Construction project land-saving optimization method and system oriented to resource saving and efficient utilization

Abstract

The invention relates to the technical field of natural resource saving, in particular to a construction project land-saving optimization method and system oriented to resource saving and efficient utilization, and provides a three-level index case library construction and dynamic weight matching technology. The method comprises the following steps: firstly, performing multi-level classification on historical cases based on industry types, regional conditions and construction scales, and extracting standardized land use data (L), unit capacity (E) and land saving effects (S) to construct a case library; secondly, dynamically allocating weight coefficients of industries, regions and scales through an analytic hierarchy process (AHP), calculating case similarity in combination with a weighted Euclidean distance, and accurately screening matched cases; according to the method, the land use nuclear reduction rate reaches 26.48% in projects such as a Minjiang old wood hole avionics hub project, and the intensive utilization level of land resources is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of natural resource conservation, and in particular to a land-saving optimization method and system for construction projects oriented to resource conservation and efficient utilization. Background Art

[0002] Currently, land conservation evaluation for construction projects relies primarily on manual experience and comparisons of static indicators. This leads to core issues such as crude case library construction, a lack of data standardization, rigid similarity calculation models, and a disconnect between optimization solutions and industry design specifications. Existing technologies often classify cases using a single dimension (such as industry type) without establishing a multi-level indexing system encompassing "industry-terrain-scale." This leads to the misuse of dissimilar data when matching cases. For example, a mountainous highway project could be incorrectly matched with a plains case, resulting in a land reduction rate of less than 5%. Furthermore, data standardization is insufficient, with key parameters such as land scale and production capacity indicators not standardized and noise removed. In particular, there is a lack of effective analytical tools for unstructured data (such as the text of the land conservation chapter), resulting in low data credibility.

[0003] At the technical model level, existing methods mostly use basic algorithms such as Euclidean distance or cosine similarity, and do not introduce a dynamic weight allocation mechanism. For example, the impact of production capacity scale on land use in mining projects is significantly higher than building density, and the traditional model's equal weighting of features causes the optimization results to deviate from actual needs. What's more serious is that existing technologies ignore mandatory industry design specifications (such as safety spacing for chemical projects and navigation standards for navigating power hubs) and simply aim to reduce the scale of land use, which may cause safety hazards or functional defects. Taking the Panzhihua Honggenan Mine Project as an example, due to the lack of a detailed case library of "vanadium-titanium magnetite-southwestern mountainous areas-extra-large projects", the traditional method mistakenly referred to the data of small and medium-sized mines in shallow hilly areas, limiting the optimization effect; This invention specifically proposes an intelligent optimization system that integrates multi-source data governance, dynamic weight allocation and probabilistic cloud models. By constructing a three-level index case library and defining fuzzy membership functions and constraint rule encoding, it achieves a technological leap from experience-driven to data-driven. In empirical projects such as the Minjiang Laomukong Navigation and Hydropower Project, the land reduction rate has increased to 26.48%, providing a quantifiable and reusable technical path for resolving the contradiction between land resource protection and development. Summary of the Invention

[0004] The purpose of the present invention is to provide a land-saving optimization method and system for construction projects aimed at resource conservation and efficient utilization. By constructing a three-level case library of industry, terrain and scale, using the hierarchical analysis method to dynamically assign weights to match similar cases, and combining land use constraints to generate a multi-objective optimization plan, efficient utilization of land resources and compliant land saving can be achieved.

[0005] In order to achieve the above technical objectives and the above technical effects, the present invention is implemented through the following technical solutions: A land saving optimization method for construction projects oriented towards resource conservation and efficient utilization comprises the following steps: S1: Construct a multi-level indexed land-saving case database: Classify historical construction project cases into three levels based on industry type, terrain conditions, and construction scale, extract the total land use data (L), unit production capacity data (E), and land-saving effect index (S) of each case, and generate a standardized case vector; S2: Determine the land-saving constraints of the target project: Based on the industry type of the target project, the upper limit of land use indicators in the relevant industry design specifications and the minimum land size to ensure production capacity; S3: Data preprocessing and fuzzification: Normalize the land use data and production capacity data of the target project to eliminate dimensional differences and generate input feature vectors; S4: Dynamic weighted similarity matching: The analytic hierarchy process (AHP) is used to assign weight coefficients for industry type, terrain conditions, and construction scale, and the similarity between the target project and the cases in the case library is calculated using weighted Euclidean distance; S5: Generation of multi-objective optimization model: With the optimization goals of minimizing land use scale and maximizing production efficiency, a land-saving optimization plan is generated in combination with constraints. The plan includes functional zoning adjustment suggestions and process flow optimization paths.

[0006] Furthermore, step S1 includes the following sub-steps: S1.1: The construction project to be queried is input into the land-saving case database. Using a multi-level indexing method, the same industry and the same sub-industry as the candidate construction project are found in the land-saving case database;

[0007] in, Indicates the industry type of the database; Indicates that there are already cases in the industry type; represents the K-th industry; Represents a specific case in the system case library.

[0008] S1.2: Find cases with similar characteristics in the sub-industries of the construction project land saving case database;

[0009] in, The characteristics of the construction project case being queried; The case features already exist in the current construction project land saving case database; is the dot product; is the Euclidean distance.

[0010] At the same time, in order to find similar cases more conveniently and quickly, a weight is given to the characteristic value of the alternative construction project case.

[0011]

[0012] Among them, wi is the weight of the i-th feature.

[0013] Furthermore, the step S2 includes: The first values to be set are L for land area, C for production capacity, and S for land-saving optimization implementation effect.

[0014] Data indicating land occupied by construction projects; Indicates unit production scale data; Represents the optimization result of the land-saving scheme.

[0015] The function of the optimization objective is, It represents the worst land-saving efficiency; It represents the best land-saving effect.

[0016] Then the mathematical logic of this part is, ,

[0017] Secondly, some industries have construction land scale standards. Even industries that do not have construction land scale standards will have design standards. Therefore, the land scale of construction projects cannot be continuously optimized, and there will be a constraint on land saving.

[0018] , Represents the upper limit of the total production scale and land area of the construction project.

[0019] , It represents the minimum land size required for the total production scale of the construction project and is the land size that must be guaranteed.

[0020] 、 The logical relationship between the two is that the constraint The total land use cannot exceed the upper limit ; The unit land area cannot be less than the minimum area required to ensure the production capacity .

[0021] Furthermore, step S3 includes the following sub-steps: S3.1: Perform data clarification work, standardize the candidate case data, remove the dimension, and process all between.

[0022]

[0023]

[0024]

[0025] is the standardized data of land occupied by construction project i; is the standardized unit production scale data of construction project i; It is the land-saving effect data of the construction project i after standardization.

[0026] S3.2: Using fuzzy theory through function processing

[0027] .

[0028] Specifically, we use product calculation, for example

[0029] Furthermore, the step S4 includes: using the analytic hierarchy process to determine weights for construction projects in different industries and terrains.

[0030] Establish a judgment matrix based on pairwise comparisons, and use the quantitative relative importance to compare the importance of the i-th element and the j-th element relative to a factor in the previous layer. To express, assuming that there are n elements involved in the comparison, the matrix is expressed as =(a ij ) n×n If evidence , where a ij >0,a ij = This is called proof and counter proof. Use fuzzy cluster analysis or multi-level fuzzy reasoning;

[0031] It is based on expected value; is a probability distribution; The big data cloud model in the land saving case library is

[0032] Furthermore, the step S5 includes the following sub-steps: Based on the above, start reasoning by similarity

[0033]

[0034] Sort, select (L j , E j , S j ) to make inferences and finally get the result of inference

[0035] R is the final reasoning result.

[0036] In another area, the present invention proposes a land conservation optimization system for construction projects, oriented towards resource conservation and efficient utilization. This system achieves efficient land resource utilization by constructing a multi-level indexed case library, a dynamic weight allocation mechanism, and an intelligent optimization model. The system comprises a multi-level case library construction module, a constraint analysis module, a data preprocessing and fuzzification module, a dynamic weighted similarity matching module, a multi-objective optimization generation module, and a solution execution feedback module, operating collaboratively to form a closed-loop management process from data collection to solution implementation.

[0037] Beneficial effects of the present invention: The present invention significantly improves the case matching accuracy and solution adaptability through the collaborative design of a multi-level indexing system and dynamic weight allocation. Traditional methods are prone to mismatching due to single-dimensional classification, such as misapplying plain mine cases to mountainous projects. The three-level indexing system of "industry-topography-scale" constructed by the present invention refines case classification layer by layer based on industry type (such as mining, transportation), terrain conditions (such as topography, climate) and construction scale (such as large, medium). Taking the Panzhihua Honggenan Mine Project as an example, by accurately locating the "vanadium titanium magnetite-southwest mountainous area-extra-large" case set and combining it with the hierarchical analysis method to dynamically allocate feature weights, the case matching error rate is reduced and the land reduction rate is improved. This combination of fine classification and dynamic weighting enables the system to quickly filter out non-similar interference cases, while quantifying feature differences through mathematical modeling to ensure a high degree of adaptability between the solution and the actual needs of the project.

[0038] This invention achieves a dual improvement in safety and economy by converting mandatory clauses in industry design specifications into computable mathematical constraints and coupling them with a multi-objective optimization model. Taking the Minjiang River Laomukong Navigation and Power Hub Project as an example, the system automatically parses the navigation standards in the "Inland Navigation Standard" (GB50139-2014), converting requirements such as lock spacing and channel width into hard constraints such as land use limits and minimum functional area requirements. Through a multi-objective optimization algorithm, the functional zoning layout and process flow were reorganized, reducing land use by 26.48% and increasing unit land production efficiency by 19.7%. This deep integration of constraint coding and optimization models not only avoids the risk of omissions in manual review, but also automatically explores design boundaries through algorithms, significantly improving the level of intensive land use.

[0039] This invention eliminates dimensional differences through normalization and maps terrain features (such as slope and precipitation) into fuzzy membership, ultimately generating a standardized input vector. This combination of technologies reduces the proportion of noise data in the case library from 17.3% to 2.1%, compresses the fluctuation range of land use prediction results by 68%, and effectively avoids the risk of deviation from the plan due to ambiguous textual expressions or complex terrain features. In addition, the closed-loop feedback and dynamic learning mechanism constructed by this invention drives continuous iterative optimization of the system through blockchain evidence storage and case library self-update functions. With each new project case, the system automatically triggers retraining of the weight model. The design of a learning rate that dynamically decays with the number of cases ensures the rapid absorption of new data while avoiding model overfitting. This enables the system to continuously adapt to complex scenarios such as new processes and policy changes, forming a self-evolving capability, and providing dynamic and stable technical support for the full life cycle management of land resources.

[0040] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. DETAILED DESCRIPTION

[0041] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention. Example 1

[0042] The land saving optimization method for construction projects oriented to resource conservation and efficient utilization described in this embodiment includes the following steps: S1: Classify the construction projects by industry type and construction terrain, collecting industry-specific technical design specifications, construction specifications, production scale, functional zoning, investment intensity, volume ratio, building density, project process flow, construction technology and techniques, land area, and detailed data. This data is then used to construct a land-saving case model and a land-saving case database based on the land-saving optimization of the construction projects. S2: Determine the objectives and constraints for land-saving optimization of the construction projects.

[0043] S3: Preprocess and fuzzify the land-saving data of construction projects, and build a big data cloud model using the design type expectation value, design strength and probability distribution function of the industry to which it belongs.

[0044] S4: Calculate the similarity between the current construction project and the land-saving cases in the land-saving case library, select similar land-saving cases based on the described similarity, and obtain case reasoning results by analyzing similar land-saving cases.

[0045] S5: Under the conditions of satisfying the described land-saving optimization objectives and constraints, a multi-dimensional and multi-attribute optimization plan for land saving in construction projects is formulated based on the described case reasoning results and the described big data cloud model.

[0046] S6: Optimize the construction project according to the above multi-dimensional and multi-attribute optimization scheme.

[0047] In this embodiment, step S1 includes: entering the construction project being searched into the land-saving case database, and then using a multi-level indexing approach to find projects in the same industry or sub-industry as the candidate construction project. (For example, if the candidate construction project is a highway, the transportation project should be selected first, and then the highway project should be selected from the transportation project.)

[0048] in, Indicates the industry type of the database; Indicates that there are already cases in the industry type; represents the K-th industry; Represents a specific case in the system case library.

[0049] In this embodiment, step S2 includes searching for cases with similar characteristics in the sub-industry of the construction project land saving case library. (For example, for the same sub-industry, the same construction area or interest, and the same construction scale as in S1, the method of enhancing similarity is adopted.)

[0050] in, The characteristics of the construction project case being queried; The case features already exist in the current construction project land saving case database; is the dot product; is the Euclidean distance.

[0051] At the same time, in order to find similar cases more conveniently and quickly, a weight is given to the characteristic value of the alternative construction project case.

[0052]

[0053] Among them, wi is the weight of the i-th feature.

[0054] In this embodiment, step S3 includes: optimizing land saving for the registered construction project.

[0055] First, there must be an upper limit for construction projects, that is, a certain limit that cannot be exceeded. The first set of values is L for land occupation scale, C for production capacity scale, and S for land conservation optimization implementation effect.

[0056] Data indicating land occupied by construction projects; Indicates unit production scale data; Represents the optimization result of the land-saving scheme.

[0057] The function of the optimization objective is, It represents the worst land-saving efficiency; It represents the best land-saving effect.

[0058] Then the mathematical logic of this part is, ,

[0059] Secondly, some industries have construction land scale standards. Even industries that do not have construction land scale standards will have design standards. Therefore, the land scale of construction projects cannot be continuously optimized, and there will be a constraint on land saving.

[0060] , Represents the upper limit of the total production scale and land area of the construction project.

[0061] , It represents the minimum land size required for the total production scale of the construction project and is the land size that must be guaranteed.

[0062] 、 The logical relationship between the two is that the constraint The total land use cannot exceed the upper limit ; The unit land area cannot be less than the minimum area required to ensure the production capacity .

[0063] In this embodiment, step S4 includes: starting the data clearing work, standardizing the candidate case data, removing the dimension, and processing all between.

[0064]

[0065]

[0066]

[0067] is the standardized data of land occupied by construction project i; is the standardized unit production scale data of construction project i; It is the land-saving effect data of the construction project i after standardization.

[0068] Through function processing, using fuzzy theory

[0069] .

[0070] Specifically, we use product calculation, for example

[0071] In this embodiment, step S5 includes: weights are manually set, and weights can be different for different industries and different terrains. Therefore, the analytic hierarchy process is used here to determine weights for construction projects in different industries and different terrains.

[0072] Establish a judgment matrix based on pairwise comparisons, and use the quantitative relative importance to compare the importance of the i-th element and the j-th element relative to a factor in the previous layer. To express, assuming that there are n elements involved in the comparison, the matrix is expressed as =(a ij ) n×n If evidence , where a ij >0,a ij = It's called proof and counter proof.

[0073] Use fuzzy cluster analysis or multi-level fuzzy reasoning.

[0074] It is based on expected value; is a probability distribution; The big data cloud model in the land saving case library is

[0075] In this embodiment, step S6 includes: based on the above, start reasoning, start by similarity

[0076]

[0077] Sort, select (L j , E j , S j ) to make inferences and finally get the result of inference

[0078] R is the final reasoning result. Example 2

[0079] The land conservation optimization system for construction projects described in this embodiment, based on multi-source data fusion and dynamic optimization decision-making, achieves efficient land resource utilization by building a multi-level index case library, a dynamic weight allocation mechanism, and an intelligent optimization model. The system is comprised of a multi-level case library construction module, a constraint analysis module, a data preprocessing and fuzzification module, a dynamic weight similarity matching module, a multi-objective optimization generation module, and a solution execution feedback module, operating collaboratively to form a closed-loop management process from data collection to solution implementation. This includes: Multi-level case library building module Data on industry type, terrain characteristics, construction scale, land use indicators (L), unit production capacity (E), and land conservation effects (S) is collected from natural resource databases and historical projects to construct a three-level index structure: the first level divides industry categories (such as mining and transportation) according to the National Economic Industry Classification; the second level subdivides terrain by terrain conditions (mountainous areas, plains) and climate type; and the third level categorizes investment scale (large, medium, and small). Semantic parsing is used to extract key parameters such as floor area ratio and process flow from unstructured data (such as design documents and land conservation chapters). This data is then used to generate standardized case vectors, which support dynamic updates and categorized storage. For example, mining projects are indexed by the tags "mine - southwestern mountainous area - large" to ensure subsequent matching accuracy.

[0080] Constraint parsing module According to the industry type of the target project, the module automatically links the industry design specification database (such as "Highway Land Use Standards" and "Inland Navigation Standards") to extract the upper limit of land use scale. and minimum land requirements to ensure production capacity The qualitative constraints in the specification (such as the safety distance of chemical projects and the navigation width of navigation and power hubs) are encoded as mathematical constraints and embedded in the optimization model. The total land use exceeds the upper limit ; The unit land area is less than the minimum area required to ensure production capacity , the system triggers an early warning and recommends adjusting the process or production capacity to avoid the plan being unfeasible.

[0081] Data preprocessing and fuzzification module The target project's land use size L, production capacity data E, and land-saving effect data S are normalized to eliminate dimensional differences and uniformly mapped to the [0, 1] range. Topographic features are processed using fuzzy membership functions. For example, slope parameters are classified into "plain," "shallow hill," "deep hill," and "mountainous" categories, and their membership degrees are quantified. Unstructured text is parsed into a land-saving potential score, and a probabilistic cloud model is used to address parameter uncertainty, such as the fluctuations in land-saving effects of the "continuous mining" process in a mining project.

[0082] Dynamic weighted similarity matching module The Analytic Hierarchy Process (AHP) dynamically assigns weight coefficients (wi) for industry, terrain, and scale, constructs a judgment matrix, and passes a consistency check. The similarity between the target project and cases in the case library is calculated based on weighted Euclidean distance, and the case with the highest similarity is selected as the optimization benchmark. The weight assignment supports adaptive adjustment to ensure that the matching results meet actual needs.

[0083] Multi-objective optimization generation module With the dual objectives of minimizing the land use scale L and maximizing the land saving effect data S, combined with the constraints 、 Generate an optimization plan. The model outputs recommendations for functional zoning reorganization (e.g., merging storage and production areas), process optimization paths (vertical stacking instead of flat layouts), and technological alternatives (underground storage, prefabricated buildings). The optimization process strictly adheres to mandatory industry regulations. For example, in navigation and power hub projects, the spacing between ship locks must meet navigation standards to avoid functional deficiencies caused by simply compressing land.

[0084] Program execution feedback module After the optimization plan is implemented, the system verifies its compliance and calculates the land reduction rate. For example, the Minjiang Laomukong Navigation and Hydropower Project reduced land use by 26.48%. The optimized project data is encrypted and stored on the blockchain and updated to the case library according to a three-level indexing rule, forming a closed-loop feedback mechanism. This module continuously iterates the case library data and weighting model, improving the system's adaptability to complex scenarios. For example, the addition of a southwestern mountainous mining case significantly improved the matching accuracy of subsequent similar projects. Example 3

[0085] Taking the land saving of the Honggenan Vanadium Titanium Magnetite Development and Utilization Project (mining and dressing tail project) in Yanbian County, Sichuan Province as an example, the cases of Yanbian County Panxi Hongge Mining Co., Ltd., Yanbian County Zhongtian Mining Co., Ltd., Longbai Sichuan Mining and Metallurgy Co., Ltd., Sichuan Jiangxi Copper Rare Earth Co., Ltd., Yanbian County Detian Mining Co., Ltd., and Panzhihua New Baima Mining Co., Ltd. of Panzhihua Iron and Steel Group, which have the same construction content, comparable production or construction scale, similar production process or technical standards, and similar construction conditions in the Panzhihua area, were optimized. Compared with the land scale in the feasibility study stage, the final land scale submitted for review was reduced from the initial 2097.8422 hectares to the current 1804.9254 hectares, a reduction of 293.1872 hectares, and a reduction ratio of 13.98%.

[0086] Taking the Minjiang River Laomukong Navigation and Hydropower Project as an example, three cascade hydropower stations in the Minjiang River basin—the Dongfengyan Navigation and Hydropower Project, the Longxikou Navigation and Hydropower Project, and the Qianwei Navigation and Hydropower Project—were selected for optimization. The final proposed land area was reduced from 2,271.526 hectares before the land conservation evaluation to the current 2,218.0683 hectares, saving a total of 53.4577 hectares, a savings rate of 2.41%. Specifically, the proposed land area of the hub area was reduced from 309.2783 hectares before the land conservation evaluation to the current 227.3628 hectares, saving a total of 81.9155 hectares, a savings rate of 26.48%.

[0087] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A land saving optimization method for construction projects oriented towards resource conservation and efficient utilization, characterized in that: The following steps are involved: S1: Construct a multi-level indexed land-saving case database, classify historical construction project cases into three levels according to industry type, terrain conditions, and construction scale, extract the total land use data L, unit production capacity data E, and land-saving effect index S of each case, and generate a standardized case vector; S2: Determine the land-saving constraints of the target project, and associate the upper limit of land use indicators in the industry design specifications and the minimum land size to ensure production capacity based on the industry type of the target project; S3: Normalize the land use data and production capacity data of the target project to eliminate dimensional differences, generate input feature vectors, and perform fuzzification using fuzzy membership functions; S4: The analytic hierarchy process (AHP) is used to assign weight coefficients to industry types, terrain conditions, and construction scales, and the similarity between the target project and the cases in the case library is calculated using weighted Euclidean distance. S5: Taking minimizing the land use scale and maximizing the land saving effect index as the optimization goal, a land saving optimization plan is generated in combination with the constraint conditions. The plan includes functional zoning adjustment suggestions and process flow optimization paths.

2. The method according to claim 1, wherein The step S1 includes the following sub-steps: S1.1: Classify the case library by industry type, including but not limited to mining, transportation, and water conservancy projects; S1.2: Under the same industry type, a secondary classification is made based on terrain conditions, including plains, shallow hills, deep hills, mountains, and climate types; S1.3: Under the same terrain conditions, the construction scale is classified into three levels, including but not limited to extra-large, large, medium and small projects; S1.4: Extract the total land use data, unit production capacity data, and land-saving effect indicators for each case to form a standardized case vector.

3. The method according to claim 1, wherein The constraints in step S2 include: S2.1: Define the total land size limit based on industry design specifications , constraints The total land use cannot exceed the upper limit ; S2.2: Define the minimum unit land area based on the guaranteed production capacity scale , constraints The unit land area cannot be less than the minimum area required to ensure the production capacity ; S2.3: Encode mandatory provisions in industry design specifications as constraints, including safety spacing, industry standards, or process flow layout requirements.

4. The method according to claim 1, wherein The step S3 comprises: S3.1: Normalize the land use data, production capacity data, and land saving effect data of the target project to eliminate dimensional differences; S3.2: Fuzzy processing of regional characteristic parameters is performed through fuzzy membership function.

5. The method according to claim 1, wherein The specific steps of dynamic weight allocation in step S4 are: S4.1: Construct a judgment matrix for the analytic hierarchy process, compare the importance of industry types, terrain conditions, and construction scales, and generate weight coefficients; S4.2: Calculate the similarity between the target project and the case database case using the weighted Euclidean distance formula. The formula is: Among them, wi is the weight of the i-th feature, and are the characteristic values of the target project and the case base cases, respectively.

6. The method according to claim 1, wherein The generation of the multi-objective optimization model in step S5 includes: S5.1: Minimizing the total land use scale and maximizing the land-saving effect index are the dual goals; S5.2: Generate optimization solutions based on the constraints, including functional zoning mergers, vertical stacking processes, or underground space utilization.

7. The method according to claim 1, wherein The step S6 is also included: S6: Feedback the implementation results of the optimization plan to the land-saving case library, update the case data and dynamically adjust the weight coefficient to form a closed-loop learning mechanism.

8. A land-saving optimization system for implementing the method according to any one of claims 1 to 7, characterized in that: include: Multi-level case library building module Data on industry types, regional characteristics, construction scale, land use indicators, unit production capacity, and land conservation effects are collected from natural resource databases and historical projects to construct a three-level index structure: the first level divides industry categories according to the National Economic Industry Classification, the second level subdivides regions by terrain conditions and climate types, and the third level is classified by investment scale. Unstructured data is semantically parsed to extract key parameters such as floor area ratio and process flow, generating standardized case vectors that support dynamic updates and classified storage. Constraint parsing module According to the industry type of the target project, automatically link the industry design specification database to extract the upper limit of land use scale and minimum land requirements to ensure production capacity ; The qualitative constraints in the specification are encoded as mathematical constraints and embedded in the optimization model; when the constraints are detected The total land use exceeds the upper limit ; The unit land area is less than the minimum area required to ensure production capacity , the system triggers an early warning and recommends adjusting the process or production capacity to avoid the plan being unfeasible; Data preprocessing and fuzzification module The target project's land use scale L, production capacity data E, and land-saving effect data S are normalized to eliminate dimensional differences and uniformly mapped to the [0, 1] interval. Regional characteristics are processed using fuzzy membership functions. Unstructured text is parsed into a land-saving potential score, and a probabilistic cloud model is used to handle parameter uncertainty. Dynamic weighted similarity matching module The analytic hierarchy process (AHP) is used to dynamically assign weight coefficients wi for industry, region, and scale, construct a judgment matrix, and pass a consistency test. The similarity between the target project and the cases in the case library is calculated based on the weighted Euclidean distance, and the case with the highest similarity is selected as the optimization benchmark. Weight distribution supports adaptive adjustment to ensure that the matching results meet actual needs; Multi-objective optimization generation module With the dual objectives of minimizing the land use scale L and maximizing the land saving effect data S, combined with the constraints 、 Generate optimization solutions; The model outputs functional zoning reorganization suggestions, process optimization paths, and technical alternatives. The optimization process strictly adheres to industry mandatory regulations to avoid functional defects caused by simply compressing land use. Program execution feedback module After the optimization plan is implemented, the system verifies its compliance and calculates the land reduction rate; the optimized project data is encrypted and stored on the blockchain and updated to the case library according to the three-level index rules, forming a closed-loop feedback mechanism; the case library data and weight model are continuously iterated to improve the system's adaptability to complex scenarios.

Images