A kind of integrated infiltration storage integration confined space green infrastructure layout collaborative decision method, electronic equipment and storage medium

Abstract

The present application belongs to the technical field of urban rain flood management and sponge city construction, and provides a kind of infiltration storage integrated limited space green infrastructure layout collaborative decision-making method, electronic equipment and storage medium, method includes: data collection and rain flood model construction, green infrastructure type matching, multi-objective optimization model construction and non-dominated solution set generation, optimal green infrastructure layout scheme screening;The present application realizes the decision-making closed loop of GI layout from multi-objective optimization to optimal scheme landing implementation by introducing infiltration storage integrated green infrastructure system into the GI optimization layout of urban limited space and the deep integration of multi-objective optimization process and entropy weight-TOPSIS decision-making method, improves the reduction efficiency of total runoff and pollution load and the feasibility and implementation of layout scheme, ensures that rain flood management effect meets the standard.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of urban rain flood management and sponge city construction, and particularly relates to a seepage-retention-integrated limited space green infrastructure layout collaborative decision-making method, an electronic device and a storage medium. BACKGROUND

[0002] The rain flood management in the limited space of the current high-density built-up area is faced with multiple constraints such as the shortage of available land, the high complexity of runoff and pollution control, and the need to consider landscape coordination and economy. The traditional gray infrastructure reconstruction is difficult to implement in the limited space, while the green infrastructure has multiple benefits such as runoff reduction and pollution control, but also faces several core bottlenecks. First, the functional limitation, single green infrastructure is limited to a single function, and it is difficult to meet the needs of seepage, retention and storage under complex underlying surface conditions, often leading to substandard runoff total and pollution load reduction effect. Second, the lack of spatial adaptation, the available space in the highly built-up area is fragmented, and there is a mismatch between the fragmented available space and the discrete facility layout mode. The existing layout method does not fully consider the adaptability of facility function integration and land constraints. Third, the perception lag of the optimization process, the termination criteria of the algorithm in the multi-objective optimization process of green infrastructure spatial layout mainly depend on passive threshold judgment, which can only reflect the historical state of the algorithm and cannot prospectively perceive and respond to the convergence state of the algorithm. Finally, the broken layout decision chain, although different types of multi-objective optimization algorithms can optimize multiple objectives such as runoff control, pollution reduction and economic investment, they can only output the optimized non-dominated solution set and lack quantitative decision-making mechanisms and methods.

[0003] Although multi-objective optimization algorithms are currently applied to the optimal layout of green infrastructure, the above bottlenecks result in the lack of systematic solutions for the synergy of seepage-retention-storage functions and spatial constraints in the existing technology and method. In the optimization layout process, it is difficult to dynamically respond to the convergence state of the algorithm, and none of the non-dominated solution set decision-making closed-loop problems can be solved. Therefore, the project is forced to rely on artificial experience decision-making, which seriously restricts the scientificity and effectiveness of the rain flood management in the limited space. SUMMARY

[0004] In order to overcome the deficiencies of the prior art, the present application aims to provide a seepage-retention-integrated limited space green infrastructure layout collaborative decision-making method, an electronic device and a storage medium, which solves the problems of single functional limitation, poor matching precision, only reflecting the historical state of the algorithm and only outputting the non-dominated solution set in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following solutions.

[0006] A seepage-retention-integrated limited space green infrastructure layout collaborative decision-making method, comprising:

[0007] collecting basic information of a research area, generalizing sub- catchments and a rainwater pipe network of the research area by a SWMM software according to the basic information, setting model parameters, and obtaining a rain flood model; the basic information includes: terrain data, rainwater pipe network information, land use type data, and historical measured rainfall runoff data;

[0008] defining a seepage-storage integrated green infrastructure system for a city limited space, matching a seepage-storage integrated green infrastructure type of the seepage-storage integrated green infrastructure system to a surface distribution and characteristics of the research area, and obtaining a green infrastructure type matching result;

[0009] coupling a non-dominated sorting genetic algorithm with the rain flood model, taking a layout area corresponding to the green infrastructure type matching result as a decision variable, taking a total runoff reduction rate, a pollution load reduction rate, and a life cycle cost as objective functions, setting area constraints and effect constraints, obtaining a multi-objective optimization model, and performing multi- return period rainfall scenario optimization calculation on the multi-objective optimization model to obtain a non-dominated solution set;

[0010] evaluating and screening the non-dominated solution set by using an entropy weight- superior and inferior solution distance method to obtain an optimal green infrastructure layout scheme.

[0011] Preferably, basic information of a research area is collected, sub- catchments and a rainwater pipe network of the research area are generalized by a SWMM software according to the basic information, model parameters are set, and a rain flood model is obtained, including:

[0012] terrain data, rainwater pipe network information, land use type data, and historical measured rainfall runoff data of the research area are collected to obtain the basic information;

[0013] sub- catchments are divided by using a comprehensive division method, the sub- catchments and rainwater pipe network data are imported into SWMM for model generalization by a GIS platform, and the rain flood model is obtained;

[0014] catchment parameters and structure parameters of seepage-storage integrated green infrastructure are set in the rain flood model; the catchment parameters include: characteristic width, average slope, impervious rate, Manning coefficient, depression storage, and Horton infiltration coefficient.

[0015] Preferably, a seepage-storage integrated green infrastructure system for a city limited space is defined, a seepage-storage integrated green infrastructure type of the seepage-storage integrated green infrastructure system is matched to a surface distribution and characteristics of the research area, and a green infrastructure type matching result is obtained, including:

[0016] The integrated infiltration, retention, and storage green infrastructure system is defined; the integrated infiltration, retention, and storage green infrastructure system includes: infiltration facilities, retention facilities, and storage facilities; the infiltration facilities include: permeable pavement and green roofs; the retention facilities include: rain gardens, constructed wetlands, and bioretention ponds; the storage facilities include: bioretention ponds, rainwater ponds, and green roofs;

[0017] The ordinary roads, squares, and parking lots in the study area will be transformed into the permeable pavement, and the roofs in the study area will be transformed into the green roofs.

[0018] The retention facilities are used to slow down the runoff accumulation process in the study area;

[0019] When surface runoff cannot be discharged, the aforementioned storage facilities are used to store the runoff in the study area.

[0020] Preferably, the non-dominated solution set is evaluated and screened using the entropy weight-superiority solution distance method to obtain the optimal green infrastructure deployment scheme, including:

[0021] Construct an initial indicator matrix; the initial indicator matrix includes: the total runoff reduction rate, the pollution load reduction rate, and the life cycle cost;

[0022] The initial index matrix is ​​sequentially subjected to index weight calculation, information entropy calculation, and weight calculation to obtain the matrix weight; the expression for the matrix weight is: ;in, ; ;in, For the first The final weight of each indicator; For the first Information entropy of each indicator; The total number of evaluation indicators; For the first The first scheme is in the The proportion under each indicator; This represents the total number of solutions in the non-dominated solution set. This represents the j-th index value of the i-th scheme after dimensionless standardization.

[0023] The initial index matrix is ​​weighted using the matrix weights to obtain a weighted decision matrix.

[0024] The positive ideal solution and the negative ideal solution are determined based on the weighted decision matrix.

[0025] The proximity is calculated based on the distances of each solution in the non-dominated solution set from the positive ideal solution and the negative ideal solution, yielding a relative proximity. The expression for the relative proximity is: ; wherein, is a relative closeness degree of the th scheme; is a Euclidean distance of the th scheme to the negative ideal solution; is a Euclidean distance of the th scheme to the positive ideal solution;

[0026] determining the scheme with the largest relative closeness degree as the optimal green infrastructure layout scheme.

[0027] Preferably, the non-dominated sorting genetic improvement algorithm is coupled with the rainwater model, the layout area corresponding to the green infrastructure type matching result is taken as a decision variable, the runoff total reduction rate, the pollution load reduction rate and the life cycle cost are taken as objective functions, the area constraint and the effect constraint condition are set, a multi-objective optimization model is obtained, and the multi-objective optimization model is subjected to multiple return period rainfall scenario optimization calculation to obtain a non-dominated solution set, including:

[0028] constructing an original non-dominated sorting genetic algorithm;

[0029] replacing the termination criterion of the generation improvement rate determination of the hypervolume index in the original non-dominated sorting genetic algorithm with an intelligent termination criterion of a multi-objective evolutionary algorithm based on hypervolume foresight prediction to obtain the non-dominated sorting genetic improvement algorithm;

[0030] setting the layout area of each type of green infrastructure in each sub-converging water area as a decision variable;

[0031] defining a general objective function according to the decision variable; the expression of the general objective function is: ; wherein, ; ; ; ; is a general objective function vector; , , are the runoff total reduction rate, the pollution load reduction rate and the life cycle cost, respectively; is the runoff total amount without GI layout; is the runoff total amount after GI layout; is the pollution load amount without GI layout; is the pollution load amount after GI layout; is the cost of a sub-converging water area; is the dimension of the decision variable; is the layout area; is the construction cost; is the operation and maintenance cost; is a growth rate coefficient; is a service life;

[0032] setting area constraints and effect constraints; the area constraints include: a proportion of the permeable pavement layout area to the total road area, a proportion of the bioretention pool to the total green area, a proportion of the green roof to the total roof area; the effect constraints include: a lower limit of the total runoff reduction rate, a lower limit of the pollution load reduction rate;

[0033] fusing the non-dominated sorting genetic improvement algorithm, the overall objective function, the area constraints, and the effect constraints to obtain the multi-objective optimization model;

[0034] performing a combined facility scenario optimization calculation on the rainwater model according to the multi-objective optimization model by using an NSGA-II or NSGA-III algorithm to obtain the non-dominated solution set.

[0035] Preferably, the bioretention pool is laid on the green land in the study area; the green roof is laid on the roof in the study area; the permeable pavement is laid on the road and open space in the study area; the parameter setting of the bioretention pool includes: a dam height of 100 mm, a water storage layer thickness of 100 mm, and a soil layer thickness of 400 mm; the parameter setting of the green roof includes: a dam height of 150 mm, a soil layer thickness of 100 mm, and a drainage cushion layer thickness of 100 mm; the parameter setting of the permeable pavement includes: a surface layer dam height of 20 mm, a road surface layer thickness of 150 mm, and a water storage layer thickness of 150 mm.

[0036] Preferably, a termination criterion of a generation improvement rate determination of a hypervolume indicator in the original non-dominated sorting genetic algorithm is replaced by an intelligent termination criterion of a multi-objective evolutionary algorithm based on hypervolume forward prediction to obtain the non-dominated sorting genetic improvement algorithm, including:

[0037] collecting hypervolume indicator data;

[0038] when the algorithm runs to the t th generation, data in the historical hypervolume indicator data within a range of less than or equal to W generations from the t th generation iteration are set as an observation window;

[0039] a linear regression model is trained by using data in the observation window to obtain a linear regression hypervolume prediction model;

[0040] a future indicator value after N generations is predicted by using the linear regression hypervolume prediction model, and a prediction benefit is calculated by using the indicator value after the N generations to obtain a prediction relative benefit; an expression of the prediction relative benefit is: ; wherein, the predicted relative benefit; the index value after the N generations; the current index value;

[0041] a minimum start generation number is set, the predicted relative benefit is calculated when the iteration number of the algorithm is greater than the minimum start generation number, and the iteration is stopped when the predicted relative benefit is less than a set threshold.

[0042] Preferably, an electronic device comprises at least one processor and a memory connected in communication with the processor, wherein the memory stores instructions executable by the processor, and the instructions are executed by the processor to enable the processor to perform the aforementioned integrated stagnation and integration limited space green infrastructure layout collaborative decision-making method.

[0043] Preferably, a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the aforementioned integrated stagnation and integration limited space green infrastructure layout collaborative decision-making method.

[0044] The present application discloses the following technical effects:

[0045] The present application provides an integrated stagnation and integration limited space green infrastructure layout collaborative decision-making method, an electronic device and a storage medium, which solves the problem of single function limitation of traditional green infrastructure by integrating stagnation and integration design, realizes multifunctional collaborative optimization under complex underlying surface conditions, solves the problem of spatial fragmentation in highly developed areas through facility function integration and dynamic adaptation method of discrete space layout, realizes efficient use of limited land resources, improves multi-objective optimization algorithm and establishes a regression model to realize active prediction and intelligent decision-making of the optimization process, solves the defect that the termination criterion of the existing multi-objective optimization algorithm depends on passive threshold judgment, realizes forward-looking intelligent decision-making, and solves the problem that the existing technology can only output a non-dominated solution set, realizes a decision-making closed loop. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0047] Figure 1 The integrated stagnation and integration limited space green infrastructure layout collaborative decision-making process schematic diagram provided by the embodiments of the present application;

[0048] Figure 2 A flow chart of a percolation storage integrated limited space green infrastructure layout collaborative decision-making method is provided for the embodiment of the present application.

[0049] Figure 3 An entropy weight-ideal and poor solution distance method decision result graph is provided for the embodiment of the present application. DETAILED DESCRIPTION

[0050] 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. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0051] The present application aims to provide a percolation storage integrated limited space green infrastructure layout collaborative decision-making method, electronic equipment and storage medium, to solve the problems of single function limitation, poor matching precision, only reflecting the historical state of the algorithm and only outputting the non-dominated solution set in the prior art.

[0052] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0053] Figure 1 A flow chart of a percolation storage integrated limited space green infrastructure layout collaborative decision-making method is provided for the embodiment of the present application. Figure 1 As shown in the flow chart of the percolation storage integrated limited space green infrastructure layout collaborative decision-making method provided by the present application, the present application provides a percolation storage integrated limited space green infrastructure layout collaborative decision-making method, which comprises:

[0054] Step 100: Collecting basic data of a research area, generalizing sub- catchment areas and rainwater pipe networks of the research area by SWMM software according to the basic data, setting model parameters, and obtaining a rain flood model; the basic data includes topographic data, rainwater pipe network information, land use type data and historical measured rainfall runoff data;

[0055] Step 200: Defining a percolation storage integrated green infrastructure system for urban limited space, matching the percolation storage integrated green infrastructure type of the percolation storage integrated green infrastructure system according to the underlying surface distribution and characteristics of the research area, and obtaining a green infrastructure type matching result;

[0056] Step 300: coupling the non-dominated sorting genetic improvement algorithm with the stormwater model, taking the layout area corresponding to the green infrastructure type matching result as the decision variable, taking the total runoff reduction rate, pollution load reduction rate and life cycle cost as the objective function, setting the area constraint and effect constraint condition, obtaining a multi-objective optimization model, and performing multi-return period rainfall scenario optimization calculation on the multi-objective optimization model to obtain a non-dominated solution set;

[0057] Step 400: evaluating and screening the non-dominated solution set by using the entropy weight-advantage and disadvantage solution distance method to obtain an optimal green infrastructure layout scheme.

[0058] Further, basic data of a research area is collected, and sub-catchment areas and rainwater pipe networks of the research area are generalized by SWMM software according to the basic data, model parameters are set, and a stormwater model is obtained, including:

[0059] The basic data is obtained by collecting topographic data, rainwater pipe network information, land use type data and historical measured rainfall runoff data of the research area;

[0060] The sub-catchment areas are divided by using a comprehensive division method, and the sub-catchment areas and rainwater pipe network data are imported into SWMM for model generalization by using a GIS platform to obtain the stormwater model;

[0061] In the stormwater model, catchment area parameters and structure parameters of the infiltration and storage integrated green infrastructure are set. The catchment area parameters include characteristic width, average slope, impervious rate, Manning coefficient, depression storage and Horton infiltration coefficient.

[0062] Specifically, an infiltration and storage integrated green infrastructure system oriented to a city limited space is defined, and an infiltration and storage integrated green infrastructure type of the infiltration and storage integrated green infrastructure system is matched according to the distribution and characteristics of the underlying surface of the research area to obtain a green infrastructure type matching result, including:

[0063] The infiltration and storage integrated green infrastructure system is set. The infiltration and storage integrated green infrastructure system includes infiltration facilities, retention facilities and storage facilities. The infiltration facilities include permeable pavement and green roof. The retention facilities include rainwater garden, artificial wetland and biological retention pond. The storage facilities include biological retention pond, rainwater pond and green roof.

[0064] The ordinary roads, squares and parking lots in the research area are transformed into the permeable pavement, and the roofs in the research area are transformed into the green roof.

[0065] The retention facilities are used to delay the collection process of runoff in the research area.

[0066] When surface runoff cannot be discharged, the aforementioned storage facilities are used to store the runoff in the study area.

[0067] Furthermore, the non-dominated solution set is evaluated and screened using the entropy weight-superiority solution distance method to obtain the optimal green infrastructure deployment scheme, including:

[0068] Construct an initial indicator matrix; the initial indicator matrix includes: the total runoff reduction rate, the pollution load reduction rate, and the life cycle cost;

[0069] The initial index matrix is ​​sequentially subjected to index weight calculation, information entropy calculation, and weight calculation to obtain the matrix weight; the expression for the matrix weight is: ;in, ; ;in, For the first The final weight of each indicator; For the first Information entropy of each indicator; The total number of evaluation indicators; For the first The first scheme is in the The proportion under each indicator; This represents the total number of solutions in the non-dominated solution set. This represents the j-th index value of the i-th scheme after dimensionless standardization.

[0070] The initial index matrix is ​​weighted using the matrix weights to obtain a weighted decision matrix.

[0071] The positive ideal solution and the negative ideal solution are determined based on the weighted decision matrix.

[0072] The proximity is calculated based on the distances of each solution in the non-dominated solution set from the positive ideal solution and the negative ideal solution, yielding a relative proximity. The expression for the relative proximity is: ;in, For the first The relative similarity of the two solutions; For the first The Euclidean distance from each solution to the negative ideal solution; For the first The Euclidean distance from each solution to the ideal solution;

[0073] The scheme with the highest relative proximity is determined as the optimal green infrastructure deployment scheme.

[0074] Specifically, the non-dominated sorting genetic improvement algorithm is coupled with the rain flood model, a layout area corresponding to the green infrastructure type matching result is taken as a decision variable, a runoff total amount reduction rate, a pollution load reduction rate and a life cycle cost are taken as objective functions, an area constraint and an effect constraint condition are set, a multi-objective optimization model is obtained, and the multi-objective optimization model is subjected to multiple return period rainfall scenario optimization calculation, so that a non-dominated solution set is obtained, including:

[0075] An original non-dominated sorting genetic algorithm is constructed;

[0076] A termination criterion of a generation improvement rate determination of a hypervolume index in the original non-dominated sorting genetic algorithm is replaced by an intelligent termination criterion of a multi-objective evolutionary algorithm based on hypervolume forward prediction, so that the non-dominated sorting genetic improvement algorithm is obtained;

[0077] Layout areas of various green infrastructures in various sub-converging water areas are taken as decision variables;

[0078] A general objective function is defined according to the decision variables; an expression of the general objective function is: ; wherein, ; ; ; ; is a general objective function vector; 、 、 are the runoff total amount reduction rate, the pollution load reduction rate and the life cycle cost respectively; is a runoff total amount without GI layout; is a runoff total amount after GI layout; is a pollution load amount without GI layout; is a pollution load amount after GI layout; is a cost of a sub-converging water area; is a dimension of a decision variable; is a layout area; is a construction cost; is an operation and maintenance cost; is a growth rate coefficient; is a service life;

[0079] An area constraint and an effect constraint are set; the area constraint includes a proportion of the permeable pavement layout area to the total road area, a proportion of the bioretention pool to the total green area, and a proportion of the green roof to the total roof area; the effect constraint includes a lower limit of the runoff total amount reduction rate and a lower limit of the pollution load reduction rate;

[0080] fuse the non-dominant sorting genetic improvement algorithm, the overall objective function, the area constraint, and the effect constraint to obtain the multi-objective optimization model;

[0081] According to the multi-objective optimization model, NSGA-II or NSGA-III algorithm is used for combined facility scenario optimization calculation of the rain flood model to obtain the non-dominant solution set.

[0082] Preferably, the bioretention pool is arranged on the green land of the research area; the green roof is arranged on the roof of the research area; the permeable pavement is arranged on the road and open space of the research area; the parameter setting of the bioretention pool includes: a dam height of 100 mm, a water storage layer thickness of 100 mm, and a soil layer thickness of 400 mm; the parameter setting of the green roof includes: a dam height of 150 mm, a soil layer thickness of 100 mm, and a drainage cushion layer thickness of 100 mm; the parameter setting of the permeable pavement includes: a surface layer dam height of 20 mm, a road surface layer thickness of 150 mm, and a water storage layer thickness of 150 mm.

[0083] Further, the termination criterion of the generation improvement rate determination of the hypervolume index in the original non-dominant sorting genetic algorithm is replaced by an intelligent termination criterion of a multi-objective evolutionary algorithm based on hypervolume forward prediction to obtain the non-dominant sorting genetic improvement algorithm, including:

[0084] Collecting hypervolume index data;

[0085] When the algorithm runs to the tth generation, data in the historical hypervolume index data within a range of less than or equal to W generations from the tth generation iteration are set as an observation window;

[0086] Training a linear regression model using data in the observation window to obtain a linear regression hypervolume prediction model;

[0087] Predicting an index value after N generations in the future using the linear regression hypervolume prediction model, and calculating a prediction benefit using the index value after N generations to obtain a prediction relative benefit; the expression of the prediction relative benefit is: ; wherein, is the prediction relative benefit; is the index value after N generations; is the current index value;

[0088] Setting a minimum start generation number, when the iteration number of the algorithm is greater than the minimum start generation number, calculating the prediction relative benefit, and stopping iteration when the prediction relative benefit is less than a set threshold.

[0089] Specifically, in view of the problems and deficiencies of the prior art, the embodiment provides a collaborative decision-making method for urban green infrastructure layout, which realizes optimal configuration of green infrastructure "cost-effectiveness" through construction of an integrated framework of "rain flood model-multi-objective optimization-comprehensive evaluation". The method combines physical mechanism model and data-driven algorithm to solve the problems of facility function limitation, spatial adaptation deficiency, perception lag and broken decision chain in the existing GI layout, and specifically includes the following core steps:

[0090] S1, constructing a rain flood model of the research area: collecting basic data of the research area, generalizing the subcatchment and rainwater pipe network of the research area based on the SWMM software, setting model parameters, and using measured rainfall-runoff data for calibration and verification to establish a reliable rain flood model;

[0091] S2, matching green infrastructure types: defining an integrated green infrastructure system for urban limited space, matching the integrated green infrastructure types according to the distribution and characteristics of the underlying surface of the research area;

[0092] S3, coupling multi-objective optimization algorithm and rain flood model: improving the termination criterion of traditional non-dominated sorting genetic algorithm, coupling the improved non-dominated sorting genetic algorithm with the rain flood model, taking the layout area of green infrastructure as the decision variable, taking the total runoff reduction rate, pollution load reduction rate and life cycle cost as the objective function, setting area constraints and effect constraints, constructing a multi-objective optimization model, and obtaining a non-dominated solution set through multiple return period rainfall scenario optimization calculation;

[0093] S4, comprehensive evaluation of green infrastructure scheme and determination of optimal scheme: using entropy weight-ideal distance method, comprehensively considering the index performance under different return periods, comprehensively evaluating the non-dominated solution set, and selecting the optimal green infrastructure layout scheme.

[0094] Further, step S1 is to construct a rain flood model of the research area, and the specific process includes:

[0095] Collect high-precision terrain data, rainwater pipe network information, land use type data and historical measured rainfall runoff data of the research area;

[0096] Subcatchment is divided by using comprehensive division method, and subcatchment and rainwater pipe network data are imported into SWMM based on GIS platform to complete model generalization;

[0097] Set subcatchment parameters and GI parameters, the subcatchment parameters include characteristic width, average slope, impervious rate, Manning coefficient, depression storage, Horton infiltration coefficient, and structure parameters of integrated green infrastructure;

[0098] The model parameters are calibrated and verified by using the measured rainfall-runoff data, and the Nash efficiency coefficient is used to evaluate the simulation accuracy to ensure the reliability of the model.

[0099] Specifically, step S2 is a matching of green infrastructure types, specifically including:

[0100] The available space for GI reconstruction in high-density built-up areas is limited. Therefore, an infiltration retention and storage integrated green infrastructure (IRS-GI) system is defined, which includes green infrastructure that typically includes infiltration, retention, and storage;

[0101] Impervious surfaces such as roads and roofs often account for a large proportion in high-density built-up areas. By infiltrating ordinary roads, squares, and parking lots into permeable pavements, or by transforming roofs into green roofs, some impervious surfaces can be effectively converted into permeable surfaces, thereby increasing the amount of infiltration of rainwater and reducing surface runoff. Common infiltration measures include permeable pavement and green roof;

[0102] A high proportion of impervious surfaces accelerates the surface runoff speed and directly discharges pollutants in rainwater into water bodies. By temporarily storing runoff through retention facilities, the runoff collection process is delayed and runoff pollution is reduced. Common retention measures include rain gardens, constructed wetlands, and bioretention ponds;

[0103] In highly urbanized areas, the drainage system often struggles to effectively cope with rainfall events with high recurrence periods. Surface runoff cannot be discharged in time. By building storage green infrastructure, excess runoff can be stored, reducing the impact on the urban drainage system and reducing pollution of receiving water bodies. Common storage measures include bioretention ponds, rainwater ponds, and green roofs;

[0104] In combination with the existing urban structure and characteristics of the study area, IRS-GI measures are flexibly selected and combined to reconstruct and integrate restricted spaces and underlying surfaces, effectively control urban runoff and non-point source pollution, and also help to improve the urban landscape.

[0105] Further, step S3 is the improvement of multi-objective optimization algorithms and their coupling with the stormwater model, specifically including:

[0106] First, the overall calculation process of the algorithm: (1) First, randomly generate a certain number of individuals as the initial population, and calculate the target vector of each individual in the initial population; (2) divide the individuals in the population into different non-dominated levels, and perform fast non-dominated sorting; (3) calculate the crowding distance of different individuals; (4) generate a new generation of individuals through selection, crossover and mutation; (5) after generating new individuals, repeat steps (2) to (4), when the termination condition is reached, the iteration is terminated and all non-dominated individuals in the algorithm are output, forming a non-dominated solution set. In this process, the population size is set to 3 times the number of features, ensuring that the non-dominated sorting genetic algorithm can cover and search the solution space as much as possible, while also ensuring the computational efficiency of the algorithm.

[0107] Second, the improvement of the termination criterion: an intelligent termination criterion for multi-objective evolutionary algorithm based on hyper-volume forward prediction is proposed. Unlike the conventional termination criterion that uses the generation improvement rate of the hyper-volume indicator, the improved termination criterion converts the algorithm termination problem into a prediction problem, improving the forward-looking, efficiency and adaptability of the multi-objective optimization algorithm. The main calculation steps of the improved termination criterion are:

[0108] (1) Hyper-volume indicator data collection:

[0109] At a certain generation t of the non-dominated sorting algorithm, calculate and record the key performance indicator hyper-volume HV t , forming a data sequence arranged in chronological order: HV1, HV2, …, HV t .

[0110] (2) Define the time series window:

[0111] When the algorithm runs to the tth generation, extract the data of the last W generations from the historical data as the observation window. According to the number of optimization variables, W can be taken as 20 to 50.

[0112] (3) Model training and prediction:

[0113] Use the data in the window to train a lightweight model. Since the hyper-volume indicator data is small and needs to be trained quickly online, the model must be as simple as possible. A linear regression-based model is established to predict the hyper-volume, and the trained model is used to predict the indicator value HV t+N after N generations.

[0114] (4) Formulate termination rules:

[0115] Calculate the predicted relative return based on the linear regression hyper-volume prediction model:

[0116]

[0117] A threshold value ε = 0.1% is set. If Gain < ε, the prediction indicates that the improvement of solution set is not obvious in the future N generations, and it is not worth continuing to run. The algorithm terminates. Otherwise, continue iteration. To ensure the robustness of the prediction, the value of prediction step N needs to balance between short-sightedness (N is too small) and uncertainty (N is too large). The strategy of N = W / 2 is adopted to ensure that the prediction is based on sufficient historical trends, while avoiding overly aggressive long-term extrapolation.

[0118] (5) The starting time of the improved termination criterion:

[0119] Because the initial results of multi-objective optimization algorithms change dramatically, the improved termination criterion does not start predicting from the first generation. By setting the minimum starting generation t min , only when the current generation t > t min , the above prediction termination process is activated. t min can be set to a certain proportion (such as 20%) of the total maximum iteration number or a fixed value (such as 100 generations) to ensure that the population has been preliminarily converged.

[0120] The layout area of each type of green infrastructure in each sub-catchment is taken as the decision variable;

[0121] The objective function is defined as follows:

[0122] Total runoff reduction rate:

[0123]

[0124] Wherein is the total runoff without GI layout, is the total runoff after layout;

[0125] Pollution load reduction rate:

[0126]

[0127] Wherein is the pollution load without GI layout, is the pollution load after layout;

[0128] Life cycle cost:

[0129]

[0130]

[0131] is the layout area, is the construction cost, is the operation and maintenance cost, is the growth rate coefficient, For the use of life;

[0132] Multi-objective collaborative optimization:

[0133] Among the three objectives, the greater the runoff reduction rate and pollution load reduction rate, the more significant the program effect; the smaller the life cycle cost, the more economically feasible the program. By taking the negative value of the first two objectives to achieve maximum conversion, the three objectives are unified into a minimization problem, and the overall objective function of the multi-objective optimization algorithm is:

[0134]

[0135] Set the constraint condition: area constraint: integrated green infrastructure system with infiltration and storage, represented by permeable pavement, bioretention, and green roof, and determine its area constraint. The area of permeable pavement is the proportion of the total area of the road, which can generally be set at 20% to 90%; the bioretention is the proportion of the total area of the green land, which can generally be set at 20% to 90%; the green roof is the proportion of the total area of the roof, which can generally be set at 20% to 90%;

[0136] Effect constraint: runoff reduction rate generally ≥ 25%, pollution load reduction rate generally ≥ 30%;

[0137] Improved multi-objective optimization algorithm coupled with rainwater model:

[0138] NSGA-Ⅱ or NSGA-Ⅲ algorithm is adopted and improved, coupled with the urban rainwater model, to optimize the calculation of the combined facility scenario, and the specific process is as follows:

[0139] First, use the urban rainwater model to simulate the runoff total , pollution load of the study area without green infrastructure.

[0140] Next, input the area of green infrastructure such as bioretention, green roof, and permeable pavement in each sub-catchment area in the urban rainwater model as the decision variable of the multi-objective optimization algorithm, and calculate the life cycle cost f3 under this layout scheme according to the life cycle cost calculation formula of each green infrastructure.

[0141] Then, use the model to simulate and calculate the runoff total , pollution load under the existing green infrastructure layout, and combine the runoff total , pollution load of the study area without LID facilities to calculate the runoff reduction rate f1 and pollution load reduction rate f2. Input the optimization objectives f1, f2, and f3 as the objective function of the improved optimization algorithm, and set the area and effect constraints.

[0142] Finally, set the parameters required for the improved optimization algorithm, get the green infrastructure optimization layout results, including the area of each solution in the non-dominated solution set in different sub-catchment and f1, f2, f3, store and draw the results. The design storm can use 1, 2, 3, 5, 10-year return period, and the rainfall process is generated based on the rainfall intensity formula of the study area and the Chicago rain type.

[0143] Preferably, in step S4, the green infrastructure scheme comprehensive evaluation and optimal scheme determination specifically includes: entropy weight method to calculate the weight:

[0144] Standardized index matrix;

[0145] Calculate the proportion of indicators:

[0146]

[0147] Calculate information entropy:

[0148]

[0149] Determine the weight:

[0150]

[0151] Calculate the closeness degree by the superior and inferior solution distance method:

[0152] Construct a weighted decision matrix;

[0153] Determine the positive and negative ideal solutions;

[0154] Calculate the distance of each scheme to the positive and negative ideal solutions ;

[0155] Calculate the relative closeness:

[0156]

[0157] Select the scheme with the largest closeness degree as the optimal scheme.

[0158] Optionally, when constructing the rainwater model in step S1, the system needs to collect 1:500 high-precision topographic data, rainwater pipe network distribution map, 10m precision land use type data and historical measured rainfall runoff data of the study area, and use the comprehensive division method combined with the GIS platform to generalize the sub-catchment. Specifically, first manually preliminarily divide the sub-catchment according to the terrain, building and road direction, take the pipe network starting point, turning point and every 400m to 600m inspection well as the node, generate the Thiessen polygon through ArcGIS and manually adjust. When setting the model parameters, the characteristic width of the sub-catchment is calculated according to the following formula:

[0159]

[0160] wherein is the target width, is the height.

[0161] The average slope is extracted by DEM digital elevation model, and the impervious rate is calculated by weighting the water body 0%, green land 5%, roof 75%, open space and square 80%, and road 90%. The Horton infiltration equation is used for model calibration:

[0162]

[0163] wherein is the target function value at t time, the initial infiltration rate , the stable infiltration rate , the attenuation coefficient After the measured flow data is calibrated, the Nash efficiency coefficient is calculated to ensure the reliability of the model.

[0164] Preferably, the matched green infrastructure parameters in step S2 refer to the Technical Guidelines for Sponge City Construction, wherein the bio-retention pool has a berm height of 100 mm, a water storage layer thickness of 100 mm, and a soil layer thickness of 400 mm; the green roof has a berm height of 150 mm, a soil layer thickness of 100 mm, and a drainage cushion layer thickness of 100 mm; and the permeable pavement surface has a berm height of 20 mm, a road surface layer thickness of 150 mm, and a water storage layer thickness of 150 mm.

[0165] Optionally, when the multi-objective optimization model is coupled with the rainwater model in step S3, the layout area of the bio-retention pool, the green roof, and the permeable pavement in the n sub-catchment areas is taken as the 3n-dimensional decision variable, and the layout area of each facility is constrained by the land use type: the permeable pavement is 20% to 90% of the total road area, the bio-retention pool is 20% to 90% of the total green area, and the green roof is 20% to 90% of the total roof area. In the objective function, the total runoff reduction rate is:

[0166]

[0167] wherein the total runoff without layout is obtained by SWMM simulation without LID;

[0168] The pollution load reduction rate is:

[0169]

[0170] The pollution load is Based on the exponential scouring function , the calculation is:

[0171]

[0172] wherein is the area width, the wash-off coefficient C1=0.008, the wash-off exponent .

[0173] Life cycle cost:

[0174]

[0175]

[0176] wherein, the specific parameters are: the construction cost of the bioretention pool is 600 yuan / m2, the maintenance cost is 18 yuan / m2·year, the service life is 15 years, the construction cost of the green roof is 205 yuan / m2, the maintenance cost is 30 yuan / m2·year, the service life is 15 years, the construction cost of the permeable pavement is 260 yuan / m2, the maintenance cost is 12 yuan / m2·year, the service life is 10 years, the growth rate coefficient The effect constraint is set as: the runoff total reduction rate is greater than or equal to 25%, and the pollution load reduction rate is greater than or equal to 30%, and the optimization is realized by coupling PySWMM and the multi-objective algorithm.

[0177] Optionally, in the step S3, when the rainfall scenario is designed, the Chicago rain type is generated to design the rainstorm, and the intensity formula is:

[0178]

[0179] wherein is the target dependent variable, T is the return period, t is the rainfall duration, 、 、 、 is a parameter, which is determined according to a statistical method.

[0180] According to the rain peak position coefficient , the peak position is determined, the rainfall processes of 1, 2, 3, 5 and 10 years are generated, and the corresponding total amount is calculated.

[0181] Specifically, in the step S4, when the entropy weight-ideal and poor solution distance method is used, an initial index matrix is constructed, and the runoff total reduction rate, the pollution load reduction rate and the life cycle cost of the selected scheme under different design rainstorm return periods are included in the matrix.

[0182] The weight is calculated by the entropy weight method:

[0183] The index matrix is standardized.

[0184] The index proportion is calculated.

[0185]

[0186] The information entropy is calculated.

[0187]

[0188] Determination of weights:

[0189]

[0190] Calculate closeness degree by TOPSIS:

[0191] Construct a weighted decision matrix;

[0192] Determine the positive and negative ideal solutions;

[0193] Calculate the distance of each scheme to the positive and negative ideal solutions ;

[0194] Calculate the relative closeness:

[0195]

[0196] Select the scheme with the largest closeness degree as the optimal scheme.

[0197] Embodiment: The embodiment provides a comprehensive optimization method for urban green infrastructure layout, as shown in the following figure, which comprises the following steps: Figure 2

[0198] S1, analysis of regional overview. Collect data such as topography, pipe network, land use, analyze geographical location, topography, hydrology and meteorology, and characteristics of river and lake system, diagnose current problems such as waterlogging and non-point source pollution, and clarify the construction conditions of green infrastructure; In the implementation process, high-precision topographic data, rainwater pipe network information, land use type data and historical measured rainfall runoff data and other materials of the research area need to be collected systematically. The accuracy of data collection needs to be ensured, such as 1:500 topographic data and 10m land use data.

[0199] S2, rainwater model construction. Based on SWMM, divide the catchment area and pipe network, set the parameters of sub-catchment area (such as impervious rate, Manning coefficient) and GI, use the measured data to calibrate and verify the model, and ensure the reliability of simulation; Taking a certain area of a city as the research object, based on SWMM software, the research area is divided into 59 sub-catchment areas, and the measured rainfall-runoff data is used to calibrate and verify the model, and the calibrated Manning coefficient of impervious surface is 0.013, the initial infiltration rate is 60mm / h, and the stable infiltration rate is 1.5mm / h. The Nash efficiency coefficient is 0.811 in verification, and the model is reliable.

[0200] ​S3, matching green infrastructure types. Define the integrated green infrastructure system of infiltration and storage facing the urban limited space, match the infiltration and storage integrated green infrastructure types combined with the distribution and characteristics of the underlying surface in the study area; in this embodiment, three different green infrastructures of green roof, permeable pavement and bioretention are selected as examples for illustration.

[0201] S4, coupling improved multi-objective optimization algorithm and rainwater model. The improved non-dominated sorting genetic algorithm is coupled with the rainwater model, the layout area of green infrastructure is taken as the decision variable, the total runoff reduction rate, pollution load reduction rate and life cycle cost are taken as the objective function, the area constraint and effect constraint condition are set, the multi-objective optimization model is constructed, and the non-dominated solution set is obtained through multiple return period rainfall scenario optimization calculation; the layout area of bioretention, green roof and permeable pavement in 59 sub-catchment areas is taken as the decision variable, which is 177 dimensions in total. The objective function includes:

[0202] Total runoff reduction rate:

[0203]

[0204] Wherein is the total runoff without LID, is the total runoff after layout;

[0205] Pollution load reduction rate:

[0206]

[0207] Load is taken as a representative;

[0208] Life cycle cost:

[0209]

[0210]

[0211] Take 5%, is the service life.

[0212] The constraint condition is that the layout area of each type of facility accounts for 20% to 90% of the used land (area constraint), such as permeable pavement not more than 90% of the road space; (effect constraint);

[0213] Couple the improved NSGA-Ⅱ algorithm to realize multi-objective optimization. Generate 1, 2, 3, 5 and 10-year design rainstorm, adopt Chicago rain type, rain peak position coefficient 0.41, and obtain non-dominated solution set under different return period.

[0214] S5, green infrastructure scheme comprehensive evaluation and optimal scheme determination. Entropy weight-pareto optimal solution distance method is adopted, and the performances of indexes under different return periods are comprehensively considered to evaluate the non-dominated solution set, and the optimal GI facility layout scheme is selected and compared;

[0215] When the entropy weight-pareto optimal solution distance method is adopted, the initial index matrix is constructed, and the total runoff reduction rate, pollution load reduction rate and life cycle cost of the alternative scheme under 5 design storm return periods are included in the matrix. The total runoff reduction rate and pollution load reduction rate are taken as positive indexes, and the life cycle cost is taken as a negative index. The Euclidean distance of the scheme to the positive and negative ideal solutions is calculated;

[0216] The entropy weight method is used to calculate the weight:

[0217] The standardized index matrix is obtained;

[0218] The index proportion is calculated:

[0219]

[0220] The information entropy is calculated:

[0221]

[0222] The weight is determined:

[0223]

[0224] The closeness degree is calculated by the pareto optimal solution distance method:

[0225] The weighted decision matrix is constructed;

[0226] The positive and negative ideal solutions are determined;

[0227] The distance of each scheme to the positive and negative ideal solutions is calculated ;

[0228] The relative closeness degree is calculated:

[0229]

[0230] The scheme with the largest relative closeness degree is selected as the optimal scheme. The larger the relative closeness degree , the closer the scheme to the ideal state, and the better the comprehensive performance. The calculation results of the embodiment are shown in Figure 3 , which respectively show the size of the relative closeness degree of all schemes, so as to determine the optimal implementation scheme.

[0231] Specifically, the results show that in the optimized results of the improved NSGA-Ⅱ algorithm, the runoff and pollution reduction rates increase with increasing cost, but the marginal benefits diminish. When the cost is below RMB 152 million / year, a 10% cost increase can lead to a 15% increase in the runoff reduction rate; after exceeding the threshold, the same investment only increases the rate by 3.8%. The life cycle cost is RMB 108 million to RMB 169 million / year, corresponding to a runoff reduction rate of 28.9% to 55.8% and a pollution reduction rate of 42.2% to 67.0%. The integrated infiltration, retention, and storage confined space green infrastructure demonstrates excellent efficiency under different return periods, effectively controlling runoff and reducing pollution while balancing cost input. However, with the increase of the return period, the runoff control and pollution reduction effects decrease under the same cost input, indicating that the GI efficiency decays with increasing rainfall intensity. The effectiveness of each calculation scheme is evaluated based on the entropy weight-superiority distance method, such as... Figure 3 As shown, the optimal solution was determined. Solution 571 ranked highest, closest to the optimal solution (0.399), and furthest from the worst solution (0.660), achieving a proximity score of 0.623. It performed best in cost-effectiveness balance optimization and was thus the optimal solution. This solution includes 37.92 hectares of bioretention ponds (30.33% of green space), 121.69 hectares of green roofs (69.14% of rooftops), and 120.44 hectares of permeable pavement (75.27% of road and open space), covering a total area of ​​280.05 hectares (60.75% of the area). Under historical rainfall on June 19, 2023, this solution achieved a runoff reduction rate of 53.69%, a pollution load reduction rate of 65.54%, and a life-cycle cost of 144 million yuan / year, achieving a balance between cost and effectiveness.

[0232] As an optional implementation, this embodiment also provides an electronic device, including: at least one processor, and a memory communicatively connected to the processor; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to enable the processor to execute the aforementioned collaborative decision-making method for the layout of green infrastructure in a confined space integrating permeability, stagnation, and storage.

[0233] As an optional implementation, this embodiment also provides a non-transitory computer-readable storage medium storing computer instructions, which are used to cause a computer to execute the aforementioned collaborative decision-making method for the layout of green infrastructure in a confined space that integrates permeation, stagnation, and storage.

[0234] The beneficial effects of this invention are as follows:

[0235] (1) Function synergy optimization, improve runoff and pollution control efficiency. Through the integration of infiltration storage integrated design, break through the single function limitation of traditional green infrastructure, realize multi-functional synergy optimization under complex underlying surface conditions. The present application can significantly improve the reduction efficiency of total runoff and pollution load, and ensure that the effect of rainwater management meets the standard.

[0236] (2) Space efficient adaptation, solve the problem of fragmented land constraints. In view of the problem of spatial fragmentation in highly developed areas, the present application proposes a dynamic adaptation method of facility function integration and discrete space layout, which realizes the efficient use of limited land resources. The present application can accurately match the available space and facility demand, improve the feasibility and implementation of the layout scheme.

[0237] (3) Prospective prediction driven, break through the limitation of lag judgment. The super volume prospective prediction mechanism proposed by the present application realizes the active prediction and intelligent decision of the optimization process by establishing a regression model, realizes the leap from passive threshold judgment to forward-looking intelligent decision, and the present application can significantly improve the intelligent level of the algorithm.

[0238] (4) Decision-making closed loop formation, promote the scientific scheme landing. The present application can automatically generate the optimal scheme considering runoff control, pollution reduction, economic cost and landscape coordination by combining multi-objective optimization algorithm and quantitative decision mechanism, reduce manual intervention, and significantly improve the scientificity and effectiveness of engineering decision-making.

[0239] Each embodiment in the specification is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between each embodiment can be referred to each other.

[0240] In this paper, specific examples are used to describe the principles and implementation methods of the present application. The above description of the embodiments is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation method and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for collaborative decision making of integrated pervious and impervious confined space green infrastructure layout, characterized in that, The method comprises the following steps: Collecting basic information of a research area, generalizing sub- catchment areas and rainwater pipe networks of the research area by SWMM software according to the basic information, setting model parameters, and obtaining a rain and flood model; the basic information comprises terrain data, rainwater pipe network information, land use type data, and historical measured rainfall runoff data; Defining a seepage and storage integrated green infrastructure system for a city limited space, matching seepage and storage integrated green infrastructure types of the seepage and storage integrated green infrastructure system according to the distribution and characteristics of the underlying surface of the research area, and obtaining a green infrastructure type matching result; Coupling a non-dominated sorting genetic algorithm with the rain and flood model, taking the layout area corresponding to the green infrastructure type matching result as a decision variable, taking a total runoff reduction rate, a pollution load reduction rate, and a life cycle cost as objective functions, setting area constraints and effect constraints, obtaining a multi-objective optimization model, and performing multi- return period rainfall scenario optimization calculation on the multi-objective optimization model to obtain a non-dominated solution set; Evaluating and screening the non-dominated solution set by using an entropy weight- superior and inferior solution distance method to obtain an optimal green infrastructure layout scheme; Defining a seepage and storage integrated green infrastructure system for a city limited space, matching seepage and storage integrated green infrastructure types of the seepage and storage integrated green infrastructure system according to the distribution and characteristics of the underlying surface of the research area, and obtaining a green infrastructure type matching result, comprising: Setting the seepage and storage integrated green infrastructure system; the seepage and storage integrated green infrastructure system comprises seepage facilities, retention facilities, and storage facilities; the seepage facilities comprise permeable pavement and green roof; the retention facilities comprise rainwater garden, artificial wetland, and biological retention pool; and the storage facilities comprise biological retention pool, rainwater pond, and green roof; Transforming ordinary roads, squares, and parking lots in the research area into the permeable pavement, and transforming roofs in the research area into the green roof; Using the retention facilities to delay the collection process of runoff in the research area; Using the storage facilities to store runoff in the research area when surface runoff cannot be discharged; Evaluating and screening the non-dominated solution set by using an entropy weight- superior and inferior solution distance method to obtain an optimal green infrastructure layout scheme, comprising: Constructing an initial index matrix; the initial index matrix comprises the total runoff reduction rate, the pollution load reduction rate, and the life cycle cost; The initial index matrix is subjected to index proportion calculation, information entropy calculation and weight calculation in sequence to obtain a matrix weight; an expression of the matrix weight is: ; wherein, ; ; wherein, is the maximum weight of the i-th index; is the maximum weight of the i-th index; is the information entropy of the i-th index; is the information entropy of the i-th index; is the total number of evaluation indexes; is the proportion of the j-th index of the i-th scheme; is the proportion of the j-th index of the i-th scheme; is the total number of schemes in the non-dominated solution set; is the total number of schemes in the non-dominated solution set; represents the j-th index value of the i-th scheme after dimensionless normalization processing. Performing weighted processing on the initial index matrix by using a matrix weight to obtain a weighted decision matrix; Determining a positive ideal solution and a negative ideal solution according to the weighted decision matrix; According to the distance of each scheme in the non-dominated solution set from the positive ideal solution and the negative ideal solution, the closeness degree is calculated to obtain a relative closeness degree; the expression of the relative closeness degree is: ; wherein, is the relative closeness degree of the i th scheme; is the Euclidean distance of the i th scheme to the positive ideal solution; is the Euclidean distance of the i th scheme to the negative ideal solution; and ​​​ Determining a scheme with the largest relative closeness degree as the optimal green infrastructure layout scheme.

2. The method of claim 1, wherein, Collecting basic information of a research area, generalizing sub- catchment areas and rainwater pipe networks of the research area by SWMM software according to the basic information, setting model parameters, and obtaining a rain and flood model, comprising: Collect topographic data, rainwater pipe network information, land use type data and historical measured rainfall runoff data of the research area to obtain the basic data; Divide sub-catchment areas by using comprehensive division method, import the sub-catchment areas and rainwater pipe network data into SWMM through GIS platform to obtain the rainwater and flood model; Set parameters of catchment area and structure parameters of integrated infiltration and detention green infrastructure in the rainwater and flood model; the parameters of catchment area include characteristic width, average slope, impervious rate, Manning coefficient, depression storage and Horton infiltration coefficient.

3. The method of claim 1, wherein, Couple non-dominated sorting genetic algorithm and the rainwater and flood model, take the layout area corresponding to the matching result of the green infrastructure type as decision variable, take total runoff reduction rate, pollution load reduction rate and life cycle cost as objective function, set area constraint and effect constraint condition to obtain a multi-objective optimization model, and perform optimization calculation on the multi-objective optimization model under multiple return period rainfall scenarios to obtain a non-dominated solution set, including: Construct original non-dominated sorting genetic algorithm; Replace the termination criterion of generation improvement rate determination of the hyper volume index in the original non-dominated sorting genetic algorithm with intelligent termination criterion of multi-objective evolutionary algorithm based on hyper volume foresight prediction to obtain the non-dominated sorting genetic algorithm; Set the layout area of each type of green infrastructure in each sub-catchment area as decision variable; defining an overall objective function according to the decision variables; an expression of the overall objective function is: ; wherein, ; ; ; ; is an overall objective function vector; , , is the runoff total reduction rate, the pollution load reduction rate, the life cycle cost respectively; is the runoff total amount without GI; is the runoff total amount after GI; is the pollution load amount without GI; is the pollution load amount after GI; is the cost of a sub-catchment; is the dimension of the decision variable; is the layout area; is the construction cost; is the operation and maintenance cost; is the growth rate coefficient; is the service life; Set area constraint and effect constraint; the area constraint includes the proportion of permeable pavement layout area to total road area, the proportion of bioretention pond to total green area, and the proportion of green roof to total roof area; the effect constraint includes the lower limit of total runoff reduction rate and the lower limit of pollution load reduction rate; Fuse the non-dominated sorting genetic algorithm, the overall objective function, the area constraint and the effect constraint to obtain the multi-objective optimization model; Perform combined facility scenario optimization calculation on the rainwater and flood model according to the multi-objective optimization model by using NSGA-II or NSGA-III algorithm to obtain the non-dominated solution set.

4. The method of claim 1, wherein, The bioretention pond is laid on the green land in the research area, the green roof is laid on the roof in the research area, the permeable pavement is laid on the road and open space in the research area, the parameter setting of the bioretention pond includes a dam height of 100 mm, a water storage layer thickness of 100 mm and a soil layer thickness of 400 mm, the parameter setting of the green roof includes a dam height of 150 mm, a soil layer thickness of 100 mm and a drainage cushion layer thickness of 100 mm, and the parameter setting of the permeable pavement includes a surface layer dam height of 20 mm, a road surface layer thickness of 150 mm and a water storage layer thickness of 150 mm.

5. The method of claim 3, wherein, Replace the termination criterion of generation improvement rate determination of the hyper volume index in the original non-dominated sorting genetic algorithm with intelligent termination criterion of multi-objective evolutionary algorithm based on hyper volume foresight prediction to obtain the non-dominated sorting genetic algorithm, including: Collect hyper volume index data; When the algorithm runs to the t th generation, set the data in the historical hyper-volume index data within the range of the iteration number of the t th generation and less than or equal to the W th generation as an observation window; Train the linear regression model using the data in the observation window to obtain a linear regression hyper-volume prediction model; The linear regression hyper-volume prediction model is used to predict the index value after N generations in the future, and the predicted relative yield is obtained by calculating the predicted yield using the index value after N generations; the expression of the predicted relative yield is: ; wherein, is the predicted relative yield; is the index value after N generations; is the current index value. Set a minimum start generation number, when the iteration number of the algorithm is greater than the minimum start generation number, calculate the predicted relative return, and stop iteration when the predicted relative return is less than a set threshold.

6. An electronic device, comprising: Comprise: At least one processor, and a memory connected in communication with the processor; wherein the memory stores instructions executable by the processor, the instructions executed by the processor to enable the processor to perform the method of any one of claims 1 to 5.

7. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to enable the computer to perform the method of any one of claims 1 to 5.

Images