Method for optimizing co-regulation of water-salinity and pollutants

Abstract

The present application relates to the technical field of water-salt and pollutant collaborative regulation optimization, and particularly relates to a water-salt and pollutant collaborative regulation optimization method. Spatial position information and attribute information corresponding to strip fields, drainage ditches, ecological sinks, strong drainage stations and external drainage river outlets in a target region are acquired; a target directed graph network model is constructed according to the spatial position information and attribute information corresponding to the strip fields, drainage ditches, ecological sinks, strong drainage stations and external drainage river outlets; and a water-salt and pollutant collaborative regulation optimization scheme corresponding to the target region is determined based on the target directed graph network model. The facility operation parameters (such as the split ratio) are optimized in a targeted manner, the conflict between the salt control and pollution reduction objectives is resolved, fine configuration is achieved, the treatment efficiency is improved, and the demand for comprehensive treatment of the water environment in the irrigation area is met.

Description

Technical Field

[0001] This invention relates to the field of synergistic regulation and optimization technology of water, salt and pollutants, and specifically to a method for synergistic regulation and optimization of water, salt and pollutants. Background Technology

[0002] Traditional hydrological models are ill-suited to the flat terrain, fragmented irrigation canals, and dispersed discharge characteristics of coastal plain irrigation areas. Furthermore, they rely heavily on numerous parameters that are difficult and costly to obtain, resulting in high computational complexity and failing to meet the real-time optimization and management needs of large-scale irrigation districts. Simultaneously, existing treatment methods lack a systematic approach, often focusing on single objectives and failing to adequately consider the water-salt-nutrient coupling effect. The operating parameters of facilities such as ecological ingestion ponds are frequently set based on experience, failing to be precisely configured according to the pollution load of individual fields and the treatment capacity of the facilities, leading to low treatment efficiency.

[0003] Existing technologies fail to construct precise quantitative models by combining the spatial and attribute characteristics of core elements such as irrigation strips, drainage ditches, and ecological absorption ponds. This makes it impossible to achieve coordinated simulation and optimization of water, salt, and pollutant migration and control schemes, thus making it difficult to solve the core problems of conflict between salt control and pollution reduction and low treatment efficiency.

[0004] Therefore, developing a collaborative regulation and optimization method based on network model construction using key element information of irrigation areas has become an urgent need for the comprehensive management of water environment in coastal saline-alkali irrigation areas. Summary of the Invention

[0005] This invention provides a method for the synergistic regulation and optimization of water, salt and pollutants, in order to solve the core problems of the inability to achieve synergistic simulation and regulation scheme optimization of water, salt and pollutant migration, the difficulty in solving the conflict between salt control and pollution reduction and the low treatment efficiency.

[0006] In a first aspect, the present invention provides a method for the synergistic regulation and optimization of water, salt, and pollutants. The method includes: acquiring spatial location information and attribute information corresponding to strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and outflowing river outlets in a target area; constructing a target directed graph network model based on the spatial location information and attribute information corresponding to strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and outflowing river outlets; and determining a synergistic regulation and optimization scheme for water, salt, and pollutants in the target area based on the target directed graph network model.

[0007] In one optional implementation, a target directed graph network model is constructed based on the spatial location and attribute information corresponding to each strip of farmland, drainage ditch, ecological infiltration pond, forced drainage station, and outflow river outlet. This includes: identifying each strip of farmland in the target area as a source node; identifying each drainage ditch in the target area as a transmission node; identifying each ecological infiltration pond in the target area as a processing node; identifying each forced drainage station in the target area as a control node; identifying each outflow river outlet in the target area as a collection node; determining the node attributes corresponding to each source node, transmission node, processing node, control node, and collection node based on the spatial location and attribute information corresponding to each strip of farmland, drainage ditch, ecological infiltration pond, forced drainage station, and outflow river outlet; and determining the target directed edges corresponding to each source node, transmission node, processing node, control node, and collection node based on the spatial location and attribute information corresponding to each strip of farmland, drainage ditch, ecological infiltration pond, forced drainage station, and outflow river outlet, thus obtaining the target directed graph network model.

[0008] In one optional implementation, the transmission relationship of the target directed graph network model is as follows: each source node transmits to each transmission node, each transmission node transmits to each processing node, each transmission node transmits to each control node, each processing node transmits to each control node, and each control node transmits to each collection node. Based on the spatial location and attribute information corresponding to the strip fields, drainage ditches, ecological absorption ponds, forced drainage stations, and outflow river outlets, the target directed edges corresponding to the source nodes, transmission nodes, processing nodes, control nodes, and collection nodes are determined, including: for two types of nodes with potential transmission relationships, calculating the distance matching degree, water flow adaptability, and service matching degree between the current node and potential transmission nodes; calculating the spatial correlation degree based on the distance matching degree, water flow adaptability, and service matching degree; determining the target directed edges between the current node and potential transmission nodes based on the spatial correlation degree; and adding target edge attribute information to each target directed edge.

[0009] In one optional implementation, the target edge attribute information includes at least one of basic attribute information, dynamic attribute information, and optimized attribute information. The dynamic attribute information includes a dynamic flow threshold and an attenuation coefficient. For each target directed edge, target edge attribute information is added, including: determining the basic attribute information corresponding to each target directed edge; the basic attribute information includes at least one of the edge type, edge start point, edge end point, edge length, and edge creation time corresponding to each target directed edge; wherein the edge transmitted from each source node to each transmission node is field drainage; calculating the dynamic flow threshold corresponding to each target directed edge based on its edge type; calculating the attenuation coefficient corresponding to the transmission-type target directed edge based on its edge type; the attenuation coefficient is used to characterize the attenuation parameters of water salt and pollutants along the path; calculating the initial flow allocation ratio corresponding to each target directed edge based on its edge type; and determining the initial flow allocation ratio as the optimized attribute information corresponding to the target directed edge.

[0010] In one optional implementation, based on a target directed graph network model, an optimization scheme for the coordinated regulation of water, salt, and pollutants corresponding to a target area is determined, including: calculating the first nitrogen load, first phosphorus load, and first salt load corresponding to the source nodes according to the attribute information and hydrological scenario parameters in the target directed graph network model; calculating the nitrogen load, phosphorus load, and salt load corresponding to the transmission nodes, processing nodes, control nodes, and collection nodes in the target directed graph network model in sequence according to the target directed graph network model; using the target flow allocation ratio corresponding to each target directed edge in the target directed graph network model as a decision variable; constructing an objective function, which includes a first sub-function, a second sub-function, and a third sub-function; the first sub-function is used to characterize the minimum nitrogen load corresponding to the collection node; the second sub-function is used to characterize the minimum phosphorus load corresponding to the collection node; the third sub-function is used to characterize the maximum salt load corresponding to the collection node; solving for the target flow allocation ratio corresponding to the target directed edge based on a preset multi-objective evolutionary algorithm; and determining the optimization scheme for the coordinated regulation of water, salt, and pollutants corresponding to the target area according to the target flow allocation ratio.

[0011] In one optional implementation, the first nitrogen load, first phosphorus load, and first salinity load corresponding to the source nodes are calculated based on the attribute information and hydrological scenario parameters corresponding to the source nodes in the target directed graph network model. This includes: obtaining the hydrological scenario parameters corresponding to each source node, including rainfall, irrigation, runoff coefficient, nitrogen leaching rate, phosphorus leaching rate, and salinity concentration coefficient; for each source node, calculating the total drainage volume of the source node based on the area of ​​the source node and the rainfall, irrigation, and runoff coefficient in the hydrological scenario parameters; calculating the first nitrogen load and first phosphorus load of the source node based on the nitrogen fertilizer application rate, phosphorus fertilizer application rate, area, nitrogen leaching rate, and phosphorus leaching rate of the source node; and calculating the first salinity load of the source node based on the total drainage volume, the soil salinity attribute of the source node, and a preset salinity concentration coefficient.

[0012] In one optional implementation, based on the target directed graph network model, the nitrogen load, phosphorus load, and salt load corresponding to the transmission nodes, processing nodes, control nodes, and sink nodes in the target directed graph network model are calculated sequentially, including: for transmission nodes, determining the first nitrogen input, first phosphorus input, and first salt input corresponding to each transmission node according to the target directed graph network model; multiplying the first nitrogen input, first phosphorus input, and first salt input by the attenuation coefficient corresponding to the target directed edge between the source node and the transmission node to obtain the second nitrogen load, second phosphorus load, and second salt load corresponding to the transmission node; for processing nodes, based on the target directed graph network model... The second nitrogen input, second phosphorus input, and second salt input corresponding to each processing node are determined. Based on the second nitrogen input, second phosphorus input, nitrogen absorption rate, and phosphorus absorption rate corresponding to the processing node, the third nitrogen load and third phosphorus load corresponding to the processing node are calculated. Based on the second salt input and salt absorption rate corresponding to the processing node, the third salt load corresponding to the processing node is calculated. For control nodes, the fourth nitrogen load, fourth phosphorus load, and fourth salt load corresponding to each control node are determined according to the target directed graph network model. For pooling nodes, the fifth nitrogen load, fifth phosphorus load, and fifth salt load corresponding to each pooling node are determined according to the target directed graph network model.

[0013] In one optional implementation, the target flow allocation ratio corresponding to the target directed edge is solved based on a preset multi-objective evolutionary algorithm, including: constructing constraints based on the attribute information of each node in the target directed graph network model; the constraints include at least one of basic constraints, economic cost constraints, and operation and maintenance constraints; and using the preset multi-objective evolutionary algorithm, the objective function is solved based on the constraints to obtain the target flow allocation ratio corresponding to the target directed edge.

[0014] In one optional implementation, a preset multi-objective evolutionary algorithm is used to solve the objective function based on constraints to obtain the target flow allocation ratio corresponding to the target directed edge. This includes: randomly generating a preset number of initial decision variable vectors; the initial decision variable vectors include random decision variable vectors, empirical decision variable vectors, and boundary decision variable vectors; based on constraints, deleting initial decision variable vectors that do not meet the constraints from each initial decision variable vector to obtain the remaining decision variable vectors; calculating the first sub-function value, second sub-function value, and third sub-function value corresponding to each remaining decision variable vector based on the objective function; and determining the remaining decision variable vectors based on the first sub-function value, second sub-function value, and third sub-function value corresponding to each remaining decision variable vector. The corresponding penalty coefficient is used; remaining decision variable vectors with penalty coefficients greater than a preset threshold are marked as to be eliminated; the dominance relationship between each remaining effective decision variable vector is calculated; based on the dominance relationship, each remaining effective decision variable vector is stratified to obtain multiple levels; the smaller the level corresponding to the remaining effective decision variable vector, the better the remaining effective decision variable vector is; for each first decision variable vector in the first level, the number of subordinate decision vectors dominated by each first decision variable vector is calculated; candidate decision variable vectors with a number of subordinate decision vectors greater than a preset threshold are selected from each first decision variable vector; based on each candidate decision variable vector, the target decision variable vector is determined; based on the target decision variable vector, the target flow allocation ratio corresponding to the target directed edge is obtained.

[0015] In one optional implementation, determining the target decision variable vector based on each candidate decision variable vector includes: sorting each candidate decision variable vector in ascending order according to the values ​​of a first sub-function to obtain a first sequence; sorting each candidate decision variable vector in ascending order according to the values ​​of a second sub-function to obtain a second sequence; sorting each candidate decision variable vector in descending order according to the values ​​of a third sub-function to obtain a third sequence; calculating a first range corresponding to the values ​​of the first sub-functions based on the first sequence; calculating a second range corresponding to the values ​​of the second sub-functions based on the second sequence; calculating a third range corresponding to the values ​​of the third sub-functions based on the third sequence; calculating, for each candidate decision variable vector, the difference between the candidate decision variable vector and its adjacent candidate decision variable vector in each sequence; and dividing the difference corresponding to each sequence by the global range of that sequence to obtain the target decision variable vector. The distance components of each candidate decision variable vector in each sequence are calculated; based on the distance components of each candidate decision variable vector in each sequence, the crowding distance corresponding to the candidate decision variable vector is obtained; based on the crowding distance corresponding to each candidate decision variable vector, backup decision variable vectors with crowding distances greater than a preset distance threshold are selected; two backup decision variable vectors are selected from the backup decision variable vectors as parent decision variable vectors and cross-operated to generate two offspring decision variable vectors; the variable length is calculated based on the current generation corresponding to the offspring decision variable vectors; the variable length is used to mutate the offspring decision variable vectors to obtain mutated decision variable vectors; the parent decision variable vectors and the mutated decision variable vectors are merged into a temporary population; the above steps are repeated based on the temporary population until the preset number of iterations is obtained to obtain the target decision variable vector.

[0016] The water-salt and pollutant synergistic regulation optimization method provided in this application acquires the spatial location and attribute information of strip farmland, drainage ditches, ecological infiltration ponds, forced drainage stations, and outflow river outlets in the target area. Accurate collection of location and attribute data for core facilities such as strip farmland and drainage ditches provides a foundation for subsequent model construction, solving the problems of fuzzy parameter acquisition and poor adaptability in traditional technologies, ensuring that the model fits the actual situation of the irrigation area. Based on the spatial location and attribute information of strip farmland, drainage ditches, ecological infiltration ponds, forced drainage stations, and outflow river outlets, a target directed graph network model is constructed. This breaks through the limitations of traditional hydrological model division, adapting to the crisscrossing and dispersed discharge characteristics of plain irrigation areas. The directed graph clearly depicts the migration paths of water, salt, and pollutants, reducing model complexity, improving simulation accuracy, and providing a quantitative tool for synergistic regulation. Based on the target directed graph network model, the corresponding water-salt and pollutant synergistic regulation optimization scheme for the target area is determined. Based on the objective-oriented directed graph network model, the collaborative simulation of water, salt and pollutant migration is realized, and the facility operation parameters (such as the diversion ratio) are optimized in a targeted manner to resolve the conflict between salt control and pollution reduction objectives, achieve refined configuration, improve treatment efficiency, and meet the comprehensive water environment management needs of irrigation areas. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the first process of the synergistic regulation and optimization method of water, salt and pollutants according to an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of the second process of the water-salt and pollutant synergistic regulation and optimization method according to an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] According to an embodiment of the present invention, an embodiment of a method for synergistic regulation and optimization of water, salt and pollutants is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0022] This embodiment provides a method for synergistic regulation and optimization of water, salt, and pollutants, which can be used in electronic devices. Figure 1 This is a flowchart of the water-salt and pollutant synergistic regulation and optimization method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0023] Step S101: Obtain the spatial location information and attribute information of the strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and external discharge river outlets in the target area.

[0024] Specifically, the electronic device can receive the spatial location and attribute information of the strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and outflow river outlets in the target area input by the user.

[0025] Step S102: Based on the spatial location and attribute information of the strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and external discharge river outlets, construct a target directed graph network model.

[0026] Specifically, electronic devices can determine each node and the connection relationships between them based on the spatial location and attribute information of each strip of land, drainage ditch, ecological ingestion pond, forced drainage station, and outflow river outlet, and construct a target directed graph network model.

[0027] This step will be explained in detail below.

[0028] Step S103: Based on the target directed graph network model, determine the optimization scheme for the coordinated regulation of water, salt and pollutants in the target area.

[0029] Specifically, electronic devices can determine the pollutant load and salt load corresponding to each node based on the target directed graph network model, and then determine the water-salt and pollutant synergistic regulation optimization scheme corresponding to the target area based on the pollutant load and salt load corresponding to each node.

[0030] This step will be explained in detail below.

[0031] The water-salt and pollutant synergistic regulation optimization method provided in this application acquires the spatial location and attribute information of strip farmland, drainage ditches, ecological infiltration ponds, forced drainage stations, and outflow river outlets in the target area. Accurate collection of location and attribute data for core facilities such as strip farmland and drainage ditches provides a foundation for subsequent model construction, solving the problems of fuzzy parameter acquisition and poor adaptability in traditional technologies, ensuring that the model fits the actual situation of the irrigation area. Based on the spatial location and attribute information of strip farmland, drainage ditches, ecological infiltration ponds, forced drainage stations, and outflow river outlets, a target directed graph network model is constructed. This breaks through the limitations of traditional hydrological model division, adapting to the crisscrossing and dispersed discharge characteristics of plain irrigation areas. The directed graph clearly depicts the migration paths of water, salt, and pollutants, reducing model complexity, improving simulation accuracy, and providing a quantitative tool for synergistic regulation. Based on the target directed graph network model, the corresponding water-salt and pollutant synergistic regulation optimization scheme for the target area is determined. Based on the objective-oriented directed graph network model, the collaborative simulation of water, salt and pollutant migration is realized, the facility operation parameters are optimized in a targeted manner, the conflict between salt control and pollution reduction objectives is resolved, refined configuration is achieved, treatment efficiency is improved, and the comprehensive water environment management needs of irrigation areas are met.

[0032] This embodiment provides a method for synergistic regulation and optimization of water, salt, and pollutants, which can be used in electronic devices. Figure 2 This is a flowchart of the water-salt and pollutant synergistic regulation and optimization method according to an embodiment of the present invention, such as... Figure 2As shown, the process includes the following steps:

[0033] Step S201: Obtain the spatial location information and attribute information of the strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and outflow river outlets in the target area.

[0034] Please refer to the above description of step S101 for details on this step, which will not be repeated here.

[0035] Step S202: Based on the spatial location and attribute information of the strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and external discharge river outlets, construct a target directed graph network model.

[0036] Specifically, step S202 above may include the following steps:

[0037] Step S2021: Identify each field in the target area as a source node.

[0038] Specifically, electronic devices can extract the vector boundaries of all fields based on GIS spatial data of the target area (such as remote sensing images, land use maps, and farmland ownership data). Plots that are too small (e.g., less than 100m²) or have no cultivation function (e.g., areas occupied by field ridges or ditches) are eliminated to ensure that the source nodes correspond to actual cultivated units.

[0039] Then, the electronic equipment can assign a unique number to each field using the format "F + three digits," starting from F001 and incrementing sequentially, for example, F001, F002, ..., F020 (assuming there are 20 fields in the target area). The numbering is associated with the spatial location of the fields (e.g., numbering from west to east, or from north to south), facilitating subsequent tracing. In the directed graph network model, an independent node is created for each identified field, labeled as "source" for the node type and Level 1 for the level, thus completing the initial determination of the source node.

[0040] Step S2022: Identify each drainage ditch in the target area as a transmission node.

[0041] Specifically, the electronic device can distinguish the levels (agricultural ditches, branch ditches, tributary ditches, main ditches) and confluence relationships (upstream ditch → confluence point → downstream ditch) of drainage ditches based on GIS ditch vector data. Then, the electronic device extracts the key confluence points of the drainage ditches, with each confluence point corresponding to a transmission node. The electronic device can use the format "FD + three-digit number" to uniformly number each drainage ditch, starting from FD001 and incrementing sequentially, for example, FD001, FD002, ..., FD005 (assuming there are 5 core confluence points in the target area). The numbering is associated with the drainage ditch level (e.g., agricultural ditch confluence points are numbered earlier, and main ditch confluence points are numbered later), facilitating the differentiation of transmission levels. In the directed graph network model, the electronic device creates an independent node for each confluence point, labeled as "transport" for the node type and Level 2 for the level, completing the initial determination of the transmission node.

[0042] Step S2023: Identify each ecological absorption pond in the target area as a treatment node.

[0043] Specifically, the electronic equipment collects construction files and operation and maintenance records of ecological wastewater treatment ponds in the target area, and filters out the normally operating wastewater treatment ponds. Based on GIS spatial data, the electronic equipment obtains the accurate location coordinates of each ecological wastewater treatment pond, ensuring its spatial correlation with the drainage ditch system.

[0044] Electronic devices can use the format "EP + two digits" to uniformly number each ecological treatment pond, starting from EP01 and incrementing sequentially, such as EP01, EP02, ..., EP05. The number is associated with the corresponding drainage ditch number, facilitating subsequent edge construction. In the directed graph network model, the electronic device creates an independent node for each effective ecological treatment pond, labeled as "treatment" for the node type and Level 2.5 for the level, thus completing the initial determination of the treatment node.

[0045] Step S2024: Identify each strong drainage station in the target area as a control node.

[0046] Specifically, electronic equipment can be used in conjunction with water conservancy facility ledgers to verify the location, operational status, and service area of ​​forced drainage stations within the target area. It can also exclude backup forced drainage stations or facilities that have failed to operate, ensuring that control nodes correspond to actually available drainage power units.

[0047] Then, the electronic device can assign a unified number to each pumping station using the format "PS + two digits", starting from PS01 and incrementing sequentially, for example, PS01, PS02, ..., PS05. This number is associated with the corresponding service's transmission / processing node number (e.g., the pumping station number for services FD001 and EP01 is PS01), clarifying the service relationship. In the directed graph network model, the electronic device creates an independent node for each valid pumping station, labeled as "pumping_station" for the node type and Level 4 for the level, completing the initial determination of the control node.

[0048] Step S2025: Identify each outflow channel outlet in the target area as a collection node.

[0049] Specifically, the electronic equipment can determine the final destination of the drainage from the forced drainage station in the target area based on water system GIS data, and clarify the specific geographical coordinates of the outlet. If there are multiple discharge outlets in the target area, they need to be merged according to river affiliation to ensure the consistency of load statistics. A fixed number "OUT01" is used to uniformly number each discharge river outlet. The number is associated with the name of the discharge river for easy identification. In the directed graph network model, the electronic equipment creates nodes for the discharge river outlets, with the node type labeled "sink" and the level labeled Level5, completing the initial determination of the aggregation nodes.

[0050] Step S2026: Based on the spatial location and attribute information of the strip fields, drainage ditches, ecological absorption ponds, forced drainage stations, and external discharge river outlets, determine the node attributes corresponding to the source node, transmission node, processing node, control node, and collection node.

[0051] Specifically, the basic attributes of the source node (strip field) include: Spatial location attributes: x-coordinate, y-coordinate (latitude and longitude or planar coordinates of the geometric center of the strip field). Cultivation attributes: area (area, unit m²), crop_type (crop type, such as "rice" or "wheat"). Soil and fertilization attributes: salinity (soil salinity, unit g / kg), N_fertilizer (nitrogen fertilizer application rate, unit kg / ha), P_fertilizer (phosphate fertilizer application rate, unit kg / ha). Association attribute: pumping_station (corresponding control node number, such as "PS01").

[0052] The basic attributes of the transmission node (drainage ditch) include: Spatial location attributes: x-coordinate, y-coordinate (geographic coordinates of the confluence point). Ditch attributes: ditch_type (ditch type, such as "agricultural ditch confluence point" or "small ditch confluence point"), length (corresponding ditch length, in meters), width (ditch width, in meters). The basic attributes of the treatment node (ecological absorption pond) include: Spatial location attributes: x-coordinate, y-coordinate (center coordinates of the absorption pond). Functional attributes: volume (volume, in m³), ​​N_retention (nitrogen absorption rate, value 0-1), P_retention (phosphorus absorption rate, value 0-1), salt_retention (salt absorption rate, value 0-1), max_capacity (maximum pollutant treatment capacity, in kg / d).

[0053] The basic attributes of the control node (forced drainage station) include: Spatial location attributes: x-coordinate, y-coordinate (actual location coordinates of the forced drainage station). Operational attributes: capacity (drainage capacity, unit m³ / h), power (operating power, unit kW), cost_per_hour (operating cost per hour, unit yuan / h).

[0054] The basic attributes of the aggregation node (outlet of the discharge channel) include: Spatial location attributes: x-coordinate, y-coordinate (geographic coordinates of the outlet). Identification attribute: outlet_name (name of the discharge channel, such as "Main Discharge Channel of XX Irrigation District").

[0055] Step S2027: Based on the spatial location and attribute information of the strip fields, drainage ditches, ecological absorption ponds, forced drainage stations, and external discharge river outlets, determine the target directed edges corresponding to the source node, transmission node, processing node, control node, and collection node, and obtain the target directed graph network model.

[0056] Specifically, the transmission relationship of the target directed graph network model is as follows: each source node transmits to each transmission node, each transmission node transmits to each processing node, each transmission node transmits to each control node, each processing node transmits to each control node, and each control node transmits to each sink node. The step S2027 above, "determining the target directed edges corresponding to the source nodes, transmission nodes, processing nodes, control nodes, and sink nodes based on the spatial location and attribute information corresponding to the strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and external discharge river outlets," can include the following steps:

[0057] Step a1: For the two types of nodes with potential transmission relationships, calculate the distance matching degree, water flow adaptability degree, and service matching degree between the current node and the potential transmission nodes.

[0058] Specifically, determine the node pair type, clarify the "current node" and "potential transmission node" (e.g., the current node is the source node F001, and the potential transmission nodes are FD001 and FD002), and then determine the existence of a potential transmission relationship between the "current node" and the "potential transmission node".

[0059] Then, the electronic device can calculate the Euclidean distance between the current node and the potential transmission node based on the x and y coordinates of the two nodes. The Euclidean distance formula is as follows: Then, the electronic device can determine the maximum Euclidean distance dist_max between nodes of the same type, count the Euclidean distance dist for all pairs of nodes of the same type, and take the maximum value (e.g., the maximum dist value for all F→FD node pairs is 500m). Then, the electronic device substitutes these values ​​into the formula to calculate R. dist : The distance matching degree between the current node and the potential transmission node is calculated.

[0060] This only applies to three types of edges directly associated with the ditch transmission function: "field drainage edge (F→FD)," "entry into the influent pond edge (FD→EP)," and "direct discharge edge (FD→PS)." The electronic equipment can acquire the core parameters of the ditch and extract the slope and width of the corresponding ditch from the transmission node attributes. Then, the slope and width are normalized to the 0~1 range respectively to obtain the slope. std and width std Then, the electronic device calculates the water flow fit Rflow based on the formula: Rflow = 0.5·slope std +0.5 width std .

[0061] For the field drainage edge (F→FD), the electronic device obtains the "pumping_station" attribute (corresponding to the forced drainage station number) of the current source node (F) and checks whether it matches the number of the control node (PS) connected downstream of the potential transmission node (FD). If the results match, the service matching degree Rservice=1; otherwise, the service matching degree Rservice=0.

[0062] For the process from FD to EP, the electronic device obtains the "ditch_type" attribute (channel type, such as agricultural ditch or distribution ditch) of the current transmission node (FD) and checks whether it matches the "service channel type" attribute of the potential processing node (EP). If the results match, the service matching degree Rservice=1; otherwise, the service matching degree Rservice=0.

[0063] Step a2: Calculate the spatial correlation degree based on the distance matching degree, water flow adaptability degree, and service matching degree.

[0064] Specifically, electronic devices can perform weighted calculations on distance matching degree, water flow adaptability degree, and service matching degree to obtain spatial correlation degree. For example, substituting into the weighted formula, the calculation is: R = 0.6·Rdist + 0.3·Rflow + 0.1·Rservice.

[0065] Step a3: Based on spatial correlation, determine the target directed edge between the current node and potential transmission nodes.

[0066] Specifically, the electronic device can filter candidate transmission nodes whose spatial correlation with each potential transmission node is greater than a preset spatial correlation threshold based on the spatial correlation between the current node and each potential transmission node. If the number of candidate transmission nodes is greater than one, the candidate transmission node with the highest spatial correlation is determined as the target transmission node, and the target directed edge between the current node and the target transmission node is determined. The electronic device constructs backup edges with the remaining candidate transmission nodes. For example, if a node has multiple potential connection objects (e.g., source node F001 can connect to FD001 and FD002, and both have R ≥ 0.6), the electronic device selects the object with the largest R to construct the main edge (e.g., FD001 has R = 0.85, FD002 has R = 0.72, the main edge is F001→FD001). A backup edge (F001→FD002) is constructed for the second-best object, with a default flow. ratio =0 (no traffic allocation), activated only when the main edge is overloaded or fails.

[0067] Step a4: Add target edge attribute information for each target directed edge.

[0068] Specifically, the target edge attribute information includes at least one of basic attribute information, dynamic attribute information, and optimized attribute information, wherein the dynamic attribute information includes a dynamic flow threshold and an attenuation coefficient. Step a4 above may include the following steps:

[0069] Step a41: Determine the basic attribute information corresponding to the directed edges of each target.

[0070] The basic attribute information includes at least one of the following: edge type, edge start point, edge end point, edge length, and edge creation time for each target directed edge. The edges transmitted from each source node to each transmission node represent field drainage.

[0071] Specifically, the electronic device can determine the basic attribute information corresponding to the directed edges of each target based on the starting point and ending point of each directed edge.

[0072] In this sequence: Source node → Transmission node: Edge type is field drainage. Transmission node → Processing node: Edge type is into the treatment pond. Transmission node → Control node: Edge type is direct discharge edge. Processing node → Control node: Edge type is outflow from the treatment pond. Control node → Collection node: Edge type is forced discharge.

[0073] Step a42: Calculate the dynamic flow threshold corresponding to each target's directed edge based on the edge type.

[0074] Specifically, electronic devices can filter target directed edges according to the edge type corresponding to the target directed edge. The necessary parameters for calculation (such as area, volume, and drainage capacity) are extracted from the corresponding node attributes. Then, these parameters are substituted into the corresponding formula to calculate the dynamic flow threshold Q. edge,max Keep two decimal places. Record it in the edge's attributes and name it "Q". edge,max ".

[0075] For example, for the field drainage edge (F→FD), the calculation formula is: Q edge,ma x=S field 0.001, where S field This represents the area attribute (unit: m²) of the source node (F), with 0.001 being the drainage coefficient per unit area (m³ / (h·m²)), which can be adjusted according to actual needs. Example: If the area of ​​source node F001 is 5000m², then Q... edge,max =5000 × 0.001 = 5.00m 3 / h.

[0076] For the boundary between the inlet and outlet of the treatment pool (FD→EP), the calculation formula is: Q edge,max =VEP·0.1, where VEP is the volume attribute (unit: m³) of the endpoint treatment node (EP), and 0.1 is the unit volume influent coefficient (m³ / (h·m³)), which can be modified according to actual needs. For example, if the volume of treatment node EP01 is 200m³, then Q... edge,max =200 × 0.1 = 20.00m 3 / h.

[0077] For a straight edge (FD→PS), the calculation formula is: Q edge,max =QPS·0.3, where QPS is the capacity (drainage capacity) attribute (unit: m³ / h) of the endpoint control node (PS), and 0.3 is the upper limit of the direct discharge flow rate, which can be adjusted according to actual needs. For example, if the capacity of control node PS01 is 500 m³ / h, then QPS = QPS·0.3. edge,max =500 × 0.3 = 150.00m3 / h.

[0078] For the outflow edge of the absorption pool (EP→PS), the calculation formula is: Q edge,max =VEP·0.12, where VEP is the volume attribute (unit: m³) of the final treatment node (EP), and 0.12 is a slightly higher drainage coefficient (to ensure smooth drainage from the collection tank), which can be adjusted according to actual needs. Example: If the volume of treatment node EP01 is 200 m³, then Q edge,max =200 × 0.12 = 24.00m 3 / h.

[0079] For strong discharge from the water's edge (PS→OUT), the calculation formula is: Q edge,max =QPS, where QPS is the capacity (drainage capacity) attribute (unit: m³ / h) of the endpoint control node (PS). For example, if the capacity of control node PS01 is 500 m³ / h, then QPS... edge,max =500.00m 3 / h.

[0080] Step a43: Calculate the attenuation coefficient corresponding to the directed edge of the transmission target based on the edge type of each target's directed edge.

[0081] The attenuation coefficient is used to characterize the attenuation parameters of water salt and pollutants along the flow path.

[0082] Specifically, the electronic device can filter out three types of directed edges for transmission targets, and then extract the length attribute and the width attribute of the starting transmission node for each directed edge. Then, the length attribute and the width attribute of the starting transmission node are substituted into the corresponding formula to calculate the attenuation coefficient k. edge Keep two decimal places. Record it in the attribute of the target directed edge and name it "k". edge "This attribute is not required for non-transfer class edges."

[0083] For example, this study only considers three types of transport edges: field drainage edges, inlet / outlet disposal pond edges, and direct discharge edges, reflecting the degree of nitrogen and phosphorus pollutant loss during ditch transport. Attenuation coefficient k edge The calculation formula is: k edge=0.01·Ledge / Wditch. Where Ledge is the "length" attribute of the edge (unit: m). Wditch is the "channel width" attribute of the starting transmission node (FD) (unit: m, which needs to be added to the transmission node attributes beforehand). 0.01 is the basic attenuation coefficient (to ensure the value is reasonable) and can be adjusted according to actual needs. For example, if the length of the edge FD001→EP01 entering the absorption pool is 100m, and the channel width of FD001 is 2m, then kedge = 0.01 × (100 / 2) = 0.50.

[0084] Step a44: Calculate the initial flow allocation ratio corresponding to the directed edge of each target based on the edge type.

[0085] Specifically, the electronic device can calculate the initial flow allocation ratio corresponding to the target directed edge according to the edge type. The edge entering the digestion pool and the straight discharge edge need to be calculated in pairs to ensure that the sum is 1.

[0086] For example, the calculation formula for the inlet / outlet boundary (FD→EP) is as follows: .in, P is the initial flow allocation ratio corresponding to the edge of the receiving pool. EP Priority is assigned to the endpoint processing node (EP) (levels 1-5, with higher absorption rates indicating higher priority; for example, EP01 with an absorption rate of 45% corresponds to P). EP =5). P FD Priority levels (1-5) are assigned to the starting transmission node (FD), with higher channel levels having higher priority, such as the main canal confluence point corresponding to P. FD =5). P max =5 is the maximum priority. 0.2 is the base value (ensuring at least 20% of the flow is directly discharged to avoid overloading the collection pool; this can be modified according to actual needs). 0.6 is an adjustment coefficient that amplifies the impact of priority on the diversion ratio; this can be adjusted according to actual needs.

[0087] For straight edges (FD→PS), the calculation formula is: flow_ratio_init=1-α init `flow_ratio_init` represents the initial flow allocation ratio corresponding to the incoming straight-line edge. The sum of the flow allocation ratios of the incoming and straight-line edges of the same starting transmission node (FD) must be 1 to ensure flow balance.

[0088] Electronic devices can assign the same path_id to two edges (the edge entering the absorption pool + the straight-line edge) of the same starting transmission node (FD), in the format of "FD number + _path". Example: The path_id of FD001→EP01 and FD001→PS01 is both "FD001_path".

[0089] The flow distribution ratio for the field drainage side, the outflow side of the influent pond, and the side with strong discharge is fixed at 1.0, i.e., f low_ratio_init =1.0. This type of edge has no traffic splitting requirement; all upstream traffic is transmitted to downstream nodes through this edge.

[0090] Optionally, to limit fluctuations in the traffic allocation ratio during optimization and avoid system instability caused by extreme values, electronic devices can set a traffic allocation ratio range for each edge. For example, the edge entering the absorption pool: [0, 0.8] (up to 80% of traffic enters the absorption pool to prevent overload). The edge directly flowing through the absorption pool: [0.2, 1.0] (at least 20% of traffic flows directly through the absorption pool, complementary to the range of the edge entering the absorption pool). Other edges: [1.0, 1.0] (fixed, no fluctuation range).

[0091] Core objective: To associate the input / output pool edge and the direct output edge of the same starting point transmission node, so as to facilitate dual-path collaborative optimization (such as synchronous adjustment of the split ratio).

[0092] Step a45: Determine the initial flow allocation ratio as the optimized attribute information corresponding to the target directed edge.

[0093] Specifically, the electronic equipment can allocate the initial flow rate corresponding to the edge of the absorption pool (FD→EP). The optimized attribute information corresponding to the target directed edge is determined.

[0094] Step S203: Based on the target directed graph network model, determine the optimization scheme for the coordinated regulation of water, salt and pollutants in the target area.

[0095] Specifically, step S203 above may include the following steps:

[0096] Step S2031: Based on the attribute information corresponding to the source nodes in the target directed graph network model and the hydrological scenario parameters, calculate the first nitrogen load, the first phosphorus load and the first salt load corresponding to the source nodes.

[0097] Specifically, step S2031 above may include the following steps:

[0098] Step b1: Obtain the hydrological scenario parameters corresponding to each source node.

[0099] The hydrological scenario parameters include rainfall, irrigation, runoff coefficient, nitrogen leaching rate, phosphorus leaching rate, and salt concentration coefficient.

[0100] Specifically, the electronic device can receive user input of rainfall amount, rainfall intensity, irrigation amount, runoff coefficient, and salt concentration coefficient.

[0101] Electronic devices can calculate nitrogen and phosphorus leaching rates based on rainfall intensity. The nitrogen leaching rate function is: LN =a·Rb·f(G), where: R is rainfall intensity (mm / h), a and b are experimental fitting coefficients (e.g., a=0.02, b=0.8), and f(G) is the growth period correction factor (seedling stage=0.6, growing stage=0.8, maturity stage=1.2, fallow stage=1.5). Phosphorus leaching rate function: L P =c·Rd·f(G)·S, where S is the soil clay content (the higher the proportion of clay, the stronger the phosphorus adsorption, and S takes the value of 0.7-1.3), and c and d are the experimental fitting coefficients.

[0102] Step b2: For each source node, calculate the total drainage volume of the source node based on the area of ​​the source node and the rainfall, irrigation and runoff coefficients in the hydrological scenario parameters.

[0103] Specifically, for each source node, the electronic equipment can calculate the total drainage volume of the source node based on the area of ​​the source node and the rainfall, irrigation amount and runoff coefficient in the hydrological scenario parameters. The formula is as follows: Among them, Q total γ represents the total drainage volume of the source node (unit: m³). A represents the area of ​​the source node (strip field) (unit: m²). P represents the rainfall (unit: mm). I represents the irrigation amount (unit: mm). γ represents the runoff coefficient (dimensionless).

[0104] Step b3: Calculate the first nitrogen load and first phosphorus load of the source node based on the amount of nitrogen fertilizer applied, the amount of phosphorus fertilizer applied, the area, the nitrogen leaching rate, and the phosphorus leaching rate.

[0105] Specifically, the electronic device can calculate the first nitrogen load corresponding to the source node based on the nitrogen fertilizer application rate, phosphorus fertilizer application rate, area, nitrogen leaching rate, and phosphorus leaching rate of the source node. The calculation formula is as follows: Among them, N load1 The first nitrogen load (unit: kg). F N This represents the nitrogen fertilizer application rate at the source node (unit: kg / hm², obtained from the source node attributes, e.g., 355 kg / hm² for rice, 294 kg / hm² for wheat). L N is the nitrogen leaching rate (dimensionless, obtained in step b1). A is the source node area (unit: m²).

[0106] The electronic device can calculate the corresponding first phosphorus load of the source node based on the nitrogen fertilizer application rate, phosphorus fertilizer application rate, area, nitrogen leaching rate, and phosphorus leaching rate of the source node. The calculation formula is as follows: .in, The first phosphorus load (unit: kg). F P Phosphate fertilizer application rate at the source node (unit: kg / hm) 2 Obtained from the source node attributes, such as 270 kg / hm of rice. 2Wheat 162kg / hm 2 ). Phosphorus leaching rate (dimensionless, obtained in step b1). A is the source node area (unit: m²). 2 ).

[0107] Step b4: Calculate the first salt load of the source node based on the total drainage volume, the soil salinity attributes of the source node, and the preset salt concentration coefficient.

[0108] Specifically, the electronic device can calculate the first salt load of the source node based on the total drainage volume, the soil salinity attributes of the source node, and a preset salt concentration coefficient. The calculation formula is: Salt load1 =S soil ×K c ×Q total Salt load1 The first salt load (unit: kg). S soil The soil salinity at the source node (unit: g / kg, obtained from the source node attributes, e.g., 3.5 g / kg for coastal saline-alkali land). K c Q is the salt concentration coefficient (dimensionless, obtained in step b1). total The total drainage volume of the source node (unit: m³, calculation result of step b2).

[0109] Step S2032: Based on the target directed graph network model, calculate the nitrogen load, phosphorus load, and salt load corresponding to the transmission node, processing node, control node, and collection node in the target directed graph network model in sequence.

[0110] Specifically, step S2032 above may include the following steps:

[0111] Step c1: For each transmission node, determine the first nitrogen input, first phosphorus input, and first salt input based on the target directed graph network model.

[0112] Specifically, electronic devices can determine the current transmission node FD. j All predecessor source nodes F i (Obtained by traversing the incoming edges of the target directed graph). Then, the first nitrogen input, first phosphorus input, and first salt input corresponding to each transmission node are calculated based on the following formula.

[0113] Specifically, the first nitrogen input is: The first phosphorus input is: The first salt input amount is: The initial inflow rate is: Among them, F i For transmission node FD jThe predecessor source node (such as F001, F002, i.e., related to FD) j (A strip of land connected by the drainage edge of the field). For source node F i The first nitrogen load (kg).

[0114] For field drainage edge F i →FD j The traffic allocation ratio (0-1, with the primary edge defaulting to 1.0 and the backup edge defaulting to 0; this can be dynamically adjusted during optimization). ∑ represents the traffic allocation ratio for all nodes connected to the transmission node FD. j Summing of connected source nodes (ensuring all upstream loads are included). FD j This refers to the current transmission node being calculated (e.g., FD001).

[0115] Step c2: Multiply the first nitrogen input, the first phosphorus input, and the first salt input by the attenuation coefficient corresponding to the target directed edge between the source node and the transmission node to obtain the second nitrogen load, the second phosphorus load, and the second salt load corresponding to the transmission node.

[0116] Specifically, the dwell time t is The attenuation factor is: Where, length: is the channel length (m, taken from the length attribute of the edge). velocity: is the preset flow velocity of the water (m / s, such as 0.5m / s). k: attenuation coefficient (1 / day, the empirical value for nitrogen and phosphorus is 0.1, which can be adjusted according to the actual situation). exp() is the natural exponential function.

[0117] Specifically, the second nitrogen load is: N out2 =N in1 ×decay, the second phosphorus load is: P out2 =P in1 ×decay.

[0118] Second salt load: salt out2 =salt in1 (No salinity reduction). Second outlet water flow rate: water out2 =water in1

[0119] Step c3: For each processing node, determine the second nitrogen input, second phosphorus input, and second salt input based on the target directed graph network model.

[0120] Specifically, for each processing node, the second nitrogen input is: The second phosphorus input is: The second salt input amount is: The second inflow rate is: .

[0121] Among them, FD j For processing node EP k The predecessor transmission node (such as FD001, and EP) k (Connected via the side of the inlet / outlet pool). For FD at the edge of the disposal pool j →EP k The initial flow allocation ratio (i.e., the initial flow allocation ratio α, a decision variable, dynamically adjusted during optimization, ranging from [0, 0.8]). EP k This represents the current processing node (e.g., EP01). The meanings of the remaining characters are the same as in step c1.

[0122] Step c4: Based on the second nitrogen input, the second phosphorus input, the nitrogen absorption rate and phosphorus absorption rate corresponding to the processing node, calculate the third nitrogen load and the third phosphorus load corresponding to the processing node.

[0123] Specifically, the electronic device can receive the nitrogen and phosphorus absorption rates input by the user. Then, it multiplies the second nitrogen input by 1 and subtracts the difference in nitrogen absorption rates to obtain the third nitrogen load. At the same time, it multiplies the second phosphorus input by 1 and subtracts the difference in phosphorus absorption rates to obtain the third phosphorus load.

[0124] Step c5: Based on the second salt input and the salt absorption rate corresponding to the processing node, calculate the third salt load corresponding to the processing node.

[0125] Specifically, the electronic device can receive the salt absorption rate input by the user. Then, the difference between the second salt input and the salt absorption rate is multiplied by 1 to calculate the third salt load corresponding to the processing node.

[0126] Step c6: For each control node, determine the fourth nitrogen load, fourth phosphorus load, and fourth salt load corresponding to each control node based on the target directed graph network model.

[0127] Specifically, the fourth nitrogen load is: The fourth phosphorus load is: The fourth salt load is: The fourth inflow rate is: .

[0128] Where pred is the control node PS m Predecessor nodes, including transport nodes FD j (Straight edge) and processing node EP k (Outflow edge). The flow allocation ratio for the straight-flow side is 1 - α (complementary to α on the side entering the absorption pool), and the flow allocation ratio for the outflow side is 1.0 (all processed flow enters the forced-flow station). PSm This is the control node currently being calculated (e.g., PS01).

[0129] Step c7: For the pooling nodes, determine the fifth nitrogen load, fifth phosphorus load, and fifth salt load corresponding to each pooling node based on the target directed graph network model.

[0130] Specifically, the fifth nitrogen load is: The fifth phosphorus load is: The fifth salt load is: The total drainage volume is: .

[0131] Among them, PS m This refers to all control nodes (e.g., PS01-PS05, covering the entire target area). ∑ represents the sum of loads over all control nodes (ensuring no load omissions across the entire area). The fifth load is the output load of the aggregation node, corresponding to outlet_N, outlet_P, and outlet_salt in the optimization objective function.

[0132] Step S2033: The target flow allocation ratio corresponding to each target directed edge in the target directed graph network model is used as the decision variable.

[0133] Specifically, electronic devices can integrate all decision variables (optimizable flow allocation ratios) into a decision variable vector, in the following format: Where n is the number of optimizable edges (only edges entering the influent pool; the 1-α of straight edges is automatically derived from α and does not need to be considered as a separate decision variable). For example, if the target area has 5 edges entering the influent pool (EP01-EP05), then... Each element corresponds to the flow distribution ratio of each edge.

[0134] Step S2034: Construct the objective function.

[0135] The objective function comprises a first sub-function, a second sub-function, and a third sub-function. The first sub-function characterizes the minimum nitrogen load corresponding to the pooling node. The second sub-function characterizes the minimum phosphorus load corresponding to the pooling node. The third sub-function characterizes the maximum salt load corresponding to the pooling node.

[0136] Specifically, the mathematical expression for the first sub-function (nitrogen load minimization) is: Under the constraint of the decision variable vector X, minimize the final nitrogen load N of the pooling node. load5 (i.e., outlet_N in the optimization objective).

[0137] The mathematical expression for the second sub-function (phosphorus load minimization) is: Under the constraint of the decision variable vector X, minimize the final phosphorus load P of the pooling node. load5 (i.e., outlet_P in the optimization objective).

[0138] The mathematical expression for the third subfunction (salt load maximization) is: Under the constraint of the decision variable vector X, maximize the final salt load (salt) of the pooling node. load5 (i.e., outlet_salt in the optimization objective). Salt load is only slightly reduced in the absorption pool (salt absorption rate is usually low, such as 0.1). The smaller the α value (the larger the direct discharge ratio), the smaller the proportion of salt load reduction, and the larger the final output. Therefore, salt load needs to be maximized by reducing the α value.

[0139] The three sub-functions inherently conflict and cannot simultaneously achieve optimality. Increasing α (more flow into the treatment pond): nitrogen and phosphorus load reductions increase, f1 and f2 decrease (meeting the first two objectives), but salt load reduction increases, and f3 decreases (violating the third objective). Decreasing α (more flow directly discharged): salt load reduction decreases, f3 increases (meeting the third objective), but nitrogen and phosphorus load reductions decrease, and f1 and f2 increase (violating the first two objectives).

[0140] Therefore, electronic devices need to find a Pareto optimal set of solutions, that is, there are no other solutions that can improve one objective without worsening another.

[0141] Step S2035: Based on the preset multi-objective evolution algorithm, solve for the target flow allocation ratio corresponding to the target directed edge.

[0142] Specifically, step S2035 above may include the following steps:

[0143] Step d1: Based on the attribute information corresponding to each node in the target directed graph network model, construct the constraints.

[0144] The constraints include at least one of the following: basic constraints, economic cost constraints, and operation and maintenance constraints.

[0145] Specifically, electronic devices can construct basic constraints based on the attribute information of each node and edge in the target directed graph network model, limiting the physical boundaries of decision variables; these are the core constraints. These basic constraints include:

[0146] Traffic allocation ratio range constraints: The flow splitting ratio α at the inlet of the collection pool should not exceed 0.8 (to avoid overloading the collection pool) and should not be lower than 0 (to avoid the collection pool being idle). The flow splitting ratio at the outlet side is complementary to α to ensure flow balance.

[0147] Node carrying capacity constraints: Processing Node (EP) k Nitrogen input N in2,k Not exceeding its maximum nitrogen load capacity (To avoid overloading the disposal tank and affecting the treatment effect).

[0148] The derived formula is: , To enter EP k The split ratio.

[0149] Edge flow threshold constraint: The actual flow of directed edges for each target. Not exceeding its dynamic flow threshold (Avoid overloading of ditches and drainage stations).

[0150] The derived formula is: ,in, The flow allocation ratio for the edge.

[0151] Electronic devices can construct economic cost constraints based on the operating costs of forced drainage stations and wastewater treatment ponds, thus limiting the economic feasibility of the optimization results.

[0152] The economic cost constraint is: Among them, C PS,m For control node PS m The unit operating cost (yuan / m³, e.g., 0.05 yuan / m³). Q PS,m For PS m The inflow rate (m³). C EP,k For processing node EP k The unit operating cost (yuan / m³, e.g., 0.1 yuan / m³). Q EP,k For EP k The inflow rate (m³). C max The maximum permissible daily operating cost for the target area (RMB / day, set according to the irrigation district budget). Increasing α will increase the operating cost of the influent pond (increased processing flow), while decreasing α will increase the operating cost of the forced discharge station (increased direct discharge flow). The constraint ensures that the total operating cost does not exceed the budget.

[0153] Electronic equipment can be configured with operation and maintenance constraint functions based on the irrigation district's operation and maintenance capabilities to limit the operability of the optimization results. Specifically, the operation and maintenance constraint functions include the following:

[0154] Flow split ratio adjustment range constraint: Decision variables Optimized value and initial value The difference does not exceed the maximum adjustment range. (e.g., 0.2). Avoid sudden changes in the diversion ratio (e.g., from 0.3 to 0.8), which may prevent maintenance personnel from making timely adjustments, or cause the absorption pool and forced drainage station to be unable to adapt to sudden changes in traffic.

[0155] Forced load balancing constraints at drainage stations: Actual flow rate of each strong drainage station With its maximum drainage capacity The ratio should be between 30% and 90%. Avoid leaving the power station idle (load < 30%) or operating at full load (load > 90%) to extend equipment life and improve operational stability.

[0156] It's important to note the priority order: Basic constraints > Operational constraints > Economic cost constraints. During optimization, first ensure that basic constraints are met (physical feasibility is a prerequisite). Once basic constraints are met, then meet operational constraints (ensuring practical operability). Finally, find the optimal solution within the economic cost constraints; if this cannot be satisfied, the economic cost constraints can be appropriately relaxed (or the budget adjusted).

[0157] Step d2: Using a preset multi-objective evolutionary algorithm, the objective function is solved based on the constraints to obtain the target flow allocation ratio corresponding to the target directed edge.

[0158] Specifically, step d2 above may include the following steps:

[0159] Step d21: Randomly generate a preset number of initial decision variable vectors.

[0160] The initial decision variable vector includes a random decision variable vector, an empirical decision variable vector, and a boundary decision variable vector.

[0161] The preset number (population size) can be set in conjunction with the complexity of the optimization problem, usually between 50 and 200 (e.g., 100), to ensure that the population has sufficient diversity without causing a surge in computation due to excessive size.

[0162] Specifically, the electronic device can allocate the number of three types of vectors according to the proportions of random decision variable vectors, empirical decision variable vectors, and boundary decision variable vectors (e.g., for a population size of 100: 70 random, 20 empirical, and 10 boundary). The three types of vectors are generated separately, ensuring that the element α of each vector... i All vectors are within the range of [0, 0.8]. Integrate all vectors to form an initial set of decision variable vectors (initial population).

[0163] Specifically, for the random decision variable vector (accounting for 60%~70%), the electronic equipment can generate the diversion ratio α for each side of the inlet and outlet pool by uniform random sampling based on the range of decision variable values ​​[0, 0.8]. i .

[0164] Generating formula: α i =rand(0,0.8) (i=1,2,...,n) where rand(a,b) represents a uniformly randomized real number generated within the interval [a,b].

[0165] For the empirical decision variable vector (accounting for 20%~30%), electronic equipment can set reasonable diversion ratio combinations based on the irrigation district's historical operation and maintenance experience, node priorities, or expert knowledge to ensure that the vector closely approximates the actual feasible solution. Generation basis: Higher priority absorption pools (larger PEP) correspond to higher α. i (e.g., 0.6~0.8), fully utilizing its processing capacity. Low-priority disposal pools (small PEP) correspond to medium α. i (e.g., 0.3~0.5), balancing processing and straight-line requirements. For example, if EP01 and EP03 are high priority, the generated vector Xexp=[0.72,0.45,0.68,0.38,0.55].

[0166] For the boundary decision variable vector (accounting for 10%), the electronic device can take boundary values ​​(0, 0.8) or near-boundary values ​​(0.1, 0.7) within the range of decision variable values ​​to cover the feasible region boundary and avoid the algorithm missing the boundary optimal solution. Example: X bound1 =[0.8,0.0,0.8,0.0,0.8],X bound2 =[0.1,0.7,0.1,0.7,0.1].

[0167] Step d22: Based on the constraints, delete the initial decision variable vectors that do not meet the constraints from each initial decision variable vector to obtain the remaining decision variable vectors.

[0168] Specifically, the electronic device can extract each initial decision variable vector from the initial population one by one. Electronic device verification Does the system satisfy all constraints? The electronic device retains all vectors that satisfy the constraints, forming a set of remaining decision variable vectors (feasible population). If the number of feasible population vectors is less than the preset number, randomly generated vectors are added and the population is re-selected to ensure that the population size is not lower than the minimum requirement (e.g., 50).

[0169] Step d23: Based on the objective function, calculate the first sub-function value, the second sub-function value, and the third sub-function value corresponding to each remaining decision variable vector.

[0170] Specifically, electronic devices can extract the vector of remaining decision variables. Each element α in i(The diversion ratio at the edge of the influent collection pool). Then, based on the target directed graph network model, the electronic equipment calculates the second nitrogen load, second phosphorus load, and second salt load corresponding to the transmission node in sequence according to the load transfer process. The third nitrogen load, third phosphorus load, and third salt load corresponding to the processing node are then calculated. The fourth nitrogen load, fourth phosphorus load, and fourth salt load corresponding to the control node are calculated. Finally, the fifth nitrogen load, fifth phosphorus load, and fifth salt load corresponding to the collection node are obtained. Then, the fifth nitrogen load, fifth phosphorus load, and fifth salt load corresponding to the collection node are used as three sub-function values ​​respectively: .

[0171] Step d24: Determine the penalty coefficients corresponding to the remaining decision variable vectors based on the first sub-function values, second sub-function values, and third sub-function values ​​corresponding to each remaining decision variable vector.

[0172] Specifically, electronic devices can be penalized using a two-dimensional approach: "degree of constraint violation + deviation of objective function value," as shown in the following formula:

[0173] .

[0174] Among them, the degree of constraint violation For vectors that satisfy the constraints, For minor violations of constraints (such as α) i Vectors with a value of 0.81 are quantized according to the degree of violation: (Normalized to [0,1]). Objective function value deviation. Calculate the deviation between the vector and the mean of the objective function value of the current population, normalize it, and use it as a penalty term. .in, Let be the mean of the k-th sub-function of the current population. and This is weight information.

[0175] Step d25: Mark the remaining decision variable vectors with penalty coefficients greater than a preset threshold as pending elimination.

[0176] Specifically, the electronic device can set a preset threshold for the penalty coefficient (e.g., 0.3, which can be adjusted according to the actual situation). Then, the electronic device marks the remaining decision variable vectors with penalty coefficients greater than the threshold as to be eliminated, further screening the population and retaining high-quality vectors.

[0177] Step d26: Calculate the dominance relationships among the remaining effective decision variable vectors.

[0178] Specifically, the electronic device identifies the remaining decision variable vectors that are not marked as to be eliminated as the remaining valid decision variable vectors. For two remaining valid decision variable vectors... and If the following conditions are met, then it is called Dominate (recorded as) ):

[0179] 1. Strictly superior to at least one subfunction :

[0180] (Minimize the target).

[0181] (Maximize the goal).

[0182] 2. Strictly superior to at least one subfunction :

[0183] or .

[0184] The electronic device records the dominance relationships among all remaining decision variable vectors based on the above calculation results (such as constructing a dominance matrix).

[0185] For example, if there are 4 valid vectors ( Dominance matrix middle express Dominate 0 indicates no control:

[0186]

[0187] The above Characterization Dominate , . Dominate , , .

[0188] Step d27: Based on the dominance relationship, the remaining effective decision variable vectors are stratified to obtain multiple levels.

[0189] Among them, the smaller the level corresponding to the remaining effective decision variable vector, the better the remaining effective decision variable vector is.

[0190] Specifically, the electronic device performs non-dominated sorting of the remaining effective decision variable vectors based on dominance relationships, dividing the population into multiple levels. The smaller the level (such as the first level), the better the vector (Pareto optimal), providing a basis for subsequent screening of candidate decision variable vectors.

[0191] Specifically, initialize the first level (Rank=1): collect all vectors that are not dominated by any other vector (i.e., vectors whose rows in the dominating matrix are all 0). These vectors constitute the first level (Pareto optimal front). Initialize the second level (Rank=2): after deleting the vectors from the first level, collect the vectors that are not dominated by any other remaining vectors, forming the second level. Repeat the above steps until all valid vectors have been assigned a level.

[0192] For example, based on the dominance matrix above: First level (Rank=1): (Not dominated by any vector). Second level (Rank=2): (only by) Dominated (not dominated by other residual vectors). Third level (Rank=3): , (quilt , control).

[0193] Step d28: For each first decision variable vector in the first level, calculate the number of lower-level decision vectors dominated by each first decision variable vector.

[0194] Specifically, the electronic device can extract all the first decision variable vectors in the first level (such as...). Then, based on the dominance matrix, count the number of lower-level vectors dominated by each first-level vector (i.e., the number of 1s in the corresponding column of the lower-level vector in the dominance matrix), and record the dominance count of each first-level vector.

[0195] For example, the first-level vector The dominance matrix has row vectors of [1,1,0,1] and lower-level vectors of [1,1,0,1]. , , ,but The number of units controlled is 3.

[0196] Step d29: Select candidate decision variable vectors from each first decision variable vector where the number of lower-level decision vectors is greater than a preset threshold.

[0197] Specifically, the electronic device can compare the number of subordinate decision vectors dominated by each first decision variable vector with a preset threshold. Then, it can select candidate decision variable vectors from the first decision variable vectors whose number of subordinate decision vectors exceeds the preset threshold.

[0198] Step d210: Determine the target decision variable vector based on each candidate decision variable vector.

[0199] Specifically, step d210 above may include the following steps:

[0200] Step d2101: Sort the candidate decision variable vectors in ascending order according to the values ​​of the first sub-functions to obtain the first sequence.

[0201] Specifically, the electronic device can sort the candidate decision variable vectors in ascending order of the first sub-function values ​​corresponding to each candidate decision variable vector to obtain the first sequence.

[0202] Step d2102: Sort the candidate decision variable vectors in ascending order according to the values ​​of the second sub-functions to obtain the second sequence.

[0203] Specifically, the electronic device can sort the candidate decision variable vectors in ascending order of the second sub-function values ​​corresponding to each candidate decision variable vector to obtain a second sequence.

[0204] Step d2103: Sort the candidate decision variable vectors in descending order according to the values ​​of the third sub-functions to obtain the third sequence.

[0205] Specifically, the electronic device can sort the candidate decision variable vectors in descending order of the third sub-function values ​​corresponding to each candidate decision variable vector to obtain a third sequence.

[0206] Step d2104: Based on the first sequence, calculate the first range corresponding to the first sub-function value.

[0207] Specifically, the electronic device obtains the maximum and minimum values ​​in the first sequence, and calculates the first range corresponding to the first sub-function value based on the maximum and minimum values ​​in the first sequence.

[0208] Step d2105: Based on the second sequence, calculate the second range corresponding to the second sub-function value.

[0209] Specifically, the electronic device obtains the maximum and minimum values ​​in the second sequence, and calculates the second range corresponding to the second sub-function value based on the maximum and minimum values ​​in the second sequence.

[0210] Step d2106: Based on the third sequence, calculate the third range corresponding to the third sub-function value.

[0211] Specifically, the electronic device obtains the maximum and minimum values ​​in the third sequence, and calculates the third range corresponding to the third sub-function value based on the maximum and minimum values ​​in the third sequence.

[0212] Step d2107: For each candidate decision variable vector, calculate the difference between the candidate decision variable vector and its adjacent candidate decision variable vectors in each sequence. Divide the difference for each sequence by the global range of that sequence to obtain the distance component of the candidate decision variable vector in each sequence.

[0213] Specifically, for each candidate decision variable vector in each sequence, identify the left and right adjacent candidate decision variable vectors. Then, calculate the difference in sub-function values ​​between adjacent vectors (note the direction of the difference for salt loading sequences). Divide the difference by the range of the corresponding sequence to obtain the distance component, and record the distance components (d1, d2, d3) of each vector in the three sequences.

[0214] Specifically, for the candidate decision variable vector in the middle of the sequence: distance component = (difference between left and right adjacent vectors + difference between right and right adjacent vectors) / corresponding range. For the sequence boundary vectors (first / last): distance component = difference between one-sided adjacent vectors / corresponding range (only the difference between existing adjacent vectors is calculated).

[0215] Step d2108: Based on the distance components of each candidate decision variable vector in each sequence, obtain the crowding distance corresponding to the candidate decision variable vector.

[0216] Specifically, the electronic device can sum the distance components of each candidate decision variable vector in each sequence to calculate the crowding distance corresponding to the candidate decision variable vector.

[0217] Step d2109: Based on the crowding distances corresponding to each candidate decision variable vector, select the backup decision variable vector whose crowding distance is greater than a preset distance threshold.

[0218] The preset distance threshold can be adjusted according to the number and diversity requirements of candidate decision variable vectors, and is usually set to 0.8~1.2 (e.g., 1.0).

[0219] Specifically, the electronic device can compare the crowding distance corresponding to the candidate decision variable vector with a preset distance threshold. Then, based on the comparison result, it selects a backup decision variable vector from among the candidate decision variable vectors whose crowding distance is greater than the preset distance threshold.

[0220] Step d21010: Select two alternative decision variable vectors from each alternative decision variable vector as parent decision variable vectors and perform a cross operation to generate two child decision variable vectors.

[0221] Specifically, the electronic device can randomly select two alternative decision variable vectors from the alternative decision variable vectors as parent decision variable vectors. The electronic device can perform cross calculations on each alternative decision variable vector according to the cross probability, and perform boundary corrections on the cross-generated offspring decision variable vectors to generate two effective offspring decision variable vectors.

[0222] Wherein, the crossover probability is set as follows: highly correlated decision variables (such as the entry / exit pool edge α corresponding to the same transmission node): crossover probability p c=0.95, enhancing synergy. Low-correlation decision variable: cross probability p c =0.8, balancing exploration and convergence.

[0223] For each decision variable α i Generate random numbers ,like If so, then crossover is performed according to the following formula: , where β is the distribution exponent (usually taken as 2.0 to control the crossover strength). After crossover, if... If the value exceeds the range [0, 0.8], force truncation to the boundary (e.g., α). i1 (The value of ′=0.85 was corrected to 0.8).

[0224] Step d21011: Calculate the variable asynchronous length based on the current generation corresponding to the offspring decision variable vector.

[0225] Specifically, the electronic device can calculate the variable length based on the current generation corresponding to the offspring decision variable vector, using the following formula: Where δ is the variable asynchronous length (range [0, 0.1]). This represents the current iteration number (starting from 1). Set the preset total number of iterations (e.g., 100).

[0226] Step d21012 involves performing a mutation operation on the offspring decision variable vector based on the variable asynchronous length to obtain the mutated decision variable vector.

[0227] Specifically, the electronic device can set the α corresponding to the high-priority processing node (EP). i The mutation probability p m =0.05 (reduces high-quality solution perturbation). α corresponding to low-priority processing nodes. i The mutation probability p m =0.15 (Enhanced Exploration).

[0228] The variation range of electronic equipment is calculated based on the variable asynchronous length, where the variation range = ±δ × 0.8 (0.8 is α). i The maximum value is determined to ensure that the result remains within the range of [0, 0.8] after mutation. For each offspring decision variable vector, the electronic device generates random numbers. ,like This triggers a mutation. If a mutation is triggered, α is randomly perturbed within the mutation range. i This ensures that the mutated variables are within the range of [0, 0.8], thus obtaining the mutated decision variable vector.

[0229] Step d21013: Merge the parent decision variable vectors and the mutated decision variable vectors into a temporary population.

[0230] Specifically, electronic devices can merge the parent decision variable vectors and the mutated decision variable vectors into a temporary population.

[0231] Step d21014: Based on the temporary population, repeat the above steps until the preset number of iterations is reached to obtain the target decision variable vector.

[0232] Specifically, the electronic device can repeat the above steps based on a temporary population until a preset number of iterations are reached, and then select the target decision variable vector from the final Pareto optimal solution set according to the above screening rules.

[0233] Step d211: Based on the target decision variable vector, obtain the target flow allocation ratio corresponding to the target directed edge.

[0234] Specifically, the electronic device is based on all α values ​​in the target decision variable vector. i (Flow distribution ratio at the inlet / outlet pool). Derive the flow distribution ratio at the straight outlet (1-α). i Fix the flow allocation ratio of other edges to 1.0. Organize the target flow allocation ratios of all target directed edges to form an edge-allocation ratio mapping table.

[0235] Step S2036: Based on the target flow allocation ratio, determine the optimization scheme for the coordinated regulation of water, salt and pollutants in the target area.

[0236] Specifically, the electronic equipment determines the optimized scheme for the coordinated regulation of water, salt, and pollutants in the target area based on the target flow allocation ratio.

[0237] The water-salt and pollutant synergistic regulation and optimization method provided in this application clearly defines the model roles of each core facility through node type classification, and clearly defines the starting point, path, treatment unit, control unit, and endpoint of water-salt and pollutant migration, laying the foundation for accurate simulation. It determines node attributes and integrates the spatial and attribute data of each facility to make the model parameters more realistic, solving the problems of vague parameters and poor adaptability in traditional models. Then, it calculates spatial correlation to scientifically determine the transmission relationship between nodes, avoiding subjective assumptions and ensuring that directed edges conform to the water flow and pollutant migration patterns in the irrigation area. It determines the target directed edges and basic attributes, clearly depicting the migration paths of water-salt and pollutants, adapting to the crisscrossing characteristics of ditches in plain irrigation areas, and breaking through the limitations of traditional model division. It calculates dynamic flow thresholds to set reasonable flow upper limits for each edge, avoiding overload of ditches and forced drainage stations, and ensuring the physical feasibility of the regulation scheme. It calculates attenuation coefficients to quantify the pollutant attenuation pattern along the path, improving the accuracy of load calculation and compensating for the deficiency of traditional models in ignoring pollutant migration losses. It sets an initial flow allocation ratio to provide a reasonable initial value for the optimization algorithm, shortening the iteration convergence time and improving optimization efficiency.

[0238] Next, the first load at the source node is calculated. Based on attribute information and hydrological scenarios, the pollution load source is accurately quantified, providing accurate input data for subsequent migration simulations. The nitrogen, phosphorus, and salt loads corresponding to the transmission, treatment, control, and collection nodes are calculated sequentially, and the migration path is deduced layer by layer to fully simulate the transmission, attenuation, and absorption process of water, salt, and pollutants. This achieves full-chain quantification from "source to path to terminal," solving the problem of the one-sidedness of single-node calculations. Using the flow allocation ratio as the decision variable, the core control parameters are focused on and directly linked to facility operation strategies to ensure the optimization results are feasible. Then, a three-objective function is constructed to specifically resolve the conflict between "minimizing nitrogen and phosphorus and maximizing salt," aligning with the synergistic needs of salt control and pollution reduction, and overcoming the limitations of single-objective optimization. Multi-dimensional constraints are constructed, taking into account physical feasibility, economic cost, and operational capabilities, avoiding "mathematically optimal but engineering infeasible," and ensuring the practicality of the solution. A diverse initial population is generated, combining randomness, experience, and boundary vectors to ensure the population covers the feasible region, improving the algorithm's global search capability and avoiding local optima. Constraint-based screening is performed, and penalty coefficients are calculated for elimination. Rapidly eliminate low-quality vectors to improve the overall quality of the population and shorten the algorithm iteration cycle. Calculate dominance relationships and perform stratification based on these relationships. Screen high-quality solutions through non-dominated sorting to accurately identify the Pareto optimal front, providing a scientific basis for subsequent screening. Candidate vector screening and target vector determination, combined with the number of dominant vectors and congestion distance, select the comprehensive optimal decision variables to ensure that the optimization results are both effective and diverse. Obtain the target flow allocation ratio: transform the optimization results into specific facility operation parameters to achieve refined configuration and solve the problem of extensive settings based on traditional experience. Determine a coordinated control optimization scheme, integrate the model and optimization results to form an operable governance strategy, ultimately resolving the contradiction between salt control and pollution reduction, and improving the comprehensive efficiency of irrigation area water environment management.

[0239] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for synergistic regulation and optimization of water, salt, and pollutants, characterized in that, The method includes: Acquire spatial location and attribute information of strip fields, drainage ditches, ecological ingestion ponds, forced drainage stations, and external discharge river outlets in the target area; Based on the spatial location and attribute information of the strip fields, drainage ditches, ecological absorption ponds, forced drainage stations, and outlets of the external discharge rivers, a target directed graph network model is constructed. Based on the target directed graph network model, an optimization scheme for the coordinated regulation of water, salt and pollutants in the target area is determined. Specifically, based on the target directed graph network model, the optimization scheme for the coordinated regulation of water, salt, and pollutants corresponding to the target region is determined, including: Based on the attribute information and hydrological scenario parameters corresponding to the source nodes in the target directed graph network model, calculate the first nitrogen load, the first phosphorus load, and the first salt load corresponding to the source nodes; Based on the target directed graph network model, the nitrogen load, phosphorus load, and salt load corresponding to the transmission node, processing node, control node, and collection node in the target directed graph network model are calculated sequentially. The target flow allocation ratio corresponding to each target directed edge in the target directed graph network model is used as the decision variable; An objective function is constructed, comprising a first sub-function, a second sub-function, and a third sub-function; the first sub-function characterizes the minimum nitrogen load corresponding to the collection node; the second sub-function characterizes the minimum phosphorus load corresponding to the collection node; and the third sub-function characterizes the maximum salt load corresponding to the collection node. Based on a preset multi-objective evolutionary algorithm, the target flow allocation ratio corresponding to the target directed edge is calculated; Based on the target flow distribution ratio, determine the optimization scheme for the coordinated regulation of water, salt and pollutants in the target area; The step of solving the target flow allocation ratio corresponding to the target directed edge based on a preset multi-objective evolution algorithm includes: Based on the attribute information corresponding to each node in the target directed graph network model, constraints are constructed; the constraints include at least one of basic constraints, economic cost constraints, and operation and maintenance constraints. Using the preset multi-objective evolution algorithm, based on the constraints, the objective function is solved to obtain the objective flow allocation ratio corresponding to the objective directed edge; The step of using the preset multi-objective evolutionary algorithm to solve the objective function based on the constraints to obtain the target flow allocation ratio corresponding to the target directed edge includes: A predetermined number of initial decision variable vectors are randomly generated; the initial decision variable vectors include random decision variable vectors, empirical decision variable vectors, and boundary decision variable vectors. Based on the constraints, delete the initial decision variable vectors that do not satisfy the constraints from each of the initial decision variable vectors to obtain the remaining decision variable vectors; Based on the objective function, calculate the first sub-function value, the second sub-function value, and the third sub-function value corresponding to each of the remaining decision variable vectors; Based on the first sub-function value, the second sub-function value, and the third sub-function value corresponding to each of the remaining decision variable vectors, determine the penalty coefficient corresponding to the remaining decision variable vector; The remaining decision variable vectors whose penalty coefficients are greater than a preset threshold are marked as to be eliminated; Calculate the dominance relationships among the remaining effective decision variable vectors; Based on the dominance relationship, the remaining effective decision variable vectors are stratified to obtain multiple levels; wherein, the smaller the level corresponding to the remaining effective decision variable vector, the better the remaining effective decision variable vector is. For each first decision variable vector in the first level, calculate the number of lower-level decision vectors dominated by each first decision variable vector. From each of the first decision variable vectors, select candidate decision variable vectors whose number of lower-level decision vectors is greater than a preset threshold. Based on the candidate decision variable vectors, determine the target decision variable vector; Based on the target decision variable vector, the target flow allocation ratio corresponding to the target directed edge is obtained.

2. The method according to claim 1, characterized in that, The construction of a target directed graph network model based on the spatial location and attribute information of the farmland, drainage ditch, ecological infiltration pond, forced drainage station, and outflow river outlet includes: Each of the fields in the target area is identified as a source node; Each of the drainage ditches in the target area is identified as a transmission node; Each of the ecological absorption ponds in the target area is designated as a treatment node; Each of the forced drainage stations in the target area is identified as a control node; Each of the outflowing river outlets in the target area is identified as a convergence node; Based on the spatial location and attribute information of the strip fields, the drainage ditches, the ecological absorption ponds, the forced drainage stations, and the outlets of the external discharge rivers, the node attributes corresponding to the source node, the transmission node, the processing node, the control node, and the collection node are determined. Based on the spatial location and attribute information of the fields, drainage ditches, ecological absorption ponds, forced drainage stations, and outflow river outlets, the target directed edges corresponding to the source node, transmission node, processing node, control node, and collection node are determined, thus obtaining the target directed graph network model.

3. The method according to claim 2, characterized in that, The transmission relationship of the target directed graph network model is as follows: each source node transmits to each transmission node, each transmission node transmits to each processing node, each transmission node transmits to the control node, each processing node transmits to each control node, and each control node transmits to each collection node; the target directed edges corresponding to the source nodes, transmission nodes, processing nodes, control nodes, and collection nodes are determined based on the spatial location information and attribute information corresponding to the strip fields, drainage ditches, ecological infiltration ponds, forced drainage stations, and outflow river outlets, respectively. For two types of nodes with potential transmission relationships, calculate the distance matching degree, water flow adaptability degree, and service matching degree between the current node and the potential transmission nodes; Calculate the spatial correlation degree based on the distance matching degree, the water flow adaptability degree, and the service matching degree; Based on the spatial correlation degree, the target directed edge between the current node and the potential transmission node is determined; For each of the aforementioned target directed edges, add target edge attribute information.

4. The method according to claim 3, characterized in that, The target edge attribute information includes at least one of basic attribute information, dynamic attribute information, and optimized attribute information, wherein the dynamic attribute information includes a dynamic flow threshold and an attenuation coefficient. Adding target edge attribute information for each of the target directed edges includes: Determine the basic attribute information corresponding to each of the directed edges of the targets; the basic attribute information includes at least one of the following: edge type, edge start point, edge end point, edge length, and edge creation time; Calculate the dynamic flow threshold corresponding to each directed edge of the target based on the edge type of each directed edge of the target; Based on the edge type corresponding to the directed edge of each target, the attenuation coefficient corresponding to the directed edge of the transmission target is calculated; the attenuation coefficient is used to characterize the attenuation parameters of water salt and pollutants along the path; Calculate the initial flow allocation ratio corresponding to each of the directed edges of the target based on the edge type; The initial flow allocation ratio is determined as the optimized attribute information corresponding to the target directed edge.

5. The method according to claim 1, characterized in that, The step of calculating the first nitrogen load, first phosphorus load, and first salt load corresponding to the source node based on the attribute information and hydrological scenario parameters of the source node in the target directed graph network model includes: Obtain the hydrological scenario parameters corresponding to each of the source nodes, including rainfall, irrigation amount, runoff coefficient, nitrogen leaching rate, phosphorus leaching rate, and salt concentration coefficient; For each of the aforementioned source nodes, the total drainage volume of the source node is calculated based on the area of ​​the source node and the rainfall, irrigation volume and runoff coefficient in the hydrological scenario parameters. The first nitrogen load and the first phosphorus load of the source node are calculated based on the nitrogen fertilizer application rate, phosphorus fertilizer application rate, area, nitrogen leaching rate, and phosphorus leaching rate of the source node, respectively. Based on the total drainage volume, the soil salinity attributes of the source node, and the preset salt concentration coefficient, the first salt load of the source node is calculated.

6. The method according to claim 1, characterized in that, The step of calculating the nitrogen load, phosphorus load, and salt load corresponding to the transmission node, processing node, control node, and sink node in the target directed graph network model in sequence, based on the target directed graph network model, includes: For each transmission node, the first nitrogen input, first phosphorus input, and first salt input are determined according to the target directed graph network model. Multiply the first nitrogen input, the first phosphorus input, and the first salt input by the attenuation coefficient corresponding to the target directed edge between the source node and the transmission node to obtain the second nitrogen load, the second phosphorus load, and the second salt load corresponding to the transmission node. For each processing node, the second nitrogen input, second phosphorus input, and second salt input are determined according to the target directed graph network model. Based on the second nitrogen input, the second phosphorus input, the nitrogen absorption rate and phosphorus absorption rate corresponding to the processing node, the third nitrogen load and the third phosphorus load corresponding to the processing node are calculated. Based on the second salt input and the salt absorption rate corresponding to the processing node, the third salt load corresponding to the processing node is calculated. For each control node, the fourth nitrogen load, fourth phosphorus load, and fourth salt load are determined according to the target directed graph network model. For each of the aggregation nodes, the fifth nitrogen load, the fifth phosphorus load, and the fifth salt load are determined according to the target directed graph network model.

7. The method according to claim 1, characterized in that, The step of determining the target decision variable vector based on each of the candidate decision variable vectors includes: The candidate decision variable vectors are sorted in ascending order according to the values ​​of the first sub-function to obtain the first sequence; The candidate decision variable vectors are sorted in ascending order according to the values ​​of the second sub-functions to obtain the second sequence; The candidate decision variable vectors are sorted in descending order according to the values ​​of the third sub-function to obtain the third sequence; Based on the first sequence, calculate the first range corresponding to the first sub-function value; Based on the second sequence, calculate the second range corresponding to the second sub-function value; Based on the third sequence, calculate the third range corresponding to the third sub-function value; For each of the candidate decision variable vectors, calculate the difference between the candidate decision variable vector and its adjacent candidate decision variable vector in each sequence; Divide the difference between each sequence by the global range of that sequence to obtain the distance component of the candidate decision variable vector in each sequence; Based on the distance components of each candidate decision variable vector in each sequence, the crowding distance corresponding to the candidate decision variable vector is obtained; Based on the crowding distance corresponding to each of the candidate decision variable vectors, select the backup decision variable vectors whose crowding distance is greater than a preset distance threshold. Two alternative decision variable vectors are selected from each of the alternative decision variable vectors as parent decision variable vectors and cross-operated to generate two child decision variable vectors. Calculate the variable length based on the current generation corresponding to the offspring decision variable vector; Based on the variable asynchronous length, a mutation operation is performed on the offspring decision variable vector to obtain the mutated decision variable vector; The parent decision variable vectors and the mutated decision variable vectors are merged into a temporary population. Based on the temporary population, the above steps are repeated until the preset number of iterations is reached to obtain the target decision variable vector.

Images