A method and system for constructing a rule for collaborative zoning of mineral resource development and ecological protection

Abstract

The application discloses a kind of mineral resources exploitation and ecological protection collaborative zoning rule construction method and system.The method includes: collecting the multi-source basic data of target area, extracts resource endowment, ecological constraint feature and spatial constraint feature, and obtains spatial feature standard set by normalization;Policy planning text is processed by natural language, and the control provisions of prohibition and restriction are extracted, which are converted into hard rule set;With historical zoning sample as training data, combined with spatial feature standard set, the initial decision rule set is obtained by machine learning algorithm training;Based on resource dimension and ecological dimension, develop utility measure and protection utility measure are constructed, rule scoring function is constructed by collaborative weighting, and double-target conflict resolution evaluation model is formed;The initial decision rule set is substituted into the evaluation model score, and the rules with low score or imbalance are eliminated.The optimized soft rule set is fused with the hard rule set to form a collaborative zoning rule library.The application automatically processes the development and protection conflict through collaborative scoring mechanism, and outputs a scientific and reasonable mineral resources collaborative zoning scheme.

Description

Technical Field

[0001] This invention relates to the field of computer-aided decision-making and spatial planning technology, and in particular to a method and system for constructing zoning rules for coordinated mineral resource development and ecological protection. Background Technology

[0002] In territorial spatial planning and mineral resource planning, the scientific delineation of functional zones such as "key development zones," "restricted development zones," and "prohibited development zones" is a core challenge. There is an inherent contradiction between mineral resource development and ecological protection: on the one hand, national economic development requires a stable supply of mineral resources, necessitating the designation of key development zones to ensure resource security; on the other hand, ecological civilization construction demands strict protection of ecological space, requiring the designation of prohibited and restricted development zones. Finding a balance between these two aspects and achieving synergy between development and protection is a critical issue that urgently needs to be addressed in the field of spatial planning.

[0003] The existing technology has the following drawbacks: Highly subjective: Traditional methods rely heavily on manually establishing partitioning rules, lacking objective data support.

[0004] Conflicts in regulations: Delineating development zones solely based on resource endowments often conflicts with ecological protection red lines; conversely, emphasizing protection at the expense of national strategic resource security needs. Existing automated zoning methods lack a quantitative mechanism to resolve the conflict between the dual objectives of "development and protection."

[0005] Sample scarcity: Compared to big data scenarios such as image recognition, the number of samples in typical "key mineral development zones" across the country is limited, which cannot support the training of large-scale deep neural network models.

[0006] Unstructured policy texts: A large number of national and provincial plans and protection policies exist in the form of unstructured texts, which are difficult to transform into computer-executable constraints. Currently, they mostly rely on manual interpretation by experts, which is inefficient and prone to omissions.

[0007] Therefore, providing a method and system for constructing zoning rules for the coordinated development of mineral resources and ecological protection to solve the above problems is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a method and system for constructing collaborative zoning rules for mineral resource development and ecological protection. This method combines rigid policy constraints with flexible patterns discovered through data mining, and uses a collaborative scoring mechanism to automatically handle the contradiction between development and protection, thereby outputting a scientific and reasonable collaborative zoning scheme for mineral resources.

[0009] This invention provides a method for constructing zoning rules for coordinated mineral resource development and ecological protection, comprising the following steps: Collect multi-source basic data of the target area, extract resource endowment features, ecological constraint features and spatial constraint features from the multi-source basic data, and perform normalization processing to obtain a spatial feature standard set. The multi-source basic data includes mineral resource data, ecological environment data and spatial location data of the target area. Obtain the policy planning text, perform natural language processing on the policy planning text to extract prohibitive and restrictive control clauses, and convert them into a structured set of hard rules; Obtain historical zoning samples, use the historical zoning samples as training data, combine them with the spatial feature standard set, and train them using machine learning algorithms to obtain a preliminary selection decision rule set; Based on the resource and ecological dimensions of the aforementioned spatial feature standard set, a development utility metric for measuring regional development value and a protection utility metric for measuring regional ecological value are constructed respectively. A rule-based scoring function is then constructed based on the synergistic weighting of the development utility metric and the protection utility metric to form a dual-objective conflict resolution evaluation model. The initial decision rule set is substituted into the dual-objective conflict resolution evaluation model. Each rule in the initial decision rule set is scored according to the rule scoring function. Rules with scores below a preset threshold and rules that cause the difference between the development utility measure and the protection utility measure to exceed the tolerance range are eliminated to obtain an optimized soft rule set. The optimized soft rule set is then merged with the hard rule set and sorted by priority to form a collaborative zoning rule library.

[0010] Preferably, the extraction of resource endowment features, ecological constraint features, and spatial constraint features from the multi-source basic data includes the following steps: Using a regular grid as the evaluation unit, the resource endowment characteristics, ecological constraint characteristics, and spatial constraint characteristics of each evaluation unit are extracted from the multi-source basic data through spatial overlay analysis; The resource endowment characteristics include mineral resource reserves and resource grade; The ecological constraint characteristics include the ecological function importance index, the ecological sensitivity index, and vegetation coverage. The spatial constraint features include distance from towns, distance from main transportation routes, and terrain slope, wherein the distance from main transportation routes is used to characterize transportation convenience, and the terrain slope is used to characterize development difficulty.

[0011] Preferably, the step of performing natural language processing on the policy planning text to extract prohibitive and restrictive control clauses and convert them into a structured set of hard rules includes the following steps: Construct a core control dimension dictionary, which includes spatial entity terms, development behavior terms, and constraint operation terms; Based on the core control dimension dictionary, the policy planning text is subjected to part-of-speech tagging and entity recognition to identify the spatial entity words, development behavior words and constraint operation words in the text; Based on predefined space-dominant templates and attribute composite templates, dependency parsing analysis is performed on sentences that identify the space entity words, the development behavior words, and the constraint operation words to extract key components including space entities, constraints, and development behaviors, and these key components are used as the prohibitive and restrictive control clauses. Transform the aforementioned prohibitive and restrictive control clauses into a standardized four-tuple rule structure: ; in, For spatial entity names, For spatial buffer range, For constraint type, Based on the mapping zone type, a candidate hard rule set is obtained; The conflicting rules extracted from different policy planning texts in the candidate hard rule set are resolved according to the principles of priority of legal effect, priority of publication time, or priority of protection, thus forming the structured hard rule set.

[0012] Preferably, the step of using the historical zoning samples as training data, combined with the spatial feature standard set, and training with a machine learning algorithm to obtain the initial selection decision rule set includes the following steps: The collection includes samples from key development zones and resource reserve zones extracted from official planning documents, as well as samples from general development zones obtained through stratified random sampling, to form an initial sample set; The initial sample set is validated by removing samples that conflict with the structured set of hard rules, thus obtaining a validated sample set. The validated sample set is adjusted for class balance using oversampling or undersampling methods to form a training sample set; The spatial location of each sample in the training sample set is spatially superimposed with the spatial feature standard set. The feature vector corresponding to the location of each sample is extracted from the spatial feature standard set. Each sample is assigned the corresponding resource endowment feature, ecological constraint feature and spatial constraint feature to form a training sample set with features. The feature-rich training sample set is input into the decision tree algorithm for training to obtain the decision tree model; Traverse the path from the root node to each leaf node in the decision tree model, take the splitting condition on each path as the rule antecedent, and take the zoning type corresponding to the leaf node as the rule consequent to generate the initial selection decision rule set in IF-THEN form.

[0013] Preferably, the step of constructing a development utility metric for measuring regional development value and a protection utility metric for measuring regional ecological value based on the resource dimension and ecological dimension of the spatial feature standard set includes the following steps: The mineral resource reserves and the resource grade are weighted and aggregated to obtain a comprehensive resource endowment value; Based on the comprehensive value of resource endowment, the distance from the main transportation route, and the terrain slope, a resource development potential index is constructed using a weighted aggregation method to quantify the resource development value of the evaluation unit. The distance from the main transportation route is used to determine the weight of the transportation convenience, and the terrain slope is used to determine the weight of the development difficulty. Based on the ecological function importance index, the ecological sensitivity index, and the vegetation coverage, an ecological constraint intensity index is constructed using a weighted aggregation method to quantify the ecological protection importance of the evaluation unit. For each rule in the preliminary decision rule set, the average value of the resource development potential index of the samples covered by the rule is used as the development utility measure of the rule, and the average value of the ecological constraint intensity index of the samples covered by the rule is used as the protection utility measure of the rule.

[0014] Preferably, the step of constructing a rule-based scoring function based on the collaborative weighting of the development utility measure and the protection utility measure to form a bi-objective conflict resolution evaluation model includes the following steps: For each rule in the initial decision rule set, a rule scoring function is constructed based on the development utility metric and the protection utility metric of that rule; The formula for the rule scoring function is as follows: ; in, Rate the rule for this rule. For the development utility metric described in this rule, The measure of the protective utility of this rule. and These are the co-weighting coefficients for the development utility measure and the protection utility measure, respectively. Conflict penalty item; The conflict penalty item is determined according to the following rules: if the conclusion of the rule is a key development area and the mean value of the ecological sensitivity index of the samples covered by the rule is higher than the preset ecological threshold, then the conflict penalty item is a value greater than the weighted sum of the development utility measure and the protection utility measure of the rule, so that the rule score of the rule is negative; otherwise, the conflict penalty item is zero. The bi-objective conflict resolution evaluation model is constructed from the rule scoring function and is used to quantitatively evaluate each rule in the initial decision rule set.

[0015] Preferably, the collaborative weighting coefficient is determined through the following steps: Retrieve national-level policy texts released within a preset time period; The frequency of occurrence of development-oriented terms and protection-oriented terms were statistically analyzed from the policy text. The development-oriented terms were selected from the constraint operators to represent positive constraints on development; the protection-oriented terms were selected from the constraint operators to represent negative constraints on development. The TF-IDF method is used to calculate the weighted frequency of the development-oriented terms and the frequency of the protection-oriented terms to obtain the weighted frequency of the development-oriented terms and the weighted frequency of the protection-oriented terms for each policy text, which are used as the development-oriented score and the protection-oriented score for each policy text. According to the release time order of the policy texts, a time decay factor is introduced to calculate the weighted average of the development orientation score and the protection orientation score of each policy text. The policy texts released later are given higher weights. The comprehensive development orientation strength and comprehensive protection orientation strength are calculated. The sum of the comprehensive development guidance intensity and the comprehensive protection guidance intensity is used as the normalization benchmark. The proportion of the comprehensive development guidance intensity to the normalization benchmark is used as the synergistic weighting coefficient of the development utility measure. The proportion of the comprehensive protection guidance intensity to the normalization benchmark is used as the synergistic weighting coefficient of the protection utility measure.

[0016] Preferably, the step of merging the optimized soft rule set with the hard rule set and sorting them by priority to form a collaborative zoning rule base includes the following steps: The hard rule set is merged with the optimized soft rule set, and each merged rule is assigned a priority. The rules in the hard rule set have the highest priority, and the rules in the optimized soft rule set are assigned corresponding secondary priorities from high to low according to their rule scores, thus obtaining the initial fused rule set. Redundancy is eliminated from the initial fusion rule set. When two rules in the initial fusion rule set have an inclusion relationship in their preconditions and the same conclusion, the rule with the simpler precondition is retained to obtain the simplified rule set. A standardized data structure is constructed for each rule in the simplified rule set. The data structure includes rule number, priority, preconditions, conclusion type, rule score, and rule source. The data structures of all rules constitute the collaborative zoning rule base.

[0017] Preferably, after constructing the collaborative zoning rule base, the following steps are also included: Obtain the target spatial unit to be judged, and extract its corresponding resource endowment feature, ecological constraint feature and spatial constraint feature from the spatial feature standard set to form the feature vector to be judged; The feature vector to be determined is matched with the rules in the collaborative zoning rule base, and the rules are judged in descending order of priority to determine whether the preconditions of the rules are met. If any of the hard rules are met, the zone type corresponding to that hard rule is directly output, and the matching is terminated. If none of the hard rules are met, the matching degree between the feature vector to be determined and each of the soft rules is calculated. For rules with a matching degree greater than a preset matching threshold, a weighted vote is performed based on the rule score, and the voting result is used as the output partition type. If the matching degree of all the soft rules is not greater than the preset matching threshold, then the preset default division type is output.

[0018] A system for constructing rules for coordinated zoning of mineral resource development and ecological protection includes: The spatial feature acquisition module is used to collect multi-source basic data of the target area, extract resource endowment features, ecological constraint features and spatial constraint features from the multi-source basic data, and perform normalization processing to obtain a spatial feature standard set. The multi-source basic data includes mineral resource data, ecological environment data and spatial location data of the target area. The hard rule extraction module is used to acquire policy planning texts, perform natural language processing on the policy planning texts to extract prohibitive and restrictive control clauses, and convert them into a structured set of hard rules; The soft rule learning module is used to acquire historical zoning samples, use the historical zoning samples as training data, combine them with the spatial feature standard set, and use machine learning algorithms to train the initial selection decision rule set. The evaluation model construction module is used to construct a development utility measure for measuring the development value of a region and a protection utility measure for measuring the ecological value of a region, respectively, based on the resource dimension and ecological dimension of the spatial feature standard set. Based on the synergistic weighting of the development utility measure and the protection utility measure, a rule scoring function is constructed to form a dual-objective conflict resolution evaluation model. The rule optimization and fusion module is used to substitute the initial decision rule set into the dual-objective conflict resolution evaluation model, score each rule in the initial decision rule set according to the rule scoring function, eliminate rules with scores below a preset threshold, and rules that cause the difference between the development utility metric and the protection utility metric to exceed the tolerance range, to obtain an optimized soft rule set, and then merge the optimized soft rule set with the hard rule set, and sort them by priority to form a collaborative zoning rule library.

[0019] Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: Strong synergy: By constructing a dual-objective conflict resolution evaluation model (corresponding to step S4 of claim 1), the contradiction between "development" and "protection" is quantitatively resolved in the rule construction stage, avoiding the one-sidedness of unilateral decision-making and realizing effective synergy between mineral resource development and ecological protection.

[0020] Adapting to small samples: By combining policy text mining (corresponding to claim 3) and machine learning (corresponding to claim 4), highly reliable zoning rules can still be generated even when the sample size is insufficient, thus solving the problem of traditional deep learning's dependence on big data.

[0021] Good interpretability: The final generated IF-THEN rule (corresponding to claim 4) is intuitive in form, easy for decision-makers to understand and implement, meets the needs of government planning and management, and overcomes the shortcomings of the "black box" model that is difficult to interpret.

[0022] Policy text structuring: Through natural language processing and dependency parsing (corresponding to claim 3), unstructured policy planning texts are automatically transformed into a set of structured hard rules, improving the efficiency and completeness of rule extraction.

[0023] Dynamic adaptability: By using a dynamic weight determination method based on policy text quantification (corresponding to claim 7), the collaborative weight coefficients can be automatically adjusted as the national strategic orientation changes, ensuring the timeliness and adaptability of the model.

[0024] In summary, this invention effectively solves the coordination problem between development and protection in mineral resource planning by constructing a dual-objective conflict resolution evaluation model and combining policy text mining and machine learning. It has significant theoretical value and practical significance. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart of the method for constructing zoning rules for coordinated mineral resource development and ecological protection in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the system for constructing collaborative zoning rules for mineral resource development and ecological protection in an embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The embodiments of this invention are written in a progressive manner.

[0029] like Figure 1 As shown, a method for constructing zoning rules for coordinated mineral resource development and ecological protection includes the following steps: S1. Collect multi-source basic data of the target area, extract resource endowment characteristics, ecological constraint characteristics and spatial constraint characteristics from the multi-source basic data, and perform normalization processing to obtain a spatial feature standard set. The multi-source basic data includes mineral resource data, ecological environment data and spatial location data of the target area.

[0030] S2. Obtain the policy planning text, perform natural language processing on the policy planning text to extract the prohibitive and restrictive control clauses, and convert them into a structured set of hard rules.

[0031] S3. Obtain historical zoning samples, use the historical zoning samples as training data, combine them with a spatial feature standard set, and use machine learning algorithms to train the initial selection decision rule set.

[0032] S4. Based on the resource and ecological dimensions of the spatial feature standard set, a development utility measure for measuring the development value of a region and a protection utility measure for measuring the ecological value of a region are constructed respectively. A rule-based scoring function is constructed based on the synergistic weighting of the development utility measure and the protection utility measure to form a dual-objective conflict resolution evaluation model.

[0033] S5. Substitute the initial decision rule set into the dual-objective conflict resolution evaluation model, score each rule in the initial decision rule set according to the rule scoring function, remove rules with scores below the preset threshold, and rules that cause the difference between development utility measurement and protection utility measurement to exceed the tolerance range, to obtain the optimized soft rule set, and merge the optimized soft rule set with the hard rule set, and sort them by priority to form a collaborative zoning rule library.

[0034] The spatial feature standard set refers to a standardized feature dataset after normalization, with each evaluation unit corresponding to a feature vector containing resource endowment features, ecological constraint features, and spatial constraint features. The structured hard rule set refers to a set of rules extracted from policy planning texts, expressed in standardized quadruplets, and possessing mandatory binding force. The preliminary selection decision rule set refers to an initial set of rules learned from historical zoning samples using machine learning algorithms, expressed in IF-THEN form. The dual-objective conflict resolution evaluation model refers to a scoring model built based on development utility measures and protection utility measures, used to quantify the synergistic effect of evaluation rules. The collaborative zoning rule base refers to a rule database for spatial unit discrimination, formed by merging the hard rule set and the optimized soft rule set, and sorting them by priority.

[0035] Example: Construction of Collaborative Zoning Rules for a Mineral Resource-Rich Area The technical solution of the present invention will be fully described below with reference to a specific mineral resource-rich area.

[0036] I. Data Acquisition and Feature Space Construction (Step S1) The following three types of basic data were collected covering the target area: Mineral resource data came from the 2021 version of the National Mineral Deposits Database, acquiring data on 152 mineral deposits in the region, including deposit names, geographical coordinates, resource reserves (10,000 tons), resource grade (%), deposit type, and current utilization status; Ecological and environmental data came from the Resource and Environmental Science and Data Platform of the Chinese Academy of Sciences, acquiring vector data of nature reserve boundaries (including 3 national and 2 provincial nature reserves), remote sensing data of national land use status (30m resolution), vegetation index remote sensing inversion dataset (NDVI), and ecological function importance index and ecological sensitivity index from the China Ecosystem Assessment and Ecological Security Pattern Database; Spatial location data came from road network data from the transportation department, vector data of urban built-up areas, and spatial distribution data of slope in China (30m resolution) from the Resource and Environmental Science and Data Platform of the Chinese Academy of Sciences.

[0037] Using a 1km×1km regular grid as the evaluation unit, the target area was divided into 5000 evaluation units. Features of each evaluation unit were extracted through spatial overlay analysis: resource endowment features included mineral resource reserves and resource grade; ecological constraint features included ecological function importance index, ecological sensitivity index, and vegetation cover; spatial constraint features included distance from towns, distance from main transportation routes, and terrain slope, with distance from main transportation routes representing transportation convenience and terrain slope representing development difficulty.

[0038] It should be noted that the "ecological constraint characteristics" in this application is a comprehensive concept used to fully describe the degree of ecological constraints on mineral resource development in a region, specifically including three dimensions: ecological function importance index, ecological sensitivity index, and vegetation cover. Among them, the ecological function importance index reflects the ecological function value of the region in terms of water conservation, biodiversity protection, and soil and water conservation; Ecological sensitivity index: reflects the degree of sensitivity of a region to human disturbance; Vegetation coverage: Reflects the vegetation cover of the ground surface.

[0039] The above three dimensions together constitute the basic indicator system of ecological constraint characteristics, providing data support for the subsequent calculation of the Ecological Constraint Intensity Index (ECII).

[0040] All features are subjected to min-max normalization to ensure comparability between features of different dimensions. The normalization formula is: ; in, The original values ​​of the features, and These are the minimum and maximum values ​​of the feature in the entire dataset, respectively. The value is the normalized value, ranging from [0, 1]. Outliers outside the normal range are truncated.

[0041] This process makes features of different dimensions comparable, which facilitates subsequent rule mining and evaluation.

[0042] According to the "Technical Regulations for the Compilation of Provincial Mineral Resources Master Plans" and regional ecological protection requirements, the study area is divided into five functional zones: (Prohibited Development Zones): Areas where mineral resource development is strictly prohibited (such as the core areas of nature reserves and within ecological red lines); (Restricted Development Zone): Areas where limited development is permitted, but the scale and environmental impact must be strictly controlled (such as general areas of ecological functional zones and areas surrounding basic farmland). (General Development Zone): Areas with average resource conditions and low ecological sensitivity, where development according to conventional standards is permitted; (Key Development Zones): Areas rich in resources, with favorable mining conditions, and playing a vital supporting role in the macroeconomy; (Resource Reserve Area): Areas with great resource potential but insufficient current exploration, or areas that are not yet subject to large-scale commercial exploitation as national strategic reserves.

[0043] II. Hard rule extraction (step S2) Eighteen policy and planning documents related to mineral resource development and ecological environmental protection in the target area were collected, including national, provincial, and municipal planning documents.

[0044] The dictionary of core control dimensions is constructed as follows: the dictionary of spatial entities includes 86 terms such as nature reserves, drinking water sources, basic farmland, ecological protection red lines, railways, and expressways; the dictionary of development behavior terms includes 24 terms such as mining, exploration, prospecting, construction, and production; and the dictionary of constraint operation terms includes 32 terms such as prohibition (prohibited, strictly prohibited, not allowed), restriction (restricted, controlled, avoided), permission (allowed, may), and encouragement (encouraged, supported, prioritized).

[0045] The LTP (Language Technology Platform) was used to perform word segmentation, part-of-speech tagging, and dependency parsing on the policy text.

[0046] Example sentence 1: "Mineral resource exploration and development activities are strictly prohibited within the core area of ​​the nature reserve." Match the spatial-dominant template to extract the spatial entity "core area of ​​the nature reserve", the constraint operation "strictly prohibited", and the development behavior "exploration and development".

[0047] Example sentence 2: "Within the concentrated mining area, large and medium-sized sand and gravel mining is permitted." Matching attribute composite template, extract the spatial entity "concentrated mining area", the condition "large and medium-sized sand and gravel mine", the constraint operation "permit", and the development behavior "mining".

[0048] Transform prohibitive and restrictive control clauses into a standardized four-tuple rule structure: ; in, For spatial entity names, For spatial buffer range, For constraint type, The mapping area type has a value range of 100%. These correspond to prohibited development zones, restricted development zones, general development zones, key development zones, and resource reserve zones, respectively, ultimately resulting in a set of candidate hard rules.

[0049] For example, rules : <Core area of ​​nature reserve, 0 meters, prohibited> > Rules Railway, 1000 meters, prohibited. >

[0050] The conflicting rules extracted from different policy planning texts were resolved based on the principles of priority of legal effect, priority of publication time, or priority of protection, forming a structured set of 47 rules.

[0051] Among them, the handling of hard rule conflicts follows the following general principles: Legal force takes precedence: national laws and regulations > local policies > industry standards; Publication time takes precedence: if the effects are of the same level, the latest published policy document shall prevail; Protection First, Last Resort: If it is impossible to determine, the rule with stricter constraints (leaning towards ecological protection) will be adopted by default.

[0052] III. Learning Soft Rules (Step S3) Samples were extracted from official planning documents: Based on the list of "key mining areas" in the "Overall Mineral Resources Plan of XX Province (2021-2025)," 25 mining areas were extracted as samples of key development areas. Based on the list of "key exploration areas" in the plan, 18 mining areas were selected as samples for resource reserve areas. In the remaining areas, 22 general development zone samples were drawn using a stratified random sampling method. This forms the initial sample set.

[0053] The initial sample set is validated, and samples that conflict with the set of hard rules are removed, resulting in a validated sample set (where, type...). The quantity is 23, and the type is The quantity is 18, and the type is... The number of samples is 22. The SMOTE oversampling method is used to adjust the class balance of the samples, forming a training sample set containing 75 samples (25 samples per class).

[0054] The spatial location of each sample in the training sample set is spatially superimposed with the standard set of spatial features for analysis. Each sample is assigned a corresponding feature vector to form a training sample set with features.

[0055] The decision tree was trained using the DecisionTreeClassifier library from Python's scikit-learn library, with a maximum depth of 5 and a minimum number of samples per leaf node of 3. After training, the decision tree was traversed from the root node to each leaf node to generate IF-THEN rules. For example, the rule... : IF (Resource development potential index ≥ 0.78) AND (Ecological sensitivity ≤ 0.35) AND (Distance from main transportation line ≤ 10.2km) THEN (Zoning type = z4) [Confidence level = 0.92], where confidence level refers to the proportion of correctly classified training samples covered by the rule, used to measure the reliability of the rule.

[0056] A total of 15 initial soft rules were generated, and after cleaning, 10 soft rules were obtained, which served as the initial selection decision rule set.

[0057] IV. Construction of a Dual-Objective Conflict Resolution Evaluation Model (Step S4) Comprehensive Index Calculation: The Resource Development Potential Index (RDPI) comprehensively reflects the resource development value of the evaluation unit. It is a weighted aggregation of resource characteristics and spatial characteristics, and the calculation formula is as follows: ; in, The comprehensive value of resource endowment is obtained by weighted average after normalization of reserves, grade, etc. For traffic convenience, the value is negatively normalized from the distance to the transportation route (the closer the distance, the larger the value). To account for the development difficulty, which negatively impacts development potential, a subtractive approach is adopted, normalizing the results based on factors such as slope and ore deposit depth. As weight, satisfying In this embodiment, .

[0058] The Ecological Constraint Intensity Index (ECII) comprehensively reflects the ecological importance and vulnerability of the evaluation unit. It is calculated by aggregating ecological characteristics, and the formula is as follows: ; in, As an index of ecological function importance; It is an ecological sensitivity index; As a normalized value of vegetation cover, this index weights and aggregates the three dimensions of ecological constraint characteristics to achieve a quantitative evaluation of regional ecological protection needs. As the weight, in this embodiment, .

[0059] The aforementioned weight values ​​were determined based on a combination of expert experience and the analytic hierarchy process, reflecting the relative importance of resource reserves, transportation convenience, and development difficulty to development potential, as well as the relative importance of ecological function, ecological sensitivity, and vegetation coverage to ecological protection constraints.

[0060] For example, a sample of a key development zone : =0.815, =0.72, =0.42; Therefore, RDPI = 0.6 × 0.815 + 0.3 × 0.72 - 0.1 × 0.42 = 0.489 + 0.216 - 0.042 = 0.663; =0.32, =0.28, =0.45; Then ECII = 0.3 × 0.32 + 0.4 × 0.28 + 0.3 × 0.45 = 0.096 + 0.112 + 0.135 = 0.343.

[0061] Utility measurement calculation: For each rule in the initial decision rule set, calculate the average RDPI and ECII of the samples it covers. For example, if this rule covers 8 samples, then: ; .

[0062] Dynamic determination of collaborative weight coefficients Eight national-level policy documents released in the past five years (2019-2023) were collected. The frequency of occurrence of development-oriented and protection-oriented terms was counted, and the development-oriented score and protection-oriented score of each policy document were calculated using the TF-IDF method.

[0063] Introducing a time decay factor =0.8, calculate the comprehensive development guidance intensity: ; Comprehensive protection guidance intensity The calculation formula is: ; After normalization, the collaborative weight coefficients are obtained: ; ; Construction of rule-based scoring function The formula for the rule-based scoring function is: ; In this embodiment, It is 0.535. The value is 0.465. (Conflict penalty item) The following rule applies: if the rule concludes that the area is a key development zone and the mean ecological sensitivity index of the samples covered by the rule is higher than the preset ecological threshold (0.4 in this embodiment), then... The value is greater than The value is set such that the rule score is negative; otherwise, it is zero.

[0064] Calculation rules Rating: =0.535×0.710+0.465×0.323 0 = 0.380 + 0.150 = 0.530.

[0065] Evaluation model formation A bi-objective conflict resolution evaluation model, composed of rule scoring functions, is used to quantitatively evaluate each rule in the initial decision rule set.

[0066] V. Rule-based pruning and conflict resolution (Step S5) The 10 rules in the initial decision rule set were substituted into the evaluation model to calculate the score. All rule scores were higher than the preset threshold of 0.4, and the difference between development utility and protection utility did not exceed the tolerance range of 0.3. All of them were retained to obtain the optimized soft rule set.

[0067] The set of hard rules (47 rules) and the optimized set of soft rules (10 rules) are merged: the hard rules are given the highest priority, and the soft rules are given the second highest priority according to the scores from high to low, resulting in an initial fusion rule set of 57 rules.

[0068] Redundancy removal is performed on the initial fusion rule set. The specific judgment method for redundancy removal is as follows: For two rules with the same conclusion, if the condition set of rule A contains all the conditions of rule B (that is, each condition of rule B can find the exact same feature, comparison operator and threshold in rule A), then rule A is determined to contain rule B. In this case, rule B with more concise conditions is retained and rule A is removed.

[0069] For example, rules (IF (Resource Development Potential Index ≥ 0.72) AND (Ecological Sensitivity ≤ 0.30) AND (Distance from Main Transportation Line ≤ 8.5km) AND (Resource Reserves ≥ 5 million tons) THEN) ) and rules (IF (Resource Development Potential Index ≥ 0.78) AND (Ecological Sensitivity ≤ 0.35) AND (Distance from Main Transportation Line ≤ 10.2km) THEN) There is an inclusion relationship, so it is retained. Eliminate After removing redundancies, a simplified rule set of 52 rules was obtained.

[0070] Build a standardized data structure for each rule: ; The collaborative zoning rule base consists of the data structure of all rules: ; in, For the collaborative zoning rule base, A set of inviolable hard rules. This is the set of soft rules optimized through conflict resolution.

[0071] Rule base application verification This section corresponds to the application steps of the preferred collaborative zoning rule base: Ten untrained grid cells were selected as test samples. Taking test cell T01 as an example: its resource development potential index is 0.68, ecological sensitivity is 0.31, and its distance from the main transportation route is 8.5 km, placing it outside the constraints of any hard rules. A soft rule was applied: priority 3 rule. All conditions (resource development potential index ≥ 0.65, ecological sensitivity ≤ 0.40, and distance from town ≥ 3.0 km) are met, and the output conclusion is z4 (key development zone). After expert review, the result is deemed reasonable.

[0072] For fuzzy regions where feature values ​​are near rule thresholds (e.g., ecological sensitivity indices between 0.35 and 0.40), a fuzzy membership function is introduced for refinement. For continuous feature conditions in each soft rule, they are transformed into trapezoidal or Gaussian membership functions, defining a gradual membership degree for feature values ​​near the threshold. For the spatial unit to be determined, its matching degree with each soft rule (the minimum or product of the membership degrees of each condition) is calculated, and the rule with the highest matching degree is selected as the output. When there is more than one rule with the highest matching degree, the centroid method is used for defuzzification: the weighted average of the matching degree of each rule and the corresponding zoning type code is calculated, and then this average is rounded and mapped to the closest discrete zoning type.

[0073] Fuzzy rules can be generated by expanding the threshold boundary of the decision tree generated in step S3 (for example, expanding the hard threshold "ecological sensitivity index < 0.4" into a fuzzy interval), or they can be predefined by domain experts in combination with fuzzy expressions in policy texts (such as "development in ecologically fragile areas should be strictly controlled").

[0074] Implementation effect The implementation of this embodiment has achieved the following results: In terms of rule extraction efficiency, the manual policy interpretation work that originally required 2-3 weeks can be completed in 2 hours through natural language processing, extracting 47 hard rules covering more than 95% of key control clauses; In terms of conflict resolution, 18% of the key development zones delineated by traditional methods conflict with the ecological red line, while this method reduces the conflict rate to less than 3%; In terms of identification accuracy, 9 out of 10 test samples are consistent with the expert judgment, with an accuracy rate of 90%.

[0075] System Implementation Examples like Figure 2 As shown, the present invention also provides a system for constructing collaborative zoning rules for mineral resource development and ecological protection, comprising: The spatial feature acquisition module is used to collect multi-source basic data of the target area, extract resource endowment features, ecological constraint features, and spatial constraint features from the multi-source basic data, and perform normalization processing to obtain a spatial feature standard set. The multi-source basic data includes mineral resource data, ecological environment data, and spatial location data of the target area. This module corresponds to step S1 in the method embodiment, and its specific implementation can be referred to the data acquisition, evaluation unit division, feature extraction, and normalization processing processes described above.

[0076] The hard rule extraction module is used to acquire policy planning texts, perform natural language processing on the policy planning texts to extract prohibitive and restrictive control clauses, and convert them into a structured set of hard rules. This module corresponds to step S2 in the method embodiment, and its specific implementation can be referred to the core control dimension dictionary construction, dependency parsing, template matching, quadruple transformation, and conflict resolution process described above.

[0077] The soft rule learning module is used to acquire historical zoning samples. Using these samples as training data, and combined with a spatial feature standard set, a machine learning algorithm is employed to train the initial selection decision rule set. This module corresponds to step S3 in the method embodiment, and its specific implementation can be found in the preceding processes of sample acquisition, data verification, category balancing, feature association, decision tree training, and rule generation.

[0078] The evaluation model construction module is used to construct development utility metrics for measuring regional development value and protection utility metrics for measuring regional ecological value, based on the resource and ecological dimensions of the spatial feature standard set. It then constructs a rule-based scoring function based on the synergistic weighting of the development and protection utility metrics to form a dual-objective conflict resolution evaluation model. This module corresponds to step S4 in the method embodiment, and its specific implementation can be found in the preceding sections on comprehensive index calculation, utility metric calculation, dynamic determination of synergistic weight coefficients, rule-based scoring function construction, and conflict penalty term setting.

[0079] The rule optimization and fusion module is used to input the initial selection decision rule set into the dual-objective conflict resolution evaluation model. It scores each rule in the initial selection decision rule set according to the rule scoring function, eliminating rules with scores below a preset threshold, and rules that cause the difference between the development utility metric and the protection utility metric to exceed the tolerance range. This yields an optimized soft rule set, which is then merged with the hard rule set and sorted by priority to form a collaborative zoning rule base. This module corresponds to step S5 in the method embodiment, and its specific implementation can be found in the preceding sections on rule scoring and elimination, rule fusion and priority assignment, redundancy elimination, data structure construction, and rule base formation.

[0080] The connections between the modules are as follows: the output of the spatial feature acquisition module is connected to the input of the hard rule extraction module and the soft rule learning module; the output of the hard rule extraction module and the soft rule learning module is connected to the input of the evaluation model construction module; the output of the evaluation model construction module is connected to the input of the rule optimization and fusion module; and the rule optimization and fusion module outputs the collaborative zoning rule base.

[0081] In this system embodiment, all functional modules can be integrated into a single processor, or each module can be a separate device, or two or more modules can be integrated into a single device. Each functional module can be implemented in hardware or as a combination of hardware and software functional units.

[0082] The system can combine rigid policy constraints with flexible patterns discovered through data mining, automatically address the conflict between development and protection through a collaborative scoring mechanism, and output a scientific and reasonable collaborative zoning scheme for mineral resources.

[0083] The foregoing has provided a detailed description of a method and system for constructing collaborative zoning rules for mineral resource development and ecological protection, as provided in this application. The above description of the disclosed embodiments enables those skilled in the art to implement or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing zoning rules for coordinated mineral resource development and ecological protection, characterized in that, Includes the following steps: Collect multi-source basic data of the target area, extract resource endowment features, ecological constraint features and spatial constraint features from the multi-source basic data, and perform normalization processing to obtain a spatial feature standard set. The multi-source basic data includes mineral resource data, ecological environment data and spatial location data of the target area. Obtain the policy planning text, perform natural language processing on the policy planning text to extract prohibitive and restrictive control clauses, and convert them into a structured set of hard rules; Obtain historical zoning samples, use the historical zoning samples as training data, combine them with the spatial feature standard set, and train them using machine learning algorithms to obtain a preliminary selection decision rule set; Based on the resource and ecological dimensions of the aforementioned spatial feature standard set, a development utility metric for measuring regional development value and a protection utility metric for measuring regional ecological value are constructed respectively. A rule-based scoring function is then constructed based on the synergistic weighting of the development utility metric and the protection utility metric to form a dual-objective conflict resolution evaluation model. The initial decision rule set is substituted into the dual-objective conflict resolution evaluation model. Each rule in the initial decision rule set is scored according to the rule scoring function. Rules with scores below a preset threshold and rules that cause the difference between the development utility measure and the protection utility measure to exceed the tolerance range are eliminated to obtain an optimized soft rule set. The optimized soft rule set is then merged with the hard rule set and sorted by priority to form a collaborative zoning rule library. The step of constructing a rule-based scoring function based on the collaborative weighting of the development utility measure and the protection utility measure to form a bi-objective conflict resolution evaluation model includes the following steps: For each rule in the initial decision rule set, a rule scoring function is constructed based on the development utility metric and the protection utility metric of that rule; The formula for the rule scoring function is as follows: ； in, Rate the rule for this rule. The development utility metric described in this rule, The measure of the protective utility of this rule. and These are the co-weighting coefficients for the development utility measure and the protection utility measure, respectively. Conflict penalty item; The conflict penalty item is determined according to the following rules: if the conclusion of the rule is a key development area and the mean ecological sensitivity index of the samples covered by the rule is higher than the preset ecological threshold, then the conflict penalty item is a value greater than the weighted sum of the development utility measure and the protection utility measure of the rule, so that the rule score of the rule is negative; otherwise, the conflict penalty item is zero. The rule scoring function constitutes the bi-objective conflict resolution evaluation model, which is used to quantitatively evaluate each rule in the initial decision rule set. The collaborative weighting coefficient is determined through the following steps: Retrieve national-level policy texts released within a preset time period; The frequency of occurrence of development-oriented terms and protection-oriented terms were statistically analyzed from the policy text. The development-oriented terms were selected from the constraint operators to represent positive constraints on development; the protection-oriented terms were selected from the constraint operators to represent negative constraints on development. The TF-IDF method is used to calculate the weighted frequency of the development-oriented terms and the frequency of the protection-oriented terms to obtain the weighted frequency of the development-oriented terms and the weighted frequency of the protection-oriented terms for each policy text, which are used as the development-oriented score and the protection-oriented score for each policy text. According to the release time order of the policy texts, a time decay factor is introduced to calculate the weighted average of the development orientation score and the protection orientation score of each policy text. The policy texts released later are given higher weights. The comprehensive development orientation strength and comprehensive protection orientation strength are calculated. The sum of the comprehensive development guidance intensity and the comprehensive protection guidance intensity is used as the normalization benchmark. The proportion of the comprehensive development guidance intensity to the normalization benchmark is used as the synergistic weighting coefficient of the development utility measure. The proportion of the comprehensive protection guidance intensity to the normalization benchmark is used as the synergistic weighting coefficient of the protection utility measure.

2. The method for constructing zoning rules for coordinated mineral resource development and ecological protection according to claim 1, characterized in that, The step of extracting resource endowment features, ecological constraint features, and spatial constraint features from the multi-source basic data includes the following steps: Using a regular grid as the evaluation unit, the resource endowment characteristics, ecological constraint characteristics, and spatial constraint characteristics of each evaluation unit are extracted from the multi-source basic data through spatial overlay analysis; The resource endowment characteristics include mineral resource reserves and resource grade; The ecological constraint characteristics include the ecological function importance index, the ecological sensitivity index, and vegetation coverage. The spatial constraint features include distance from towns, distance from main transportation routes, and terrain slope, wherein the distance from main transportation routes is used to characterize transportation convenience, and the terrain slope is used to characterize development difficulty.

3. The method for constructing zoning rules for coordinated mineral resource development and ecological protection according to claim 2, characterized in that, The process of performing natural language processing on policy planning texts to extract prohibitive and restrictive control clauses and convert them into a structured set of hard rules includes the following steps: Construct a core control dimension dictionary, which includes spatial entity terms, development behavior terms, and constraint operation terms; Based on the core control dimension dictionary, the policy planning text is subjected to part-of-speech tagging and entity recognition to identify the spatial entity words, development behavior words and constraint operation words in the text; Based on predefined space-dominant templates and attribute composite templates, dependency parsing analysis is performed on sentences that identify the space entity words, the development behavior words, and the constraint operation words to extract key components including space entities, constraints, and development behaviors, and these key components are used as the prohibitive and restrictive control clauses. Transform the aforementioned prohibitive and restrictive control clauses into a standardized four-tuple rule structure: ； in, For spatial entity names, For spatial buffer range, For constraint type, Based on the mapping zone type, a candidate hard rule set is obtained; The conflicting rules extracted from different policy planning texts in the candidate hard rule set are resolved according to the principles of priority of legal effect, priority of publication time, or priority of protection, thus forming the structured hard rule set.

4. The method for constructing zoning rules for coordinated mineral resource development and ecological protection according to claim 3, characterized in that, The step of using the historical zoning samples as training data, combined with the spatial feature standard set, and training with a machine learning algorithm to obtain a preliminary selection decision rule set includes the following steps: The collection includes samples from key development zones and resource reserve zones extracted from official planning documents, as well as samples from general development zones obtained through stratified random sampling, to form an initial sample set; The initial sample set is validated by removing samples that conflict with the structured set of hard rules, thus obtaining a validated sample set. The validated sample set is adjusted for class balance using oversampling or undersampling methods to form a training sample set; The spatial location of each sample in the training sample set is spatially superimposed with the spatial feature standard set. The feature vector corresponding to the location of each sample is extracted from the spatial feature standard set. Each sample is assigned the corresponding resource endowment feature, ecological constraint feature and spatial constraint feature to form a training sample set with features. The feature-rich training sample set is input into the decision tree algorithm for training to obtain the decision tree model; Traverse the path from the root node to each leaf node in the decision tree model, take the splitting condition on each path as the rule antecedent, and take the zoning type corresponding to the leaf node as the rule consequent to generate the initial selection decision rule set in IF-THEN form.

5. The method for constructing zoning rules for coordinated mineral resource development and ecological protection according to claim 4, characterized in that, Based on the resource and ecological dimensions of the aforementioned spatial feature standard set, development utility metrics for measuring regional development value and metrics for measuring regional development value are respectively constructed. The measurement of the conservation efficacy of the domain's ecological value includes the following steps: The mineral resource reserves and the resource grade are weighted and aggregated to obtain a comprehensive resource endowment value; Based on the comprehensive value of resource endowment, the distance from the main transportation route, and the terrain slope, a resource development potential index is constructed using a weighted aggregation method to quantify the resource development value of the evaluation unit. The distance from the main transportation route is used to determine the weight of the transportation convenience, and the terrain slope is used to determine the weight of the development difficulty. Based on the ecological function importance index, the ecological sensitivity index, and the vegetation coverage, an ecological constraint intensity index is constructed using a weighted aggregation method to quantify the ecological protection importance of the evaluation unit. For each rule in the preliminary decision rule set, the average value of the resource development potential index of the samples covered by the rule is used as the development utility measure of the rule, and the average value of the ecological constraint intensity index of the samples covered by the rule is used as the protection utility measure of the rule.

6. The method for constructing zoning rules for coordinated mineral resource development and ecological protection according to claim 1, characterized in that, The step of merging the optimized soft rule set with the hard rule set and sorting them by priority to form a collaborative zoning rule base includes the following steps: The hard rule set is merged with the optimized soft rule set, and each merged rule is assigned a priority. The rules in the hard rule set have the highest priority, and the rules in the optimized soft rule set are assigned corresponding secondary priorities from high to low according to their rule scores, thus obtaining the initial fused rule set. Redundancy is eliminated from the initial fusion rule set. When two rules in the initial fusion rule set have an inclusion relationship in their preconditions and the same conclusion, the rule with the simpler precondition is retained to obtain the simplified rule set. A standardized data structure is constructed for each rule in the simplified rule set. The data structure includes rule number, priority, preconditions, conclusion type, rule score, and rule source. The data structures of all rules constitute the collaborative zoning rule base.

7. The method for constructing zoning rules for coordinated mineral resource development and ecological protection according to claim 6, characterized in that, After constructing the collaborative zoning rule base, the following steps are also included: Obtain the target spatial unit to be judged, and extract its corresponding resource endowment feature, ecological constraint feature and spatial constraint feature from the spatial feature standard set to form the feature vector to be judged; The feature vector to be determined is matched with the rules in the collaborative zoning rule base, and the rules are judged in descending order of priority to determine whether the preconditions of the rules are met. If any of the hard rules are met, the zone type corresponding to that hard rule is directly output, and the matching is terminated. If none of the hard rules are met, the matching degree between the feature vector to be determined and each of the soft rules is calculated. For rules with a matching degree greater than a preset matching threshold, a weighted vote is performed based on the rule score, and the voting result is used as the output partition type. If the matching degree of all the soft rules is not greater than the preset matching threshold, then the preset default division type is output.

8. A system for constructing rules for coordinated zoning of mineral resource development and ecological protection, characterized in that, include: The spatial feature acquisition module is used to collect multi-source basic data of the target area and extract resource endowment features from the multi-source basic data. Ecological and spatial constraint characteristics are identified and normalized to obtain a spatial feature standard set. The multi-source basic data includes... Mineral resource data, ecological environment data, and spatial location data for the target area; The hard rule extraction module is used to acquire policy planning texts, perform natural language processing on the policy planning texts to extract prohibitive and restrictive control clauses, and convert them into a structured set of hard rules; The soft rule learning module is used to acquire historical zoning samples, use the historical zoning samples as training data, combine them with the spatial feature standard set, and use machine learning algorithms to train the initial selection decision rule set. The evaluation model construction module is used to construct a development utility measure for measuring the development value of a region and a protection utility measure for measuring the ecological value of a region, respectively, based on the resource dimension and ecological dimension of the spatial feature standard set. Based on the synergistic weighting of the development utility measure and the protection utility measure, a rule scoring function is constructed to form a dual-objective conflict resolution evaluation model. The step of constructing a rule-based scoring function based on the collaborative weighting of the development utility measure and the protection utility measure to form a dual-objective conflict resolution evaluation model includes the following steps: For each rule in the initial decision rule set, a rule scoring function is constructed based on the development utility metric and the protection utility metric of that rule; The formula for the rule scoring function is as follows: ； in, Rate the rule for this rule. The development utility metric described in this rule, The measure of the protective utility of this rule. and These are the co-weighting coefficients for the development utility measure and the protection utility measure, respectively. Conflict penalty item; The conflict penalty item is determined according to the following rules: if the conclusion of the rule is a key development area and the mean ecological sensitivity index of the samples covered by the rule is higher than the preset ecological threshold, then the conflict penalty item is a value greater than the weighted sum of the development utility measure and the protection utility measure of the rule, so that the rule score of the rule is negative; otherwise, the conflict penalty item is zero. The rule scoring function constitutes the bi-objective conflict resolution evaluation model, which is used to quantitatively evaluate each rule in the initial decision rule set. The collaborative weighting coefficient is determined through the following steps: Retrieve national-level policy texts released within a preset time period; The frequency of occurrence of development-oriented terms and protection-oriented terms were statistically analyzed from the policy text. The development-oriented terms were selected from the constraint operators to represent positive constraints on development; the protection-oriented terms were selected from the constraint operators to represent negative constraints on development. The TF-IDF method is used to calculate the weighted frequency of the development-oriented terms and the frequency of the protection-oriented terms to obtain the weighted frequency of the development-oriented terms and the weighted frequency of the protection-oriented terms for each policy text, which are used as the development-oriented score and the protection-oriented score for each policy text. According to the release time order of the policy texts, a time decay factor is introduced to calculate the weighted average of the development orientation score and the protection orientation score of each policy text. The policy texts released later are given higher weights. The comprehensive development orientation strength and comprehensive protection orientation strength are calculated. The sum of the comprehensive development guidance intensity and the comprehensive protection guidance intensity is used as the normalization benchmark. The proportion of the comprehensive development guidance intensity to the normalization benchmark is used as the synergistic weighting coefficient of the development utility measure. The proportion of the comprehensive protection guidance intensity to the normalization benchmark is used as the synergistic weighting coefficient of the protection utility measure. The rule optimization and fusion module is used to substitute the initial decision rule set into the dual-objective conflict resolution evaluation model, score each rule in the initial decision rule set according to the rule scoring function, eliminate rules with scores below a preset threshold, and rules that cause the difference between the development utility metric and the protection utility metric to exceed the tolerance range, to obtain an optimized soft rule set, and then merge the optimized soft rule set with the hard rule set, and sort them by priority to form a collaborative zoning rule library.

Images