE intelligent agent framework, integrates agricultural hydrological models, multi-objective decision-making methods, and a large language model to construct a multi-objective collaborative decision-making intelligent agent for "water-food-ecology-economy." This invention uses a data reading plugin to acquire water, food, ecological, and economic indicators from different irrigation nitrogen application scenarios simulated by the agricultural hydrological model as input data. It employs multiple multi-objective decision-making methods for parallel decision analysis to obtain several preliminary suitable first irrigation nitrogen application strategies. Then, a large language model is used to further evaluate these preliminary suitable first irrigation nitrogen application strategies to obtain a final suitable second irrigation nitrogen application strategy. This method does not rely on human analysis or the results of a single decision model, significantly improving the efficiency, accuracy, and intelligence level of agricultural water and nitrogen management strategy formulation.

[0016] According to an embodiment of the present invention, the simulated decision variables are extracted using the COZE intelligent agent framework, and the data type of the simulated decision variables is converted into floating-point type to obtain simulated decision variable data. This includes: extracting decision variables using a file reading plugin in the COZE intelligent agent framework, and then using a code plugin to write data processing code to remove spaces and commas from each decision variable file and convert the data type into floating-point type.

[0017] According to an embodiment of the present invention, the three decision-making methods VIKOR, AHP, and TOPSIS are deployed in the code plugin to establish corresponding multi-objective decision-making plugins, forming VIKOR decision-making models, AHP decision-making models, and TOPSIS decision-making models.

[0018] According to an embodiment of the present invention, the core algorithm of the VIKOR decision model is as follows:

[0019] ;

[0020] ;

[0021] ;

[0022] In the formula: x ij Let be the value of the j-th decision indicator under the i-th irrigation nitrogen application scenario; A j + The j-th decision indicator is the optimal solution across all scenarios; A j - The j-th decision indicator is the worst solution in all scenarios; S i Let be the distance between scenario i and the optimal solution; R i Let be the distance between scenario i and the worst solution; w jLet be the weight of the j-th decision indicator; v The decision coefficient represents the decision-maker's preference; Q i The decision benefit ratio for each scenario is given, with the minimum value corresponding to the optimal scenario.

[0023] According to an embodiment of the present invention, the decision coefficient v can be 0.5.

[0024] According to an embodiment of the present invention, the core algorithm of the TOPSIS decision model is as follows:

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] In the formula: C i Let be the score for the i-th irrigation nitrogen application scenario, with the maximum value corresponding to the best scenario; D i + Let be the distance between the i-th scenario and the ideal solution; D i - Let be the distance between the i-th scenario and the worst solution; w j Let be the weight of the j-th decision indicator; r ij Let be the standardized value of the j-th decision indicator in scenario i; x ij Let be the original value of the j-th decision indicator for scenario i.

[0030] According to an embodiment of the present invention, under the premise that the consistency of the decision matrix is ​​acceptable, the core calculation formula of the AHP decision model is as follows:

[0031] ;

[0032] ;

[0033] In the formula: CR For consistency ratio, when CR < 0.1, the consistency of the decision matrix is ​​acceptable; λ max It is the largest eigenvalue of the decision matrix; S i This is the score for scenario i, with the highest value corresponding to the best scenario; n The number of scenarios involving nitrogen application during irrigation; wj Let be the weight of the j-th decision indicator; x ij Let be the original value of the j-th decision indicator for scenario i.

[0034] According to an embodiment of the present invention, the first irrigation nitrogen application strategy includes the VIKOR first irrigation nitrogen application strategy, the AHP first irrigation nitrogen application strategy, and the TOPSIS first irrigation nitrogen application strategy;

[0035] The first irrigation nitrogen application decision variable data includes VIKOR first irrigation nitrogen application decision variable data, AHP first irrigation nitrogen application decision variable data, and TOPSIS first irrigation nitrogen application decision variable data.

[0036] According to an embodiment of the present invention, the process of judging the first irrigation nitrogen application strategy and the first irrigation nitrogen application decision variable data until a second irrigation nitrogen application strategy is obtained is as follows:

[0037] Based on three judgment conditions, the data of the VIKOR first irrigation nitrogen application strategy and the VIKOR first irrigation nitrogen application decision variables are sequentially evaluated.

[0038] Data on AHP first irrigation nitrogen application strategy and AHP first irrigation nitrogen application decision variables;

[0039] In addition, the TOPSIS first irrigation nitrogen application strategy and the TOPSIS first irrigation nitrogen application decision variable data are used for judgment;

[0040] The first irrigation nitrogen application strategy that simultaneously meets all three judgment conditions is determined as the second irrigation nitrogen application strategy;

[0041] If the number of first irrigation nitrogen application strategies that simultaneously meet all three judgment conditions is greater than 1, then the first irrigation nitrogen application strategy that achieves the highest yield will be used as the second irrigation nitrogen application strategy.

[0042] According to an embodiment of the present invention, a large language model, such as an AI model like the Deep Seek model, can be used in the process of judging the first irrigation nitrogen application strategy and the first irrigation nitrogen application decision variable data until a second irrigation nitrogen application strategy is obtained.

[0043] According to an embodiment of the present invention, the Deep Seek model is used to determine the first irrigation nitrogen application strategy and the first irrigation nitrogen application decision variable data until a second irrigation nitrogen application strategy is obtained. The process is as follows:

[0044] First, the VIKOR first irrigation nitrogen application strategy and the VIKOR first irrigation nitrogen application decision variable data are evaluated:

[0045] 1. The crop yield reduction of the first irrigation nitrogen application strategy compared to the crop yield reduction of the local irrigation nitrogen application strategy is less than a preset threshold;

[0046] 2. The water use efficiency, positive ecosystem service value, and economic benefits corresponding to the first irrigation nitrogen application strategy are higher than those of the local control irrigation nitrogen application strategy;

[0047] 3. The negative ecosystem service value corresponding to the first irrigation nitrogen application strategy is lower than that of the local control irrigation nitrogen application strategy;

[0048] If the above three judgment conditions are met, then the VIKOR first irrigation nitrogen application strategy generated by the VIKOR method decision is determined to be the final suitable second irrigation nitrogen application strategy;

[0049] If any of the above three judgment conditions are not met, the same judgment process is then applied to the results generated by the AHP and TOPSIS decision-making methods in turn until the final suitable second irrigation nitrogen application strategy is obtained.

[0050] If the irrigation nitrogen application strategies and corresponding results generated by the three decision-making methods all meet the above judgment conditions, then the irrigation nitrogen application strategy that yields the highest output will be taken as the final suitable second irrigation nitrogen application strategy.

[0051] According to an embodiment of the present invention, the three judgment conditions include:

[0052] The crop yield reduction of the first irrigation nitrogen application strategy compared to the local irrigation nitrogen application strategy is less than a preset threshold;

[0053] The water use efficiency, positive ecosystem service value, and economic benefits of the first irrigation nitrogen application strategy were higher than those of the local control irrigation nitrogen application strategy.

[0054] The negative ecosystem service value corresponding to the first irrigation nitrogen application strategy was lower than that of the local control irrigation nitrogen application strategy.

[0055] According to an embodiment of the present invention, the preset threshold can be 0.05.

[0056] According to an embodiment of the present invention, the local control irrigation nitrogen application strategy is set as scenario 1.

[0057] According to an embodiment of the present invention, calculating the percentage change of the second irrigation nitrogen application decision variable data compared to the local control irrigation nitrogen application decision variable data of the local control irrigation nitrogen application strategy includes:

[0058] Using the DeepSeek model, we extracted the yield, water use efficiency, positive ecosystem service value, negative ecosystem service value, and economic benefits corresponding to the final suitable second irrigation nitrogen application strategy from the simulated decision variable data, and calculated the percentage change of these indicators compared with the local control irrigation nitrogen application strategy.

[0059] According to an embodiment of the present invention, outputting the irrigation amount, nitrogen application amount, and percentage change corresponding to the second irrigation nitrogen application strategy includes: using the agent's termination plugin to output the irrigation amount, nitrogen application amount, second irrigation nitrogen application decision variable data, and percentage change compared to the local control irrigation nitrogen application decision variable data corresponding to the final suitable second irrigation nitrogen application strategy.

[0060] According to an embodiment of the present invention, the simulation decision variables include water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value, and economic benefits.

[0061] According to an embodiment of the present invention, the preset irrigation nitrogen application scenario is a scenario with different irrigation and nitrogen application rates determined based on the actual conditions of the study area.

[0062] According to an embodiment of the present invention, the water use efficiency, the crop yield, the positive ecosystem service value, and the economic benefits are set as positive decision variables, and the negative ecosystem service value is set as a negative decision variable.

[0063] According to a second aspect of the present invention, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method.

[0064] According to a third aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the steps of the method.

[0065] Compared to traditional multi-objective decision-making methods, this invention provides a user-friendly interface through a computer program. The decision-making process can be initiated with a single click, and the output of the decision results is concise and clear, making it easier for users to understand. This effectively reduces the user's learning curve and improves decision-making efficiency.

[0066] The beneficial effects of this invention are as follows:

[0067] This invention combines the advantages of artificial intelligence technology in intelligent agents and user-friendly interaction with the advantages of multi-objective decision-making methods in objective system decision-making. At the same time, it fully considers the ability of agricultural hydrological models to simulate various decision indicators such as water, food, ecology, and economy, thereby improving the efficiency and reliability of agricultural irrigation nitrogen management decisions and reducing the cost and threshold for users to learn and use it.

[0068] This invention solves the problems of traditional irrigation nitrogen application decision-making methods, such as difficulty in obtaining relevant decision indicators for water, food, ecology, and economy, incomplete consideration of indicators, and lack of coordinated regulation of ecosystem service value. It can determine appropriate irrigation nitrogen application strategies based on comprehensive consideration of coordinated regulation of multiple objectives of water, food, ecology, and economy.

[0069] Compared with traditional irrigation nitrogen application decision-making methods, this invention greatly reduces time costs and mitigates the impact of human subjectivity and uncertainty caused by single decision-making methods. The decision-making process has a high level of intelligence and automation. In the management of agricultural irrigation nitrogen application in large irrigation areas with complex planting structures, it can achieve efficient and high-precision irrigation nitrogen application strategy decision-making, making refined and efficient water and nitrogen management in large irrigation areas possible. Attached Figure Description

[0070] The present invention includes the following figures:

[0071] Figure 1 This is a flowchart of the multi-objective collaborative irrigation nitrogen application intelligent decision-making method according to an embodiment of the present invention;

[0072] Figure 2 This refers to the irrigation water volume and nitrogen application volume data for different irrigation nitrogen application scenarios in this embodiment of the invention;

[0073] Figure 3 These are the simulated decision variables obtained under different irrigation nitrogen application scenarios in embodiments of the present invention;

[0074] Figure 4 This is a user-friendly interface showing the start and output results of the multi-objective collaborative irrigation nitrogen application decision-making method based on AI agents for water, food, ecology, and economy in this embodiment of the invention.

[0075] Figure 5 This invention illustrates the regulatory effect of a suitable irrigation nitrogen application strategy on the "water-food-ecology-economy" system obtained in the embodiments of the present invention. Detailed Implementation

[0076] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings. Example 1:

[0077] like Figure 1 As shown, a multi-objective collaborative irrigation nitrogen application intelligent decision-making method includes the following steps:

[0078] S1. Scenario simulation is performed using the AHC model, and then the file reading plugin in the COZE agent framework is used to obtain the simulation decision variables:

[0079] S1.1. Using the AHC model, water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value, and economic benefits under the preset irrigation nitrogen application scenario are used as simulation decision variables, such as... Figure 2 and Figure 3 As shown; Figure 2 The data shows irrigation and nitrogen application scenarios for spring wheat in the Hetao Irrigation District. Data on different irrigation and nitrogen application scenarios for other crop types are also available. Figure 3 Examples are shown in the simulation of water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value, and economic benefits of wheat farmland in the Hetao Irrigation District in spring 2020 under different irrigation and nitrogen application scenarios. Data for different irrigation and nitrogen application scenarios for other crop types in different years can also be obtained through simulation.

[0080] S1.2. Use the file reading plugin in the COZE agent framework to extract the simulated decision variables, and then use the code plugin to write data processing code to remove spaces and commas from each decision variable file and convert the data type to floating point.

[0081] S2. Parallel analysis of decision variables using three multi-objective decision-making methods was conducted to obtain preliminary suitable irrigation nitrogen application strategies for the three methods:

[0082] S2.1 Deploy the three decision-making methods VIKOR, AHP, and TOPSIS in the code plugin and establish the corresponding multi-objective decision-making plugin;

[0083] The core algorithm of the VIKOR decision model is as follows:

[0084] ;

[0085] ;

[0086] ;

[0087] In the formula: x ij Let be the value of the j-th decision indicator under the i-th irrigation nitrogen application scenario; A j + The j-th decision indicator is the optimal solution across all scenarios; A j - The j-th decision indicator is the worst solution in all scenarios; S i Let be the distance between scenario i and the optimal solution; R i Let be the distance between scenario i and the worst solution; w j Let be the weight of the j-th decision indicator;v The decision coefficient represents the decision-maker's preference; Q i The decision benefit ratio for each scenario is given, with the minimum value corresponding to the optimal scenario.

[0088] The decision coefficient is 0.5.

[0089] The core algorithm of the TOPSIS decision model is as follows:

[0090] ;

[0091] ;

[0092] ;

[0093] ;

[0094] In the formula: C i Let be the score for the i-th irrigation nitrogen application scenario, with the maximum value corresponding to the best scenario; D i + Let be the distance between the i-th scenario and the ideal solution; D i - Let be the distance between the i-th scenario and the worst solution; w j Let be the weight of the j-th decision indicator; r ij Let be the standardized value of the j-th decision indicator in scenario i; x ij Let be the original value of the j-th decision indicator for scenario i.

[0095] Assuming acceptable consistency of the decision matrix, the core calculation formula of the AHP decision model is as follows:

[0096] ;

[0097] ;

[0098] In the formula: CR For consistency ratio, when CR < 0.1, the consistency of the decision matrix is ​​acceptable; λ max It is the largest eigenvalue of the decision matrix; S i This is the score for scenario i, with the highest value corresponding to the best scenario; n The number of scenarios involving nitrogen application during irrigation; w j Let be the weight of the j-th decision indicator; x ijLet be the original value of the j-th decision indicator for scenario i.

[0099] Among the three decision-making methods, the weights of the decision variables—water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value, and economic benefits—can be determined according to the actual situation of the study area and the needs of the managers.

[0100] S2.2. The simulated decision variables formatted in step S2.1 are used as input data for three decision algorithms to perform parallel decision analysis, obtain the corresponding first irrigation nitrogen application strategy, and the first irrigation nitrogen application decision variable data (water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value and economic benefits) for each strategy, and output the irrigation amount, nitrogen application amount and corresponding first irrigation nitrogen application decision variable data of the first irrigation nitrogen application strategy.

[0101] In the three decision-making methods, water use efficiency, crop yield, positive ecosystem service value, and economic benefits are set as positive decision variables, while negative ecosystem service value is set as a negative decision variable.

[0102] S3. Using a large language model, make a one-step decision on the first irrigation nitrogen application strategy obtained in step S2 to obtain the final suitable second irrigation nitrogen application strategy:

[0103] S3.1 Use the DeepSeek model to evaluate the first irrigation nitrogen application decision variable data (water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value, and economic benefits) corresponding to the first irrigation nitrogen application strategy obtained by the VIKOR, AHP, and TOPSIS methods in step S2.

[0104] The judgment process of the DeepSeek model is as follows:

[0105] First, the VIKOR first irrigation nitrogen application strategy and the VIKOR first irrigation nitrogen application decision variable data are evaluated:

[0106] 1. The crop yield corresponding to the first irrigation nitrogen application strategy is less than the crop yield of the local control irrigation nitrogen application strategy by a preset threshold of 0.05. The local control irrigation nitrogen application strategy is preset to scenario 1.

[0107] 2. The water use efficiency, positive ecosystem service value, and economic benefits corresponding to the first irrigation nitrogen application strategy are higher than those of the local control irrigation nitrogen application strategy;

[0108] 3. The negative ecosystem service value corresponding to the first irrigation nitrogen application strategy is lower than that of the local control irrigation nitrogen application strategy;

[0109] If the above three judgment conditions are met, then the VIKOR first irrigation nitrogen application strategy generated by the VIKOR method decision is determined to be the final suitable second irrigation nitrogen application strategy;

[0110] If any of the above three judgment conditions are not met, the same judgment process is then applied to the results generated by the AHP and TOPSIS decision-making methods in turn until the final suitable second irrigation nitrogen application strategy is obtained.

[0111] If the irrigation nitrogen application strategies and corresponding results generated by the three decision-making methods all meet the above judgment conditions, then the irrigation nitrogen application strategy that obtains the highest yield will be taken as the final suitable second irrigation nitrogen application strategy.

[0112] S3.2 Using the DeepSeek model, extract the second irrigation nitrogen application decision variable data (crop yield, water use efficiency, positive ecosystem service value, negative ecosystem service value, and economic benefits) corresponding to the final suitable second irrigation nitrogen application strategy from the decision variable data obtained in step S1, and calculate the percentage change of these indicators compared with the local control irrigation nitrogen application strategy.

[0113] S3.3. Using the agent's termination plugin, output the irrigation amount, nitrogen application amount, and decision variable data extracted in step S3.2 corresponding to the final suitable second irrigation nitrogen application strategy, as well as the percentage change compared to the local control irrigation nitrogen application strategy. The output interface is as follows: Figure 4 As shown, the regulatory effect of appropriate irrigation nitrogen application strategies on the "water-food-ecology-economy" system is as follows: Figure 5 As shown.

[0114] The above embodiments are only used to illustrate the present invention and are not intended to limit the present invention. Those skilled in the art can make various changes and modifications without departing from the essence and scope of the present invention. Therefore, all equivalent technical solutions also fall within the protection scope of the present invention.

[0115] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A multi-objective collaborative intelligent decision-making method for nitrogen application in irrigation, characterized in that, Includes the following steps: The AHC model was used to simulate the decision variables under a pre-defined irrigation nitrogen application scenario. The simulated decision variables are extracted using the COZE agent framework, and the data type of the simulated decision variables is converted into floating point type to obtain the simulated decision variable data; Within the COZE agent framework, the simulated decision variable data are input into the VIKOR decision model, AHP decision model, and TOPSIS decision model respectively for parallel decision analysis, to obtain the first irrigation nitrogen application strategy under the best scenario and the corresponding first irrigation nitrogen application decision variable data. Within the COZE agent framework, a large language model is used to judge the first irrigation nitrogen application strategy and the first irrigation nitrogen application decision variable data until the second irrigation nitrogen application strategy is obtained. Extract the decision variable data for the second irrigation nitrogen application strategy; Calculate the percentage change in the second irrigation nitrogen application decision variable data compared to the local control irrigation nitrogen application strategy's local control nitrogen application decision variable data; Output the irrigation amount, nitrogen application amount, and percentage change corresponding to the second irrigation nitrogen application strategy; The simulation decision variables include water use efficiency, crop yield, positive ecosystem service value, negative ecosystem service value, and economic benefits. The water use efficiency, crop yield, positive ecosystem service value, and economic benefits are set as positive decision variables, and the negative ecosystem service value is set as a negative decision variable. The preset irrigation nitrogen application scenarios are scenarios with different irrigation and nitrogen application rates determined based on the actual conditions of the study area.

2. The intelligent decision-making method for multi-objective collaborative irrigation nitrogen application according to claim 1, characterized in that: The core algorithm of the VIKOR decision model is as follows: ； ； ； In the formula: x ij Let be the value of the j-th decision indicator under the i-th irrigation nitrogen application scenario; A j + The j-th decision indicator is the optimal solution across all scenarios; A j - The j-th decision indicator is the worst solution in all scenarios; S i Let be the distance between scenario i and the optimal solution; R i Let be the distance between scenario i and the worst solution; w j Let be the weight of the j-th decision indicator; v The decision coefficient represents the decision-maker's preference; Q i The decision benefit ratio for each scenario is given, with the minimum value corresponding to the optimal scenario.

3. The intelligent decision-making method for multi-objective collaborative irrigation nitrogen application according to claim 1, characterized in that: The core algorithm of the TOPSIS decision model is as follows: ； ； ； ； In the formula: C i Let be the score for the i-th irrigation nitrogen application scenario, with the maximum value corresponding to the best scenario; D i + Let be the distance between the i-th scenario and the ideal solution; D i - Let be the distance between the i-th scenario and the worst solution; w j Let be the weight of the j-th decision indicator; r ij Let be the standardized value of the j-th decision indicator in scenario i; x ij Let be the original value of the j-th decision indicator for scenario i.

4. The intelligent decision-making method for multi-objective collaborative irrigation nitrogen application according to claim 1, characterized in that, Assuming acceptable consistency of the decision matrix, the core calculation formula of the AHP decision model is as follows: ； ； In the formula: CR For consistency ratio, when CR < 0.1, the consistency of the decision matrix is ​​acceptable; λ max It is the largest eigenvalue of the decision matrix; i>S i This is the score for scenario i, with the highest value corresponding to the best scenario; n The number of scenarios involving nitrogen application during irrigation; w j Let be the weight of the j-th decision indicator; x ij Let be the original value of the j-th decision indicator for scenario i.

5. The intelligent decision-making method for multi-objective collaborative irrigation nitrogen application according to claim 1, characterized in that: The first irrigation nitrogen application strategy includes the VIKOR first irrigation nitrogen application strategy, the AHP first irrigation nitrogen application strategy, and the TOPSIS first irrigation nitrogen application strategy; The first irrigation nitrogen application decision variable data includes VIKOR first irrigation nitrogen application decision variable data, AHP first irrigation nitrogen application decision variable data, and TOPSIS first irrigation nitrogen application decision variable data.

6. The intelligent decision-making method for multi-objective collaborative irrigation nitrogen application according to claim 5, characterized in that: The first irrigation nitrogen application strategy and the first irrigation nitrogen application decision variable data are evaluated until a second irrigation nitrogen application strategy is obtained. The process is as follows: Based on three judgment conditions, the data of the VIKOR first irrigation nitrogen application strategy and the VIKOR first irrigation nitrogen application decision variables are sequentially evaluated. Data on AHP first irrigation nitrogen application strategy and AHP first irrigation nitrogen application decision variables; In addition, the TOPSIS first irrigation nitrogen application strategy and the TOPSIS first irrigation nitrogen application decision variable data are used for judgment; The first irrigation nitrogen application strategy that simultaneously meets all three judgment conditions is determined as the second irrigation nitrogen application strategy; If the number of first irrigation nitrogen application strategies that simultaneously meet all three judgment conditions is greater than 1, then the first irrigation nitrogen application strategy that achieves the highest yield will be used as the second irrigation nitrogen application strategy.

7. The intelligent decision-making method for multi-objective collaborative irrigation nitrogen application according to claim 6, characterized in that: The three judgment conditions include: The crop yield reduction of the first irrigation nitrogen application strategy compared to the local irrigation nitrogen application strategy is less than a preset threshold; The water use efficiency, positive ecosystem service value, and economic benefits of the first irrigation nitrogen application strategy are higher than those of the local irrigation nitrogen application strategy. The negative ecosystem service value corresponding to the first irrigation nitrogen application strategy is lower than that of the local irrigation nitrogen application strategy.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-7.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-7.

Images