Collaborative optimization method and system for improving resilience of industrial chain based on low-carbon constraint

Abstract

The application provides a low-carbon constraint-based industrial chain resilience improvement collaborative optimization method and system, relating to the technical field of industrial chain optimization. The method comprises: obtaining historical carbon emission data of multiple links of a target industrial chain; constructing a multi-scenario model of the target industrial chain based on industrial-link background information and an annual historical carbon emission data set; obtaining current industrial-environment background information, matching the target scenario model, and performing carbon emission prediction; performing collaborative optimization of multiple links based on Q-link carbon emission prediction results and the target scenario model with the target carbon emission index as a constraint to obtain a target collaborative optimization scheme. The application solves the technical problem of resource waste and poor efficiency of an industrial chain due to the fact that carbon emission management methods are often limited to optimization of a single link and lack overall optimization of links of an industrial chain, thereby improving resource utilization efficiency and low-carbon resilience of the entire industrial chain through multi-scenario model construction and collaborative optimization.

Description

Technical Field

[0001] This application relates to the field of supply chain optimization technology, and in particular to a collaborative optimization method and system for enhancing supply chain resilience based on low-carbon constraints. Background Technology

[0002] As a complex system comprised of multiple links in production, transportation, processing, and sales, the overall carbon emission level is determined by the collaborative operation of each link. Traditional carbon emission management methods mostly focus on optimizing single links in the industrial chain, such as improving production processes, optimizing transportation routes, or adjusting energy use structures to achieve localized carbon emission reductions. While these methods can achieve some carbon reduction effects in certain areas, they often fail to reach the optimal overall emission reduction target because they ignore the close connections between links in the industrial chain and the systemic nature of carbon emissions. Due to the close connections between links in the industrial chain and the systemic nature of carbon emissions, optimizing a single link cannot fully realize the potential for carbon emission reduction and may lead to resource waste and poor efficiency. Furthermore, when facing uncertainties such as production fluctuations, changes in energy supply, or policy tightening, problems such as imbalanced resource allocation, chaotic production planning, or carbon emission rebounds can easily occur, severely restricting the overall low-carbon efficiency and risk resistance of the industrial chain.

[0003] In summary, existing technologies suffer from the problem that carbon emission management methods are often limited to optimizing a single link, lacking overall optimization of the coordination of all links in the industrial chain, resulting in waste of industrial chain resources, poor efficiency, and weak risk resistance. Summary of the Invention

[0004] The purpose of this application is to provide a collaborative optimization method and system for improving the resilience of the industrial chain based on low-carbon constraints, in order to solve the technical problems in the prior art, where carbon emission management methods are often limited to the optimization of a single link and lack overall optimization of the collaboration of all links in the industrial chain, resulting in waste of industrial chain resources, poor efficiency and weak risk resistance.

[0005] In view of the above problems, this application provides a collaborative optimization method and system for improving the resilience of the industrial chain based on low-carbon constraints.

[0006] Firstly, this application provides a collaborative optimization method for enhancing the resilience of industrial chains based on low-carbon constraints. This method is implemented through a collaborative optimization system for enhancing the resilience of industrial chains based on low-carbon constraints. The method includes: acquiring a set of Q annual historical carbon emission data for Q links in a target industrial chain, where each annual historical carbon emission data includes industry-link background information, and Q is a positive integer; constructing a multi-scenario model of the target industrial chain based on the industry-link background information and the set of Q annual historical carbon emission data, obtaining a multi-scenario model library, where each scenario model includes industry-environmental background feature labels; acquiring current industry-environmental background information and matching it with the industry-environmental background feature labels of each scenario model in the multi-scenario model library to obtain a target scenario model, and predicting carbon emissions for the Q links based on the target scenario model to obtain carbon emission prediction results for the Q links; and using a target carbon emission index as a constraint, collaboratively optimizing the Q links of the target industrial chain based on the carbon emission prediction results for the Q links and the target scenario model to obtain a target collaborative optimization scheme.

[0007] Optionally, each scenario model in the multi-scenario model library includes a centralized carbon emission data chain for each stage, Q stage carbon emission margins, and Q annual carbon emission correction factors.

[0008] Optionally, taking the same industry-segment background as the aggregation target, the Q sets of historical carbon emission data for each year are aggregated separately based on industry-segment background information to obtain Q aggregated annual historical carbon emission data sets. Each aggregated annual historical carbon emission data set is a combination of historical data from a single segment within the same industry-environment background and is identified with industry-environment background feature tags. Outlier cleaning is performed on the Q aggregated annual historical carbon emission data sets to determine Q segment-specific concentrated carbon emission data sets and Q segment-specific carbon emission margins. Annual carbon emission growth trend analysis is performed on the Q segment-specific concentrated carbon emission data sets to determine Q annual carbon emission correction factors. Scenario integration is performed on the Q segment-specific concentrated carbon emission data sets based on industry-environment background feature tags to determine M segment-specific concentrated carbon emission data chains. Based on the M segment-specific concentrated carbon emission data chains, the Q segment-specific carbon emission margins, and the Q annual carbon emission correction factors, M scenario models are constructed, and the M scenario models are labeled using corresponding industry-environment background feature tags to obtain the multi-scenario model library.

[0009] Optionally, extract any one of the Q aggregated annual historical carbon emission data sets; count the frequency of occurrence of each aggregated annual historical carbon emission data in the aggregated annual historical carbon emission data set to obtain the group center with the highest frequency; perform outlier cleaning on the aggregated annual historical carbon emission data set based on the group center to obtain the stage-specific concentrated carbon emission data set and stage-specific carbon emission margin; and so on, determine the group center for each of the Q aggregated annual historical carbon emission data sets and perform outlier cleaning to obtain the Q stage-specific concentrated carbon emission data sets and Q stage-specific carbon emission margins.

[0010] Optionally, the average distance between the aggregated annual historical carbon emission data set and the population center is calculated, and one-tenth of the average distance is used as the outlier removal scale; an initial neighborhood of the population center is constructed using the outlier removal scale; the initial neighborhood of the population center is expanded twice according to the outlier removal scale to obtain a first expanded neighborhood and a second expanded neighborhood; it is determined whether both the first expanded neighborhood and the second expanded neighborhood have a neighborhood density greater than or equal to the neighborhood density of the initial neighborhood of the population center. If so, the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center is calculated. When the neighborhood density difference is less than a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, the expanded neighborhood of the second group center is taken as the expanded neighborhood of the target group center, and the expanded neighborhood of the target group center is taken as the centralized carbon emission data set of the process. The difference between the historical carbon emission data of the largest aggregated year and the historical carbon emission data of the smallest aggregated year in the expanded neighborhood of the target group center is taken as the carbon emission margin of the process.

[0011] Optionally, the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center is calculated. When the neighborhood density difference is greater than or equal to a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, the expanded neighborhood of the second group center is expanded once again according to the outlier cleaning scale until the preset expansion stop condition is met, and the expanded neighborhood of the target group center is obtained.

[0012] Optionally, the preset expansion stop condition is: if the expanded neighborhood density is greater than or equal to the original neighborhood density and the expansion count meets the preset expansion count, then expansion stops; if the expanded neighborhood density is less than the original neighborhood density, then expansion stops immediately.

[0013] Optionally, the carbon emission margins of the Q target stages in the target scenario model are obtained; with the target carbon emission index as a constraint and the carbon emission margins of the Q target stages as optimization bandwidth constraints, the carbon emission prediction results of the Q stages are identified using a collaborative optimization scheme identifier to determine the target collaborative optimization scheme.

[0014] Optionally, a stage emission monitoring window is set; carbon emissions are monitored for Q stages in the stage emission monitoring window to determine whether the target carbon emission indicators are met; if not, a collaborative optimization correction instruction is generated.

[0015] Secondly, this application also provides a collaborative optimization system for enhancing the resilience of industrial chains based on low-carbon constraints, used to execute the collaborative optimization method for enhancing the resilience of industrial chains based on low-carbon constraints as described in the first aspect. The collaborative optimization system for enhancing the resilience of industrial chains based on low-carbon constraints includes: a data acquisition module for acquiring a set of Q annual historical carbon emission data for Q links of the target industrial chain, wherein each annual historical carbon emission data includes industry-link background information, and Q is a positive integer; and a model building module for performing multi-scenario modeling of the target industrial chain based on the industry-link background information and the set of Q annual historical carbon emission data. The system constructs a multi-scenario model library, where each scenario model includes industry-environment background feature labels. A model matching module acquires current industry-environment background information and matches it with the industry-environment background feature labels of each scenario model in the multi-scenario model library to obtain a target scenario model. Based on the target scenario model, it predicts carbon emissions for Q stages, obtaining the carbon emission prediction results for the Q stages. A collaborative optimization module, constrained by the target carbon emission index, performs collaborative optimization on the Q stages of the target industrial chain based on the carbon emission prediction results for the Q stages and the target scenario model, obtaining a target collaborative optimization scheme.

[0016] One or more technical solutions provided in this application have at least the following beneficial effects: By acquiring a set of historical carbon emission data for Q segments of the target industrial chain over Q years, where each year's historical carbon emission data includes industry-segment background information and Q is a positive integer; a multi-scenario model of the target industrial chain is constructed based on the industry-segment background information and the set of historical carbon emission data for Q years, resulting in a multi-scenario model library, where each scenario model includes industry-environment background feature labels; the current industry-environment background information is acquired and matched with the industry-environment background feature labels of each scenario model in the multi-scenario model library to obtain the target scenario model, and carbon emission predictions for the Q segments are performed based on the target scenario model to obtain the carbon emission prediction results for the Q segments; with the target carbon emission index as a constraint, the Q segments of the target industrial chain are collaboratively optimized based on the carbon emission prediction results for the Q segments and the target scenario model to obtain the target collaborative optimization scheme. In other words, by collecting annual historical carbon emission data for Q links in the target industrial chain, and combining this with industry-link background information to construct a set of carbon emission data for each link in different years, a multi-scenario model library is built based on the collected historical data and background labels. According to the current industry-environment background information, the target scenario model that best fits the current context is determined, and carbon emission predictions are made for the Q links. Based on the prediction results, the Q links in the target industrial chain are collaboratively optimized to obtain a specific target collaborative optimization scheme. Based on this scheme, production plans, resource allocation, and carbon emission control strategies for each link in the industrial chain are adjusted to ultimately achieve the low-carbon emission target of the industrial chain and minimize carbon emissions. This not only improves the resource utilization efficiency of the entire industrial chain but also significantly enhances the response speed and recovery capability of the industrial chain when facing external disturbances, thereby comprehensively improving the low-carbon resilience and collaborative operation efficiency of the industrial chain.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in 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 merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1This is a flowchart illustrating the collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints proposed in this application. Figure 2 This is a schematic diagram of the collaborative optimization system for improving the resilience of the industrial chain based on low-carbon constraints, as proposed in this application.

[0020] Figure labeling: Data acquisition module 11, model building module 12, model matching module 13, collaborative optimization module 14. Detailed Implementation

[0021] This application provides a collaborative optimization method and system for enhancing the resilience of the industrial chain based on low-carbon constraints. This addresses the technical problems in existing technologies where carbon emission management methods are often limited to optimizing individual links, lacking overall optimization of the entire industrial chain, leading to resource waste, poor efficiency, and weak risk resistance. By collecting annual historical carbon emission data for Q links in the target industrial chain and combining this with industry-link background information to construct a carbon emission data set for each link in different years, a multi-scenario model library is built based on the collected historical data and background labels. According to the current industry-environment background information, the target scenario model that best fits the current context is determined, and carbon emission predictions are made for the Q links. Based on the prediction results, the Q links of the target industrial chain are collaboratively optimized to obtain a specific target collaborative optimization scheme. Based on this scheme, production plans, resource allocation, and carbon emission control strategies for each link in the industrial chain are adjusted to ultimately achieve the low-carbon emission target of the industrial chain and minimize carbon emissions. This not only improves the resource utilization efficiency of the entire industrial chain but also significantly enhances the response speed and recovery capability of the industrial chain when facing external disturbances, thereby comprehensively improving the low-carbon resilience and collaborative operation efficiency of the industrial chain.

[0022] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0023] Example 1, please refer to the appendix. Figure 1 This application provides a collaborative optimization method for enhancing the resilience of industrial chains based on low-carbon constraints. This method is executed through a collaborative optimization system for enhancing the resilience of industrial chains based on low-carbon constraints. The specific steps of this collaborative optimization method include: S100: Obtain the set of Q annual historical carbon emission data for Q links in the target industrial chain, where each annual historical carbon emission data includes industry-link background information, and Q is a positive integer.

[0024] Specifically, the target industrial chain is the specific industrial chain of the research object, including several interrelated links such as raw material procurement, production and processing, packaging and transportation, and sales and distribution. The target industrial chain is broken down to identify Q links, and annual historical carbon emission data sets for each link are collected through channels such as corporate annual reports, industry reports, and statistical data. Each year's historical carbon emission data includes industry-link background information, namely data on the overall industry situation and the overall situation of the links, such as industry scale, industry structure, link characteristics, and energy consumption.

[0025] The Q-year historical carbon emission dataset comprises carbon emission data from multiple years across various stages of the target industry chain, including upstream, downstream, and downstream segments (such as suppliers, manufacturers, distributors, and exporters). This includes, but is not limited to, carbon emissions from energy consumption (such as electricity and fuel), carbon emissions during production processes (such as the energy efficiency of production equipment and the carbon emission characteristics of production processes), carbon emissions from transportation and logistics (such as raw material transportation and finished product distribution), and carbon emissions from waste treatment and recycling. By acquiring the Q-year historical carbon emission dataset for each of the Q segments of the target industry chain, a comprehensive understanding of the carbon emission situation at each stage of the industry chain can be achieved.

[0026] S200: Based on industry-link background information and Q sets of historical carbon emission data for each year, a multi-scenario model of the target industrial chain is constructed to obtain a multi-scenario model library. Each scenario model includes industry-environment background feature labels.

[0027] Each scenario model in the multi-scenario model library includes a centralized carbon emission data chain for each stage, Q stage carbon emission margins, and Q annual carbon emission correction factors.

[0028] Furthermore, this application S200 includes: Using the same industry-segment background as the aggregation target, Q sets of historical carbon emission data for different years are aggregated according to industry-segment background information to obtain Q aggregated annual historical carbon emission data groups. Each aggregated annual historical carbon emission data group represents a combination of historical data from a single segment within the same industry-environment background and is labeled with industry-environment background feature tags. Outlier cleaning is performed on the Q aggregated annual historical carbon emission data groups to determine Q segment-specific concentrated carbon emission data sets and Q segment-specific carbon emission margins. Annual carbon emission growth trend analysis is performed on the Q segment-specific concentrated carbon emission data sets to determine Q annual carbon emission correction factors. Based on the industry-environment background feature tags, scenario integration is performed on the Q segment-specific concentrated carbon emission data sets to determine M segment-specific concentrated carbon emission data chains. Based on the M segment-specific concentrated carbon emission data chains, the Q segment-specific carbon emission margins, and the Q annual carbon emission correction factors, M scenario models are constructed. These M scenario models are then labeled with corresponding industry-environment background feature tags to obtain the multi-scenario model library.

[0029] Specifically, the data collection targets data from the same industry and process segment. Based on this industry and process segment background information, the Q sets of historical carbon emission data for different years are clustered into categories, forming a set of historical carbon emission samples under the same background. The Q sets of historical carbon emission data are then categorized and summarized according to background characteristics, with process data belonging to the same background category (i.e., the same label) grouped together. Each of the Q aggregated sets of historical carbon emission data for different years reflects the historical carbon emission data of a single process segment under the same environmental background conditions. Each aggregated data group is labeled with an industry-environment background feature tag for easy reference and identification later. The industry-environment background feature tag is used to identify processes or data groups with similar industry and environmental characteristics.

[0030] Outlier cleaning is performed on Q aggregated annual historical carbon emission datasets. This involves identifying and extracting outliers—values ​​that deviate significantly from the normal data—to ensure the accuracy and reliability of the data. After outlier cleaning, Q aggregated carbon emission datasets and Q carbon emission margins for each stage are obtained. The aggregated carbon emission datasets are the clean datasets after removing outliers, accurately reflecting the typical carbon emission behavior of each stage under the given circumstances. The carbon emission margins represent the additional load or capacity that each stage can withstand in terms of carbon emissions, providing a margin for error.

[0031] To analyze the annual growth trend of carbon emissions across Q aggregated carbon emission datasets, we need to map the relationship between carbon emissions at each stage and time onto a graph, obtaining the corresponding functional relationship. In other words, we construct a two-dimensional coordinate system with time on the x-axis and carbon emission values ​​on the y-axis. Mapping the carbon emission values ​​for each time period onto this coordinate system yields a functional relationship, typically a linear function. Based on the trend of this linear function, we can obtain the annual correction factor for each stage, i.e., the Q annual carbon emission correction factors. A positive slope indicates an upward trend in carbon emissions, and the factor is positive; a negative slope indicates a downward trend, and the factor is negative; a slope close to 0 indicates that carbon emissions at that stage remain stable. The annual carbon emission correction factor is a value derived from trend analysis and is used to adjust the carbon emission data in future scenario forecasts to better align with historical trend evolution. For example, if carbon emissions at a certain stage increase by an average of 2% annually, its correction factor is +2%; if it decreases by an average of 1.5% annually, it is -1.5%.

[0032] For example, the carbon emissions in the first year are 960 tons, the second year is 975 tons, the third year is 990 tons, the fourth year is 1005 tons, and the fifth year is 1020 tons. After performing a linear regression, the slope obtained is 15, and it is found that the carbon emissions increase by 1.5% per month each year. The correction factor is 1 + 1.5% = 1.015. Based on the carbon emissions in the fifth year multiplied by the correction factor, the predicted carbon emissions in the sixth year can be obtained as 1035 tons.

[0033] Based on the industry-environment background feature labels, the carbon emission data sets of Q links are integrated in chronological order. In other words, the carbon emission data sets of links belonging to the same industry-environment background feature label are integrated in chronological order to obtain M links of carbon emission data chains, each data chain corresponding to a specific industry-environment background feature label.

[0034] Based on M centralized carbon emission data chains across various stages, Q carbon emission margins across different stages, and Q annual carbon emission correction factors, M scenario models are constructed. Each scenario model includes a centralized carbon emission data chain across stages, Q carbon emission margins across stages, and Q annual carbon emission correction factors. Each scenario model represents a carbon emission behavior prediction and adjustment model under a specific industry-environmental background and is identified by its background feature label. All models are aggregated to form a callable multi-scenario model library. Each scenario model considers different industry-environmental background characteristics and analyzes carbon emission characteristics and carbon emission margins under different backgrounds through multi-scenario modeling, thereby providing accurate carbon emission predictions for each stage under different scenarios.

[0035] Furthermore, this application also includes the following steps: Extract any one of the Q aggregated annual historical carbon emission data sets; count the frequency of occurrence of each aggregated annual historical carbon emission data in the aggregated annual historical carbon emission data set to obtain the cluster center with the highest frequency; perform outlier cleaning on the aggregated annual historical carbon emission data set based on the cluster center to obtain the stage-specific concentrated carbon emission data set and stage-specific carbon emission margin; and so on, determine the cluster center for each of the Q aggregated annual historical carbon emission data sets and perform outlier cleaning to obtain the Q stage-specific concentrated carbon emission data sets and Q stage-specific carbon emission margins.

[0036] Specifically, from the set of Q aggregated annual historical carbon emission data sets, any one aggregated data set is selected as the aggregated annual historical carbon emission data set, meaning the aggregated annual historical carbon emission data set corresponding to any specific stage is selected. Statistical analysis is performed on each aggregated annual historical carbon emission data set within the selected aggregated annual historical carbon emission data set, calculating the frequency of each data point. This identifies the cluster center within the dataset, i.e., the carbon emission data point with the highest frequency in the dataset. The cluster center typically represents the typical carbon emission level of that data set, or the mean of most data points, reflecting the main trend of the data.

[0037] Based on the identified cluster centers, outlier cleaning is performed on the aggregated annual historical carbon emission data set for this stage. Data with significant discrepancies or biases from the cluster centers are removed to ensure that the dataset used in the analysis does not contain excessive outliers or invalid extreme values. After outlier cleaning, the stage-specific aggregated carbon emission data set is obtained, with outliers removed, representing the true carbon emission data for this stage under a specific context. Furthermore, the stage-specific carbon emission margin is calculated based on this data, i.e., the additional load or capacity that this stage can withstand in terms of carbon emissions.

[0038] For each of the Q aggregated annual historical carbon emission data sets, repeat the above steps, calculating the cluster center for each data set, performing outlier cleansing, and obtaining the aggregated carbon emission data set and carbon emission margin for each stage. Outlier cleansing ensures the accuracy and reliability of carbon emission data and removes noisy data. Determining the cluster center identifies the representative carbon emission level for each stage, ensuring more representative data. Determining the carbon emission adjustment space (margin) for each stage helps assess the potential and flexibility of each stage in terms of carbon emissions.

[0039] Furthermore, this application also includes the following steps: Calculate the average distance between the aggregated annual historical carbon emission data set and the population center, and use one-tenth of the average distance as the outlier removal scale; construct the initial neighborhood of the population center using the outlier removal scale; expand the initial neighborhood of the population center twice according to the outlier removal scale to obtain a first expanded neighborhood and a second expanded neighborhood; determine whether both the first expanded neighborhood and the second expanded neighborhood have a neighborhood density greater than or equal to the neighborhood density of the initial neighborhood of the population center. If so... Then, the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center is calculated. When the neighborhood density difference is less than a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, the expanded neighborhood of the second group center is taken as the expanded neighborhood of the target group center, and the expanded neighborhood of the target group center is taken as the centralized carbon emission data set of the link. The difference between the historical carbon emission data of the largest aggregated year and the historical carbon emission data of the smallest aggregated year in the expanded neighborhood of the target group center is taken as the carbon emission margin of the link.

[0040] Specifically, the distance from the group center to the aggregated annual historical carbon emission data set is calculated, i.e., the distance between the group center and each aggregated annual historical carbon emission data set. The average of all distances is then taken to reflect the overall density of the data distribution relative to the group center. One-tenth of this average distance is set as the outlier cleaning scale to determine which data points are considered outliers. Based on the outlier cleaning scale, an initial neighborhood of the group center is constructed, containing a series of data points whose distance to the group center is less than the outlier cleaning scale.

[0041] The initial neighborhood is expanded twice at the same scale, constructing a first expanded neighborhood and a second expanded neighborhood of the group center. Each expansion increases the range of the neighborhood to include more data points near the group center. The first expanded neighborhood of the group center is larger than the initial neighborhood of the group center and contains more peripheral data. In the second expansion, the range is expanded again to form the second expanded neighborhood of the group center.

[0042] Calculate the neighborhood density (i.e., the number of data points within each neighborhood) for each of the three neighborhoods to obtain the neighborhood density of the initial neighborhood of the group center, the neighborhood density of the expanded neighborhood of the first group center, and the neighborhood density of the expanded neighborhood of the second group center. Compare the neighborhood densities of the expanded neighborhoods of the first and second group centers with the neighborhood density of the initial neighborhood of the group center. If both are greater than the neighborhood density of the initial neighborhood of the group center, the expansion process is considered effective, and the difference in neighborhood density between the expanded neighborhoods of the first and second group centers is calculated.

[0043] The preset difference threshold is a threshold set to determine whether the densities of two neighboring regions are close. If the density difference is less than this value, the two neighboring regions are considered to have similar densities. Then, it is determined whether the neighborhood density difference is less than the preset difference threshold, and whether the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center. When both conditions are met, the expanded neighborhood of the second group center is taken as the expanded neighborhood of the target group center and used as the centralized carbon emission data set for this stage.

[0044] The difference between the historical carbon emission data of the largest and smallest aggregated annual periods in the expanded neighborhood of the target group center is used as the carbon emission margin of the segment, which represents the additional load or capacity that this segment can withstand in terms of carbon emissions.

[0045] For Q aggregated annual historical carbon emission data sets, the above calculations are performed to obtain the Q aggregated carbon emission data sets and Q carbon emission margins for each stage. Through outlier cleaning, the carbon emission data set and carbon emission margin for each stage are accurately determined, eliminating the influence of outliers and thus providing accurate carbon emission predictions for each stage under different scenarios.

[0046] Furthermore, this application also includes the following steps: Calculate the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center. When the neighborhood density difference is greater than or equal to a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, then the expanded neighborhood of the second group center continues to be expanded once according to the outlier cleaning scale until the preset expansion stop condition is met, and the expanded neighborhood of the target group center is obtained.

[0047] Furthermore, this application also includes the following steps: The preset expansion stop condition is that if the expanded neighborhood density is greater than or equal to the original neighborhood density, and the expansion count meets the preset expansion count, expansion will stop; if the expanded neighborhood density is less than the original neighborhood density, expansion will stop immediately.

[0048] Specifically, if the difference in neighborhood density between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center is greater than or equal to a preset difference threshold, it indicates that the density difference between the two neighborhoods is too large, and it is still necessary to continue expanding the data coverage to obtain a more representative dataset. When the neighborhood density of the expanded neighborhood of the second group center is greater than that of the expanded neighborhood of the first group center, the expanded neighborhood of the second group center is expanded again according to the outlier cleaning scale to obtain a larger neighborhood dataset. This process is progressive, and the new neighborhood density should be recalculated each time it is expanded.

[0049] Expansion will stop when a preset termination condition is met. The expanded neighborhood at this point will be used as the target population center expansion neighborhood, forming the final centralized carbon emission data set for this stage. The preset termination condition includes: if the expanded neighborhood density is greater than or equal to the pre-expansion neighborhood density, it indicates that the current expansion is still adding representative data, and expansion will continue until the preset maximum number of expansions is reached; or, if the expanded neighborhood density is less than the pre-expansion neighborhood density, it indicates that the expansion is introducing sparse or invalid data, and expansion should be terminated immediately, retaining the previous neighborhood. The preset number of expansions is a pre-set maximum number of expansions to avoid over-expansion introducing noise or unrepresentative data.

[0050] S300: Obtain the current industry-environment background information, match it with the industry-environment background feature labels of each scenario model in the multi-scenario model library to obtain the target scenario model, and perform carbon emission prediction for Q links based on the target scenario model to obtain the carbon emission prediction results for Q links.

[0051] Specifically, the current industry-environment background information is a data set representing the external environment and internal operational status of the target industry chain when conducting carbon emission forecasting. This includes industry category, production scale, energy type, geographical region, and emission policy level. The current industry-environment background information is matched with the industry-environment background feature labels of each scenario model in the multi-scenario model library. A similarity algorithm is used to quantify the degree of closeness, and the scenario model with the highest similarity score is selected as the target scenario model. For example, if the background of the target industry has a label similarity of 0.92 with model A, while other models are all below 0.9, then model A is selected as the target scenario model.

[0052] By utilizing the centralized carbon emission data chain included in the target scenario model, and combining it with the annual carbon emission correction factors for Q links, the historical carbon emission data for each link is trend-corrected to obtain the predicted carbon emission value for each environment. For example, if the correction factor for a certain link is 1.04, it indicates that the annual carbon emission is growing at a rate of 4%, and the future predicted value will be obtained based on this correction. The carbon emission prediction results for the Q links are obtained by predicting the possible future carbon emissions of each link in the industrial chain based on the target scenario model.

[0053] Carbon emission forecasts are made based on scenario models that best match the current industry and environmental context, thereby improving the accuracy and reliability of the forecasts. Furthermore, by using annual carbon emission correction factors to adjust the forecast results, the accuracy of the forecasts is further improved, which helps to formulate more effective carbon emission management strategies, optimize supply chain operations, reduce resource waste, and improve operational efficiency and competitiveness.

[0054] S400: Using the target carbon emission target as a constraint, based on the carbon emission prediction results of the Q links and the target scenario model, the Q links of the target industrial chain are collaboratively optimized to obtain the target collaborative optimization scheme.

[0055] Furthermore, this application S400 includes: Obtain the carbon emission margins of Q target stages in the target scenario model; using the target carbon emission index as a constraint and the carbon emission margins of the Q target stages as optimization bandwidth constraints, use a collaborative optimization scheme identifier to identify the carbon emission prediction results of the Q stages, and determine the target collaborative optimization scheme.

[0056] Specifically, the carbon emission margins of the Q target stages in the target scenario model are obtained. These Q stage carbon emission margins represent the amount of carbon emissions that can be reduced or increased at each stage while maintaining normal operation of the industrial chain; they are the emission elasticity within a reasonable fluctuation range and serve as the boundary conditions for the optimization search. The target carbon emission index is introduced, representing the total carbon emission reduction that the entire industrial chain needs to achieve. Using a collaborative optimization scheme identifier, with the target carbon emission index as a constraint and the carbon emission margins of the Q target stages as optimization bandwidth constraints, the carbon emission prediction results for the Q stages are identified and optimized to select the optimal carbon emission management strategy.

[0057] The optimization scheme identifier utilizes a multi-objective programming algorithm to search for the optimal solution among all possible combinations of emission adjustments at each stage. In short, starting with the predicted carbon emissions of Q stages, combined with a clear low-carbon target value (i.e., the target carbon emission index), it guides the entire industrial chain to optimize resource allocation, energy structure, and production scheduling under the premise of controlled emissions through a constraint-driven collaborative optimization process.

[0058] Resource allocation optimization refers to the rational allocation of resources such as manpower, equipment, and raw materials while meeting production needs, thereby reducing unnecessary energy consumption or redundant emissions. Production scheduling optimization refers to controlling emission timing and reducing carbon intensity under peak load through scheduling, load shifting, and off-peak production. Energy use optimization refers to optimizing the energy structure (such as increasing the proportion of renewable energy and reducing dependence on fossil fuels) to improve carbon efficiency.

[0059] The process obtains the predicted carbon emissions for Q stages based on the target scenario model, describing the emissions at each stage of the industrial chain under current conditions. This serves as the baseline data for the current unoptimized state. An externally defined target carbon emission indicator, such as "total annual emissions not exceeding 9,000 tons of CO2," is introduced as an upper limit constraint that must be strictly adhered to during the optimization process, forming the core boundary of the entire optimization model. The target carbon emission indicator is compared with the current prediction results to determine if there is pressure to exceed the limit. For example, if the current prediction value is 9,700 tons, exceeding the target by 700 tons, then a suitable adjustment combination needs to be found among the Q stages to reduce the total value to within 9,000 tons. By constructing a multi-objective constrained optimization model, it is necessary to ensure that the total carbon emissions are less than the target carbon emission indicator, that the adjustment range of each stage falls within the historical carbon emission margin range, and that resource allocation does not lead to production interruptions. Additional factors such as control costs, energy switching feasibility, and inter-stage coupling are comprehensively considered to avoid inefficiency caused by drastic emission reduction in a single stage.

[0060] Finally, a multi-dimensional collaborative optimization scheme is output, including recommended target carbon emission values ​​for each stage; recommended resource allocation change strategies; suggestions for adjusting production scheduling or energy use methods; and deviations from the original forecast data. This achieves a controllable and implementable collaborative optimization path, minimizing carbon emissions or maximizing efficiency across the entire industrial chain. Production plans, resource allocation, and carbon emission control strategies at each stage of the industrial chain are adjusted based on the collaborative optimization scheme to ultimately achieve the low-carbon emission target of the entire industrial chain.

[0061] By coordinating and optimizing Q links in the target industrial chain, the problems of traditional carbon emission reduction methods that only focus on local or single links are solved. This ensures that while meeting the overall carbon emission reduction target, the emission reduction potential and operational constraints of each link are fully considered, so as to achieve low-carbon coordinating optimization of the entire industrial chain.

[0062] Furthermore, this application also includes the following steps: Set a stage emission monitoring window; monitor carbon emissions for Q stages in the stage emission monitoring window, determine whether the target carbon emission indicators are met, and if not, generate a collaborative optimization correction instruction.

[0063] Specifically, phased emission monitoring windows are set up to regularly monitor carbon emissions at each key link (Q links) in the industrial chain. Based on the implemented collaborative optimization plan, the compliance of actual carbon emissions with target values ​​is dynamically tracked. If an emission exceeding trend or deviation from the target is detected, a collaborative optimization correction instruction is issued in a timely manner, thereby achieving dynamic closed-loop management and adaptive adjustment of carbon emissions.

[0064] The phased emission monitoring window can be flexibly configured according to management needs, such as monthly, quarterly, or per key process completion cycle. During this monitoring window, the actual carbon emissions of Q key links in the industrial chain are monitored, and the monitored actual emission data are compared and analyzed with the target carbon emission indicators set in the target collaborative optimization scheme. Once it is found that the target carbon emission indicators are not met, collaborative optimization correction instructions are generated, including measures such as adjusting production processes, optimizing energy structure, improving energy efficiency, and introducing carbon capture technology. Through collaborative optimization correction instructions, on-the-fly feedback adjustment is achieved, thereby building a highly reliable and adaptive carbon emission control network.

[0065] Through phased emission monitoring windows, the implementation effects of collaborative optimization schemes can be effectively monitored and evaluated, and adjustments and optimizations can be made in a timely manner based on actual conditions to ensure the achievement of target carbon emission targets.

[0066] In summary, the collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints provided in this application has the following beneficial effects: By acquiring a set of historical carbon emission data for Q segments of the target industrial chain over Q years, where each year's historical carbon emission data includes industry-segment background information and Q is a positive integer; a multi-scenario model of the target industrial chain is constructed based on the industry-segment background information and the set of historical carbon emission data for Q years, resulting in a multi-scenario model library, where each scenario model includes industry-environment background feature labels; the current industry-environment background information is acquired and matched with the industry-environment background feature labels of each scenario model in the multi-scenario model library to obtain the target scenario model, and carbon emission predictions for the Q segments are performed based on the target scenario model to obtain the carbon emission prediction results for the Q segments; with the target carbon emission index as a constraint, the Q segments of the target industrial chain are collaboratively optimized based on the carbon emission prediction results for the Q segments and the target scenario model to obtain the target collaborative optimization scheme. In other words, by collecting annual historical carbon emission data for Q links in the target industrial chain, and combining this with industry-link background information to construct a set of carbon emission data for each link in different years, a multi-scenario model library is built based on the collected historical data and background labels. According to the current industry-environment background information, the target scenario model that best fits the current context is determined, and carbon emission predictions are made for the Q links. Based on the prediction results, the Q links in the target industrial chain are collaboratively optimized to obtain a specific target collaborative optimization scheme. Based on this scheme, production plans, resource allocation, and carbon emission control strategies for each link in the industrial chain are adjusted to ultimately achieve the low-carbon emission target of the industrial chain and minimize carbon emissions. This not only improves the resource utilization efficiency of the entire industrial chain but also significantly enhances the response speed and recovery capability of the industrial chain when facing external disturbances, thereby comprehensively improving the low-carbon resilience and collaborative operation efficiency of the industrial chain.

[0067] Example 2: Based on the same inventive concept as the low-carbon constraint-based collaborative optimization method for enhancing supply chain resilience in Example 1, this application also provides a low-carbon constraint-based collaborative optimization system for enhancing supply chain resilience. Please refer to the appendix. Figure 2 The collaborative optimization system for enhancing supply chain resilience based on low-carbon constraints includes: The data acquisition module 11 is used to acquire a set of historical carbon emission data for Q annual periods for Q links in the target industrial chain, wherein each annual historical carbon emission data includes industry-link background information, and Q is a positive integer; the model building module 12 is used to construct a multi-scenario model of the target industrial chain based on the industry-link background information and the set of historical carbon emission data for Q annual periods, and obtain a multi-scenario model library, wherein each scenario model includes industry-environment background feature labels; the model matching module 13 is used to acquire the current industry-environment background information, match it with the industry-environment background feature labels of each scenario model in the multi-scenario model library to obtain the target scenario model, and perform carbon emission prediction for Q links based on the target scenario model to obtain the carbon emission prediction results for Q links; the collaborative optimization module 14 is used to perform collaborative optimization for the Q links of the target industrial chain based on the carbon emission prediction results for Q links and the target scenario model, with the target carbon emission index as a constraint, to obtain the target collaborative optimization scheme.

[0068] Furthermore, the model building module 12 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: Each scenario model in the multi-scenario model library includes a centralized carbon emission data chain for each stage, Q stage carbon emission margins, and Q annual carbon emission correction factors.

[0069] Furthermore, the model building module 12 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: Using the same industry-segment background as the aggregation target, Q sets of historical carbon emission data for different years are aggregated according to industry-segment background information to obtain Q aggregated annual historical carbon emission data groups. Each aggregated annual historical carbon emission data group represents a combination of historical data from a single segment within the same industry-environment background and is labeled with industry-environment background feature tags. Outlier cleaning is performed on the Q aggregated annual historical carbon emission data groups to determine Q segment-specific concentrated carbon emission data sets and Q segment-specific carbon emission margins. Annual carbon emission growth trend analysis is performed on the Q segment-specific concentrated carbon emission data sets to determine Q annual carbon emission correction factors. Based on the industry-environment background feature tags, scenario integration is performed on the Q segment-specific concentrated carbon emission data sets to determine M segment-specific concentrated carbon emission data chains. Based on the M segment-specific concentrated carbon emission data chains, the Q segment-specific carbon emission margins, and the Q annual carbon emission correction factors, M scenario models are constructed. These M scenario models are then labeled with corresponding industry-environment background feature tags to obtain the multi-scenario model library.

[0070] Furthermore, the model building module 12 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: Extract any one of the Q aggregated annual historical carbon emission data sets; count the frequency of occurrence of each aggregated annual historical carbon emission data in the aggregated annual historical carbon emission data set to obtain the cluster center with the highest frequency; perform outlier cleaning on the aggregated annual historical carbon emission data set based on the cluster center to obtain the stage-specific concentrated carbon emission data set and stage-specific carbon emission margin; and so on, determine the cluster center for each of the Q aggregated annual historical carbon emission data sets and perform outlier cleaning to obtain the Q stage-specific concentrated carbon emission data sets and Q stage-specific carbon emission margins.

[0071] Furthermore, the model building module 12 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: Calculate the average distance between the aggregated annual historical carbon emission data set and the population center, and use one-tenth of the average distance as the outlier removal scale; construct the initial neighborhood of the population center using the outlier removal scale; expand the initial neighborhood of the population center twice according to the outlier removal scale to obtain a first expanded neighborhood and a second expanded neighborhood; determine whether both the first expanded neighborhood and the second expanded neighborhood have a neighborhood density greater than or equal to the neighborhood density of the initial neighborhood of the population center. If so... Then, the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center is calculated. When the neighborhood density difference is less than a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, the expanded neighborhood of the second group center is taken as the expanded neighborhood of the target group center, and the expanded neighborhood of the target group center is taken as the centralized carbon emission data set of the link. The difference between the historical carbon emission data of the largest aggregated year and the historical carbon emission data of the smallest aggregated year in the expanded neighborhood of the target group center is taken as the carbon emission margin of the link.

[0072] Furthermore, the model building module 12 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: Calculate the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center. When the neighborhood density difference is greater than or equal to a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, then the expanded neighborhood of the second group center continues to be expanded once according to the outlier cleaning scale until the preset expansion stop condition is met, and the expanded neighborhood of the target group center is obtained.

[0073] Furthermore, the model building module 12 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: The preset expansion stop condition is that if the expanded neighborhood density is greater than or equal to the original neighborhood density, and the expansion count meets the preset expansion count, expansion will stop; if the expanded neighborhood density is less than the original neighborhood density, expansion will stop immediately.

[0074] Furthermore, the collaborative optimization module 14 in the low-carbon constraint-based supply chain resilience enhancement collaborative optimization system is also used for: Obtain the carbon emission margins of Q target stages in the target scenario model; using the target carbon emission index as a constraint and the carbon emission margins of the Q target stages as optimization bandwidth constraints, use a collaborative optimization scheme identifier to identify the carbon emission prediction results of the Q stages, and determine the target collaborative optimization scheme.

[0075] Furthermore, the collaborative optimization system for enhancing supply chain resilience based on low-carbon constraints also includes an emission monitoring module, which is further used for: Set a stage emission monitoring window; monitor carbon emissions for Q stages in the stage emission monitoring window, determine whether the target carbon emission indicators are met, and if not, generate a collaborative optimization correction instruction.

[0076] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Figure 1 The collaborative optimization method and specific examples for improving the resilience of the industrial chain based on low-carbon constraints in Example 1 are also applicable to the collaborative optimization system for improving the resilience of the industrial chain based on low-carbon constraints in this example. Through the foregoing detailed description of the collaborative optimization method for improving the resilience of the industrial chain based on low-carbon constraints, those skilled in the art can clearly understand the collaborative optimization system for improving the resilience of the industrial chain based on low-carbon constraints in this example. Therefore, for the sake of brevity, it will not be described in detail here.

[0077] The above description of the disclosed embodiments enables those skilled in the art to make 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.

[0078] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A collaborative optimization method for enhancing the resilience of the industrial chain based on low-carbon constraints, characterized in that, include: Obtain the set of Q annual historical carbon emission data for Q links in the target industrial chain, where each annual historical carbon emission data includes industry-link background information, and Q is a positive integer; Based on industry-link background information and Q sets of historical carbon emission data for each year, a multi-scenario model of the target industrial chain is constructed to obtain a multi-scenario model library. Each scenario model includes industry-link background feature labels. Obtain the current industry-environment background information, match it with the industry-link background feature tags of each scenario model in the multi-scenario model library to obtain the target scenario model, and predict the carbon emissions of Q links based on the target scenario model to obtain the carbon emission prediction results of Q links. Using the target carbon emission target as a constraint, the Q links of the target industrial chain are collaboratively optimized based on the carbon emission prediction results of the Q links and the target scenario model to obtain the target collaborative optimization scheme; Among these, a multi-scenario model of the target industrial chain is constructed based on industry-link background information and a set of Q annual historical carbon emission data, resulting in a multi-scenario model library, including: Using the same industry-segment background as the aggregation target, the historical carbon emission data sets of Q years are aggregated separately according to the industry-segment background information to obtain Q aggregated annual historical carbon emission data sets. Each aggregated annual historical carbon emission data set is a combination of historical data of one segment under the same industry-environment background, and is marked with industry-segment background feature tags. Outlier cleaning is performed on Q aggregated annual historical carbon emission data sets to determine Q centralized carbon emission data sets for each stage and Q stage carbon emission margins. The carbon emission margin is the additional load or capacity that each stage can withstand in terms of carbon emissions, which is used to provide fault tolerance space. Analyze the annual growth trend of carbon emissions from the Q aggregated carbon emission datasets to determine the Q annual carbon emission correction factors. Based on industry-link background feature tags, the centralized carbon emission data sets of the Q links are integrated to determine the centralized carbon emission data chains of M links. Based on the M centralized carbon emission data chains of the links, Q carbon emission margins of the links, and Q annual carbon emission correction factors, M scenario models are constructed, and the M scenario models are identified by the corresponding industry-link background feature tags to obtain the multi-scenario model library.

2. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 1, characterized in that, Each scenario model in the multi-scenario model library includes a centralized carbon emission data chain for each stage, Q stage carbon emission margins, and Q annual carbon emission correction factors.

3. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 1, characterized in that, Outlier cleaning is performed on Q aggregated annual historical carbon emission data sets to determine Q aggregated carbon emission data sets for different stages and Q stages of carbon emission margins, including: Extract any one aggregated annual historical carbon emission data set from the set of Q aggregated annual historical carbon emission data sets; The frequency of occurrence of historical carbon emission data for each aggregated year in the aggregated annual historical carbon emission data group is counted to obtain the group center with the highest frequency of occurrence; Based on the group center, outlier cleaning is performed on the aggregated annual historical carbon emission data group to obtain the stage-centralized carbon emission data set and stage carbon emission margin. Similarly, for each of the Q aggregated annual historical carbon emission data sets, a group center is determined, and outlier cleaning is performed to obtain the Q aggregated carbon emission data sets and the Q carbon emission margins for each stage.

4. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 3, characterized in that, Based on the aforementioned group center, outlier cleaning is performed on the aggregated annual historical carbon emission data set to obtain the stage-specific carbon emission data set and stage-specific carbon emission margin, including: Calculate the average distance between the aggregated annual historical carbon emission data set and the center of the group, and use one-tenth of the average distance as the outlier removal scale; The initial neighborhood of the population center is constructed using the outlier cleanup scale. The initial neighborhood of the group center is expanded twice according to the outlier cleanup scale to obtain the first expanded neighborhood of the group center and the second expanded neighborhood of the group center. Determine whether both the first group center expanded neighborhood and the second group center expanded neighborhood have a neighborhood density greater than or equal to the neighborhood density of the initial neighborhood of the group center. If so, calculate the neighborhood density difference between the first group center expanded neighborhood and the second group center expanded neighborhood. When the neighborhood density difference is less than a preset difference threshold and the neighborhood density of the second group center expanded neighborhood is greater than the neighborhood density of the first group center expanded neighborhood, then the second group center expanded neighborhood is taken as the target group center expanded neighborhood, and the target group center expanded neighborhood is taken as the segment-centralized carbon emission data set. The difference between the historical carbon emission data of the largest and smallest aggregated annual period in the expanded neighborhood of the target group center is used as the carbon emission margin of the process.

5. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 4, characterized in that, Also includes: Calculate the neighborhood density difference between the expanded neighborhood of the first group center and the expanded neighborhood of the second group center. When the neighborhood density difference is greater than or equal to a preset difference threshold and the neighborhood density of the expanded neighborhood of the second group center is greater than the neighborhood density of the expanded neighborhood of the first group center, then the expanded neighborhood of the second group center continues to be expanded once according to the outlier cleaning scale until the preset expansion stop condition is met, and the expanded neighborhood of the target group center is obtained.

6. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 5, characterized in that, The preset expansion stop condition is that if the expanded neighborhood density is greater than or equal to the original neighborhood density, and the expansion count meets the preset expansion count, expansion will stop; if the expanded neighborhood density is less than the original neighborhood density, expansion will stop immediately.

7. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 2, characterized in that, Constrained by the target carbon emission target, the Q links of the target industrial chain are collaboratively optimized based on the carbon emission prediction results of the Q links and the target scenario model to obtain a target collaborative optimization scheme, including: Obtain the carbon emission margins for the Q target stages of the target scenario model; Using the target carbon emission index as a constraint and the carbon emission margin of Q target stages as an optimization bandwidth constraint, the collaborative optimization scheme identifier is used to identify the carbon emission prediction results of the Q stages and determine the target collaborative optimization scheme.

8. The collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints as described in claim 1, characterized in that, include: Set a phased emission monitoring window; During the stage emission monitoring window, carbon emissions are monitored for Q stages to determine whether the target carbon emission indicators are met. If not, a collaborative optimization correction instruction is generated.

9. A collaborative optimization system for enhancing the resilience of the industrial chain based on low-carbon constraints, characterized in that: The steps for implementing the collaborative optimization method for enhancing supply chain resilience based on low-carbon constraints according to any one of claims 1 to 8, wherein the collaborative optimization system for enhancing supply chain resilience based on low-carbon constraints comprises: The data acquisition module is used to acquire a set of Q annual historical carbon emission data for Q links in the target industrial chain. Each annual historical carbon emission data includes industry-link background information, where Q is a positive integer. The model building module is used to build multi-scenario models of the target industrial chain based on industry-link background information and Q sets of historical carbon emission data for each year, and to obtain a multi-scenario model library. Each scenario model includes industry-link background feature labels. The model matching module is used to obtain the current industry-environment background information, match it with the industry-link background feature labels of each scenario model in the multi-scenario model library to obtain the target scenario model, and perform carbon emission prediction for Q links based on the target scenario model to obtain the carbon emission prediction results for Q links. The collaborative optimization module is used to collaboratively optimize the Q links of the target industrial chain based on the carbon emission prediction results of the Q links and the target scenario model, with the target carbon emission index as a constraint, to obtain the target collaborative optimization scheme.

Images