Preferred method for stereoscopic water network networking scheme

Abstract

The application relates to the technical field of three-dimensional water network networking, and discloses an optimization method of a three-dimensional water network networking scheme. The optimization method comprises the following steps: generating a plurality of candidate schemes of a target region through a multi-objective optimization model, and inputting the candidate schemes into a pre-constructed system dynamics model to obtain corresponding original index values; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data and a plurality of functions expected to be realized at each level; each original index value is fed back to the multi-objective optimization model, and iterative optimization is performed on the multi-objective optimization model to generate a Pareto non-inferior solution set; the final index values in the Pareto non-inferior solution set are converted into single comprehensive index values through a dimension reduction decision model; and each candidate scheme is sorted according to the single comprehensive index values of the candidate schemes, and the highest candidate scheme is taken as an optimal networking scheme of the target region. The application can construct an optimal three-dimensional water network networking scheme for complex terrains.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of three-dimensional water network networking, in particular to a preferred method of a three-dimensional water network networking scheme. BACKGROUND

[0002] The existing water network system technology is undergoing a phased change from "linear transmission and distribution" to "spatial three-dimensional networking and digital collaboration". In terms of spatial structure construction, research has been able to combine the characteristics of complex terrain (such as karst topography), adopt node encryption, link strengthening, and supplementary short board, and construct a "surface-underground" "backbone-branch" collaborative spatial structure, thereby enhancing the overall connectivity and operational stability; in terms of control level, digital twin and uncertainty analysis technology are introduced into the field of water network simulation and risk identification, through the establishment of a mapping system of "physical water network-information water network-management water network-digital twin water network", visual modeling and operation simulation are realized to assist scheduling and planning decision-making. However, the current research and practice still have three deficiencies: first, the water network networking in complex terrain areas lacks systematic quantitative analysis in node distribution and link adaptation, making it difficult to fully depict the influence of factors such as elevation difference and flow direction constraints; second, existing models are mostly based on local optimization or empirical parameters, lacking a unified solution mechanism for multiple objectives under global constraints; third, simulation analysis can describe the structure characteristics, but lacks a self-consistent closed loop in scheme optimization and result verification. The above problems result in a low coupling degree between water network spatial configuration, operation scheduling and scheme evaluation, making it difficult to obtain a truly optimal three-dimensional water network networking scheme under multiple constraints. SUMMARY

[0003] The purpose of the embodiments of the present application is to provide a preferred method of a three-dimensional water network networking scheme, to solve the problem that the prior art is difficult to construct an optimal three-dimensional water network networking scheme in complex terrain.

[0004] To achieve the above-mentioned purpose, the first aspect of the present application provides a preferred method of a three-dimensional water network networking scheme, which comprises:

[0005] A multi-objective optimization model is used to generate a plurality of candidate water network networking schemes for a target region, and each candidate water network networking scheme is input into a pre-constructed system dynamics model to obtain a corresponding original index value; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data, and a plurality of functions expected to be realized by each level;

[0006] The original index value corresponding to each candidate water network networking scheme is fed back to the multi-objective optimization model, and a non-dominated sorting genetic algorithm is used to iteratively optimize it to generate a Pareto non-inferior solution set;

[0007] The final index value of each candidate water network scheme in the Pareto non-inferior solution set is converted into a single comprehensive index value by a dimension reduction decision model;

[0008] The multiple candidate water network schemes are ranked according to the single comprehensive index values of the multiple candidate water network schemes, and the candidate water network scheme with the highest single comprehensive index value among the single comprehensive index values of the multiple candidate water network schemes is taken as the optimal networking scheme of the target region.

[0009] In the embodiment of the present application, the step of constructing the system dynamics model comprises: in response to the three-dimensional water network networking generation instruction of the target region, parsing the multiple functions expected to be implemented by each level of the target region from the three-dimensional water network networking generation instruction; obtaining the water network data and the multiple engineering basic data of each level of the target region; determining the standardized index set of the target region based on the water network data, the multiple functions expected to be implemented by each level, and the factor analysis method; the standardized index set contains the correspondence between each level and the multiple functions; based on the standardized index set, mapping the multiple functions into corresponding multiple modules, and determining the core variables of each module; constructing the engineering data table of each level based on the multiple engineering basic data and the preset keyword library; determining the variable mapping rule between all key parameters in the engineering data table of each level and the core variables of all modules corresponding to the level, and forming the dynamic interaction link between the key parameters and the core variables; and constructing the system dynamics model according to the variable mapping rule and the dynamic interaction link.

[0010] In the embodiment of the present application, the step of determining the standardized index set of the target region based on the water network data, the multiple functions expected to be implemented by each level, and the factor analysis method comprises: establishing an initial index pool containing the multiple functions expected to be implemented by each level based on the water network data; performing standardization processing on the initial index pool, and determining the correlation coefficient matrix between the multiple indexes in the initial index pool after the standardization processing; performing eigenvalue decomposition on the correlation coefficient matrix to obtain a component list sorted by eigenvalue size; determining an initial factor loading matrix based on the factor analysis method, the component list, and a preset threshold value; performing factor rotation on the initial factor loading matrix to obtain a rotated factor loading matrix; and determining the standardized index set based on the factor loading matrix.

[0011] In the embodiment of the present application, the step of determining the initial factor loading matrix based on the factor analysis method, the component list, and the preset threshold value comprises: determining a first component list according to the eigenvalue of each component in the component list and a preset eigenvalue threshold value; the preset threshold value is the preset eigenvalue threshold value; and constructing the initial factor loading matrix based on the first component list.

[0012] In the embodiment of the present application, the step of determining the initial factor loading matrix based on the factor analysis method, the component list and the preset threshold value comprises: sequentially determining the cumulative variance contribution rate of each component in the component list; in the case that the cumulative variance contribution rate of the current component is the first to reach or exceed the preset variance contribution rate threshold value among the plurality of components, determining the serial number of the current component in the component list as the candidate factor number; the preset threshold value is the preset variance contribution rate threshold value; and constructing the initial factor loading matrix by sequentially selecting the components of the candidate factor number from the component list.

[0013] In the embodiment of the present application, the step of constructing the engineering data table of each level based on the plurality of engineering basic data and the preset keyword library comprises: parsing the engineering name, the functional attribute and the running feature of each engineering basic data from the plurality of engineering basic data; performing semantic analysis on each engineering basic data based on the preset keyword library, and outputting the key parameters of each engineering basic data; determining at least one module corresponding to each engineering basic data based on the functional attribute and the running feature; and constructing the engineering data table of each level based on the engineering name, the functional attribute, the running feature, the key parameters and the at least one corresponding module of the plurality of engineering basic data of each level.

[0014] In the embodiment of the present application, the preferred method further comprises: obtaining historical observation data of the water network of each level in the target region; determining an initial value range of the system dynamics model; constructing a posterior distribution of the parameters of the system dynamics model based on the system dynamics model, the initial value range and the historical observation data; sampling in the posterior distribution by the Hamilton Monte Carlo algorithm to determine an error value between the simulation value of the system dynamics model and the observed value in the historical observation data; adjusting the initial value range based on the error value, and repeatedly iterating the system dynamics model until the error value is lower than a preset precision threshold value.

[0015] The second aspect of the application provides a preferred system of a three-dimensional water network networking scheme, comprising: a generation module, which generates a plurality of candidate water network networking schemes of a target region through a multi-objective optimization model, and inputs each candidate water network networking scheme into a pre-constructed system dynamics model to obtain corresponding original index values; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data, and a plurality of functions expected to be realized at each level; a feedback module, which feeds back the original index values corresponding to each candidate water network networking scheme to the multi-objective optimization model, and iteratively optimizes the same through a non-dominated sorting genetic algorithm to generate a Pareto non-inferior solution set; a conversion module, which converts final index values of each candidate water network networking scheme in the Pareto non-inferior solution set into a single comprehensive index value through a dimension reduction decision model; and a sorting module, which sorts the plurality of candidate water network networking schemes according to the single comprehensive index values of the candidate water network networking schemes, and selects the highest candidate water network networking scheme among the single comprehensive index values of the candidate water network networking schemes as the optimal networking scheme of the target region.

[0016] The third aspect of the application provides a computer program product, comprising a computer program which, when executed by a processor, implements the preferred method described above.

[0017] The fourth aspect of the application provides a machine-readable storage medium having instructions stored thereon for causing a machine to perform the preferred method described above.

[0018] Through the above technical solution, first, a plurality of candidate water network networking schemes are batch-generated through a multi-objective optimization model, and each candidate water network networking scheme is input into a pre-constructed system dynamics model to obtain corresponding original index values, so that the physical interpretable calculation results are used as evaluation basis to improve the comparability and reliability of the evaluation; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data, and a plurality of functions expected to be realized at each level; the original index values corresponding to each candidate water network networking scheme are fed back to the multi-objective optimization model and iteratively optimized through a non-dominated sorting genetic algorithm to automatically converge to obtain a Pareto non-inferior solution set, thereby retaining effective candidates under cross-objective trade-off from the source; on this basis, the final index values of each candidate water network networking scheme are extracted from the Pareto non-inferior solution set, and the final index values are converted into a single comprehensive index value through a dimension reduction decision model, so as to unify the dimension scale and reduce the artificial subjectivity; finally, each candidate water network networking scheme is sorted according to the single comprehensive index value, and the highest one is selected as the optimal networking scheme of the target region, so as to ensure that the output result has clear executability and uniqueness, thereby realizing fast, standardized, traceable and reproducible optimal three-dimensional water network networking scheme acquisition.

[0019] Other features and advantages of the embodiments of the present application will be described in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings are included to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used to explain the embodiments of the present application together with the following detailed description, but do not limit the embodiments of the present application. In the drawings:

[0021] Figure 1 A flow chart of a preferred method of a three-dimensional water network networking scheme according to an embodiment of the present application is schematically shown;

[0022] Figure 2 A structural diagram of a computer device according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. It should be understood that the specific embodiments described herein are merely used to explain and illustrate the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0024] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are merely used to explain the relative positional relationship, movement condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directional indications also change accordingly.

[0025] In addition, if the embodiments of the present application involve descriptions of “first”, “second”, etc., the descriptions of “first”, “second”, etc. are merely for description purposes, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by “first”, “second” can explicitly or implicitly include at least one of the features. In addition, the technical solutions of the various embodiments can be combined with each other, but it must be based on the fact that a person of ordinary skill in the art can realize it, and when the combination of the technical solutions contradicts each other or cannot be realized, it should be considered that the combination of the technical solutions does not exist, and is not within the scope of protection claimed by the present application.

[0026] Figure 1 A flow chart of a preferred method of a three-dimensional water network networking scheme according to an embodiment of the present application is schematically shown. As shown in the figure, Figure 1As shown, the embodiment of the present application provides a preferred method of three-dimensional water network networking scheme, which can include the following steps.

[0027] Step 101: generating a plurality of candidate water network networking schemes of the target region by the multi-objective optimization model, inputting each candidate water network networking scheme into a pre-constructed system dynamics model to obtain corresponding original index values; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data and a plurality of functions expected to be realized at each level.

[0028] In the embodiment of the present application, the plurality of candidate water network networking schemes refer to a candidate configuration set formed after the instantiation of scheme parameters such as node settings, connection relationships, engineering scales, operation and distribution rules in the target region, which belong to the specific value combination of decision variables of the multi-objective optimization model in the feasible region; the pre-constructed system dynamics model refers to a model system that has completed structure setting and parameter caliber unification and can calculate the time sequence of water supply, flood control and ecological related states under the driving of scheme parameters; the original index value refers to a set of numerical evaluation quantities consistent with the function side measurement caliber obtained by calculating each candidate water network networking scheme through the pre-constructed system dynamics model, which is used to reflect the state performance and comparability of the scheme under the given caliber. The multi-objective optimization model generates a plurality of candidate water network networking schemes satisfying the constraints, and maps the scheme parameters to the original index values that can be directly used for evaluation, thereby completing the standardized calculation conversion and structured expression of the scheme to the evaluation quantity.

[0029] Step 102: feeding the original index values corresponding to each candidate water network networking scheme to the multi-objective optimization model, and iteratively optimizing it by the non-dominated sorting genetic algorithm to generate a Pareto non-inferior solution set.

[0030] Step 103: converting the final index values of each candidate water network networking scheme in the Pareto non-inferior solution set into a single comprehensive index value by the dimension reduction decision model.

[0031] In the embodiment of the present application, the Pareto non-inferior solution set refers to a set of schemes that are not dominated by each other in multi-objective optimization, that is, there is no other scheme that is at least not worse in all objectives and better in at least one objective, thereby retaining effective candidates under cross-objective trade-off; the extraction of the final index value of each candidate water network networking scheme refers to reading and collecting the evaluation set of the scheme in the unified standard only for the scheme in the Pareto non-inferior solution set, so as to serve as the effective data basis for subsequent discrimination; the dimension reduction decision model includes a multi-index comprehensive evaluation method, which is used to map the final index value to one-dimensional measurement of the same dimension under the premise of maintaining the distinguishability and stability; the single comprehensive index value is the numerical representation of the overall performance of the scheme after the transformation of the dimension reduction decision model, which can be directly compared on the same scale. Through this step, the final index value of the effective candidate scheme is extracted and converted to the same scale, forming a single comprehensive index value set that can be directly used for consistency evaluation and ordered comparison.

[0032] Step 104: According to the single comprehensive index value of each candidate water network networking scheme, the multiple candidate water network networking schemes are sorted, and the candidate water network networking scheme with the highest single comprehensive index value among the single comprehensive index values of the multiple candidate water network networking schemes is taken as the optimal networking scheme of the target region.

[0033] In the embodiment of the present application, the sorting of each candidate water network networking scheme refers to taking the single comprehensive index value as the only sorting basis, and performing ordered arrangement from high to low on all candidate schemes under the same dimension and caliber, so as to ensure consistent evaluation scale and clear comparison relationship; the highest candidate water network networking scheme refers to the candidate scheme that ranks first and has the highest single comprehensive index value in the sorting result, which is taken as the decision object with the best performance; the optimal networking scheme of the target region refers to the only preferred result determined for the target region under the current evaluation caliber. Through the above processing, the comprehensive evaluation result of multiple schemes is converted into a clear decision conclusion, and the optimal networking scheme that can be directly used for engineering landing and management deployment is output.

[0034] By the technical solution, firstly, a plurality of candidate water network networking schemes are generated in batches through the multi-objective optimization model, and each candidate water network networking scheme is input into the pre-constructed system dynamics model to obtain corresponding original index values, and the physical interpretable calculation results are used as evaluation basis to improve the comparability and reliability of the evaluation; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data and a plurality of functions expected to be realized at each level; the original index values corresponding to each candidate water network networking scheme are fed back to the multi-objective optimization model and iteratively optimized through the non-dominated sorting genetic algorithm to automatically converge to obtain a Pareto non-inferior solution set, thereby retaining effective candidates under cross-target trade-off from the source; on this basis, the final index values of each candidate water network networking scheme are extracted from the Pareto non-inferior solution set, and the final index values are converted into a single comprehensive index value through the dimension reduction decision model, thereby unifying the dimension scale and reducing the artificial subjectivity; finally, the single comprehensive index values are used to sort the candidate water network networking schemes and select the highest one as the optimal networking scheme of the target region, thereby ensuring that the output result has clear executability and uniqueness, so as to realize rapid, standardized, traceable and reproducible optimal three-dimensional water network networking scheme acquisition.

[0035] In the embodiment of the present application, the step of constructing the system dynamics model can include: in response to a three-dimensional water network networking generation instruction of a target region, parsing a plurality of functions expected to be realized at each level of the target region from the three-dimensional water network networking generation instruction; obtaining water network data of each level of the target region and a plurality of engineering basic data; determining a standardized index set of the target region based on the water network data, the plurality of functions expected to be realized at each level and the factor analysis method; the standardized index set contains a corresponding relationship between each level and the plurality of functions; based on the standardized index set, mapping the plurality of functions into corresponding modules and determining core variables of each module; constructing an engineering data table of each level based on the plurality of engineering basic data and a preset keyword library; determining a variable mapping rule between all key parameters in the engineering data table of each level and the core variables of all modules corresponding to the level, and forming a dynamic interaction link between the key parameters and the core variables; and constructing the system dynamics model according to the variable mapping rule and the dynamic interaction link.

[0036] In this embodiment, the target region refers to the geographically and managerially integrated spatial unit covered by the three-dimensional water network scheme to be generated; the three-dimensional water network generation instruction is a task-triggered information carrier for this spatial unit, carrying a machine-readable expression of the user's or system's network requirements; each level is a hierarchical management or spatial partitioning unit used for organization and governance within the target region; the multiple functions to be achieved are structured expressions of functional requirements such as water resource utilization, flood control safety, and ecological maintenance at each level; parsing refers to semantic recognition and field extraction of the three-dimensional water network generation instruction, transforming the elements related to the target region and the multiple functions to be achieved at each level into standardized requirement items and parameter specifications. Through the above processing, the requirement points of the three-dimensional water network generation instruction are extracted and the specifications are unified, clarifying the target set of the target region and each level on the functional side, laying a consistent functional input foundation for subsequent modeling, evaluation, and optimization using the same specifications.

[0037] In the embodiments of the present application, the water network data refers to a structured and semi-structured data set for reflecting the pattern, node position, connection relationship, capacity boundary and operation caliber of the water system in the target region, which can include topographic hydrology, transmission and distribution capacity and scheduling caliber information; the multiple engineering basic data refers to design documents, operation accounts, surveying and mapping data and monitoring records and other data carriers for describing the engineering properties and states related to specific engineering entities; the factor analysis method refers to a statistical method for dimensionality reduction and structure identification of multi-index data, which is used to extract representative public factors from preliminary indexes and select indexes accordingly; the standardized index set refers to an index set that can directly participate in modeling calculation after caliber unification and dimensionless processing, and records the corresponding relationship between each level and multiple functions in a clear data structure to ensure caliber consistency; the multiple function mappings are multiple modules corresponding to the abstract requirements on the function side, which are merged into a module set that can be calculated on the model side according to a unified rule, and are used to carry the conversion of functions to the calculation process; the core variables of each module refer to the minimum necessary variable list within the corresponding module as the basis for calculation and transmission, which are unified in name and dimension and are constrained in source and destination; the preset keyword library refers to a domain word table and rule set for semantic analysis and element extraction of multiple engineering basic data, so that key elements can be automatically identified and regularized; the engineering data table refers to a structured table with engineering as rows and engineering name, function attribute, operation feature, key parameter and corresponding module as columns, which is used to carry the engineering side input; the key parameter refers to a parameter item that can affect the module calculation result and is explicitly extracted or calculated in the engineering data table; the variable mapping rule refers to a constraint set that establishes one-to-one or many-to-many relationship between the key parameters in the engineering data table and the core variables of all modules corresponding to this level, which specifies the value caliber, update direction and dependency relationship; the dynamic interaction link refers to the time sequence and feedback connection between parameters and core variables formed according to the variable mapping rule, which is used to describe the action path of parameter change on the core variable and its feedback. Through the above processing, the continuous conversion and regularization from the water network data and the multiple engineering basic data to the standardized index set, the module and the core variable, the engineering data table, the variable mapping rule and the dynamic interaction link are completed, ensuring that the system dynamics model constructed accordingly has the characteristics of caliber unification, clear element source and clear calculation connection, thereby providing consistent, traceable and directly executable calculation basis for subsequent model operation and evaluation.

[0038] In the embodiment of the present application, the step of determining the standardized index set of the target region based on the water network data, the multiple functions expected to be achieved at each level, and the factor analysis method can include: establishing an initial index pool containing the multiple functions expected to be achieved at each level based on the water network data; performing standardization processing on the initial index pool, and determining a correlation coefficient matrix between the multiple indexes in the initial index pool after the standardization processing; performing eigenvalue decomposition on the correlation coefficient matrix to obtain a component list sorted by eigenvalue size; determining an initial factor loading matrix based on the factor analysis method, the component list, and a preset threshold; performing factor rotation on the initial factor loading matrix to obtain a rotated factor loading matrix; and determining the standardized index set based on the factor loading matrix.

[0039] In the embodiment of the present application, the initial index pool refers to a candidate index set established based on the water network data and covering the multiple functions expected to be achieved at each level, for carrying an original measurement space for subsequent screening and compression; the standardization processing refers to caliber unification and dimensionless conversion on each index in the initial index pool to eliminate dimension difference and improve comparability; the correlation coefficient matrix refers to a matrix description of the correlation degree between each two indexes calculated on the initial index pool after the standardization processing, for determining the linear association structure between the indexes; the component list sorted by eigenvalue size refers to a component sequence arranged by information contribution after the eigenvalue decomposition on the correlation coefficient matrix, which is an important basis for determining the amount of retained information; the initial factor loading matrix refers to a loading relationship table of indexes on potential common factors formed based on the factor analysis method, the component list, and a preset threshold, for expressing the explanation strength of each index on each common factor; the factor rotation refers to orthogonal or oblique rotation performed on the initial factor loading matrix to obtain a clearer loading aggregation structure; the rotated factor loading matrix refers to a result of the index-factor correspondence relationship after the factor rotation, which is convenient for explanation and screening; and the standardized index set refers to a list of indexes with strong explanation, low redundancy, and consistent caliber selected according to the rotated factor loading matrix, serving as a unified index input set for subsequent modeling. Through the above technical processing, a robust conversion from the initial index pool with a wide coverage to the standardized index set with concentrated information, clear structure, and direct calculation is realized, ensuring the consistency of the index caliber, the controllability of the correlation, and the sufficiency of the explanation.

[0040] In the embodiment of the present application, the step of determining the initial factor loading matrix based on the factor analysis method, the component list, and the preset threshold can include: determining a first component list according to the eigenvalue of each component in the component list and a preset eigenvalue threshold; the preset threshold is the preset eigenvalue threshold; and constructing the initial factor loading matrix based on the first component list.

[0041] In the embodiments of the present application, the preset threshold is a preset judgment boundary for screening components, the preset characteristic threshold is a specific form of the preset threshold in the characteristic value dimension, and is used to measure whether the explanatory power of the components reaches the retention standard; the first component list refers to a passing set screened from the component list by comparing the characteristic values of the components with the preset characteristic threshold, and represents a component subset recognized as having sufficient information contribution. Thus, the first component list is determined according to the characteristic values of each component in the component list and the preset characteristic threshold, and the initial factor loading matrix is formed by combining the directional information corresponding to the first component list according to the predetermined caliber. Through the above processing, the selection of the retained components has a unified standard and repeatability, the components with insufficient information contribution are eliminated, and the redundancy and noise are reduced, so that the initial factor loading matrix used for modeling is stably obtained without changing the original caliber.

[0042] In the embodiments of the present application, the step of determining the initial factor loading matrix based on the factor analysis method, the component list and the preset threshold can include: sequentially determining the cumulative variance contribution rate of each component in the component list; in the case that the cumulative variance contribution rate of the current component first reaches or exceeds the preset variance contribution rate threshold among the plurality of components, determining the serial number of the current component in the component list as the candidate factor number; the preset threshold is the preset variance contribution rate threshold; and selecting the components of the candidate factor number from the component list in sequence to construct the initial factor loading matrix.

[0043] In the embodiments of the present application, the cumulative variance contribution rate refers to a cumulative proportion formed by sequentially accumulating the explanatory proportions of each component to the overall variance information, and is used to measure the data explanatory degree covered by the first several components in total; the preset variance contribution rate threshold is a judgment boundary set for the cumulative variance contribution rate, and is used to determine the minimum number of components required to reach the information coverage requirement; the candidate factor number is the serial number of the current component in the component list when the cumulative variance contribution rate first reaches or exceeds the preset variance contribution rate threshold, and is used to limit the number of components participating in subsequent construction; and the sequential selection refers to strictly following the predetermined order of the component list to continuously select from the first component to the component corresponding to the candidate factor number without skipping or supplementing. Through the above processing, the number of components participating in construction is directly determined based on the cumulative variance information coverage as the only criterion without changing the caliber, and the corresponding components are sequentially selected from the component list to construct the initial factor loading matrix, so that the initial factor loading result with sufficient information explanatory degree and controlled redundancy is obtained.

[0044] In the embodiment of the present application, the step of constructing the engineering data table of each level based on the plurality of engineering basic data and the preset keyword library can include: parsing the engineering name, the function attribute and the operation characteristic of each engineering basic data from the plurality of engineering basic data; performing semantic analysis on each engineering basic data based on the preset keyword library, and outputting the key parameters of each engineering basic data; determining at least one module corresponding to each engineering basic data based on the function attribute and the operation characteristic; and constructing the engineering data table of each level based on the engineering name, the function attribute, the operation characteristic, the key parameters and the at least one corresponding module of the plurality of engineering basic data of each level.

[0045] In the embodiment of the present application, the structured field extraction and standardized input are performed on the plurality of engineering basic data, the engineering name, the function attribute and the operation characteristic are accurately positioned from the original text or table and converted into uniform data items, the semantic analysis is performed based on the preset keyword library to perform rule matching and context discrimination on the text and expression of the engineering basic data, thereby automatically identifying and outputting the key parameters, the at least one corresponding module emphasizes the one-to-one or one-to-many module attribution annotation according to the consistent caliber when the function attribute and the operation characteristic indicate that the engineering simultaneously undertakes multiple functions, and the construction of the engineering data table of each level indicates that the engineering name, the function attribute, the operation characteristic, the key parameters and the at least one corresponding module are respectively collected according to the level to form a structured table with complete fields, uniform caliber and traceability. Through the above processing, the standardized analysis and semantic extraction of the plurality of engineering basic data are completed, and the attribution relationship between the engineering and the module and the key parameters are fixed in the same data structure, thereby providing a consistent, calculable and auditable input carrier for subsequent variable mapping and model assembly based on the engineering data table.

[0046] In the embodiment of the present application, the preferred method can further include: obtaining historical observation data of the water network of each level in the target region; determining an initial value range of the system dynamics model; constructing a posterior distribution of the parameters of the system dynamics model based on the system dynamics model, the initial value range and the historical observation data; sampling in the posterior distribution by a Hamilton Monte Carlo algorithm to determine an error value between a simulation value of the system dynamics model and an observation value in the historical observation data; adjusting the initial value range based on the error value, and repeatedly iterating the system dynamics model until the error value is lower than a preset precision threshold.

[0047] In the embodiments of the present application, the historical observation data refers to the measured time series covering each level of the target region and consistent with the functional side measurement range, which is used for direct comparison with the model output; the initial value range refers to the feasible interval and prior boundary setting of the to-be-calibrated parameters in the system dynamics model, which is used for limiting the parameter search space and constraining the physical rationality; the posterior distribution of the parameters refers to the statistical characterization of the parameter uncertainty under the condition of the given model structure and the historical observation data, which is used for reflecting the matching degree of the parameters and the data; the Hamilton Monte Carlo algorithm refers to the Markov chain Monte Carlo method based on the Hamilton dynamics to realize efficient sampling in high-dimensional parameter space, which is used for obtaining representative samples from the posterior distribution of the parameters; the error value refers to the deviation measurement result constructed by taking the model simulation value and the observation value in the historical observation data as input, which is used for quantitatively evaluating the model fitting degree and serving as the basis for parameter updating; and the preset accuracy threshold refers to the quantitative limit value for determining whether the calibration is up to standard, which is used for triggering the iteration termination and result solidification. Through the above processing, the parameter calibration closed loop is completed, which takes the historical observation data as the benchmark, takes the initial value range as the constraint, takes the posterior distribution of the parameters as the core description, and takes the Hamilton Monte Carlo algorithm as the sampling engine. The parameter set output by the closed loop is stable and reliable when the error value continuously converges and reaches the preset accuracy threshold, so that the system dynamics model is consistent with the real observation and has interpretability, thereby providing a reliable calculation basis for subsequent evaluation and scheme generation based on the model.

[0048] In the embodiments of the present application, the multi-objective optimization model constructed based on the non-dominated sorting genetic algorithm can further include, at the same time or after the construction of the multi-objective optimization model: constraining the multi-objective optimization model based on the upper limit of the engineering investment of the target region, the geographical environment adaptability, and the ecological protection mandatory.

[0049] In the embodiments of the present application, the upper limit of the engineering investment refers to the boundary of the funds allowed for new construction, expansion and reconstruction, and operation transformation within the budget period of the target region, which covers the construction cost and necessary operation and maintenance cost, and is used for limiting the economic feasible domain of the scheme; the geographical environment adaptability refers to the implementable degree and safety margin of the scheme under natural conditions such as topography, river network pattern, and elevation difference, including the accessibility of the transmission and distribution channel, the constructability of the node layout, and the disturbance control to the existing water system, and is used for ensuring the matching of the space and engineering conditions; the ecological protection mandatory refers to the environmental management requirements with rigid constraint effect from the ecological red line, the habitat protection of key species, the water ecological environment control target, and the like, and is used for ensuring that the scheme does not break the ecological bottom line; and the constraint on the multi-objective optimization model refers to incorporating the above boundary conditions into the feasible region definition and judgment rules of the multi-objective optimization model in a calculable limiting form, so that the schemes that do not meet any boundary are excluded or penalized during generation and evaluation. Through the above technical processing, while maintaining the solving efficiency and the comprehensiveness of the target, it can be ensured that the candidate scheme is always in the intersection space of the economic affordability, the geographical implementability, and the ecological tolerability, thereby improving the landability and compliance of the solution set from the source and reducing invalid search.

[0050] In the embodiments of the present application, the multi-objective optimization model can be constructed based on the non-dominated sorting genetic algorithm; the multi-objective optimization model includes a plurality of objective functions, and the plurality of objective functions one-to-one correspond to a plurality of functions expected to be realized by each level.

[0051] In the embodiments of the present application, the non-dominated sorting genetic algorithm is a multi-objective optimization solving method in the evolutionary computation category, which divides levels based on the non-dominated relationship between individuals and maintains the diversity of the population in combination with the congestion degree, and is suitable for simultaneously processing evaluation objectives that conflict or cooperate with each other; the multi-objective optimization model is an optimization problem abstraction composed of a decision variable set, a constraint set and a target set, which is used to quantitatively measure the pros and cons of a scheme under a unified standard; the plurality of objective functions are specific measurement mappings of the target set, which respectively measure measurable indexes on the function side of supply, scheduling, ecology and the like, and serve as numerical basis for evaluating the performance of a scheme in the multi-objective optimization model, and the number of the plurality of objective functions is the number of each level multiplied by the number of the plurality of functions; one-to-one correspondence means that an item-by-item matching relationship is established between the plurality of objective functions and the plurality of functions expected to be realized by each level, so that each function of each level is represented by a unique objective function and there is no missing item or repeated measurement. The establishment of the multi-objective optimization model based on the non-dominated sorting genetic algorithm and the standardized configuration of the target set form a target system and a solving framework that can be directly used for unified evaluation and trade-off analysis, and ensure that the function side demands are completely mapped into the target set that can be optimized.

[0052] By the technical solution, the automation, standardization and intelligent networking optimization process for complex terrain and multi-level water network system are realized. Specifically, by analyzing the three-dimensional water network networking generation instructions of the target region, the multiple functions expected to be realized by each level can be accurately identified, forming a unified function input caliber, providing a logical starting point for multi-level and multi-functional collaboration; the multi-objective optimization model constructed based on the non-dominated sorting genetic algorithm maps the expected functions of each level to corresponding multiple objective functions, so that the optimization process considers multiple objectives such as water supply, flood control and ecology while maintaining solution stability; the dynamic simulation calculation of multiple candidate water network networking schemes is realized through the system dynamics model, which converts the abstract structure scheme into quantifiable original index values, so that the water network scheme has physical interpretability and comparability; the posteriori rate setting of model parameters is completed by combining the Hamilton Monte Carlo algorithm, so that the system dynamics model can accurately reflect the historical observation law and real operation state under different scenarios; through the iterative feedback between the multi-objective optimization model and the system dynamics model, a Pareto non-inferior solution set with balance and effectiveness is generated, so that dynamic balance is realized among multiple objectives; then the final index value is converted into a single comprehensive index value through the dimension reduction decision model, realizing the quantitative unity of complex performance evaluation; finally, the optimal networking scheme is selected as the decision output of the target region according to the comprehensive index value, ensuring that the scheme has the optimal comprehensive benefit under the economic, geographical and ecological constraints; at the same time, the multi-objective optimization model is constrained based on the upper limit of engineering investment, geographical environment adaptability and ecological protection, so that the scheme generation result meets the engineering feasibility and environmental compliance. Through the above technical chain, the whole process closed loop from data analysis, index construction, model calibration to optimization solution and scheme optimization is realized, which significantly improves the scientificity, controllability and optimality of three-dimensional water network networking, and can quickly obtain an optimal three-dimensional water network networking scheme that takes into account functional collaboration and regional constraints under complex terrain and multi-objective conditions.

[0053] The following is an embodiment of the present application:

[0054] 1. Around the three functions of water resource supply and demand balance, flood control safety and ecological protection, a nine-category index system including provincial, municipal and county levels is constructed, forming an initial index set covering multiple levels and multiple functions, providing a quantitative basis for subsequent model construction.

[0055] 1.1, Collect water network data of the study area, including characteristic values of water resource supply and demand balance, flood control safety and ecological protection, and according to the data of provincial, municipal and county levels.

[0056] 1.2, Clearly define the core dimensions of the three functions of water resource supply and demand balance, flood control safety and ecological protection, sort out the differences among provincial, municipal and county levels, and determine the level "granularity" and direction of index design.

[0057] 1.3 Based on the results of 1.2, design indicators by function and level to form a preliminary indicator pool covering multiple levels and functions.

[0058] 1.4. Use factor analysis to screen indicators and integrate and optimize them to form a "three-level nine-category" indicator system.

[0059] 1.4.1 Standardize the data in the initial indicator pool to eliminate the impact of differences in the dimensions of different indicators.

[0060] 1.4.2 Calculate the correlation coefficient matrix between indicators and determine whether the indicators are suitable for factor analysis.

[0061] 1.4.3 Extract common factors and determine the number of factors based on the standard of eigenvalues ​​greater than 1 or the cumulative variance contribution rate threshold.

[0062] 1.4.4 Perform factor rotation to make the factor loading matrix structure clearer and easier to interpret the meaning of each factor.

[0063] 1.4.5. Based on the factor loading size, indicators are selected, and those with strong explanatory power for common factors are retained. The indicators are then integrated and optimized to form a "three-level nine-category" indicator system.

[0064] 2. Based on the above indicator set, a system dynamics model covering water resource supply and demand, flood control scheduling, and ecological response modules is built to dynamically simulate the state changes of indicators at each level under different networking schemes. The model covers provincial backbone projects, municipal connecting projects, and county-level terminal projects, and can output the quantitative results of corresponding indicators.

[0065] 2.1 Based on the indicator set, divide the water resources supply and demand, flood control scheduling, and ecological response modules, and clarify the core variables and relationships of each module.

[0066] 2.2 Connect the three-level projects (provincial backbone, municipal connection, and county terminal) with the corresponding modules to build a system dynamics model.

[0067] 2.2.1 The functional attributes and operational characteristics of provincial backbone projects, municipal connecting projects, and county-level terminal projects are sorted out, and the correspondence between each project and water resource supply and demand, flood control scheduling, and ecological response modules are clarified. At the same time, "elevation difference" is taken as a key parameter of project attributes (for example, the water supply of high-altitude reservoirs needs to consider the water pressure / flow loss caused by the elevation difference during the water conveyance process, and low-altitude projects need to consider the elevation safety threshold for flood control).

[0068] 2.2.2 Establish variable mapping rules between projects and modules, and associate the key parameters (including elevation) of the third-level project with the core variables of the corresponding module to form a dynamic interaction link.

[0069] 2.2.3, Based on the above association and mapping rules, the overall structure of the system dynamics model containing three levels of engineering and three modules is constructed, and the feedback loop and variable equation of the model are determined.

[0070] 2.3, Calibrate the system dynamics model parameters.

[0071] 2.3.1, Collect historical observation data related to the model, covering key variable measured values of three levels of engineering and corresponding modules.

[0072] 2.3.2, Set the initial value range of the parameters, and determine the reasonable interval according to engineering experience, literature data and model boundary conditions.

[0073] 2.3.3, Use Hamiltonian Monte Carlo (HMC) algorithm to compare the model simulation value with the historical observation value. By constructing the posterior distribution of the system dynamics model parameters, HMC is used to sample in the distribution to calculate the error value between the simulation value and the observation value.

[0074] 2.3.4, Adjust the parameter value based on the error feedback, and repeat the iteration until the error between the simulation value and the observation value meets the preset precision threshold.

[0075] 2.3.5, Select independent verification data set to test the stability of the calibrated parameters and the prediction ability of the model, and complete the parameter calibration.

[0076] 3, Based on the non-dominated sorting genetic algorithm NSGA-III, a multi-objective optimization model is constructed, which sets three functions of water resources supply and demand balance, flood control safety and ecological protection at provincial, municipal and county levels, a total of nine objective functions, and sets the upper limit of engineering investment, terrain adaptability (including elevation) and ecological red line and other constraint conditions.

[0077] 3.1, Introduce the core framework of non-dominated sorting genetic algorithm NSGA-III, determine the algorithm parameters and running rules, and according to the corresponding relationship of three levels and three functions, combined with the management authority and regional characteristics of each level, plan the basic structure of nine objective functions and clarify the optimization focus, to ensure the matching with the actual function demand of each level. Specifically: the three levels can refer to the provincial, municipal and county levels; the function corresponds to the objective function, combined with the three levels, a total of nine objective functions.

[0078] 3.1.1, Clarify the three objective functions at the provincial level: first, the reasonable water supply (allocation) to different users by different water sources to minimize the water shortage in the provincial region; second, to reduce the flood peak of important rivers in the provincial region to the greatest extent, combined with the regulation and storage effect of large water conservancy hub projects to reduce flood threat; third, to minimize the pollutant discharge in the provincial region.

[0079] (1) The province has the smallest water shortage:

[0080]

[0081] In the formula, This refers to the amount of water shortage; for j The industry's water demand; for i Water source to j The industry's water supply; I The quantity of water resources in the region (province); J This represents the number of industries within the region.

[0082] (2) The peak reduction rate of all key rivers within the provincial region is maximized:

[0083]

[0084] In the formula, The overall peak reduction rate (effect) within the provincial region; Within the provincial region s The natural peak flow of a river; After water network scheduling within the provincial region s Peak flow of the river; Within the provincial region s The weighting coefficient of a river reflects its importance to the flood control safety of a provincial region; N This refers to the number of key rivers within a provincial region.

[0085] (3) Maximize the ecological benefits of the provincial region, that is, minimize the emissions of pollutants (chemical oxygen demand (COD), ammonia nitrogen (NH3-N), total nitrogen (TN), and total phosphorus (TP)) within the provincial region:

[0086]

[0087] In the formula, This represents the total amount of pollutants emitted. Within the provincial region j Wastewater discharge coefficient in the water-using industry; Within the region j Concentrations of pollutants in the water-using industry; for i Water supply and water use industry j The water supply.

[0088] 3.1.2. Define three objective functions at the municipal level: first, the weighted sum of the fairness target of water resource supply and demand and the water shortage rate target at the municipal level; second, the average peak reduction rate of all cities in the region; and third, the total amount of pollutant emissions of all cities in the region.

[0089] (1) The city-level water resources supply and demand balance objective function calculation formula is:

[0090] a. The sum of all city water shortages is minimized:

[0091]

[0092] wherein, is the water shortage, is the k city j industry water demand; is the k city i water supply to j industry water supply; is the k city water source water quantity; is the k city industry water user quantity; K is the city quantity.

[0093] b. Fairness is optimal:

[0094]

[0095] wherein, is the fairness objective function, reflecting the difference in water shortage rates between different cities. The smaller the difference in water shortage rates between cities, the more fair the water supply result; is the k city water shortage rate; is the average water shortage rate of all cities in the region.

[0096] c. The total objective function of the city-level water resources supply and demand balance:

[0097]

[0098] wherein, is the city-level water resources supply and demand total objective function; is the city-level fairness objective function value; is the city-level fairness objective function weight; is the city-level water shortage rate objective function value; is the city-level water shortage rate objective function weight.

[0099] (2) City-level flood control safety objective function:

[0100]

[0101] wherein, is the average peak shaving effect of all cities in the provincial region; is the k citys The natural peak flow of a river; After water network scheduling k City s Peak flow of the river; for k City s The weighting coefficient of a river reflects its influence on... k The importance of urban flood control safety; for k The city's weighting coefficient reflects k The importance of cities to provincial-level flood control security; for k Number of important rivers within the city.

[0102] (3) Maximize the ecological benefits at the municipal level, that is, minimize the emissions of pollutants (chemical oxygen demand (COD), ammonia nitrogen (NH3-N), total nitrogen (TN), and total phosphorus (TP)) of all cities in the province:

[0103]

[0104] In the formula, This represents the total pollutant emissions from all cities within the provincial region. for k City j Wastewater discharge coefficient in the water-using industry; for k City j Concentrations of pollutants in the water-using industry; for k City i Water source j Water allocation for water-using industries; For the city k The importance of ecological protection in provincial-level regions; and Cities k The number of industries using domestic water and the number of water sources.

[0105] 3.1.3. Define three objective functions at the county (district) level: first, the weighted sum of the fairness objective of water resource supply and demand and the water shortage rate objective at the county (district) level; second, the average peak reduction rate at the county (district) level; and third, the ecological benefits at the county (district) level.

[0106] (1) The formula for calculating the objective function of water resource supply and demand balance at the county (district) level is:

[0107] a. Minimum water shortage:

[0108]

[0109] In the formula, The sum of water shortage of all counties (districts), The sum of water shortage of all counties (districts), l The sum of water shortage of all counties (districts), j The water demand of water-using industries; The sum of water shortage of all counties (districts), l The sum of water shortage of all counties (districts), i The sum of water shortage of all counties (districts), j The sum of water shortage of all counties (districts), The sum of water shortage of all counties (districts), l The sum of water shortage of all counties (districts), The sum of water shortage of all counties (districts), l The sum of water shortage of all counties (districts), L The sum of water shortage of all counties (districts),

[0110] b. Optimal fairness:

[0111]

[0112] In the formula, The fairness objective function reflects the difference in water shortage rates among different counties (districts). The smaller the difference in water shortage rates among different counties (districts), the more fair the water supply result. The water shortage rate of a county (district); l The water shortage rate of a county (district); The average water shortage rate of all counties (districts) in the region. l The average water shortage rate of all counties (districts) in the region.

[0113] c. Total objective function of water resources supply and demand balance at the county level:

[0114]

[0115] In the formula, The total objective function of water resources supply and demand at the county level; The fairness objective function value at the county level; The weight of the fairness objective function at the county level; The water shortage rate objective function value at the county level; The weight of the water shortage rate objective function at the county level.

[0116] (2) County (district) level flood control safety objective function:

[0117]

[0118] In the formula, The average peak reduction rate of a county (district); The natural peak flow of a county (district) river; l The natural peak flow of a county (district) river; s The peak flow of a county (district) river after water network regulation; The peak flow of a county (district) river after water network regulation; l The peak flow of a county (district) river after water network regulation; s The peak flow of a county (district) river after water network regulation; The peak flow of a county (district) river after water network regulation;l County (District) s The weighting coefficient of the river reflects the river's influence on... l The importance of the county (district); for l The weighting coefficient of the county (district) reflects l The importance of counties (districts) to provincial-level flood control security; for l Number of important rivers within the county (district).

[0119] (3) Maximizing county (district) level ecological benefits, that is, minimizing the emissions of pollutants (chemical oxygen demand (COD), ammonia nitrogen (NH3-N), total nitrogen (TN), and total phosphorus (TP)) in all counties (districts) within the provincial region:

[0120]

[0121] In the formula, This represents the total pollutant emissions from all counties (districts) within the province. for l County (District) j Wastewater discharge coefficient in the water-using industry; for l County (District) j Concentrations of pollutants in the water-using industry; for l County (District) i Water source j Water supply volume in water-using industries; for l The importance of counties (districts) to the ecological protection of provincial-level regions.

[0122] 3.2. Review the boundary conditions that need to be followed in the model optimization process, and refine the specific setting rules for constraints such as the upper limit of engineering investment, terrain adaptability and ecological red line from the dimensions of engineering economic feasibility, geographical environment adaptability and ecological protection mandatoryness, so as to form a complete constraint system.

[0123] 3.2.1 Clarify the constraints on the economic feasibility of the project: First, set the total investment ceiling for provincial, municipal and county-level projects, and limit the investment scale of projects at each level in combination with the regional economic development level and fiscal affordability; Second, stipulate the cost threshold for unit projects, and the cost of unit projects shall not exceed the industry benchmark value to ensure the economic efficiency of project investment.

[0124] 3.2.2. Explicitly geographical environment adaptability constraints: First, according to the terrain slope, geological structure and other conditions, set the terrain adaptability standard of engineering construction, limit the corresponding type of engineering construction in the area that does not meet the terrain conditions; Second, combined with hydrological conditions, specify the site selection range of the project, strictly avoid the river flood area, geological disaster prone area layout key facilities; Third, the elevation difference between projects needs to be considered to avoid excessive elevation difference leading to a sharp increase in water resource scheduling energy consumption cost, or affecting the water conveyance feasibility due to improper elevation connection.

[0125] 3.2.3. Explicitly ecological protection constraints: First, delineate the ecological red line range, prohibit the construction of projects that affect ecology in the core area of nature reserves and the first protection area of drinking water sources; Second, set the lower limit of ecological flow to ensure the ecological water of rivers, lakes and other water bodies is guaranteed, and the stability of aquatic habitats is maintained.

[0126] 4. Take the water network engineering layout parameters as decision variables, generate a large number of candidate water network networking schemes through the non-dominated sorting genetic algorithm NSGA-III of the multi-objective optimization model, input the large number of candidate water network networking schemes into the system dynamics model to obtain the original index values, and then use the non-dominated sorting genetic algorithm NSGA-III in the multi-objective optimization model for iterative optimization to finally generate a Pareto non-inferior solution set.

[0127] 4.1. Define the water network engineering layout parameters as decision variables of the multi-objective optimization model, and define the value range and dimension of the parameters.

[0128] 4.1.1. Sort out the core parameters involved in the water network engineering layout, including the line direction of water conveyance, the scale of engineering facilities, the connection mode of nodes, etc., and clearly define the specific content of the decision variables of the multi-objective optimization model.

[0129] 4.1.2. Combined with the functional positioning and regional characteristics of the three-level engineering (provincial backbone, municipal connection, and county terminal), define the value range of each level decision variable to ensure that the parameters meet the actual construction requirements of the project.

[0130] 4.1.3. Divide the dimension of decision variables, distinguish spatial dimension (such as line length, coverage range), scale dimension (such as facility capacity, construction intensity), etc., and form a structured decision variable system.

[0131] 4.2. Generate a large number of candidate water network networking schemes based on decision variables through the non-dominated sorting genetic algorithm NSGA-III of the multi-objective optimization model.

[0132] 4.2.1. Input the water network engineering layout decision variables that have been defined into the non-dominated sorting genetic algorithm NSGA-III of the multi-objective optimization model, and set the key operating parameters such as population size and evolution generation number.

[0133] 4.2.2, Based on the value range of decision variables and dimensions, the initial population is generated by the initialization operation of the non-dominated sorting genetic algorithm NSGA-III, and each individual corresponds to a set of candidate water network networking scheme parameter combinations.

[0134] 4.2.3, According to the selection, crossover, mutation and other evolution operation rules of the non-dominated sorting genetic algorithm NSGA-III, the initial population is iteratively evolved to constantly generate new scheme individuals and expand the number of schemes.

[0135] 4.2.4, Continue to iterate until the preset evolution number or scheme number threshold is reached to form a large set of candidate water network networking schemes covering multiple parameter combinations.

[0136] 4.3, The generated large number of networking schemes are input into the system dynamics model one by one, and the original index values corresponding to each scheme are calculated and output.

[0137] 4.4, The original index values are fed back to the multi-objective optimization model, and the non-dominated sorting genetic algorithm NSGA-III is used for iterative optimization to select better schemes.

[0138] 4.4.1, The original index values of each networking scheme output by the system dynamics model are fed back to the multi-objective optimization model according to the corresponding relationship as the evaluation basis for the iterative optimization of the non-dominated sorting genetic algorithm NSGA-III.

[0139] 4.4.2, The non-dominated sorting genetic algorithm NSGA-III is used to non-dominantly sort the feedback original index values, combined with the congestion calculation, to select better scheme individuals in the current population.

[0140] 4.4.3, Based on the screening results, high-quality scheme individuals are retained, and new scheme individuals are generated through the crossover and mutation operations of the algorithm to form a new generation population.

[0141] 4.4.4, Repeat the above sorting, screening and updating operations until the optimization degree of the schemes in the population tends to be stable or reaches the preset iteration number, and obtain a better scheme set after multiple rounds of optimization.

[0142] (1) Alternative scheme of multi-objective optimization algorithm:

[0143] It should be noted that the above scheme calibrates the model parameters using the non-dominated sorting genetic algorithm NSGA-III and the Hamilton Monte Carlo algorithm. However, these two algorithms can be replaced by each other, or other multi-objective optimization algorithms can be used for replacement, as shown in Table 1.

[0144] Table 1

[0145]

[0146] (2) Alternative solution of using wolf pack algorithm to optimize projection pursuit:

[0147] In the projection pursuit model, the core role of the wolf pack algorithm (WPA) is to search for the optimal projection direction (i.e., find the low-dimensional projection vector that maximizes the projection index) through swarm intelligence, thereby revealing the potential structure of high-dimensional data. Therefore, other optimization algorithms can also be used to optimize the projection pursuit model, such as the particle swarm optimization algorithm, the genetic algorithm (GA) optimization projection pursuit model, the artificial bee colony algorithm (ABC) optimization projection pursuit model, the grey wolf optimization algorithm (GWO) optimization projection pursuit model, and the whale optimization algorithm (WOA) optimization projection pursuit model.

[0148] (3) Alternative solution of factor analysis method for screening indicators:

[0149] In the task of indicator screening (i.e., selecting representative and low-redundancy core indicators from multiple candidate indicators), in addition to the factor analysis method, there are various methods that can be selected according to data characteristics, research goals, and application scenarios. These methods achieve indicator reduction from different perspectives (such as correlation, information quantity, model contribution, etc.), as shown in Table 2.

[0150] Table 2

[0151]

[0152] 4.5. After multiple iterations, the final Pareto non-inferior solution set that meets the optimization goal is generated.

[0153] 5. Extract the final indicator values of each network scheme in the Pareto non-inferior solution set, convert them into single comprehensive indicator values using the dimension reduction decision model, sort the schemes based on these values, and select the scheme with the highest score as the optimal network scheme to achieve the collaborative optimization of multiple levels.

[0154] 5.1. Extract the final indicator values of each network scheme from the Pareto non-inferior solution set, and organize them into the indicator data set to be analyzed.

[0155] 5.2. Construct a wolf pack algorithm optimized projection pursuit model as the dimension reduction decision model, and determine the core parameters and running rules of the model.

[0156] 5.2.1. Build the basic framework of the projection pursuit model, clarify the core components such as the projection direction vector and the projection index function, and determine the basic principles and operation logic of projecting multi-dimensional indicator data into a low-dimensional subspace.

[0157] 5.2.2, introduce the wolf swarm algorithm to optimize the projection pursuit model, set the core parameters of the wolf swarm algorithm, including the wolf swarm size, the wolf exploration step, the wolf attack distance, the wolf attack range, etc., to ensure that the algorithm can efficiently search for the optimal projection direction.

[0158] 5.2.3, formulate the model running rules, clearly define the triggering conditions and iteration mechanisms of the wolf exploration, wolf pursuit and wolf attack in the wolf swarm algorithm, as well as the optimization target and convergence criterion of the projection index function in the projection pursuit model, and form a complete dimension reduction decision model running system.

[0159] 5.3, input the index data set corresponding to each scheme in the non-inferior solution set into the dimension reduction decision model, and convert the multi-dimensional final index value into a single comprehensive index value through model operation.

[0160] 5.4, sort each group of network schemes according to the comprehensive index value, select the scheme with the highest score as the optimal network scheme, and realize the collaborative optimization of multi-level objectives.

[0161] Through the above technical solutions, the optimization method of the three-dimensional water network networking scheme is proposed by focusing on the southwest Yunnan-Guizhou Plateau region, aiming to break through the existing networking limitations, provide technical support for the transformation of water network in the southwest Yunnan-Guizhou Plateau region (Yunnan water network and Guizhou water network) from scattered construction to system collaboration and from single function to comprehensive guarantee, improve the system and resilience of water network, and help regional water safety guarantee and ecological-economic coordinated development.

[0162] The embodiment of the present application also provides an optimization system for a three-dimensional water network networking scheme, which comprises: a generation module, which generates a plurality of candidate water network networking schemes of a target region through a multi-objective optimization model, and inputs each candidate water network networking scheme into a pre-constructed system dynamics model to obtain corresponding original index values; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data and a plurality of functions expected to be realized at each level; a feedback module, which feeds back the original index values corresponding to each candidate water network networking scheme to the multi-objective optimization model, and iteratively optimizes it through a non-dominated sorting genetic algorithm to generate a Pareto non-inferior solution set; a conversion module, which converts the final index value of each candidate water network networking scheme in the Pareto non-inferior solution set into a single comprehensive index value through a dimension reduction decision model; a sorting module, which sorts the plurality of candidate water network networking schemes according to the single comprehensive index value of each candidate water network networking scheme, and selects the candidate water network networking scheme with the highest single comprehensive index value among the single comprehensive index values of the plurality of candidate water network networking schemes as the optimal networking scheme of the target region.

[0163] Figure 2 The structure diagram of a computer device according to the embodiment of the present application is schematically shown. As shown in FIG. 1, the computer device comprises a processor 1001, a memory 1002, a bus 1003, a display module 1004 and an input module 1005. Figure 2As shown, the embodiment of the present application provides a computer device, which can include:

[0164] a memory 210 configured to store instructions; and

[0165] a processor 220 configured to invoke the instructions from the memory 210 and implement the above preferred method when executing the instructions.

[0166] The embodiment of the present application also provides a computer program product, which includes a computer program, and the computer program is executed by the processor to implement the above preferred method.

[0167] The embodiment of the present application also provides a machine readable storage medium, which has instructions stored thereon, and the instructions are used to make the machine execute the above preferred method.

[0168] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can take a form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take a form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage etc.) containing computer-usable program code.

[0169] The present application is described with reference to the flowcharts and / or block diagrams of the method, device (system), computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce the functions specified in the flowcharts and / or block diagrams for implementing the flows and / or blocks in the flowcharts and / or block diagrams. Figure 1 The functions specified in a flow or multiple flows and / or blocks Figure 1 The functions specified in a flow or multiple flows and / or blocks

[0170] These computer program instructions can also be stored in a computer readable memory capable of directing the computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce a product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams for implementing the flows and / or blocks in the flowcharts and / or block diagrams. Figure 1 The functions specified in a flow or multiple flows and / or blocks Figure 1 The functions specified in a flow or multiple flows and / or blocks

[0171] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 Figure 1

[0172] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0173] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the processor can execute instructions. The memory can also include non-volatile memory, such as read only memory (ROM), electrically programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), flash memory, or other memory technologies, about which the processor can execute instructions. The memory is an example of computer readable media.

[0174] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically programmable read only memory (EEPROM), flash memory or other memory technologies, compact disc read only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to computing devices. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0175] It should also be noted that the terms "comprising", "comprises", "including", "includes" or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article or apparatus that comprises the element.

[0176] ​​The above merely provides an example of the present application, and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall fall into the scope of claims of the present application.

Claims

1. A preferred method of stereoscopic water mesh networking scheme, characterized in that, The preferred method comprises: generating a plurality of candidate water network networking schemes of a target region through a multi-objective optimization model, inputting each candidate water network networking scheme into a pre-constructed system dynamics model to obtain corresponding original index values; the system dynamics model is constructed based on water network data of each level of the target region, a plurality of engineering basic data and a plurality of functions expected to be realized at each level; feeding back the original index values corresponding to each candidate water network networking scheme to the multi-objective optimization model, and iteratively optimizing them through a non-dominated sorting genetic algorithm to generate a Pareto non-inferior solution set; transforming the final index values of each candidate water network networking scheme in the Pareto non-inferior solution set into a single comprehensive index value through a dimension reduction decision model; ranking the plurality of candidate water network networking schemes according to the single comprehensive index values of each candidate water network networking scheme, and taking the candidate water network networking scheme with the highest single comprehensive index value among the single comprehensive index values of the plurality of candidate water network networking schemes as the optimal networking scheme of the target region; wherein the step of constructing the system dynamics model comprises: determining a standardized index set of the target region based on the water network data, the plurality of functions expected to be realized at each level and a factor analysis method; the standardized index set contains the correspondence between the levels and the plurality of functions; mapping the plurality of functions into corresponding modules based on the standardized index set, and determining the core variables of each module; constructing the engineering data table of each level based on the plurality of engineering basic data and a preset keyword library; determining the variable mapping rules between all key parameters in the engineering data table of each level and the core variables of all modules corresponding to the level, and forming dynamic interaction links between the key parameters and the core variables; constructing the system dynamics model according to the variable mapping rules and the dynamic interaction links.

2. The preferred method of claim 1, wherein, The step of determining the standardized index set of the target region comprises: in response to a three-dimensional water network networking generation instruction of the target region, parsing a plurality of functions expected to be realized at each level of the target region from the three-dimensional water network networking generation instruction; obtaining water network data and a plurality of engineering basic data of each level of the target region.

3. The preferred method of claim 2, wherein, The step of determining the standardized index set of the target region based on the water network data, the plurality of functions expected to be realized at each level and the factor analysis method comprises: establishing an initial index pool containing the plurality of functions expected to be realized at each level based on the water network data; standardizing the initial index pool and determining the correlation coefficient matrix between a plurality of indexes in the standardized initial index pool; performing eigenvalue decomposition on the correlation coefficient matrix to obtain a component list sorted by eigenvalue size; determining an initial factor loading matrix based on the factor analysis method, the component list and a preset threshold value; performing factor rotation on the initial factor loading matrix to obtain a rotated factor loading matrix; determining the standardized index set based on the factor loading matrix.

4. The preferred method of claim 3, wherein, The step of determining the initial factor loading matrix based on the factor analysis method, the component list and the preset threshold value comprises: determine a first ingredient list according to a characteristic value of each ingredient in the ingredient list and a preset characteristic threshold value; the preset threshold value is the preset characteristic threshold value; construct the initial factor loading matrix based on the first ingredient list.

5. The preferred method of claim 3, wherein, The step of determining the initial factor loading matrix based on the factor analysis method, the ingredient list and the preset threshold value includes: sequentially determine a cumulative variance contribution rate of each ingredient in the ingredient list; in a case where the cumulative variance contribution rate of a current ingredient first reaches or exceeds a preset variance contribution rate threshold value among a plurality of ingredients, determine a serial number of the current ingredient in the ingredient list as a candidate factor number; the preset threshold value is the preset variance contribution rate threshold value; select the candidate factor number of ingredients from the ingredient list in sequence to construct the initial factor loading matrix.

6. The preferred method according to any one of claims 2-5, characterized in that, The step of constructing the engineering data table of each level based on the plurality of engineering basic materials and a preset keyword library includes: parse an engineering name, a functional attribute and a running feature of each engineering basic material from the plurality of engineering basic materials; perform semantic analysis on the each engineering basic material based on the preset keyword library to output a key parameter of the each engineering basic material; determine at least one module corresponding to the each engineering basic material based on the functional attribute and the running feature; construct the engineering data table of each level based on the engineering name, the functional attribute, the running feature, the key parameter and the at least one corresponding module of the plurality of engineering basic materials of each level.

7. The preferred method according to any one of claims 2-5, characterized in that, The method further includes: acquire historical observation data of water networks of each level in the target region; determine an initial value range of the system dynamics model; construct a posterior distribution of parameters of the system dynamics model based on the system dynamics model, the initial value range and the historical observation data; determine an error value between a simulation value of the system dynamics model and an observation value in the historical observation data by sampling in the posterior distribution through a Hamilton Monte Carlo algorithm; adjust the initial value range based on the error value, and repeatedly iterate the system dynamics model until the error value is lower than a preset precision threshold value.

8. A preferred system for a three-dimensional water mesh networking solution, characterized by, include: a generation module that generates a plurality of candidate water network networking schemes of a target region through a multi-objective optimization model, inputs each candidate water network networking scheme into a pre-constructed system dynamics model to obtain a corresponding original index value; the system dynamics model is constructed based on water network data of each level in the target region, a plurality of engineering basic materials and a plurality of functions expected to be realized at each level; a feedback module that feeds back the original index value corresponding to each candidate water network networking scheme to the multi-objective optimization model, and iteratively optimizes it through a non-dominated sorting genetic algorithm to generate a Pareto non-inferior solution set; a conversion module that converts a final index value of each candidate water network networking scheme in the Pareto non-inferior solution set into a single comprehensive index value through a dimension reduction decision model; The sorting module sorts the multiple candidate water network networking schemes according to the single comprehensive index values of the multiple candidate water network networking schemes, and takes the candidate water network networking scheme with the highest single comprehensive index value among the single comprehensive index values of the multiple candidate water network networking schemes as the optimal networking scheme of the target region. The step of constructing the system dynamics model comprises: determining a standardized index set of the target region based on the water network data, the multiple functions expected to be realized at the levels, and a factor analysis method; the standardized index set comprises a correspondence between the levels and the multiple functions; mapping the multiple functions into corresponding multiple modules based on the standardized index set, and determining core variables of each module; constructing an engineering data table of each level based on the multiple engineering base data and a preset keyword library; determining a variable mapping rule between all key parameters in the engineering data table of each level and the core variables of all modules corresponding to the level, and forming a dynamic interaction link between the key parameters and the core variables; constructing the system dynamics model according to the variable mapping rule and the dynamic interaction link.

9. A computer program product comprising a computer program, characterized in that, The computer program, when executed by a processor, implements the preferred method according to any one of claims 1 to 7.

10. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing a machine to perform the preferred method according to any one of claims 1 to 7.

Images