A method for determining an ecological environment monitoring network configuration scheme and related apparatus

Abstract

The application discloses a kind of ecological environment monitoring network configuration scheme determination method and related device, it is related to ecological environment monitoring technical field, method includes: according to the monitoring target and ecological environment basic data of monitoring area, the sub-target with information entropy attribute corresponding to each grid unit covered by monitoring area is generated;Mapping function of sub-target about decision variable is established, according to the coupling relationship between sub-target, sub-target is grouped, and multiple collaborative optimization groups are formed;According to the information entropy attribute and mapping function of each sub-target in collaborative optimization group, the comprehensive weight of collaborative optimization group, group-level mapping function and reference information entropy are determined;With the sum of the weighted information entropy residual of each collaborative optimization group as the goal of minimization, multi-objective optimization iteration is carried out, and the optimal solution set of decision variable is generated, and the implementation configuration scheme is selected from the optimal solution set.The application fundamentally solves the problem that the optimization scheme space adaptability is insufficient due to the granularity mismatch between monitoring target and decision variable.

Description

Technical Field

[0001] This application relates to the field of ecological environment monitoring technology, and in particular to a method for determining the configuration scheme of an ecological environment monitoring network and related devices. Background Technology

[0002] Ecological and environmental monitoring and protection are crucial foundational tasks for ecological civilization construction. In practical applications such as watershed water quality monitoring, atmospheric environmental monitoring, forest fire prevention monitoring, and ecological protection zone planning, decision-makers often need to consider multiple conflicting or even mutually restrictive engineering indicators simultaneously. For example, under limited funding constraints, they need to simultaneously pursue the maximum coverage and highest monitoring accuracy of the monitoring network, or seek the optimal dynamic balance between ecological protection and socio-economic development. These complex decision-making problems with multiple objectives and constraints have become key challenges in the field of ecological and environmental monitoring and protection.

[0003] To address the aforementioned challenges, existing technologies have attempted to introduce multi-objective optimization methods into the field of ecological environment monitoring and protection. A common approach is to employ multi-objective evolutionary algorithms, such as the Non-Dominated Sorting Genetic Algorithm (NSGA-II) and its improved forms, to optimize decision variables such as monitoring node location and protection measure layout within a Pareto optimal framework, ultimately outputting a set of optimal solutions for decision-makers to choose from.

[0004] However, existing ecological environment monitoring methods based on multi-objective optimization still have a significant common technical bottleneck in practical engineering applications, namely the granularity mismatch between engineering monitoring objectives and decision variables. This problem means that although the optimized solution may perform "optimally" in terms of macro indicators, it may not be reasonable in terms of micro spatial layout, and cannot achieve truly accurate monitoring and adaptive protection, which seriously limits the practical engineering effectiveness of the optimized solution. Summary of the Invention

[0005] In view of this, this application provides a method and related apparatus for determining the configuration scheme of an ecological environment monitoring network, in order to solve the problem of insufficient spatial adaptability of the optimization scheme caused by the mismatch between the granularity of engineering monitoring objectives and decision variables. The technical solution is as follows:

[0006] The first aspect of this application provides a method for determining the configuration scheme of an ecological environment monitoring network, including:

[0007] Based on the engineering monitoring objectives and basic ecological and environmental data of the monitoring area, sub-objectives with information entropy attributes are generated for each grid unit covered by the monitoring area. The information entropy attribute represents the information uncertainty of the corresponding grid unit, the information uncertainty of the local area to which the corresponding grid unit belongs, and the information correlation between the corresponding grid unit and its neighboring grid units.

[0008] Establish mapping functions for multiple sub-objectives with respect to decision variables, and group the multiple sub-objectives according to the coupling relationship between them to form multiple collaborative optimization groups. The decision variables are the configuration parameters of the ecological environment monitoring network.

[0009] For each of the collaborative optimization groups, the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group are determined based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group.

[0010] Based on the weights, group-level mapping functions, and reference information entropy of each of the collaborative optimization groups, a multi-objective optimization iteration is performed with the objective of minimizing the sum of the weighted information entropy residuals of each of the collaborative optimization groups, generating the optimal solution set of the decision variables, and selecting a set of optimal solutions from the optimal solution set as the implementation configuration scheme of the ecological environment monitoring network.

[0011] In one possible implementation, generating sub-targets with information entropy attributes corresponding to each grid cell covered by the monitoring area, based on the engineering monitoring objectives and basic ecological and environmental data of the monitoring area, includes:

[0012] Based on the engineering monitoring objectives and the natural geographical features data in the ecological and environmental basic data, the monitoring area is divided into multiple sub-regions, so that each grid unit covered by the monitoring area is assigned to a unique sub-region;

[0013] Determine the information entropy of each sub-region, the mutual information between grid cells within each sub-region, and the maximum possible information entropy of each grid cell;

[0014] Generate a sub-target with an information entropy attribute corresponding to each of the grid cells. The information entropy attribute of the sub-target corresponding to any grid cell includes the maximum possible information entropy of the grid cell, the mutual information between the grid cell and its neighboring grid cells, and the information entropy of the sub-region to which the grid cell belongs.

[0015] In one possible implementation, the establishment of mapping functions for the multiple sub-objectives obtained are respectively about the decision variables, including:

[0016] For each sub-objective, a mapping function of the sub-objective with respect to the decision variables is established based on the information entropy attribute of the sub-objective.

[0017] In one possible implementation, the information entropy attribute of the sub-target includes: the maximum possible information entropy and the mutual information between the grid cell corresponding to the sub-target and its adjacent grid cells;

[0018] The step of establishing a mapping function of the sub-objective with respect to the decision variables based on the information entropy attribute of the sub-objective includes:

[0019] Based on the maximum possible information entropy in the information entropy attribute of the sub-target, establish a preliminary mapping function of the sub-target with respect to the decision variables;

[0020] By combining the mutual information in the information entropy attribute of the sub-target, the preliminary mapping function is regularized and optimized to generate a mapping function that can reflect the nonlinear correlation between the sub-target and the decision variable, thus obtaining the final mapping function of the sub-target with respect to the decision variable.

[0021] In one possible implementation, establishing a preliminary mapping function of the sub-target with respect to the decision variables based on the maximum possible information entropy in the information entropy attribute of the sub-target includes:

[0022] The maximum possible information entropy in the information entropy attribute of the sub-target is processed by kernel embedding mapping with the decision variables in the decision variable space to generate a preliminary mapping function of the sub-target with respect to the decision variables;

[0023] The step of combining the mutual information in the information entropy attribute of the sub-target to perform regularization optimization on the preliminary mapping function, generating a mapping function that can reflect the nonlinear correlation between the sub-target and the decision variable, includes:

[0024] Based on the preliminary mapping function, and combined with the mutual information in the information entropy attribute of the sub-target, the sub-target mapping values ​​at multiple candidate points in the decision variable space are fitted with least squares regression under regularization constraints to generate a mapping function that can reflect the nonlinear relationship between the sub-target and the decision variable.

[0025] In one possible implementation, grouping the multiple sub-objectives according to the coupling relationship between them to form multiple collaborative optimization groups includes:

[0026] Based on the mapping function of each sub-objective, calculate the partial derivative dependency coefficients between each pair of sub-objectives;

[0027] Based on the partial derivative dependency coefficients between each pair of sub-targets, a partial derivative dependency matrix is ​​constructed to characterize the nonlinear coupling relationship between each pair of sub-targets.

[0028] Based on the partial derivative dependency matrix, the sub-objectives are clustered to form multiple collaborative optimization groups.

[0029] In one possible implementation, calculating the pairwise partial derivative dependency coefficients between each of the sub-objectives based on the mapping function of each sub-objective includes:

[0030] For each pair of sub-targets, calculate the partial derivative of the mapping function of the first sub-target with respect to the mapping function of the second sub-target to obtain a first partial derivative value, and calculate the partial derivative of the mapping function of the second sub-target with respect to the mapping function of the first sub-target to obtain a second partial derivative value.

[0031] Calculate the average of the first partial derivative value and the second partial derivative value, and use the calculated average value as the partial derivative dependency coefficient between the two sub-objectives.

[0032] In one possible implementation, the step of clustering the sub-objectives according to the partial derivative dependency matrix to form multiple collaborative optimization groups includes:

[0033] The partial derivative dependency matrix is ​​symmetricized to obtain a symmetric dependency matrix;

[0034] Construct a Laplacian matrix based on the symmetric dependency matrix, and perform eigenvalue decomposition on the Laplacian matrix to extract the eigenvector corresponding to the smallest eigenvalue;

[0035] Each sub-target is mapped to a low-dimensional feature space using the feature vector, and the sub-targets are clustered in the low-dimensional feature space to form multiple collaborative optimization groups.

[0036] In one possible implementation, determining the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group includes:

[0037] Based on the information entropy attributes of each sub-objective within the collaborative optimization group, calculate the reference information entropy of the collaborative optimization group and the weight of each sub-objective within the collaborative optimization group;

[0038] Based on the weights of each sub-objective within the collaborative optimization group, the mapping functions of each sub-objective within the collaborative optimization group are merged to obtain the group-level mapping function of the collaborative optimization group;

[0039] The weights of each sub-objective within the collaborative optimization group are fused to obtain the comprehensive weight of the collaborative optimization group.

[0040] In one possible implementation, the step of performing multi-objective optimization iterations based on the weights, group-level mapping functions, and reference information entropy of each of the collaborative optimization groups, with the objective of minimizing the sum of weighted information entropy residuals of each of the collaborative optimization groups, to generate the optimal solution set of the decision variables, includes:

[0041] Obtain an initial candidate solution set, which includes multiple candidate solutions, each candidate solution being a set of decision variable values;

[0042] For each candidate solution in the current candidate solution set, based on the group-level mapping function, comprehensive weight and reference information entropy of each collaborative optimization group, the weighted information entropy residual of the candidate solution in each collaborative optimization group is calculated, and the weighted information entropy residual of the candidate solution in each collaborative optimization group is summed. The summation result is used as the fitness of the candidate solution.

[0043] Based on the fitness of each candidate solution in the current candidate solution set, a parent candidate solution is selected from the current candidate solution set, and a child candidate solution is generated based on the parent candidate solution.

[0044] Based on the parent candidate solutions and the child candidate solutions, a new candidate solution set is constructed. The new candidate solution set is used as the current candidate solution set. The process of calculating the sum of weighted information entropy residuals, selecting parent candidate solutions, generating child candidate solutions, and constructing a new candidate solution set is repeated until the preset convergence condition is met. The optimal solution is obtained from the final candidate solution set to form the optimal solution set.

[0045] In one possible implementation, selecting a parent candidate solution from the current candidate solution set based on the fitness of each candidate solution in the current candidate solution set includes:

[0046] Based on the fitness of each candidate solution in the current candidate solution set, the non-dominated sorting and crowding degree calculation are performed on the current candidate solution set to obtain the non-dominated level and crowding degree of each candidate solution in the current candidate solution set;

[0047] Based on the non-dominated level and congestion of each candidate solution in the current candidate solution set, parent candidate solutions are selected from the current candidate solution set in order of non-dominated level from low to high and congestion within the same level from large to small.

[0048] The step of obtaining the optimal solution from the final candidate solution set and forming the optimal solution set includes:

[0049] The candidate solutions with the first non-dominated level are selected from the final candidate solution set to form the optimal solution set.

[0050] In one possible implementation, selecting a set of optimal solutions from the optimal solution set as the implementation configuration scheme for the ecological environment monitoring network includes:

[0051] Based on the group-level mapping function, comprehensive weight, and reference information entropy of each of the collaborative optimization groups, the weighted information entropy residual of each optimal solution in the optimal solution set on each collaborative optimization group is calculated.

[0052] By using the fuzzy membership function, the weighted information entropy residual of each optimal solution in the optimal solution set on each of the collaborative optimization groups is mapped to a fuzzy membership value with a unified dimension. The fuzzy membership values ​​of the same optimal solution on each of the collaborative optimization groups are combined into a fuzzy membership value set, thus obtaining the fuzzy membership value set of each optimal solution in the optimal solution set.

[0053] Based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set, select an optimal solution from the optimal solution set.

[0054] In one possible implementation, selecting an optimal solution from the optimal solution set based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set includes:

[0055] The optimal solutions in the optimal solution set are ranked according to their fuzzy membership values ​​in the same collaborative optimization group to obtain the ranking of each optimal solution in the optimal solution set in each collaborative optimization group. Based on the ranking of each optimal solution in the optimal solution set in each collaborative optimization group, the score of each optimal solution in the optimal solution set is determined.

[0056] Based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set, the number of approval votes for each optimal solution in the optimal solution set is calculated. The number of approval votes for any optimal solution in the optimal solution set is the number of fuzzy membership values ​​in the set of fuzzy membership values ​​of that optimal solution that exceed a preset fuzzy membership threshold.

[0057] Based on the scores and approval votes of each optimal solution in the optimal solution set, an optimal solution is selected from the optimal solution set.

[0058] In one possible implementation, the method for determining the configuration scheme of the ecological environment monitoring network further includes:

[0059] Obtain real-time monitoring data from the ecological environment monitoring network deployed according to the implementation configuration scheme;

[0060] Based on the real-time monitoring data, the sub-region boundary, sub-region mutual information, and the group-level mapping function are corrected. The sub-region mutual information is the mutual information between grid cells within the sub-region.

[0061] When the rate of change of the mutual information of the corrected sub-region relative to the mutual information of the uncorrected sub-region exceeds a preset rate of change threshold, an updated optimal solution set is generated based on the optimal solution and the updated group-level mapping function.

[0062] A second aspect of this application provides an electronic device, including at least one processor and a memory connected to the processor, wherein:

[0063] The memory is used to store computer programs;

[0064] The processor is used to execute the computer program so that the electronic device implements the method for determining any of the above-described ecological environment monitoring network configuration schemes.

[0065] Using the aforementioned technical solutions, the method for determining the configuration scheme of the ecological environment monitoring network provided in this application fundamentally solves the problem of granularity mismatch between engineering monitoring targets and decision variables by decomposing macro-engineering monitoring targets into grid-level sub-targets with quantified information entropy attributes. This achieves precise spatial transmission of monitoring needs from macro-indicators to micro-decision. After obtaining multiple sub-targets, the sub-targets are intelligently grouped according to the coupling relationship between them to form a collaborative optimization group. This achieves cluster collaborative optimization of related sub-targets, greatly reducing optimization complexity and significantly improving the overall integrity of the scheme. With minimizing the sum of weighted information entropy residuals of each collaborative optimization group as the unified objective, an objective optimization criterion driving the maximization of global information acquisition is established, realizing scientific automatic trade-offs among multiple targets. The final generated optimal solution set provides high-quality alternative schemes that are optimal in terms of information utility and have different emphases. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0067] Figure 1 A flowchart illustrating the method for determining the configuration scheme of the ecological environment monitoring network provided in this application embodiment;

[0068] Figure 2 A flowchart illustrating the process of generating sub-targets with information entropy attributes corresponding to each grid unit covered by the monitoring area based on the engineering monitoring targets and basic ecological and environmental data of the monitoring area, as provided in this application embodiment;

[0069] Figure 3 A flowchart illustrating the process of establishing a mapping function of a sub-target with respect to decision variables based on the information entropy attribute of the sub-target, as provided in this embodiment of the application.

[0070] Figure 4 A flowchart illustrating the process of grouping multiple sub-targets into multiple collaborative optimization groups based on the coupling relationship between multiple sub-targets, as provided in an embodiment of this application;

[0071] Figure 5This is a flowchart illustrating the process of generating the optimal solution set for decision variables by performing multi-objective optimization iteration based on the weights of each collaborative optimization group, the group-level mapping function, and the reference information entropy, with the goal of minimizing the sum of the weighted information entropy residuals of each collaborative optimization group, according to an embodiment of this application.

[0072] Figure 6 A schematic diagram illustrating the process of selecting an optimal solution from the set of optimal solutions, provided for an embodiment of this application;

[0073] Figure 7 This is a schematic diagram of the device for determining the configuration scheme of the ecological environment monitoring network provided in the embodiments of this application. Detailed Implementation

[0074] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0075] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0076] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements, but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.

[0077] In the optimization modeling process, engineering-level monitoring objectives (such as "monitoring accuracy," "spatial coverage," and "time response delay") are typically coarse-grained macro-indicators, while the actual operational objects of optimization decisions (such as the deployment scheme of monitoring equipment in each grid unit and the selection of data transmission paths) are fine-grained micro-variables. The spatial or logical resolution of macro-objectives is far lower than the operational resolution of fine-grained decision variables. This leads to a significant amount of micro-spatial heterogeneity, local sensitivity differences, and non-uniform monitoring needs containing key information being blurred or even submerged during regional averaging or weighted summation when constructing the objective function. Consequently, important fine-grained information such as high-precision local monitoring needs and protection requirements for key sensitive points cannot be effectively and accurately transmitted and mapped to specific fine-grained decision variables such as monitoring point location and equipment selection. Ultimately, although the optimized solution may appear "optimal" in terms of macro-indicators, it may be unreasonable in terms of micro-spatial layout, failing to achieve truly accurate monitoring and adaptive protection, and severely limiting the practical engineering effectiveness of the optimized solution.

[0078] In view of this, the present invention provides a method for determining the configuration scheme of an ecological environment monitoring network, which can overcome the above-mentioned defects. The following embodiments will introduce the method for determining the configuration scheme of an ecological environment monitoring network provided in this application.

[0079] Please see Figure 1 The diagram illustrates a flowchart of a method for determining an ecological environment monitoring network configuration scheme according to an embodiment of this application. This method may include:

[0080] Step S101: Based on the engineering monitoring objectives and basic ecological and environmental data of the monitoring area, generate sub-objectives with information entropy attributes corresponding to each grid unit covered by the monitoring area.

[0081] In this embodiment, the engineering monitoring target of the monitoring area is a coarse-grained macro-indicator, such as monitoring accuracy and spatial coverage.

[0082] In this embodiment, the basic ecological and environmental data of the monitoring area may include natural characteristic data of the monitoring area (such as spatial boundary data of topographic and geomorphological zoning, land use type zoning and hydrological unit zoning) and ecological and environmental status data (such as pollutant concentration distribution data, biodiversity data, ecological function data, etc.). In addition, the basic ecological and environmental data of the monitoring area may also include existing monitoring facility data (such as monitoring equipment parameters, historical monitoring data, etc.) and engineering constraint data (such as budget constraints, technical constraints, management constraints, and regulatory constraints).

[0083] This embodiment divides the monitoring area into multiple grid units, decomposes the macroscopic engineering monitoring target into each grid unit, and generates an independent target with quantitative attributes, namely a sub-target, for each grid unit.

[0084] Step S102: Establish the mapping functions of the obtained sub-objectives with respect to the decision variables, and group the sub-objectives according to the coupling relationship between them to form multiple collaborative optimization groups.

[0085] Among them, the decision variables are the configuration parameters of the ecological environment monitoring network, such as the number of monitoring devices deployed in each grid unit of the monitoring area, the type of monitoring devices, and the data transmission frequency.

[0086] After obtaining multiple sub-objectives, we first establish the mapping function of each sub-objective with respect to the decision variables, and then analyze the degree of mutual influence (i.e., coupling relationship) between these sub-objectives. Sub-objectives with strong mutual influence are clustered together to form multiple collaborative optimization groups.

[0087] The mapping function establishes a quantitative causal relationship between "input" (what to configure) and "output" (what effect to obtain), while the coupling grouping intelligently aggregates a large number of sub-objectives into a few collaborative optimization groups with strong intra-group correlation and weak inter-group correlation based on the degree of mutual influence between sub-objectives. This greatly reduces the complexity of subsequent optimization.

[0088] Step S103: For each collaborative optimization group, determine the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group.

[0089] The comprehensive weight of the collaborative optimization group quantifies its relative importance in the entire monitoring task. The group-level mapping function of the collaborative optimization group is used to calculate the actual information entropy that the collaborative optimization group as a whole can achieve based on the values ​​of the decision variables. The reference information entropy of the collaborative optimization group represents the target information entropy value that the collaborative optimization group should achieve.

[0090] This embodiment simplifies the optimization problem from dealing with a large number of sub-objectives to dealing with a small number of collaborative optimization groups, laying the foundation for efficient multi-objective optimization.

[0091] Step S104: Based on the weights, group-level mapping functions, and reference information entropy of each collaborative optimization group, perform multi-objective optimization iteration with the goal of minimizing the sum of weighted information entropy residuals of each collaborative optimization group to generate the optimal solution set for the decision variables.

[0092] In this embodiment, the global optimization objective is defined as minimizing the sum of the weighted information entropy residuals of all collaborative optimization groups. The information entropy residual of any collaborative optimization group is the difference between the actual information entropy obtained by that group (calculated based on the group-level mapping function) and the reference information entropy of that group. Through multi-objective optimization iteration, the values ​​of decision variables are continuously adjusted to find the set of decision variable values ​​that minimizes this total global residual.

[0093] Step S105: Select a set of optimal solutions from the optimal solution set as the implementation configuration scheme for the ecological environment monitoring network.

[0094] The result obtained through multi-objective optimization iteration is not a single solution, but a Pareto optimal solution set, which contains multiple excellent solutions that are independent of each other in terms of the global objective and each has its own focus. The Pareto optimal solution set provides decision-makers with a high-quality "menu of solutions" rather than forcibly specifying a "unique answer". This step is to select a solution from this optimal solution set as the final implementation configuration.

[0095] The method for determining the configuration scheme of the ecological environment monitoring network provided in this application embodiment can be applied to various monitoring scenarios, such as watershed water quality monitoring scenarios, atmospheric environment monitoring scenarios, forest fire prevention monitoring scenarios, ecological protection zone monitoring scenarios, industrial park environmental monitoring scenarios, etc.

[0096] The method for determining the configuration scheme of the ecological environment monitoring network provided in this application fundamentally solves the problem of granularity mismatch between engineering monitoring targets and decision variables by decomposing macro-engineering monitoring targets into grid-level sub-targets with quantified information entropy attributes. This achieves precise spatial transmission of monitoring needs from macro-indicators to micro-decision. After obtaining multiple sub-targets, each sub-target is intelligently grouped into a collaborative optimization group based on the coupling relationship between them. This achieves cluster collaborative optimization of related sub-targets, greatly reducing optimization complexity and significantly improving the overall integrity of the scheme. With minimizing the sum of weighted information entropy residuals of each collaborative optimization group as the unified objective, an objective optimization criterion driving the maximization of global information acquisition is established, realizing scientific automatic trade-offs among multiple targets. The final generated optimal solution set provides high-quality alternative schemes that are optimal in terms of information utility and have different emphases.

[0097] In some embodiments of this application, the implementation process of "step S101: generating sub-targets with information entropy attributes corresponding to each grid unit covered by the monitoring area based on the engineering monitoring targets and ecological environment basic data of the monitoring area" in the above embodiment is described.

[0098] like Figure 2As shown, the process of generating sub-targets with information entropy attributes corresponding to each grid unit covered by the monitoring area, based on the engineering monitoring objectives and basic ecological and environmental data of the monitoring area, may include:

[0099] Step S201: Based on the engineering monitoring objectives and the natural geographical features data in the basic ecological and environmental data, the monitoring area is divided into multiple sub-regions, so that each grid unit covered by the monitoring area is assigned to a unique sub-region.

[0100] Based on the monitoring priorities and requirements of the engineering monitoring objectives, and combined with natural geographical feature data (such as spatial boundary data of topographic and geomorphological zoning, land use type zoning, and hydrological unit zoning), the monitoring area is divided into multiple non-overlapping sub-regions. It should be noted that when dividing the monitoring area into sub-regions, the principle of non-overlapping and full coverage is followed. Through division, it is ensured that each grid unit covered by the monitoring area is assigned to a unique sub-region, and all sub-regions together cover the entire monitoring area without omission or overlap.

[0101] After the sub-regions are divided, the weights of each sub-region can be determined based on the engineering monitoring objectives and the actual monitoring value of each sub-region. The weight of any sub-region can reflect the importance of that sub-region in the overall monitoring task.

[0102] After completing the division of sub-regions and determining the weights, a consistency check can be performed on the multiple sub-regions and their corresponding weights to identify issues such as overlapping sub-regions, missing grid cells, and unreasonable weight allocation, thus ensuring the validity of the sub-region division results and weights.

[0103] After verifying the consistency between the sub-region division results and weights, the number of grid cells contained in each sub-region can be counted and recorded. The association information between grid cells and sub-regions can also be recorded and stored in a unified format to form a sub-region-grid cell correspondence table.

[0104] Step S202: Determine the information entropy of each sub-region, the mutual information between grid cells within each sub-region, and the maximum possible information entropy of each grid cell.

[0105] For each sub-region, the information entropy of the sub-region can be determined based on the information characteristics of each grid cell covered by the sub-region (integrating the information uncertainty of each grid cell). The information entropy of the sub-region represents the information uncertainty level of the entire sub-region.

[0106] For each grid cell, the maximum possible information entropy of the grid cell can be determined based on its basic attribute information and ecological environment characteristic data. The maximum possible information entropy of the grid cell represents the upper limit of the information uncertainty of the grid cell under ideal conditions.

[0107] The mutual information between grid cells is used to quantify the degree of information correlation between grid cells. It can be obtained by analyzing the correlation of ecological and environmental characteristics and spatial location between grid cells to determine the intensity of information interaction between grid cells.

[0108] Step S203: Based on the information entropy of each sub-region, the mutual information between grid cells within each sub-region, the maximum possible information entropy of each grid cell, and the sub-region to which each grid cell belongs, generate a sub-target with attribute information corresponding to each grid cell.

[0109] The attribute information of the sub-target corresponding to any grid cell may include the information entropy attribute. The information entropy attribute of the sub-target corresponding to any grid cell may include the maximum possible information entropy of the grid cell, the mutual information between the grid cell and its neighboring grid cells, and the information entropy of the sub-region to which the grid cell belongs.

[0110] In addition to information entropy, the attribute information of a sub-target corresponding to any grid cell may include a unique identifier for the sub-target, spatial attributes, and other attributes. The spatial attributes of a sub-target corresponding to any grid cell may include the spatial coordinates of the grid cell and the identifier of the sub-region to which the grid cell belongs.

[0111] When generating attribute information for each sub-target, a unique identifier can be assigned to each sub-target to ensure its identifiability and traceability. A spatial indexing system can be established, with sub-regions as the index subjects, to associate each sub-target with its corresponding sub-region. Each sub-target is bound to the identifier of its respective sub-region, clarifying its spatial affiliation. Each sub-target can be bound to the spatial coordinates of its corresponding grid cell, achieving precise association between the sub-target and the grid cell. Each sub-target can also be bound to an index of the mutual information between its corresponding grid cell and adjacent grid cells. In this way, the mutual information between the grid cell corresponding to the sub-target and adjacent grid cells can be quickly retrieved through the index, integrating all the associated information of each sub-target to obtain its attribute information.

[0112] After obtaining multiple sub-objectives with information entropy attributes, multiple mapping functions of the sub-objectives with respect to decision variables can be generated. In some embodiments of this application, the process of generating multiple mapping functions of the sub-objectives with respect to decision variables is described.

[0113] The process of generating mapping functions for multiple sub-objectives with respect to decision variables may include: for each of the multiple sub-objectives, establishing a mapping function for the sub-objective with respect to decision variables based on the information entropy attribute of the sub-objective.

[0114] like Figure 3 As shown, the process of establishing a mapping function of the sub-objective with respect to the decision variables based on the information entropy attribute of the sub-objective may include:

[0115] Step S301: Based on the maximum possible information entropy in the information entropy attribute of the sub-objective, establish a preliminary mapping function of the sub-objective with respect to the decision variables.

[0116] Specifically, the maximum possible information entropy in the information entropy attribute of the sub-target is kernel-embedded and mapped to the decision variables in the decision variable space to generate a preliminary mapping function of the sub-target with respect to the decision variables.

[0117] It should be noted that kernel embedding mapping refers to mapping decision variables in the decision variable space to a high-dimensional feature space using a radial basis function (RBF) kernel, eliminating the nonlinear correlation between decision variables and the maximum possible information entropy of sub-objectives. Linear fitting is then performed on the mapped decision variables and the maximum possible information entropy of sub-objectives in the high-dimensional feature space to construct the response surface of the sub-objectives. The response surface characterizes the correspondence between changes in decision variables and the maximum possible information entropy of sub-objectives. Combined with the mapping results of the RBF kernel, an explicit functional expression for each sub-objective with respect to the decision variables is generated.

[0118] The expression for the radial basis function kernel is:

[0119] (1).

[0120] in, For radial basis kernel functions, and These are two different decision vectors in the decision variable space (a decision vector is a vector composed of each decision variable). These are kernel function parameters used to control the radial range of the kernel function. The squared Euclidean distance between the two decision vectors is given.

[0121] The explicit functional expression of the j-th sub-objective with respect to the decision variable is:

[0122] (2).

[0123] in, Let j be an explicit function of the j-th sub-objective. These are the linear fitting coefficients. Let be the sample decision vector in the decision variable space. This is the fitting constant term.

[0124] Step S302: Combining the mutual information in the information entropy attribute of the sub-objective, perform regularization optimization on the preliminary mapping function of the sub-objective with respect to the decision variables to generate a mapping function that can reflect the nonlinear relationship between the sub-objective and the decision variables, and obtain the final mapping function of the sub-objective with respect to the decision variables.

[0125] Specifically, based on the initial mapping function and combined with the mutual information in the information entropy attribute of the sub-objectives, a least squares regression fitting under regularization constraints is performed on the sub-objective mapping values ​​at multiple candidate points in the decision variable space to generate a nonlinear mapping function that reflects the nonlinear relationship between the sub-objectives and the decision variables. It should be noted that the fitting process aims to improve the accuracy of the function fitting and control the smoothness of the function.

[0126] The objective function for least squares regression fitting under regularization constraints is:

[0127] (3).

[0128] in, Let b be the weight vector of the fitted function, and b be the fitting constant term. For insensitive loss functions, Let i be the predicted sub-target mapping value at the i-th candidate point. Map the actual value of the sub-target at the i-th candidate point. For regularization parameters, is the insensitive loss parameter, and m is the number of candidate points.

[0129] It should be noted that the regularization parameter is used to control the smoothness of the fitted function. The larger the value of the regularization parameter, the smoother the fitted function, which can effectively avoid overfitting. The insensitive loss parameter is used to limit the tolerance range of the fitted residuals. When the residuals are within this range, no loss is generated. When they are outside this range, the loss is calculated according to the linear relationship.

[0130] By iteratively calculating and determining reasonable values ​​for the regularization parameter and the insensitive loss parameter, and combining this with a least squares regression fitting algorithm, the sub-objective mapping values ​​at different candidate points are fitted, correcting the nonlinear bias in the explicit function expression. The fitted functions are then processed to generate a complete nonlinear mapping function for each sub-objective. This nonlinear mapping function accurately reflects the nonlinear correlation characteristics between the decision variables and the sub-objectives. After generating the final mapping function of the sub-objectives with respect to the decision variables, it is validated to ensure that the function fitting accuracy meets the requirements of subsequent optimization and that there is no significant bias.

[0131] After generating mapping functions for multiple sub-objectives with respect to decision variables, the multiple sub-objectives are grouped according to the coupling relationship between them to form multiple collaborative optimization groups. This process is described in some embodiments of this application.

[0132] like Figure 4 As shown, the process of grouping multiple sub-objectives into multiple collaborative optimization groups based on the coupling relationship between them can include:

[0133] Step S401: Based on the mapping function (nonlinear mapping function) of each sub-target, calculate the partial derivative dependency coefficients between each pair of sub-targets.

[0134] The partial derivative dependency coefficient between two sub-objectives characterizes the degree of mutual influence between the two sub-objectives.

[0135] Specifically, the process of calculating the partial derivative dependency coefficients between each pair of sub-targets based on the mapping function (nonlinear mapping function) of each sub-target can include: for each pair of sub-targets, firstly calculate the partial derivative of the mapping function of the first sub-target with respect to the mapping function of the second sub-target, to obtain the first partial derivative value, and then calculate the partial derivative of the mapping function of the second sub-target with respect to the mapping function of the first sub-target, to obtain the second partial derivative value. Then calculate the average of the first and second partial derivative values, and use the average value as the partial derivative dependency coefficient between the two sub-targets.

[0136] The partial derivative of the nonlinear mapping function of the i-th sub-target with respect to the nonlinear mapping function of the j-th sub-target can be obtained by the following formula:

[0137] (4).

[0138] in, Let be the nonlinear mapping functions for the i-th sub-target and the j-th sub-target, respectively. and These are the gradient vectors of the nonlinear mapping functions for the two sub-objectives, respectively. " represents the vector dot product operation. The square of the magnitude of the gradient vector of the nonlinear mapping function for the j-th sub-objective.

[0139] Step S402: Based on the calculated partial derivative dependency coefficients, construct the partial derivative dependency matrix to characterize the nonlinear coupling relationship between each pair of sub-targets.

[0140] The partial derivative dependency matrix is ​​a square matrix, and each element in the partial derivative dependency matrix is ​​the partial derivative dependency coefficient between two sub-objectives. The matrix dimension of the partial derivative dependency matrix is ​​the same as the total number of sub-objectives.

[0141] Step S403: Based on the partial derivative dependency matrix, cluster multiple sub-objectives to form multiple collaborative optimization groups.

[0142] Specifically, the process of clustering multiple sub-objectives based on the partial derivative dependency matrix to form multiple collaborative optimization groups may include:

[0143] Step S4031: Symmetricize the partial derivative dependency matrix to obtain a symmetric dependency matrix.

[0144] The partial derivative dependency matrix is ​​subjected to bidirectional average symmetry processing, and the strengths of positive and negative dependencies in the asymmetric dependency matrix are fused by arithmetic average to generate a symmetric dependency matrix.

[0145] The partial derivative dependency matrix undergoes bidirectional averaging symmetry processing. The core of this process is to eliminate the asymmetry of the partial derivative dependency matrix, ensuring that the matrix accurately reflects the consistency of the bidirectional dependencies between sub-objectives. Specifically, asymmetric elements in the partial derivative dependency matrix are identified—that is, matrix elements where the strength of the positive dependency is not equal to the strength of the corresponding negative dependency. An arithmetic average fusion process is then performed on each group of asymmetric elements.

[0146] The formula for bidirectional average symmetry processing is:

[0147] (5).

[0148] in, It is a symmetric dependency matrix. Let be the element in the i-th row and j-th column of the partial derivative dependency matrix, representing the positive dependency strength of the i-th sub-target on the j-th sub-target. The element in the j-th row and i-th column of the partial derivative dependency matrix represents the inverse dependency strength of the j-th sub-target on the i-th sub-target. The value obtained after arithmetic mean fusion is the element at the corresponding position in the symmetric dependency matrix.

[0149] The above process is applied sequentially to all elements in the partial derivative dependency matrix until all asymmetric elements have been fused by arithmetic mean, generating a complete symmetric dependency matrix.

[0150] After obtaining the symmetric dependency matrix, it can be validated to check its symmetry, ensuring that any two corresponding elements in the matrix satisfy the symmetric relationship. At the same time, calculation errors of matrix elements are checked to ensure that the symmetric dependency matrix can accurately reflect the bidirectional dependency strength between each sub-objective, providing reliable data support for subsequent clustering processing.

[0151] Step S4032: Construct the Laplacian matrix based on the symmetric dependency matrix, and perform eigenvalue decomposition on the Laplacian matrix to extract the eigenvector corresponding to the smallest eigenvalue.

[0152] Specifically, firstly, a similarity matrix is ​​constructed based on the symmetric dependency matrix. The elements of the similarity matrix are derived from the elements of the symmetric dependency matrix and are used to characterize the degree of similarity between sub-targets. Then, a Laplacian matrix is ​​constructed based on the similarity matrix. Next, eigenvalue decomposition is performed on the Laplacian matrix to extract the smallest eigenvalues ​​and their corresponding eigenvectors. The formula for constructing the Laplacian matrix is:

[0153] (6).

[0154] in, For Laplace matrix, Let be the degree matrix corresponding to the similarity matrix. The degree matrix is ​​a diagonal matrix, where the diagonal elements are the sum of all elements in the corresponding row of the similarity matrix, and the off-diagonal elements are 0. This is a similarity matrix.

[0155] Step S4033: Using the extracted feature vectors, map multiple sub-targets to a low-dimensional feature space, and cluster the multiple sub-targets in the low-dimensional feature space to obtain multiple collaborative optimization groups.

[0156] After extracting the feature vector corresponding to the smallest feature value, the high-dimensional sub-target is mapped to the low-dimensional feature space through the feature vector, thus completing the dimensionality reduction embedding process of the sub-target. The purpose of dimensionality reduction embedding is to simplify the complexity of subsequent grouping calculations while preserving the core dependencies between sub-targets.

[0157] After mapping the sub-targets to a low-dimensional feature space, the k-means algorithm can be used to group the sub-targets in the low-dimensional feature space. Specifically, clustering is performed based on the similarity of the low-dimensional feature vectors of the sub-targets, and sub-targets with high similarity are grouped together, ultimately generating multiple collaborative optimization groups.

[0158] After obtaining multiple collaborative optimization groups, the groups can be validated to ensure that the sub-objectives within each collaborative optimization group have a high degree of dependency, the sub-objectives in different collaborative optimization groups have a low degree of dependency, and the grouping results are complete and without overlap.

[0159] After obtaining multiple collaborative optimization groups, for each collaborative optimization group, the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group are determined based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group. Some embodiments of this application describe the process of determining the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group.

[0160] The process of determining the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group may include:

[0161] Step a1: Calculate the weight of each sub-objective in the collaborative optimization group based on the information entropy attribute of each sub-objective within the collaborative optimization group.

[0162] For each sub-objective within the collaborative optimization group, the weight of the sub-objective is determined based on its information entropy attribute (including the maximum possible information entropy, the mutual information between the corresponding grid cell and its adjacent grid cells, and the information entropy of the sub-region to which the corresponding grid cell belongs) and the weight of the sub-region to which the grid cell to which the sub-objective belongs.

[0163] Specifically, the weights of sub-objectives can be calculated using the following formula:

[0164] (7).

[0165] in, For sub-targets The weight, For sub-targets Sub-area The weight, For sub-targets The maximum possible information entropy For sub-targets The mutual information between the current grid cell and its neighboring grid cells. For sub-targets Sub-area Information entropy, exponential term This is an information redundancy penalty factor.

[0166] Step a2-a: Based on the weights of each sub-objective within the collaborative optimization group, merge the mapping functions of each sub-objective within the collaborative optimization group to obtain the group-level mapping function of the collaborative optimization group.

[0167] Specifically, the weights of each sub-objective within the collaborative optimization group are normalized to obtain normalized weights for each sub-objective. Then, the mapping functions of each sub-objective within the collaborative optimization group are weighted and averaged according to their normalized weights to obtain the group-level mapping function of the collaborative optimization group. This group-level mapping function integrates the optimization requirements of all sub-objectives within the collaborative optimization group.

[0168] Step a2-b: Combine the weights of each sub-objective within the collaborative optimization group to obtain the comprehensive weight of the collaborative optimization group.

[0169] Specifically, the weights of each sub-objective within the collaborative optimization group are summed directly to obtain the overall weight of the collaborative optimization group.

[0170] Steps a2-c: Determine the reference information entropy of the collaborative optimization group based on the information entropy attributes of each sub-objective within the collaborative optimization group.

[0171] Specifically, the reference information entropy of the collaborative optimization group can be determined based on the maximum possible information entropy in the information entropy attribute of each sub-objective within the collaborative optimization group.

[0172] The comprehensive weights, group-level mapping functions, and reference information entropy of each collaborative optimization group can be obtained in the manner described above.

[0173] After obtaining the comprehensive weights, group-level mapping functions, and reference information entropy of each collaborative optimization group, multi-objective optimization iteration is performed based on the weights, group-level mapping functions, and reference information entropy of each collaborative optimization group, with the goal of minimizing the sum of weighted information entropy residuals of each collaborative optimization group, to generate the optimal solution set of the decision variables. This process is described in some embodiments of this application.

[0174] like Figure 5 As shown, the process of generating the optimal solution set for decision variables through multi-objective optimization iteration, based on the weights of each collaborative optimization group, the group-level mapping function, and the reference information entropy, with the objective of minimizing the sum of the weighted information entropy residuals of each collaborative optimization group, may include:

[0175] Step S501: Obtain the initial candidate solution set.

[0176] The initial candidate solution set includes multiple candidate solutions, each of which represents a set of decision variable values.

[0177] Step S502: For each candidate solution in the current candidate solution set, based on the group-level mapping function, comprehensive weight and reference information entropy of each collaborative optimization group, calculate the weighted information entropy residual of the candidate solution in each collaborative optimization group, and sum the weighted information entropy residuals of the candidate solution in each collaborative optimization group to obtain the fitness of the candidate solution.

[0178] Specifically, for each candidate solution in the current candidate solution set, firstly, based on the group-level mapping function of each co-optimization group, the actual information entropy of the candidate solution in each co-optimization group is calculated. Then, based on the actual information entropy of each co-optimization group under the candidate solution and the reference information entropy of each co-optimization group, the information entropy residual of the candidate solution in each co-optimization group is calculated. Next, based on the comprehensive weight of each co-optimization group, the information entropy residual of the candidate solution in each co-optimization group is weighted to obtain the weighted information entropy residual of the candidate solution in each co-optimization group. Finally, the weighted information entropy residuals of the candidate solution in each co-optimization group are summed, and the summation result is used as the fitness of the candidate solution.

[0179] The i-th candidate solution Fitness can be calculated using the following formula:

[0180] (8).

[0181] in, Let be the fitness of the i-th candidate solution. The comprehensive weight of the k-th collaborative optimization group is... Let be the actual information entropy of the i-th candidate solution in the k-th co-optimization group. Let be the reference information entropy of the k-th co-optimization group, and K be the number of co-optimization groups.

[0182] Step S503: Based on the fitness of each candidate solution in the current candidate solution set, select parent candidate solutions from the current candidate solution set, and generate child candidate solutions based on the parent candidate solutions.

[0183] Specifically, based on the fitness of each candidate solution in the current candidate solution set, the current candidate solution set is sorted by non-dominated order and the crowding degree is calculated to obtain the non-dominated level and crowding degree of each candidate solution in the current candidate solution set. Based on the non-dominated level and crowding degree of each candidate solution in the current candidate solution set, parent candidate solutions are selected from the current candidate solution set in order of non-dominated level from low to high and crowding degree from large to small within the same level, to ensure that the selected parent candidate solutions have both optimization and diversity. Then, child candidate solutions are generated based on the parent candidate solutions.

[0184] Based on the fitness of the candidate solutions, the candidate solutions in the current candidate solution set are divided into non-dominated levels. The non-dominated level is used to characterize the dominance relationship of the candidate solutions. Candidate solutions that cannot be dominated by other candidate solutions are classified into the first non-dominated level, candidate solutions dominated by the candidate solutions in the first non-dominated level are classified into the second non-dominated level, and so on.

[0185] Crowding degree is calculated for candidate solutions within each non-dominated level. Crowding degree is used to characterize the distribution density of candidate solutions in the current candidate solution set. The crowding degree calculation formula is as follows:

[0186] (9).

[0187] in, Let i be the crowding degree of the i-th candidate solution. and Let be the fitness of the neighboring candidate solutions of the i-th candidate solution on the j-th optimization objective. and are the maximum fitness and minimum fitness of the candidate solutions in the current candidate solution set on the j-th optimization objective, respectively, and m is the number of optimization objectives.

[0188] It should be noted that there are multiple parent candidate solutions. Preliminary offspring candidate solutions can be generated by simulating binary crossover of multiple parent candidate solutions. Then, polynomial mutation (changing the values ​​of some decision variables) is performed on the preliminary offspring candidate solutions to introduce diversity and obtain the final offspring candidate solutions.

[0189] Step S504: Construct a new set of candidate solutions based on the parent candidate solutions and the child candidate solutions, and use the new set of candidate solutions as the current set of candidate solutions.

[0190] The obtained parent and child candidate solutions are merged, and then a specified number of candidate solutions are selected based on fitness and crowding to form a new candidate solution set.

[0191] Repeat steps S502 to S504 until the preset convergence condition is met. Obtain the candidate solution with the first non-dominated level from the final candidate solution set to form the optimal solution set.

[0192] After each iteration, the optimal fitness value of the current candidate solution set is recorded. The optimal fitness value is the minimum fitness value of the candidate solution with the first non-dominated level in the candidate solution set.

[0193] The formula for calculating the change in the optimal fitness value of the candidate solution set after multiple iterations is as follows:

[0194] (10).

[0195] in, The range of change in the optimal fitness value. Let be the optimal fitness value of the candidate solution set obtained after the t-th iteration. Let t be the optimal fitness value of the candidate solution set obtained after the nth iteration, where n is the preset number of consecutive iterations.

[0196] By comparing the magnitude of the change in the optimal fitness value with a preset convergence threshold, it is determined whether the candidate solution set has reached a convergence state. If the magnitude of the change in the optimal fitness value of the candidate solution set obtained after multiple consecutive iterations is less than the preset convergence threshold, the iteration is stopped. Then, all candidate solutions with the first non-dominated level are extracted from the candidate solution set obtained after the last iteration to form the optimal solution set (i.e., the Pareto optimal solution set).

[0197] After obtaining the optimal solution set, an optimal solution is selected from the optimal solution set as the implementation configuration scheme of the ecological environment monitoring network. In some embodiments of this application, the process of selecting an optimal solution from the optimal solution set is described.

[0198] like Figure 6 As shown, the process of selecting an optimal solution from the set of optimal solutions may include:

[0199] Step S601: Based on the group-level mapping function, comprehensive weight, and reference information entropy of each collaborative optimization group, calculate the weighted information entropy residual of each optimal solution in the optimal solution set on each collaborative optimization group.

[0200] The process of calculating the weighted information entropy residual of each optimal solution in the optimal solution set on each collaborative optimization group, based on the group-level mapping function, comprehensive weight, and reference information entropy of each collaborative optimization group, is similar to the process of calculating the weighted information entropy residual of the candidate solution on each collaborative optimization group, and will not be described in detail here.

[0201] Step S602: Using the fuzzy membership function, the weighted information entropy residual of each optimal solution in the optimal solution set on each co-optimization group is mapped to a fuzzy membership value with a unified dimension. The fuzzy membership values ​​of the same optimal solution on each co-optimization group are combined into a fuzzy membership value set, thus obtaining the fuzzy membership value set of each optimal solution in the optimal solution set.

[0202] Specifically, the weighted information entropy residual of each optimal solution in the optimal solution set on each collaborative optimization group can be mapped to a fuzzy membership value with a unified dimension using the following formula:

[0203] (11).

[0204] in, Let be the fuzzy membership value of the i-th optimal solution in the optimal solution set on the k-th collaborative optimization group. The weighted information entropy residual of the i-th optimal solution on the k-th co-optimization group is... This is a membership function parameter used to control the rate of change of membership values.

[0205] For each optimal solution in the optimal solution set, after obtaining the fuzzy membership values ​​of the optimal solution in each co-optimization group, the fuzzy membership values ​​of the optimal solution in each co-optimization group are combined into a set to obtain the fuzzy membership value set of the optimal solution.

[0206] Step S603: Select an optimal solution from the optimal solution set based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set.

[0207] Specifically, the process of selecting an optimal solution from the optimal solution set based on the set of fuzzy membership values ​​of each optimal solution includes:

[0208] Step S6031-a: Rank each optimal solution in the optimal solution set according to its fuzzy membership value in the same collaborative optimization group, and obtain the ranking of each optimal solution in the optimal solution set in each collaborative optimization group. Based on the ranking of each optimal solution in the optimal solution set in each collaborative optimization group, determine the score of each optimal solution in the optimal solution set.

[0209] For each collaborative optimization group, the fuzzy membership values ​​of each optimal solution in the optimal solution set are sorted in descending order in the collaborative optimization group. The ranking of each optimal solution in the optimal solution set in the collaborative optimization group is determined according to the sorting result. For example, if the fuzzy membership value of optimal solution 1 is ranked first in the collaborative optimization group, then optimal solution 1 is ranked first in the collaborative optimization group. If the fuzzy membership value of optimal solution 2 is ranked second in the collaborative optimization group, then optimal solution 2 is ranked second in the collaborative optimization group, and so on.

[0210] After obtaining the ranking of each optimal solution in the optimal solution set within each collaborative optimization group, the score of each optimal solution in the optimal solution set is further determined based on its ranking within each collaborative optimization group. Specifically, the score of each optimal solution in the optimal solution set can be determined using the following formula:

[0211] (12).

[0212] in, Let K be the score of the i-th optimal solution in the optimal solution set, and K be the total number of collaborative optimization groups. The score of the i-th optimal solution in the optimal solution set is the rank of the i-th optimal solution in the k-th co-optimization group. In other words, the score of the i-th optimal solution in the optimal solution set is the sum of the inversion values ​​of the i-th optimal solution's rank in each co-optimization group.

[0213] Step S6031-b: Determine the number of approval votes for each optimal solution in the optimal solution set based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set.

[0214] In addition to determining the score of each optimal solution in the optimal solution set, this embodiment can also determine the number of approval votes for each optimal solution in the optimal solution set based on the fuzzy membership value set of each optimal solution in the optimal solution set. The number of approval votes for any optimal solution in the optimal solution set is the number of fuzzy membership values ​​in the fuzzy membership value set of that optimal solution that exceed a preset fuzzy membership threshold.

[0215] Step S6032: Select an optimal solution from the optimal solution set based on the scores of each optimal solution and the number of approved votes for each optimal solution.

[0216] Optionally, the optimal solutions in the optimal solution set are ranked by score, and the optimal solutions in the optimal solution set are ranked by the number of approved votes. Based on the two ranking results, the optimal solutions with the same ranking are determined. Then, the optimal solution with the highest ranking is selected from the optimal solutions with the same ranking, and the selected optimal solution is used as the implementation configuration scheme.

[0217] After obtaining the implementation configuration plan, the ecological environment monitoring network can be deployed according to the implementation configuration plan. After the ecological environment monitoring network is deployed, real-time monitoring data of the ecological environment monitoring network can be obtained. Then, based on the real-time monitoring data, deviation detection and iterative correction processing are performed on the partial derivative dependency matrix, group-level mapping function, sub-region boundaries, and sub-region mutual information (mutual information between grid cells within a sub-region). When the rate of change of the sub-region mutual information after correction compared with the sub-region mutual information before correction exceeds a preset rate of change threshold, an updated optimal solution set is generated according to the optimal solution and the updated group-level mapping function. Specifically, the optimal solution set before correction (Pareto optimal solution set) is recalculated through the updated group-level mapping function, retaining non-dominated solutions, and an updated optimal solution set is generated.

[0218] The rate of change of the mutual information of the sub-region after correction compared to the mutual information of the sub-region before correction is calculated by the following formula:

[0219] (13).

[0220] in, The rate of change of mutual information in the sub-region. To correct the mutual information of the sub-regions, To correct the mutual information of the sub-regions.

[0221] This application also provides a device for determining the configuration scheme of an ecological environment monitoring network, such as... Figure 7 As shown, the device may include: a sub-objective generation module 701, a mapping function establishment module 702, a sub-objective grouping module 703, a collaborative optimization group related data determination module 704, a multi-objective optimization module 705, and an optimal solution selection module 706.

[0222] The sub-target generation module 701 is used to generate sub-targets with information entropy attributes corresponding to each grid unit covered by the monitoring area, based on the engineering monitoring targets and basic ecological and environmental data of the monitoring area.

[0223] The mapping function establishment module 702 is used to establish mapping functions for the multiple sub-objectives obtained, respectively, with respect to the decision variables.

[0224] The sub-target grouping module 703 is used to group multiple sub-targets according to the coupling relationship between them, forming multiple collaborative optimization groups. The decision variables are the configuration parameters of the ecological environment monitoring network.

[0225] The collaborative optimization group related data determination module 704 is used to determine the comprehensive weight, group-level mapping function, and reference information entropy of each collaborative optimization group based on the information entropy attribute and mapping function of each sub-objective within the collaborative optimization group.

[0226] The multi-objective optimization module 705 is used to perform multi-objective optimization iteration based on the weights of each collaborative optimization group, the group-level mapping function, and the reference information entropy, with the objective of minimizing the sum of the weighted information entropy residuals of each collaborative optimization group, and to generate the optimal solution set of the decision variables.

[0227] The optimal solution selection module 706 is used to select a set of optimal solutions from the optimal solution set as the implementation configuration scheme for the ecological environment monitoring network.

[0228] In one possible implementation, when the sub-target generation module 701 generates sub-targets with information entropy attributes corresponding to each grid cell covered by the monitoring area based on the engineering monitoring targets and basic ecological and environmental data of the monitoring area, it is specifically used for:

[0229] Based on the engineering monitoring objectives and the natural geographical features data in the basic ecological and environmental data, the monitoring area is divided into multiple sub-regions, so that each grid unit covered by the monitoring area is assigned to a unique sub-region.

[0230] Determine the information entropy of each sub-region, the mutual information between grid cells within each sub-region, and the maximum possible information entropy of each grid cell.

[0231] Generate a sub-target with an information entropy attribute corresponding to each grid cell. The information entropy attribute of the sub-target corresponding to any grid cell includes the maximum possible information entropy of the grid cell, the mutual information between the grid cell and its neighboring grid cells, and the information entropy of the sub-region to which the grid cell belongs.

[0232] In one possible implementation, the mapping function establishment module 702, when establishing mapping functions for the multiple sub-objectives with respect to the decision variables, is specifically used for:

[0233] For each sub-objective, a mapping function of the sub-objective with respect to the decision variables is established based on the information entropy attribute of the sub-objective.

[0234] In one possible implementation, the information entropy attribute of the sub-target includes: the maximum possible information entropy and the mutual information between the grid cell corresponding to the sub-target and its adjacent grid cells.

[0235] The mapping function establishment module 702, when establishing a mapping function for the sub-objective with respect to the decision variables based on the information entropy attribute of the sub-objective, is specifically used for:

[0236] Based on the maximum possible information entropy in the information entropy attribute of the sub-objective, establish a preliminary mapping function of the sub-objective with respect to the decision variables;

[0237] By combining the mutual information in the information entropy attribute of the sub-objective, the initial mapping function is regularized and optimized to generate a mapping function that can reflect the nonlinear relationship between the sub-objective and the decision variables, thus obtaining the final mapping function of the sub-objective with respect to the decision variables.

[0238] In one possible implementation, when the mapping function establishment module 702 establishes the preliminary mapping function of the sub-target with respect to the decision variables based on the maximum possible information entropy in the information entropy attribute of the sub-target, it is specifically used for:

[0239] The maximum possible information entropy in the information entropy attribute of the sub-objective is kernel-embedded and mapped to the decision variables in the decision variable space to generate a preliminary mapping function of the sub-objective with respect to the decision variables.

[0240] When the mapping function establishment module 702 combines the mutual information in the information entropy attribute of the sub-objective to perform regularization optimization on the preliminary mapping function, and generates a mapping function that can reflect the nonlinear relationship between the sub-objective and the decision variables, it is specifically used for:

[0241] Based on the initial mapping function, and combined with the mutual information in the information entropy attribute of the sub-objective, the sub-objective mapping values ​​at multiple candidate points in the decision variable space are fitted by least squares regression under regularization constraints to generate a mapping function that can reflect the nonlinear relationship between the sub-objective and the decision variable.

[0242] In one possible implementation, when the sub-target grouping module 703 groups multiple sub-targets according to the coupling relationship between them to form multiple collaborative optimization groups, it is specifically used for:

[0243] Based on the mapping function of each sub-objective, calculate the partial derivative dependency coefficients between each pair of sub-objectives;

[0244] Based on the partial derivative dependency coefficients between each pair of sub-targets, a partial derivative dependency matrix is ​​constructed to characterize the nonlinear coupling relationship between each pair of sub-targets.

[0245] Based on the partial derivative dependency matrix, the sub-objectives are clustered to form multiple collaborative optimization groups.

[0246] In one possible implementation, the sub-target grouping module 703, when calculating the pairwise partial derivative dependency coefficients between each sub-target based on the mapping function of each sub-target, is specifically used for:

[0247] For each pair of sub-targets, calculate the partial derivative of the mapping function of the first sub-target with respect to the mapping function of the second sub-target to obtain the first partial derivative value, and calculate the partial derivative of the mapping function of the second sub-target with respect to the mapping function of the first sub-target to obtain the second partial derivative value.

[0248] Calculate the average of the first and second partial derivatives, and use the average as the partial derivative dependence coefficient between the two sub-objectives.

[0249] In one possible implementation, when the sub-objective grouping module 703 clusters the sub-objectives according to the partial derivative dependency matrix to form multiple collaborative optimization groups, it is specifically used for:

[0250] The partial derivative dependency matrix is ​​symmetricized to obtain a symmetric dependency matrix;

[0251] Construct the Laplacian matrix based on the symmetric dependency matrix, and perform eigenvalue decomposition on the Laplacian matrix to extract the eigenvector corresponding to the smallest eigenvalue;

[0252] Each sub-objective is mapped to a low-dimensional feature space using feature vectors, and then clustered in the low-dimensional feature space to form multiple collaborative optimization groups.

[0253] In one possible implementation, the collaborative optimization group related data determination module 704, when determining the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group, is specifically used for:

[0254] Based on the information entropy attributes of each sub-objective within the collaborative optimization group, calculate the reference information entropy of the collaborative optimization group and the weights of each sub-objective within the collaborative optimization group;

[0255] Based on the weights of each sub-objective within the collaborative optimization group, the mapping functions of each sub-objective within the collaborative optimization group are merged to obtain the group-level mapping function of the collaborative optimization group.

[0256] The weights of each sub-objective within the collaborative optimization group are combined to obtain the overall weight of the collaborative optimization group.

[0257] In one possible implementation, the multi-objective optimization module 705, when performing multi-objective optimization iterations based on the weights of each collaborative optimization group, the group-level mapping function, and the reference information entropy, with the objective of minimizing the sum of the weighted information entropy residuals of each collaborative optimization group, and generating the optimal solution set for the decision variables, is specifically used for:

[0258] Obtain an initial candidate solution set, which includes multiple candidate solutions, each of which represents a set of decision variable values;

[0259] For each candidate solution in the current candidate solution set, based on the group-level mapping function, comprehensive weight and reference information entropy of each collaborative optimization group, calculate the weighted information entropy residual of the candidate solution in each collaborative optimization group, sum the weighted information entropy residuals of the candidate solution in each collaborative optimization group, and use the summation result as the fitness of the candidate solution;

[0260] Based on the fitness of each candidate solution in the current candidate solution set, select parent candidate solutions from the current candidate solution set, and generate child candidate solutions based on the parent candidate solutions;

[0261] Based on the parent and child candidate solutions, a new candidate solution set is constructed. The new candidate solution set is used as the current candidate solution set. The process of calculating the sum of weighted information entropy residuals, selecting parent candidate solutions, generating child candidate solutions, and constructing a new candidate solution set is repeated until the preset convergence condition is met. The optimal solution is obtained from the final candidate solution set to form the optimal solution set.

[0262] In one possible implementation, when the multi-objective optimization module 705 selects a parent candidate solution from the current candidate solution set based on the fitness of each candidate solution in the current candidate solution set, it is specifically used for:

[0263] Based on the fitness of each candidate solution in the current candidate solution set, the non-dominated sorting and crowding degree calculation are performed on the current candidate solution set to obtain the non-dominated level and crowding degree of each candidate solution in the current candidate solution set;

[0264] Based on the non-dominated level and congestion of each candidate solution in the current candidate solution set, parent candidate solutions are selected from the current candidate solution set in order of non-dominated level from low to high and congestion within the same level from large to small.

[0265] In one possible implementation, the multi-objective optimization module 705, when obtaining the optimal solution from the final candidate solution set and forming the optimal solution set, is specifically used for:

[0266] The candidate solutions with the first non-dominated level are selected from the final candidate solution set to form the optimal solution set.

[0267] In one possible implementation, when the optimal solution selection module 706 selects a set of optimal solutions from the optimal solution set as the implementation configuration scheme for the ecological environment monitoring network, it is specifically used for:

[0268] Based on the group-level mapping function, comprehensive weight and reference information entropy of each collaborative optimization group, calculate the weighted information entropy residual of each optimal solution in the optimal solution set on each collaborative optimization group;

[0269] By using the fuzzy membership function, the weighted information entropy residual of each optimal solution in the optimal solution set on each co-optimization group is mapped to a fuzzy membership value with a unified dimension. The fuzzy membership values ​​of the same optimal solution on each co-optimization group are combined into a fuzzy membership value set, thus obtaining the fuzzy membership value set of each optimal solution in the optimal solution set.

[0270] Based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set, select an optimal solution from the optimal solution set.

[0271] In one possible implementation, when the optimal solution selection module 706 selects an optimal solution from the optimal solution set based on the set of fuzzy membership values ​​of each optimal solution, it is specifically used for:

[0272] The optimal solutions in the optimal solution set are ranked according to their fuzzy membership values ​​in the same collaborative optimization group. The ranking of each optimal solution in the optimal solution set in each collaborative optimization group is obtained. Based on the ranking of each optimal solution in the optimal solution set in each collaborative optimization group, the score of each optimal solution in the optimal solution set is determined.

[0273] Based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set, calculate the number of approval votes for each optimal solution in the optimal solution set. The number of approval votes for any optimal solution in the optimal solution set is the number of fuzzy membership values ​​in the set of fuzzy membership values ​​of that optimal solution that exceed the preset fuzzy membership threshold.

[0274] Based on the scores and approval votes of each optimal solution in the optimal solution set, select an optimal solution from the optimal solution set.

[0275] In one possible implementation, the device for determining the configuration scheme of the ecological environment monitoring network may further include: a monitoring data acquisition module, a parameter correction module, and an optimal solution set update module.

[0276] The monitoring data acquisition module is used to acquire real-time monitoring data from the ecological and environmental monitoring network deployed according to the implementation configuration plan.

[0277] The parameter correction module is used to correct the sub-region boundary, sub-region mutual information, and group-level mapping function based on real-time monitoring data. The sub-region mutual information is the mutual information between grid cells within the sub-region.

[0278] The optimal solution set update module is used to generate an updated optimal solution set based on the optimal solution and the updated group-level mapping function when the rate of change of the mutual information of the corrected sub-region relative to the mutual information of the uncorrected sub-region exceeds a preset rate of change threshold.

[0279] The device for determining the configuration scheme of the ecological environment monitoring network provided in this application fundamentally solves the problem of granularity mismatch between engineering monitoring targets and decision variables by decomposing macro-engineering monitoring targets into grid-level sub-targets with quantified information entropy attributes. It realizes the precise spatial transmission of monitoring needs from macro indicators to micro decisions. After obtaining multiple sub-targets, it intelligently groups the sub-targets according to the coupling relationship between them to form a collaborative optimization group. This realizes the cluster collaborative optimization of related sub-targets, greatly reduces the optimization complexity, and significantly improves the overall integrity of the scheme. With minimizing the sum of weighted information entropy residuals of each collaborative optimization group as the unified objective, it establishes an objective optimization criterion that drives the maximization of global information acquisition, realizes the scientific automatic trade-off between multiple targets, and finally generates an optimal solution set that provides high-quality alternative schemes that are optimal in terms of information utility and have different focuses.

[0280] This application also provides an electronic device, which may include at least one processor and a memory connected to the processor.

[0281] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application; the memory may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage.

[0282] The memory is used to store computer programs, and the processor is used to execute the computer programs so that the electronic device can implement the method for determining the configuration scheme of the ecological environment monitoring network provided in the above embodiments.

[0283] This application also provides a computer storage medium carrying one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device is able to implement the method for determining the configuration scheme of the ecological environment monitoring network provided in the above embodiments.

[0284] This application also provides a computer program product, including computer-readable instructions, which, when executed on an electronic device, enable the electronic device to implement the method for determining the configuration scheme of the ecological environment monitoring network provided in the above embodiments.

[0285] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0286] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0287] In the above embodiments, the implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, in the form of a computer program product.

[0288] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

Claims

1. A method for determining the configuration scheme of an ecological environment monitoring network, characterized in that, include: Based on the engineering monitoring objectives and basic ecological and environmental data of the monitoring area, sub-objectives with information entropy attributes are generated for each grid unit covered by the monitoring area. The information entropy attribute represents the information uncertainty of the corresponding grid unit, the information uncertainty of the local area to which the corresponding grid unit belongs, and the information correlation between the corresponding grid unit and its neighboring grid units. Establish mapping functions for multiple sub-objectives with respect to decision variables, and group the multiple sub-objectives according to the coupling relationship between them to form multiple collaborative optimization groups. The decision variables are the configuration parameters of the ecological environment monitoring network. For each of the collaborative optimization groups, the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group are determined based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group. Based on the weights, group-level mapping functions, and reference information entropy of each of the collaborative optimization groups, a multi-objective optimization iteration is performed with the objective of minimizing the sum of the weighted information entropy residuals of each of the collaborative optimization groups, generating the optimal solution set of the decision variables, and selecting a set of optimal solutions from the optimal solution set as the implementation configuration scheme of the ecological environment monitoring network.

2. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 1, characterized in that, The step of generating sub-targets with information entropy attributes for each grid unit covered by the monitoring area, based on the engineering monitoring targets and basic ecological and environmental data of the monitoring area, includes: Based on the engineering monitoring objectives and the natural geographical features data in the ecological and environmental basic data, the monitoring area is divided into multiple sub-regions, so that each grid unit covered by the monitoring area is assigned to a unique sub-region; Determine the information entropy of each sub-region, the mutual information between grid cells within each sub-region, and the maximum possible information entropy of each grid cell; Generate a sub-target with an information entropy attribute corresponding to each of the grid cells. The information entropy attribute of the sub-target corresponding to any grid cell includes the maximum possible information entropy of the grid cell, the mutual information between the grid cell and its neighboring grid cells, and the information entropy of the sub-region to which the grid cell belongs.

3. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 1, characterized in that, The established mapping functions of the multiple sub-objectives with respect to the decision variables include: For each sub-objective, a mapping function of the sub-objective with respect to the decision variables is established based on the information entropy attribute of the sub-objective.

4. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 3, characterized in that, The information entropy attribute of the sub-target includes: the maximum possible information entropy and the mutual information between the grid cell corresponding to the sub-target and its adjacent grid cells; The step of establishing a mapping function of the sub-objective with respect to the decision variables based on the information entropy attribute of the sub-objective includes: Based on the maximum possible information entropy in the information entropy attribute of the sub-target, establish a preliminary mapping function of the sub-target with respect to the decision variables; By combining the mutual information in the information entropy attribute of the sub-target, the preliminary mapping function is regularized and optimized to generate a mapping function that can reflect the nonlinear correlation between the sub-target and the decision variable, thus obtaining the final mapping function of the sub-target with respect to the decision variable.

5. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 4, characterized in that, The step of establishing a preliminary mapping function for the sub-target with respect to the decision variables based on the maximum possible information entropy in the information entropy attribute of the sub-target includes: The maximum possible information entropy in the information entropy attribute of the sub-target is processed by kernel embedding mapping with the decision variables in the decision variable space to generate a preliminary mapping function of the sub-target with respect to the decision variables; The step of combining the mutual information in the information entropy attribute of the sub-target to perform regularization optimization on the preliminary mapping function, generating a mapping function that can reflect the nonlinear correlation between the sub-target and the decision variable, includes: Based on the preliminary mapping function, and combined with the mutual information in the information entropy attribute of the sub-target, the sub-target mapping values ​​at multiple candidate points in the decision variable space are fitted with least squares regression under regularization constraints to generate a mapping function that can reflect the nonlinear relationship between the sub-target and the decision variable.

6. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 1, characterized in that, The step of grouping the multiple sub-objectives according to the coupling relationship between them to form multiple collaborative optimization groups includes: Based on the mapping function of each sub-objective, calculate the partial derivative dependency coefficients between each pair of sub-objectives; Based on the partial derivative dependency coefficients between each pair of sub-targets, a partial derivative dependency matrix is ​​constructed to characterize the nonlinear coupling relationship between each pair of sub-targets. Based on the partial derivative dependency matrix, the sub-objectives are clustered to form multiple collaborative optimization groups.

7. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 6, characterized in that, The calculation of the pairwise partial derivative dependency coefficients between each of the sub-objectives based on the mapping function of each sub-objective includes: For each pair of sub-targets, calculate the partial derivative of the mapping function of the first sub-target with respect to the mapping function of the second sub-target to obtain a first partial derivative value, and calculate the partial derivative of the mapping function of the second sub-target with respect to the mapping function of the first sub-target to obtain a second partial derivative value. Calculate the average of the first partial derivative value and the second partial derivative value, and use the calculated average value as the partial derivative dependency coefficient between the two sub-objectives.

8. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 6, characterized in that, The step involves clustering the sub-objectives based on the partial derivative dependency matrix to form multiple collaborative optimization groups, including: The partial derivative dependency matrix is ​​symmetricized to obtain a symmetric dependency matrix; Construct a Laplacian matrix based on the symmetric dependency matrix, and perform eigenvalue decomposition on the Laplacian matrix to extract the eigenvector corresponding to the smallest eigenvalue; Each sub-target is mapped to a low-dimensional feature space using the feature vector, and the sub-targets are clustered in the low-dimensional feature space to form multiple collaborative optimization groups.

9. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 1, characterized in that, The step of determining the comprehensive weight, group-level mapping function, and reference information entropy of the collaborative optimization group based on the information entropy attributes and mapping functions of each sub-objective within the collaborative optimization group includes: Based on the information entropy attributes of each sub-objective within the collaborative optimization group, calculate the reference information entropy of the collaborative optimization group and the weight of each sub-objective within the collaborative optimization group; Based on the weights of each sub-objective within the collaborative optimization group, the mapping functions of each sub-objective within the collaborative optimization group are merged to obtain the group-level mapping function of the collaborative optimization group; The weights of each sub-objective within the collaborative optimization group are fused to obtain the comprehensive weight of the collaborative optimization group.

10. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 1, characterized in that, The process of performing multi-objective optimization iterations based on the weights, group-level mapping functions, and reference information entropy of each of the collaborative optimization groups, with the objective of minimizing the sum of the weighted information entropy residuals of each of the collaborative optimization groups, to generate the optimal solution set for the decision variables, includes: Obtain an initial candidate solution set, which includes multiple candidate solutions, each candidate solution being a set of decision variable values; For each candidate solution in the current candidate solution set, based on the group-level mapping function, comprehensive weight and reference information entropy of each collaborative optimization group, the weighted information entropy residual of the candidate solution in each collaborative optimization group is calculated, and the weighted information entropy residual of the candidate solution in each collaborative optimization group is summed. The summation result is used as the fitness of the candidate solution. Based on the fitness of each candidate solution in the current candidate solution set, a parent candidate solution is selected from the current candidate solution set, and a child candidate solution is generated based on the parent candidate solution. Based on the parent candidate solutions and the child candidate solutions, a new candidate solution set is constructed. The new candidate solution set is used as the current candidate solution set. The process of calculating the sum of weighted information entropy residuals, selecting parent candidate solutions, generating child candidate solutions, and constructing a new candidate solution set is repeated until the preset convergence condition is met. The optimal solution is obtained from the final candidate solution set to form the optimal solution set.

11. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 10, characterized in that, The step of selecting a parent candidate solution from the current candidate solution set based on the fitness of each candidate solution in the current candidate solution set includes: Based on the fitness of each candidate solution in the current candidate solution set, the non-dominated sorting and crowding degree calculation are performed on the current candidate solution set to obtain the non-dominated level and crowding degree of each candidate solution in the current candidate solution set; Based on the non-dominated level and congestion of each candidate solution in the current candidate solution set, parent candidate solutions are selected from the current candidate solution set in order of non-dominated level from low to high and congestion within the same level from large to small. The step of obtaining the optimal solution from the final candidate solution set and forming the optimal solution set includes: The candidate solutions with the first non-dominated level are selected from the final candidate solution set to form the optimal solution set.

12. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 1, characterized in that, The step of selecting a set of optimal solutions from the optimal solution set as the implementation configuration scheme for the ecological environment monitoring network includes: Based on the group-level mapping function, comprehensive weight, and reference information entropy of each of the collaborative optimization groups, the weighted information entropy residual of each optimal solution in the optimal solution set on each collaborative optimization group is calculated. By using the fuzzy membership function, the weighted information entropy residual of each optimal solution in the optimal solution set on each of the collaborative optimization groups is mapped to a fuzzy membership value with a unified dimension. The fuzzy membership values ​​of the same optimal solution on each of the collaborative optimization groups are combined into a fuzzy membership value set, thus obtaining the fuzzy membership value set of each optimal solution in the optimal solution set. Based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set, select an optimal solution from the optimal solution set.

13. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 12, characterized in that, The step of selecting an optimal solution from the optimal solution set based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set includes: The optimal solutions in the optimal solution set are ranked according to their fuzzy membership values ​​in the same collaborative optimization group to obtain the ranking of each optimal solution in the optimal solution set in each collaborative optimization group. Based on the ranking of each optimal solution in the optimal solution set in each collaborative optimization group, the score of each optimal solution in the optimal solution set is determined. Based on the set of fuzzy membership values ​​of each optimal solution in the optimal solution set, the number of approval votes for each optimal solution in the optimal solution set is calculated. The number of approval votes for any optimal solution in the optimal solution set is the number of fuzzy membership values ​​in the set of fuzzy membership values ​​of that optimal solution that exceed a preset fuzzy membership threshold. Based on the scores and approval votes of each optimal solution in the optimal solution set, an optimal solution is selected from the optimal solution set.

14. The method for determining the configuration scheme of the ecological environment monitoring network according to claim 2, characterized in that, Also includes: Obtain real-time monitoring data from the ecological environment monitoring network deployed according to the implementation configuration scheme; Based on the real-time monitoring data, the sub-region boundary, sub-region mutual information, and the group-level mapping function are corrected. The sub-region mutual information is the mutual information between grid cells within the sub-region. When the rate of change of the mutual information of the corrected sub-region relative to the mutual information of the uncorrected sub-region exceeds a preset rate of change threshold, an updated optimal solution set is generated based on the optimal solution and the updated group-level mapping function.

15. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to enable the electronic device to implement the method for determining the configuration scheme of the ecological environment monitoring network as described in any one of claims 1 to 14.

Images