ArcEngine-based intelligent project site selection decision support system and method

Abstract

The invention relates to the technical field of project site selection decision support, in particular to an intelligent project site selection decision support system and method based on ArcEngine. The method specifically comprises the following steps: integrating site selection data, and realizing unified management of geographic information and coordinate system conversion by utilizing ArcEngine; missing values and abnormal values are processed by adopting multilayer perception interpolation, Bayesian interpolation, adaptive density clustering and random forest correction technologies, and data reliability is ensured through a standard verification code mechanism; building a comprehensive evaluation model of resource accessibility, environmental suitability and cost control, generating an optimal solution and simulating a long-term industrial evolution trend; through industrial chain modeling, multi-agent game analysis and a network optimization algorithm, node benefits and a synergistic effect are quantified, and supply chain dynamic scheduling is optimized in combination with reinforcement learning. According to the method, site selection scientificity and industrial cooperation benefits are remarkably improved, and the method is suitable for complex site selection requirements of the modern coal chemical industry heavy asset industry.

Description

Technical Field

[0001] The present invention relates to the technical field of project site selection decision support, and particularly to an intelligent project site selection decision support system and method based on ArcEngine. Background Technique

[0002] In recent years, in the context of multiple challenges such as resource depletion, environmental pollution, and intensified market competition faced by the coal chemical industry, it is urgent to achieve industrial transformation and upgrading through technological innovation while meeting the strategic requirements of national energy conservation, emission reduction, and sustainable development; traditional site selection decision methods mainly rely on manual experience and static data analysis, and it is difficult to fully cope with the comprehensive effects of factors such as geographical space complexity, multi-dimensional economic indicators, and environmental sensitivity.

[0003] The Chinese invention patent application with the publication number of CN111428986A discloses a multi-index-based intelligent layout site selection method for the coal chemical industry, decomposes the key coal quality indicators required by the coal chemical process, establishes multi-parameter constraint equations and intelligent optimization evaluation models for water resources, coal quality, transportation, environment, etc., collects mine coal quality analysis data, integrates them into an intelligent planning algorithm, and forms an optimized layout plan through iterative evaluation by inputting multi-objective constraint parameters such as coal quality parameters, transportation routes, water resources, and environmental protection. By continuously iteratively optimizing the location of the coal chemical plant, the optimal site selection plan for the production factor configuration can be found through cyclic trial and error.

[0004] In modern economic society, with the rapid development of big data, GIS technology, and artificial intelligence, project site selection decision-making faces severe challenges such as a sharp increase in the amount of data, multi-source heterogeneous information, and complex and variable influencing factors. Traditional site selection methods have difficulty fully considering multiple factors such as geographical space, economic indicators, environmental assessment, and policy orientation. Therefore, there is an urgent need for a decision support system that integrates data collection, preprocessing, standardization, intelligent analysis, and multi-objective optimization to solve practical problems. Summary of the Invention

[0005] The object of the present invention is to propose an intelligent project site selection decision support system and method based on ArcEngine in view of the problems existing in the background technique.

[0006] The technical solution of the present invention: An intelligent project site selection decision support method based on ArcEngine includes the following specific implementation steps:

[0007] S1. Obtain site selection data, including: geographical space data, economic and industrial data, and infrastructure data;

[0008] S2. Build a data cleaning mechanism to repair missing values through multi-layer perceptron interpolation and Bayesian interpolation, correct outliers through adaptive density clustering and random forest, unify the coordinate system and verify topological consistency, and perform standardization processing, and generate a standard verification code;

[0009] S3. Verify data compliance based on the standard verification code, build a spatial data model in combination with GIS, conduct site selection ranking, optimize the site selection plan through a multi-objective optimization algorithm, combine system dynamics simulation for long-term suitability, and introduce deep reinforcement learning to optimize decisions, and finally output an intelligent optimized site selection plan;

[0010] S4. Build an optimization model for the coal chemical industry chain, quantify node benefits with a weighted directed graph, analyze the optimal strategies of the government, enterprises and the market through multi-agent game, improve the collaborative benefits of industrial clusters by network optimization, introduce reinforcement learning to optimize supply chain scheduling, optimize the industrial layout plan, and output a decision support plan;

[0011] S5. Use ArcEngine to render candidate site selection points, display the site selection suitability in the form of a heat map and a spatial distribution map, and allow users to adjust the site selection weight parameters, update the site selection results in real time, and combine Web GIS technology to interactively select a decision support plan online.

[0012] Preferably, the specific implementation process of the data cleaning mechanism is as follows:

[0013] S21. Combine the multi-layer perceptron interpolation method and Bayesian interpolation to repair missing values:

[0014] Build an interpolation model based on the multi-layer perceptron MLP, and use other complete variables Y to predict the missing value X missing :

[0015] X missing = f(Y; θ) + ε;

[0016] In the formula, f(Y; θ) represents the mapping function trained based on the MLP; θ represents the neural network parameters; Y represents the complete input variables; ε represents the noise term;

[0017] Bayesian interpolation:

[0018]

[0019] In the formula, X missing represents the missing value; X observed represents the observed data; P(X missing ) represents the prior probability of the missing value; P(X observed |X missing ) represents the likelihood function; P(X missing |X observed ) represents the posterior probability;

[0020] Calculate the optimal interpolation by maximum a posteriori (MAP) estimation

[0021]

[0022] where represents the optimal missing value obtained by maximum a posteriori estimation;

[0023] S22. Adopt the method of adaptive density clustering and random forest correction to detect and correct outliers;

[0024] S23. Unify the coordinate system of the location data and use the adjacency matrix method to verify the topological consistency;

[0025] S24. Use the Z-score normalization method for normalization processing to obtain the normalized data.

[0026] Preferably, the correction process of correcting outliers is as follows:

[0027] S31. Detect outliers based on adaptive density clustering, identify outlier data points with too low local density, and define the local density as follows:

[0028]

[0029] In the formula, ρ i represents the density of data point i; d ij represents the Euclidean distance between data point i and data point j; h' represents the density estimation window size; K(d ij , h') represents the Gaussian kernel function; X ik and X jk represent the coordinate values of data point i and data point j in the k-th dimension respectively; e represents the natural constant;

[0030] S32. The determination condition for outliers is: ρ i < μ ρ - λ'σ ρ , and accordingly output the outlier X outlier ;

[0031] where, μ ρ represents the mean value of the density; σ ρ represents the standard deviation of the density; λ' represents the threshold for controlling outliers;

[0032] S33. Correct the detected outlier X outlier and use the random forest model to predict its correct value:

[0033]

[0034] Wherein, X correctd represents the corrected data value; T i (Y) represents the predicted value of the i-th decision tree; M represents the number of decision trees.

[0035] Preferably, the generation process of the standard verification code is as follows:

[0036] S41. Select a random integer θ ∈ [1, q - 1], and calculate the first-level meta-code eⅠ = (Bm × θ mod p) mod q;

[0037] Wherein, q is a predefined prime number of 160 bits; p is a predefined prime number of 1024 bits, and p - 1 is an integer multiple of q; Bm is the meta-code benchmark, Bm = [e × (p - 1) / q] mod p; e represents a predefined element in the group on an element;

[0038] S42. Convert the standardized data data into a binary string datab;

[0039] S43. Calculate the second-level meta-code eⅡ = {(θ - 1) × [H(datab) + ec × eⅠ]} mod q;

[0040] Wherein, ec is a predefined meta-code generation element; H is a predefined hash function;

[0041] S44. Generate the standard verification code CS = (eⅠ, eⅡ).

[0042] Preferably, the verification process for verifying data compliance based on the standard verification code is as follows:

[0043] S51. Calculate the auxiliary meta-code audit element Aea = (eⅡ - 1) mod q;

[0044] S52. Convert the standardized data data into a binary string datac;

[0045] S53. Calculate the following meta-code audit elements:

[0046] The first-level meta-code audit element eaⅠ = [H(datac) × Aea] mod q;

[0047] The second-level meta-code audit element eaⅡ = (eⅠ × Aea) mod q;

[0048] Wherein, H is a predefined hash function;

[0049] S54. Calculate the reference code CB = (Bm × eaⅠ × ep × eaⅡ mod p) mod q;

[0050] Among them, ep represents a predefined meta-code parsing element;

[0051] S55. If CB = eⅠ, then the data data meets the standard; otherwise, an alarm is given immediately.

[0052] Preferably, the sorting process of site selection sorting is as follows:

[0053] S61. Construct a spatial data model of GIS, including a vector model and a raster model:

[0054] Vector data modeling, stored in the form of points, lines, and surfaces: P i =(x i , y i , A i , C i , E i );

[0055] Among them, P i represents the attribute set of area i; (x i , y i ) represents the geographical coordinates of area i; A i represents the developable land area; C i represents the coal resource reserve; E i represents the environmental suitability score;

[0056] Raster data modeling: represented by raster data, each raster contains a spatial feature value;

[0057] S62. Use network analysis method to calculate the resource accessibility score of different candidate sites:

[0058]

[0059] In the formula, R i represents the resource accessibility score of candidate area i; W j represents the importance weight of the preset resource point j; d ij represents the traffic distance from candidate area i to resource point j;

[0060] S63. Use buffer analysis to evaluate environmental suitability:

[0061]

[0062] In the formula, E i represents the environmental suitability score of area i; d ik represents the distance from area i to environmental sensitive point k; S k represents the environmental importance score of the preset sensitive point k; α represents the distance attenuation index;

[0063] S64. Calculate the comprehensive suitability score based on the analytic hierarchy process + GIS weighted overlay analysis: S i = w1R i + w2E2 + w3Cost i ;

[0064] In the formula, S i represents the comprehensive suitability score of region i; R i represents the resource accessibility score; E i represents the environmental suitability score; Cost i represents the estimated construction cost score; w1, w2, and w3 represent the preset weights;

[0065] S65. Sort the comprehensive suitability scores of all candidate regions and select the Top-K suitable regions:

[0066]

[0067] Among them, represents the site selection plan; Ω represents the set of all candidate regions;

[0068] S66. Output the site selection ranking list List.

[0069] Preferably, the implementation process of the multi-objective optimization algorithm is as follows:

[0070] S71. Initialize the population: Use the site selection ranking list List as the initial population {P1, P2,..., P K};

[0071] S72. Non-dominated sorting: Evaluate each candidate solution P i in the population for the multi-objective function, calculate its performance in all objective functions, and sort the candidate solutions according to the dominance relationship to form multiple non-dominated levels, that is, the Pareto front;

[0072] S73. Crowding degree calculation: Calculate the crowding degree distance for each candidate solution in each non-dominated layer;

[0073] S74. Selection, crossover, and mutation: Use the crowding degree sorting and roulette wheel selection mechanism to select parent individuals from the current population and perform crossover and mutation operations to generate the offspring population;

[0074] S75. Population merging and new generation generation: Merge the parent and offspring populations, re-perform non-dominated sorting and crowding degree calculation, and select the top K excellent candidate solutions to form a new generation population;

[0075] S76. Convergence and stopping conditions: When the preset number of iterations is reached or the change in the optimal solution tends to be stable, the algorithm terminates and outputs the final Pareto optimal solution set, that is, the output site selection set List'.

[0076] Preferably, the objective function of the multi-objective intelligent optimization algorithm is:

[0077] maxF(P i ) = (f1(P i ), f2(P i ), f3(P i ));

[0078] f1(P i ) = ∑w k R ik ;

[0079] f2(P i ) = -∑w m E im ;

[0080] f3(P i ) = -Cost i ;

[0081] In the formula, P i represents the i-th candidate site selection point; f1(P i ) represents the maximization of resource accessibility; f2(P i ) represents the minimization of environmental impact; f3(P i ) represents the minimization of cost; R ik represents the score of the i-th site selection point on the k-th resource factor; w k represents the resource factor weight; E im represents the score of the impact of the i-th site selection point on the m-th environmental factor; w m represents the environmental factor weight; Cost i represents the comprehensive cost of the i-th site selection point.

[0082] Preferably, the optimization process of the industrial chain optimization model is as follows:

[0083] S91. Establish a data model for the coal chemical industry chain, clarify the relationships between industrial nodes, and evaluate the comprehensive benefits for each node. The industrial chain is represented by a weighted directed graph G: G = (N, E, W);

[0084] Among them, N represents the key facilities on the industrial chain; E represents the industrial flow path; W represents the weight of the edge;

[0085] The comprehensive benefit of each node is represented by the following formula: U i = βRe i - γEei +λPe i ;

[0086] Among them, U i represents the comprehensive benefit of any node U i ; Re i represents the resource utilization rate; Ee i represents the environmental impact, Pe i represents the economic benefit; β, γ, λ represent the weight coefficients;

[0087] S92. Model the game among the government, enterprises, and the market in the industrial layout, and solve the optimal strategy through Nash equilibrium;

[0088] S93. Adopt a network optimization algorithm to reflect the synergy effect among enterprises;

[0089] Industrial cluster network model: H = (V, L);

[0090] The synergy benefit function is:

[0091] Among them, V represents the enterprise node, and L is the cooperation relationship among enterprises; A ij represents whether enterprises i and j cooperate; W ij is the cooperation benefit;

[0092] S94. Dynamic supply chain scheduling, adopt a reinforcement learning algorithm to optimize supply chain scheduling, and adjust each link of the supply chain. The supply chain optimization goal is: min(C trans +C storage -R delivery );

[0093] Among them, C trans is the transportation cost; C storage is the inventory cost; R delivery is the on-time delivery benefit;

[0094] Use reinforcement learning to optimize the scheduling strategy. The reward function for the state St and the action At is: R t = -(C trans +C storage ) + εD t ;

[0095] Among them, D t is the on-time delivery rate; ε is the adjustment coefficient;

[0096] S95. Output the decision support plan.

[0097] Technical solution of the present invention: An intelligent project site selection decision support system based on ArcEngine, which is used to execute the above-mentioned intelligent project site selection decision support method based on ArcEngine, includes:

[0098] A data collection and fusion module, which is used to collect site selection data and perform geographic information data management and fusion using ArcEngine;

[0099] A data preprocessing and standardization module, which is used to preprocess the collected site selection data;

[0100] An intelligent analysis and site selection optimization module, which is used to perform spatial analysis based on GIS, construct a site selection evaluation model, and optimize the site selection plan based on a multi-objective optimization algorithm;

[0101] An industrial layout decision support module, which is used to score and evaluate candidate site selection points in combination with a multi-criteria decision analysis method, perform industrial matching degree analysis in combination with government policies, market demands, and resource distributions, and then generate a site selection plan;

[0102] A visualization and interaction module, which is used to provide map visualization display, heat map analysis, and site selection simulation functions using ArcEngine and Web GIS technologies, and provide the capabilities of interactive data query, dynamic adjustment of site selection parameters, and comparison of different site selection plans.

[0103] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0104] The present invention designs an intelligent project site selection decision support system and method based on ArcEngine, covering the entire process of data collection, preprocessing, intelligent analysis, decision support, and visualization:

[0105] (1) Significantly improved data quality and reliability: Adopting various preprocessing methods such as multi-layer perceptron interpolation, Bayesian interpolation, density clustering, and random forest correction to effectively solve data missing and abnormal problems, ensuring the accuracy and consistency of input data, and providing a solid data foundation for subsequent intelligent decision-making;

[0106] (2) Precise spatial analysis and intelligent site selection: Using the ArcEngine platform to achieve efficient GIS data management and fusion, constructing a detailed spatial data model; Combining multi-objective optimization algorithms, deep reinforcement learning, and dynamic simulation technologies to comprehensively consider multiple factors such as resource accessibility, environmental suitability, and construction costs, and improving the scientificity and accuracy of the site selection plan;

[0107] (3) Sustainability and forward-looking nature of long-term decision-making: By introducing system dynamics models and dynamic simulation technologies, it is possible to simulate long-term development trends under different policy and market demand scenarios, providing forward-looking verification for site selection decisions and ensuring the continuous applicability and stability of the project over the next 10 to 20 years;

[0108] (4) Optimization of comprehensive industrial layout and synergy benefits: Combining methods such as industrial chain optimization, game analysis, and network collaboration, in-depth analysis is carried out on the multi-agent game problems among the government, enterprises, and the market to achieve a reasonable industrial layout and optimal allocation of resources, thereby enhancing the synergy benefits and competitive advantages of the overall industrial chain;

[0109] (5) Quick verification of standard verification codes: The standard verification code CS is generated through random numbers and hash functions to ensure the uniqueness and integrity of the data; The verification process verifies the standard type of the transmitted data and, to a certain extent, ensures the consistency between the transmitted data and the original data, effectively preventing data tampering and incorrect transmission, and enhancing the security and credibility of the system. Brief Description of the Drawings

[0110] Figure 1 It is a system architecture diagram of an intelligent project site selection decision-making support system based on ArcEngine proposed by the present invention;

[0111] Figure 2 It is a method flow diagram of an intelligent project site selection decision-making support method based on ArcEngine proposed by the present invention. Detailed Description of the Invention

[0112] Example 1, as Figure 1 shown, the intelligent project site selection decision-making support system based on ArcEngine proposed by the present invention includes: a data collection and fusion module, a data preprocessing and standardization module, an intelligent analysis and site selection optimization module, an industrial layout decision-making support module, and a visualization and interaction module.

[0113] The data collection and fusion module collects site selection data, including but not limited to geospatial data (including but not limited to terrain, landform, meteorology, and environmental assessment), economic and industrial data (including but not limited to enterprise distribution, economic indicator consumption levels), and infrastructure data (including but not limited to transportation networks, public service facilities, and land use), and uses ArcEngine for geographic information data management and fusion;

[0114] The data preprocessing and standardization module preprocesses the collected site selection data, including but not limited to data cleaning, missing value filling, and standardization;

[0115] The intelligent analysis and site selection optimization module conducts spatial analysis based on GIS, constructs a site selection evaluation model, and optimizes the site selection plan based on a multi-objective optimization algorithm;

[0116] The industrial layout decision support module combines the multi-criteria decision analysis (MCDM) method to score and evaluate candidate site selection points, and combines government policies, market demand, and resource distribution to conduct industrial matching degree analysis, and then generates site selection plans under different scenarios;

[0117] The visualization and interaction module uses ArcEngine and Web GIS technologies to provide functions including but not limited to map visualization display, heat map analysis, and site selection simulation, and provides the ability of interactive data query, dynamic adjustment of site selection parameters, and comparison of different site selection plans.

[0118] Embodiment 2, as Figure 2 shown, the intelligent project site selection decision support method based on ArcEngine proposed by the present invention is applied to the intelligent project site selection decision support system proposed in Embodiment 1, and its specific implementation steps are as follows:

[0119] S1. The data collection and fusion module obtains site selection data, specifically:

[0120] Geospatial data: Call remote sensing images, topographic features, meteorological data, and land use data through ArcEngine, parse vector and raster format geographic data, and use spatial data matching technology to convert data in different coordinate systems to a unified projection coordinate system to improve data consistency;

[0121] Economic and industrial data: Collect data including but not limited to industrial development data, population statistics data, economic growth indicators, market demand, and government planning policy information, and connect to external data sources including but not limited to government statistical platforms and market analysis reports through APIs or databases to form a complete industrial information database;

[0122] Infrastructure data: Collect data on road traffic, public service facilities (including but not limited to hospitals, schools), energy supply networks, etc., and use the topological analysis function of ArcEngine to clean the road network data to ensure connectivity and topological correctness.

[0123] It should be noted that ArcEngine is a component-based GIS development platform, mainly used for building independent custom geographic information system applications or extending existing applications; developed based on ArcObjects (Esri's core component library), allowing developers to utilize its rich GIS function modules for secondary development, supporting multiple programming languages such as C#, VB.NET, Java, etc.; the core advantage of ArcEngine lies in its high flexibility and customizability, and developers can directly call the underlying GIS function interfaces to implement complex operations such as map display, spatial analysis, data editing, and geoprocessing.

[0124] S2. The data preprocessing and standardization module performs standardization processing on the data to ensure the consistency and availability of the data for subsequent intelligent analysis and optimization calculations. The specific implementation process is as follows:

[0125] S21. Combine missing value repair based on multi-layer perception, adaptive outlier detection, and semantic consistency checking to build a data cleaning mechanism to provide high-quality input data for subsequent intelligent site selection decisions. Specifically:

[0126] S2101. Missing value processing: Use the multi-layer perception interpolation method (MLP-Impute) to repair missing values and combine Bayesian interpolation to improve the accuracy of interpolation. Specifically:

[0127] (1) Assume that some variables X in the dataset have missing values. Build an interpolation model based on a multi-layer perceptron (MLP) to predict the missing values X using other complete variables Y missing :

[0128] X missing = f(Y; θ) + ε;

[0129] In the formula, f(Y; θ) represents the mapping function trained based on MLP; θ represents the neural network parameters; Y represents the complete input variables; ε represents the noise term;

[0130] Specifically: MLP adopts a three-layer structure:

[0131] Input layer: Receive non-missing variables Y;

[0132] Hidden layer: Adopt the ReLU activation function to extract data patterns: h i = max(0, W i Y + b i );

[0133] In the formula, h i represents the output of the i-th hidden layer; W i represents the input weight matrix; b i represents the bias term;

[0134] Output layer: Predict X missing , and perform regression: X missing = W o h + b o ;

[0135] Among them, X missing represents the missing value; W o represents the output layer weight matrix; b o represents the output layer bias term; h represents the output of the hidden layer, that is, the value passed to the output layer for final prediction;

[0136] It should be noted that the model adopts an adaptive training strategy. When the missing rate is high, L2 regularization is increased to prevent overfitting, and the loss function is in the following form:

[0137] Among them, L represents the loss function; X true and X pred represent the true value and the predicted value respectively; N represents the number of samples in the training data; λ||θ|| 2 represents the L2 regularization term; λ represents the L2 regularization coefficient; ||θ|| 2 represents the square of the L2 norm of the parameter vector θ, that is, the sum of the squares of all weights, as the regularization term;

[0138] (2) For data with strong spatio-temporal correlation, Bayesian interpolation is adopted:

[0139]

[0140] In the formula, X missing represents the missing value; X observed represents the observed data; P(X missing ) represents the prior probability of the missing value, estimated based on historical data; P(X observed |X missing ) represents the likelihood function, that is, the probability of the observed data when X missing takes a certain value; P(X missing |X observed ) represents the posterior probability, indicating the probability of the missing value X observed under the condition that the observed data X missing is known;

[0141] Calculate the optimal interpolation through the maximum a posteriori estimation (MAP)

[0142]

[0143] Among them, represents the optimal missing value obtained through the maximum a posteriori estimation;

[0144] S2102. Outliers in the data may be caused by sensor errors, manual input errors, or environmental changes. The method of adaptive density clustering (ADC) and random forest correction is used to detect and correct outliers, specifically as follows:

[0145] (1) Adaptive density clustering (ADC) is used to detect outliers and identify outlier data points with too low local density. The local density is defined as follows:

[0146]

[0147]

[0148] In the formula, ρ i represents the density of data point i; d ij represents the Euclidean distance between data point i and data point j; h' represents the density estimation window size; K(d ij , h') represents the Gaussian kernel function; X ik and X jk represent the coordinate values of data point i and data point j in the k-th dimension respectively; e represents the natural constant;

[0149] (2) The determination condition for outliers is: ρ i < μ ρ - λ'σ ρ , and based on this, the outlier X outlier is output;

[0150] Among them, μ ρ represents the mean of the density; σ ρ represents the standard deviation of the density; λ' represents the threshold for controlling outliers (set to 2.5 in this embodiment);

[0151] (3) Random forest corrects outliers: The detected outliers X outlier need to be corrected, and the random forest model is used to predict its correct value:

[0152] In the formula, X correctd represents the corrected data value; T i (Y) represents the predicted value of the i-th decision tree; M represents the number of decision trees;

[0153] S22. Spatial data conversion: Unify the format and coordinate system of geographic data and ensure topological consistency, specifically as follows:

[0154] S2201. Coordinate system conversion: Since different data sources may use different coordinate systems (including but not limited to WGS84, UTM, GCJ-02), conversion is required: (x', y') = Tr(x, y);

[0155]

[0156] Among them, (x, y) represents the original coordinates; (x', y') represents the transformed coordinates; Tr() represents the coordinate transformation function, and the Helmert 7-parameter method is adopted in this embodiment; R represents the rotation matrix; (t x , t y , t z ), t

[0157] S2202. Topological consistency verification: The adjacency matrix method is used to verify the topological consistency of geographical data;

[0158] If there are isolated points or unclosed areas, automatic correction is performed;

[0159] S23. Eliminate the influence of different data scales, and use the Z-score normalization method to perform feature normalization to obtain the normalized data data;

[0160] S24. Generate a standard verification code for the normalized data data, and the generation process of the standard verification code is as follows:

[0161] S2401. Select a random integer θ ∈ [1, q - 1], and calculate the first-level meta-code eⅠ = (Bm × θ mod p) mod q;

[0162] Among them, q is a predefined prime number of 160 bits; p is a predefined prime number of 1024 bits, and p - 1 is an integer multiple of q; Bm is the meta-code benchmark, Bm = [e × (p - 1) / q] mod p; e represents a predefined element in the group Z p * on the previous element;

[0163] S2402. Convert the normalized data data into a binary string datab;

[0164] S2403. Calculate the second-level meta-code eⅡ = {(θ - 1) × [H(datab) + ec × eⅠ]} mod q;

[0165] Among them, ec is a predefined meta-code generation element; H is a predefined hash function;

[0166] S2404. Generate the standard verification code CS = (eⅠ, eⅡ);

[0167] S25. Transmit {the standard verification code CS = (eⅠ, eⅡ), the normalized data data} to the intelligent analysis and site selection optimization module, and transmit the normalized data data to the visualization and interaction module.

[0168] S3. The intelligent analysis and site selection optimization module conducts spatial analysis based on GIS, constructs a site selection evaluation model, and optimizes the site selection plan based on a multi-objective optimization algorithm. Specifically:

[0169] S31. Receive the {standard verification code CS = (eⅠ, eⅡ), standardized data data}, extract the standard verification code CS = (eⅠ, eⅡ) and the standardized data data, and verify whether the received standardized data data meets the standard. The verification process is as follows:

[0170] S3101. Calculate the auxiliary meta-code verification element Aea = (eⅡ - 1) mod q;

[0171] S3102. Convert the received standardized data data into a binary string datac;

[0172] S3103. Calculate the following meta-code verification elements:

[0173] The primary meta-code verification element eaⅠ = [H(datac) × Aea] mod q;

[0174] The secondary meta-code verification element eaⅡ = (eⅠ × Aea) mod q;

[0175] Where H is a predefined hash function;

[0176] S3104. Calculate the reference code CB = (Bm × eaⅠ × ep × eaⅡ mod p) mod q;

[0177] Where ep represents a predefined meta-code parsing element;

[0178] S3105. If CB = eⅠ, the received standardized data data meets the standard; otherwise, an alarm is immediately issued;

[0179] S32. Conduct spatial analysis for the site selection of modern coal chemical industries in combination with GIS (Geographic Information System), and realize intelligent site selection decision support through spatial data modeling, resource accessibility analysis, environmental suitability analysis, comprehensive suitability evaluation, and optimal site selection recommendation. Specifically:

[0180] S3201. Construct a spatial data model for GIS, including a vector model (points, lines, surfaces) and a raster model (gridded data). Specifically:

[0181] Vector data modeling, stored in the form of points, lines, and surfaces: P i =(x i , y i , A i , C i , E i );

[0182] Among them, P i represents the set of attributes of region i; (x i , y i ) represents the geographical coordinates of region i; A i represents the area of developable land; C i represents the coal resource reserve; E i represents the environmental suitability score;

[0183] Grid data modeling: It is represented by grid data (Grid), and each grid contains a spatial feature value;

[0184] S3202. Considering the accessibility of coal resources, transportation routes, and market demand locations, use the network analysis method to calculate the resource accessibility scores of different candidate locations:

[0185]

[0186] In the formula, R i represents the resource accessibility score of candidate region i; W j represents the importance weight of the preset resource point j; d ij represents the transportation distance from candidate region i to resource point j;

[0187] S3203. For coal chemical projects, it is necessary to consider ecological red lines, meteorological conditions, and water resource distribution, and use buffer analysis (Buffer Analysis) to evaluate environmental suitability:

[0188]

[0189] In the formula, E i represents the environmental suitability score of region i; d ik represents the distance from region i to environmental sensitive point k; S k represents the environmental importance score of the preset sensitive point k; α represents the distance decay index, and the range set in this embodiment is [1.5, 2.0];

[0190] S3204. Based on AHP (Analytic Hierarchy Process) + GIS weighted overlay analysis (Weighted Overlay Analysis), calculate the comprehensive suitability score: S i = w1R i + w2E2 + w3Cost i ;

[0191] In the formula, S i represents the comprehensive suitability score of region i; R i represents the resource accessibility score; E i represents the environmental suitability score; Costi Indicates the estimated construction cost score; w1, w2, w3 represent preset weights;

[0192] It should be noted that the settings of the weights w1, w2, w3 are combined with AHP and machine learning optimization to avoid human subjective biases, improve the accuracy of site selection, and introduce a dynamic weight adjustment mechanism to adapt to different project requirements and enhance the level of intelligent decision-making;

[0193] S3205. Sort the comprehensive suitability scores of all candidate areas and select the Top-K suitable areas:

[0194]

[0195] Among them, represents the site selection plan; Ω represents the set of all candidate areas;

[0196] S3206. Output the site selection ranking list List;

[0197] S33. Introduce intelligent analysis and optimized site selection methods, and use multi-objective optimization (MOO), machine learning (ML), and dynamic simulation (DS) technologies to further optimize the site selection results and improve the scientificity and adaptability of decision-making. Specifically:

[0198] S3301. Since the modern coal chemical industry involves multiple objectives including but not limited to resource acquisition, environmental protection, and economy, use a multi-objective intelligent optimization algorithm (MOO) for optimal site selection optimization, and set the objective function:

[0199] maxF(P i )=(f1(P i ), f2(P i ), f3(P i ));

[0200] f1(P i )=∑w k R ik ;

[0201] f2( P i )=-∑w m E im ;

[0202] f3(P i )=-Cost i ;

[0203] In the formula, P i represents the i-th candidate site selection point; f1(P i ) represents the maximization of resource accessibility; f2(P i) represents minimizing environmental impact; f3(P i ) represents minimizing cost; R ik represents the score of the i-th site selection point on the k-th resource factor (including but not limited to coal supply, transportation); w k represents the resource factor weight calculated by the AHP+ML method; E im represents the score of the impact of the i-th site selection point on the m-th environmental factor (including but not limited to pollution, carbon emissions); w m represents the environmental factor weight calculated by the AHP+ML method; Cost i represents the comprehensive cost of the i-th site selection point;

[0204] Accordingly, output the site selection set List';

[0205] S3302. To improve the long-term feasibility of the site selection plan, establish a dynamic simulation model for the development of the coal chemical industry to simulate the development of different site selection plans in the next 10-20 years:

[0206] (1) Dynamic simulation model (based on system dynamics SD):

[0207] P(t + 1)=P(t)+ΔP 资源 (t)-ΔP 消耗 (t)-ΔP 政策 (t);

[0208] In the formula, P(t) represents the site selection suitability in the t-th year; ΔP 资源 (t) represents the growth of resource supply; ΔP 消耗 (t) represents the consumption of coal resources; Δ P 政策 (t) represents the impact of environmental protection policies;

[0209] (2) The simulation steps are as follows:

[0210] Initialize parameters: Based on the output site selection set List', set the initial state;

[0211] Run the dynamic simulation: Evaluate the long-term suitability of the site selection;

[0212] Screen out the optimal site selection plan: Ensure the sustainable development of the industry;

[0213] S3303. Introduce deep reinforcement learning (DQN) to fine-tune the site selection results: Take the optimal site selection plan selected in step S3302 as the initial state, simulate uncertain factors including but not limited to different policies and market demands, generate an optimized site selection strategy, and continuously learn historical data to improve the intelligence level of site selection;

[0214] S34. Output the optimized site selection plan and transmit the optimized site selection plan to the industrial layout decision support module and the visualization and interaction module.

[0215] S4. Based on the optimized site selection plan, the industrial layout decision support module uses industrial chain modeling, game analysis, network optimization, and dynamic scheduling methods to construct an optimized model for the modern coal chemical industry layout, providing scientific decision support to ensure a reasonable industrial layout, optimal allocation of resources, and improved collaborative benefits of the industrial chain. Specifically:

[0216] S41. Construct an industrial chain optimization model (ICM) as follows:

[0217] S4101. Establish a data model for the coal chemical industry chain, clarify the relationships between industrial nodes, and evaluate the comprehensive benefits for each node. The industrial chain is represented by a weighted directed graph G: G = (N, E, W);

[0218] Among them, N represents the key facilities in the industrial chain (including but not limited to coal mines, chemical industrial parks); E represents the industrial flow paths (including but not limited to materials and energy); W represents the weight of the edge, measuring the transportation cost and resource consumption;

[0219] The comprehensive benefit of each node is represented by the following formula: U i = βRe i - γEe i + λPe i ;

[0220] Among them, U i represents the comprehensive income of any node U i ; Re i represents the resource utilization rate; Ee i represents the environmental impact, Pe i represents the economic income; β, γ, λ represent the weight coefficients;

[0221] S42. Multi-agent game analysis (GTE): Model the game among the government, enterprises, and the market in the industrial layout. The income functions of each agent of the government, enterprises, and the market are U G , U E , U M :

[0222]

[0223] In the formula, T G represents the government tax policy obtained; I i represents the enterprise investment set by the enterprise; C sub represents the government subsidy obtained; C prod represents the enterprise cost set; Pr iDenote the market price obtained from the survey; S i Denote the market demand obtained from the survey;

[0224] Based on this, solve for the optimal strategy through Nash equilibrium: argmax(U G , U E , U M );

[0225] S43. Regard the industrial cluster as a complex network and adopt a network optimization algorithm to enhance the synergy among enterprises;

[0226] Industrial cluster network model: H = (V, L);

[0227] Among them, V represents the enterprise nodes, and L is the cooperation relationship among enterprises;

[0228] The synergy benefit function is:

[0229] Among them, A ij Indicates whether enterprises i and j cooperate; W ij Is the cooperation benefit;

[0230] Based on this, improve the overall synergy benefit by optimizing the cluster structure;

[0231] S44. Dynamic supply chain scheduling (DSC), adopt a reinforcement learning (RL) algorithm to optimize supply chain scheduling, adjust each link of the supply chain to adapt to market demand fluctuations and reduce operating costs. The supply chain optimization goal is: min(C trans + C storage - R delivery );

[0232] Among them, C trans Is the transportation cost; C storage Is the inventory cost; R delivery Is the on-time delivery benefit;

[0233] Use reinforcement learning to optimize the scheduling strategy. The reward function for state St and action At is: R t = -(C trans + C storage ) + εD t ;

[0234] Among them, D t Is the on-time delivery rate; ε is the adjustment coefficient;

[0235] S45. Combine the results of steps S41 - S44, and conduct the final industrial layout decision by integrating the industrial chain optimization model, game analysis, cluster network optimization, and supply chain scheduling. Through optimizing policies, resource allocation, and synergy effects, form an optimal industrial layout plan to provide practical decision - making support for the government, enterprises, and the market.

[0236] S46. Output the decision - making support plan and transmit the decision - making support plan to the visualization and interaction module.

[0237] S5. The visualization and interaction module uses ArcEngine to render the candidate site selection points, and displays the site selection suitability in the form of a heat map and a spatial distribution map. It also allows users to adjust the site selection weight parameters and updates the site selection results in real - time. Combining Web GIS technology, it realizes online interactive site selection decision - making support.

[0238] Embodiment 3. The intelligent project site selection decision - making support method based on ArcEngine proposed by the present invention further includes a multi - objective intelligent optimization algorithm (MOO), and its specific implementation steps are as follows:

[0239] S1. Initialize the population: Use the site selection ranking list List as the initial population {P1, P2, …, P K}.

[0240] S2. Non - dominated sorting: Evaluate each candidate solution P in the population for the multi - objective function, calculate its performance in all objective functions, and sort the candidate solutions according to the dominance relationship to form multiple non - dominated levels (i.e., Pareto fronts). i

[0241] S3. Crowding degree calculation: Calculate the crowding degree distance for the candidate solutions in each non - dominated layer to ensure population diversity.

[0242] S4. Selection, crossover, and mutation: Use the crowding degree sorting and roulette wheel selection mechanism to select parent individuals from the current population, and perform crossover (including but not limited to single - point or multi - point crossover) and mutation operations to generate offspring populations.

[0243] S5. Population merging and new generation generation: Merge the parent and offspring populations, re - conduct non - dominated sorting and crowding degree calculation, and select the top K excellent candidate solutions to form a new generation population.

[0244] S6. Convergence and stopping conditions: When the preset number of iterations is reached or the change in the optimal solution tends to be stable, the algorithm terminates, and outputs the final Pareto optimal solution set, that is, outputs the site selection set List'.

[0245] ​The embodiments of the present invention have been described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited thereto, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those skilled in the art.

Claims

1. An intelligent project site selection decision support method based on ArcEngine, characterized in that, It includes the following specific implementation steps: S1. Obtain site selection data, including: geospatial data, economic and industrial data, and infrastructure data; S2. Build a data cleaning mechanism to repair missing values through multi-layer perceptron interpolation and Bayesian interpolation, correct outliers through adaptive density clustering and random forest, unify the coordinate system and verify topological consistency, and perform standardization processing, and generate a standard verification code; S3. Verify data compliance based on the standard verification code, build a spatial data model in combination with GIS, perform site selection ranking, optimize the site selection plan through a multi-objective optimization algorithm, combine system dynamics simulation for long-term suitability, and introduce deep reinforcement learning to optimize decisions, and finally output an intelligent optimized site selection plan; S4. Build an optimization model for the coal chemical industry chain, quantify node benefits with a weighted directed graph, analyze the optimal strategies of the government, enterprises, and the market through multi-agent game, improve the collaborative benefits of industrial clusters by using network optimization, introduce reinforcement learning to optimize supply chain scheduling, optimize the industrial layout plan, and output a decision support plan; S5. Use ArcEngine to render candidate site selection points, display the site selection suitability in the form of a heat map and a spatial distribution map, and allow users to adjust the site selection weight parameters, update the site selection results in real time, and combine Web GIS technology to interactively select a decision support plan online.

2. The intelligent project site selection decision support method based on ArcEngine according to claim 1, wherein The specific implementation process of the data cleaning mechanism is as follows: S21. Repair missing values by combining multi-layer perceptron interpolation method and Bayesian interpolation: Construct an interpolation model based on the multi-layer perceptron (MLP) to predict the missing value X using other complete variables Y missing : X missing = f(Y; θ) + ε; In the formula, f(Y; θ) represents the mapping function trained based on MLP; θ represents the neural network parameters; Y represents the complete input variables; ε represents the noise term; Bayesian interpolation: where X missing represents the missing value; X observed represents the observed data; P(X missing ) represents the prior probability of the missing value; P(X observed |X missing ) represents the likelihood function; P(X missing |X observed ) represents the posterior probability; Calculate the optimal interpolation through maximum a posteriori (MAP) estimation Among them, represents the optimal missing value obtained by maximum a posteriori estimation; S22. Adopt the method of adaptive density clustering and random forest correction to detect and correct outliers; S23. Unify the coordinate system of the site selection data and use the adjacency matrix method to verify topological consistency; S24. Use the Z-score standardization method for standardization processing to obtain standardized data.

3. The intelligent project site selection decision support method based on ArcEngine according to claim 2, wherein, The correction process of correcting outliers is as follows: S31. Detect outliers based on adaptive density clustering, identify outlier data points with too low local density, and define the local density as follows: Where ρ i represents the density of data point i; d ij represents the Euclidean distance between data point i and data point j; h' represents the density estimation window size; K(d ij , h') represents the Gaussian kernel function; X ik and X jk represent the coordinate values of data point i and data point j in the k-th dimension respectively; e represents the natural constant; S32. The determination condition for the abnormal point is: ρ i <μ ρ -λ'σ ρ , and the abnormal value X is output accordingly outlier ; Among them, μ ρ represents the mean value of the density; σ ρ represents the standard deviation of the density; λ' represents the threshold for controlling outliers; S33. Correct the detected outlier value X outlier and use the random forest model to predict its correct value: where X correctd represents the corrected data value; T i (Y) represents the predicted value of the i-th decision tree; M represents the number of decision trees.

4. A method for intelligent project site selection decision support based on ArcEngine according to claim 1, characterized in that, The generation process of the standard verification code is as follows: S41. Select a random integer θ∈[1, q-1], and calculate the first-level meta-code eⅠ = (Bm×θ mod p) mod q; Among them, q is a predefined prime number of 160 bits; p is a predefined prime number of 1024 bits, and p - 1 is an integer multiple of q; Bm is the meta-code benchmark, Bm = [e×(p - 1) / q] mod p; e represents the predefined element in the group Z p * the previous element; S42. Convert the standardized data data into a binary string datab; S43. Calculate the second-level meta-code eⅡ = {(θ - 1)×[H(datab)+ec×eⅠ]} mod q; Among them, ec is a predefined meta-code generation element; H is a predefined hash function; S44. Generate the standard verification code CS = (eⅠ, eⅡ).

5. The intelligent project site selection decision support method based on ArcEngine according to claim 4, characterized in that The verification process of verifying data compliance based on the standard verification code is as follows: S51. Calculate the auxiliary meta-code audit element Aea = (eⅡ - 1) mod q; S52. Convert the standardized data data into a binary string datac; S53. Calculate the following meta-code audit elements: The first-level meta-code audit element eaⅠ = [H(datac)×Aea] mod q; The secondary meta-code verification element eaⅡ = (eⅠ × Aea) mod q; Where, H is a predefined hash function; S54. Calculate the reference code CB = (Bm × eaⅠ × ep × eaⅡ mod p) mod q; Where, ep represents a predefined meta-code parsing element; S55. If CB = eⅠ, the data data meets the standard; otherwise, an alarm is given immediately.

6. The intelligent project site selection decision support method based on ArcEngine according to claim 1, characterized in that The sorting process of site selection sorting is as follows: S61. Build the spatial data model of GIS, including vector model and raster model: Vector data modeling, stored in the form of points, lines, and surfaces: P i =(x i , y i , A i , C i , E i ); Among them, P i represents the set of attributes of region i; (x i , y i ) represents the geographical coordinates of region i; A i represents the developable land area; C i represents the coal resource reserve; E i represents the environmental suitability score. Raster data modeling: Represented by raster data, each raster contains a spatial feature value; S62. Use the network analysis method to calculate the resource accessibility scores of different candidate sites; Wherein, R i represents the resource reachability score of candidate area i; W j represents the importance weight of preset resource point j; d ij represents the traffic distance from candidate area i to resource point j. S63. Use buffer analysis to evaluate the environmental suitability; Where, E i represents the environmental suitability score of area i; d ik represents the distance from area i to environmental sensitive point k; S k represents the environmental importance score of the preset sensitive point k; α represents the distance attenuation index; S64. Calculate the comprehensive suitability score based on the analytic hierarchy process + GIS weighted overlay analysis: S i = w1R i + w2E2 + w3Cost i ; Wherein, S i represents the comprehensive suitability score of area i; R i represents the resource accessibility score; E i represents the environmental suitability score; Cost i represents the estimated construction cost score; w1, w2, and w3 represent the preset weights; S65. Sort the comprehensive suitability scores of all candidate areas and select the Top-K suitable areas; Among them, represents the site selection scheme; Ω represents the set of all candidate areas; S66. Output the site selection ranking list List.

7. A method for intelligent project site selection decision support based on ArcEngine according to claim 6, characterized in that, The implementation process of the multi-objective optimization algorithm is as follows: S71. Initialize the population: Use the site selection ranking list List as the initial population {P1, P2, …, P K}; S72, Non-dominated sorting: For each candidate solution P in the population i perform multi-objective function evaluation, calculate its performance in all objective functions, sort the candidate solutions according to the domination relationship, and form multiple non-dominated levels, i.e., the Pareto frontiers; S73. Crowding degree calculation: Calculate the crowding degree distance for each candidate solution in each non-dominated layer; S74. Selection, crossover and mutation: Use the crowding degree sorting and roulette wheel selection mechanism to select parent individuals from the current population, and perform crossover and mutation operations to generate offspring populations; S75. Population merging and new generation generation: Merge the parent and offspring populations, re-perform non-dominated sorting and crowding degree calculation, and select the top K excellent candidate solutions to form a new generation population; S76. Convergence and stopping conditions: When the preset number of iterations is reached or the change of the optimal solution tends to be stable, the algorithm terminates, and the final Pareto optimal solution set is output, that is, the site selection set List' is output.

8. A method for intelligent project site selection decision support based on ArcEngine according to claim 7, characterized in that, The objective function of the multi-objective intelligent optimization algorithm is: maxF(P i ) = (f1(P i ), f2(P i ), f3(P i )); f1(P i ) = ∑w k R ik ; f2(P i ) = -∑w m E im ; f3(P i ) = -Cost i ; Wherein, P i represents the i-th candidate site; f1(P i ) represents the maximization of resource accessibility; f2(P i ) represents the minimization of environmental impact; f3(P i ) represents the minimization of cost; R ik represents the score of the i-th site selection point on the k-th resource factor; w k represents the resource factor weight; E im represents the score of the influence of the i-th site selection point on the m-th environmental factor; w m represents the environmental factor weight; Cost i represents the comprehensive cost of the i-th site selection point.

9. The intelligent project site selection decision support method based on ArcEngine according to claim 1, wherein The optimization process of the industrial chain optimization model is as follows: S91. Establish a data model of the coal chemical industry chain, clarify the relationship between industrial nodes, and evaluate the comprehensive benefits for each node. The industrial chain is represented by a weighted directed graph G: G = (N, E, W); Where, N represents the key facilities on the industrial chain; E represents the industrial flow path; W represents the weight of the edge; The comprehensive benefit of each node is expressed by the following formula: U i = βRe i - γEe i + λPe i ; Among them, U i represents the comprehensive benefit of any node U i ; Re i represents the resource utilization rate; Ee i represents the environmental impact, Pe i represents the economic benefit; β, γ, λ represent the weight coefficients; S92. Model the game among the government, enterprises and the market in the industrial layout, and solve the optimal strategy through the Nash equilibrium; S93. Use the network optimization algorithm to reflect the synergy effect among enterprises; Industrial cluster network model: H = (V, L); The synergy benefit function is as follows: Among them, V represents enterprise nodes, and L represents the cooperation relationship between enterprises; A ij indicates whether enterprises i and j cooperate; W ij is the cooperation benefit; S94. Dynamic supply chain scheduling. The reinforcement learning algorithm is used to optimize the supply chain scheduling and adjust each link of the supply chain. The supply chain optimization objective is: min(C trans +C storage -R delivery ); Among them, C trans is the transportation cost; C storage is the inventory cost; R delivery is the revenue from on-time delivery; Using reinforcement learning to optimize the scheduling policy, the reward function for state St and action At is: R t = -(C trans + C storage ) + εD t ; Among them, D t is the on-time delivery rate; ε is the adjustment coefficient; S95. Output the decision support plan.

10. An intelligent project site selection decision support system based on ArcEngine, which is used to execute an intelligent project site selection decision support method according to any one of claims 1 to 9, characterized in that, Including: Data collection and fusion module, used to collect site selection data and manage and fuse geographic information data using ArcEngine; Data preprocessing and standardization module, used to preprocess the collected site selection data; Intelligent analysis and site selection optimization module, used to perform spatial analysis based on GIS, build a site selection evaluation model, and optimize the site selection plan based on the multi-objective optimization algorithm; Industrial layout decision support module, used to score and evaluate candidate site selection points in combination with the multi-criteria decision analysis method, and perform industrial matching degree analysis in combination with government policies, market demands and resource distributions, and then generate a site selection plan; Visualization and Interaction Module, which uses ArcEngine and Web GIS technologies to provide map visualization display, heat map analysis, and site selection simulation functions, and also provides the capabilities of interactive data query, dynamic adjustment of site selection parameters, and comparison of different site selection schemes.

Images