A global supply chain index model AI analysis method

Abstract

This invention discloses an AI analysis method for a global supply chain index model, belonging to the field of computer software technology, including the following steps: S01, based on the SCOR supply chain operation reference model, construct a full-scale indicator library from preset core dimensions, and filter core indicators through indicator importance scoring to form a core indicator set; S02, collect business performance data from historical adjustment cycles, mine the correlation rules between indicators, dynamically adjust indicator weights and establish a benchmark year, and record the original data and corrected indicator weights of the core indicators in the benchmark year; S03, collect internal and external supply chain data for the target year, perform standardization processing using a dynamic parameter adjustment mechanism, and calculate the global supply chain comprehensive index and sub-indices of each core dimension in combination with the corrected indicator weights; S04, based on the data of the target year and the benchmark year, and in combination with the corrected indicator weights, conduct supply chain index diagnosis and optimization analysis, identify weak indicators, and formulate optimization plans.

Description

Technical Field

[0001] This invention relates to the field of computer software technology, and in particular to an AI analysis method for a global supply chain index model. Background Technology

[0002] Currently, the global industrial landscape is undergoing profound restructuring, and enhancing the resilience and competitiveness of supply chains has become a core issue. Against this backdrop, the 2025 edition of the "Global Supply Chain Facilitation Report" released at the 3rd China International Supply Chain Expo, along with its proposed global supply chain index matrix encompassing dimensions such as resilience, connectivity, innovation, and facilitation, provides a new assessment perspective for supply chain analysis. Meanwhile, the SCOR model (Supply-Chain Operations Reference model) from the International Supply Chain Association, as a globally accepted supply chain operations reference framework, has evolved into a standardized system that is measurable, replicable, and optimizable, becoming a crucial foundation for supply chain analysis. The evolution of these tools and methods lays a solid theoretical and practical foundation for building an intelligent, multi-dimensional supply chain evaluation system.

[0003] However, existing supply chain analysis technologies still have significant shortcomings. Enterprise-level analysis is often scattered across KPIs (Key Performance Indicators) of various business segments, lacking integration with high-dimensional strategic indices and hindering comprehensive assessment. Analysis largely focuses on internal structured data, failing to effectively integrate external data from a global supply chain perspective, and lacks automated acquisition and dynamic analysis capabilities. Furthermore, the lack of traceability from the supply chain to the industrial chain makes it impossible to predict the cascading impact of external environmental changes on upstream and downstream operations. Due to the lack of objective quantitative indicators, enterprises struggle to accurately identify their own supply chain management strengths and weaknesses, and management cannot intuitively identify operational vulnerabilities and potential risks, thus restricting the efficiency of process optimization and resource integration.

[0004] For example, Chinese invention patent CN113962418A discloses a system and method for enhancing the accuracy of supply chain forecasting in global trade finance, comprising: a centralized cloud network and website connected to multiple existing transportation management software systems. The system utilizes existing scanning methods to monitor goods from the point of manufacture to the point of sale. Historical supply chain data for each product (or product group) is collected from scanners along the shipping route and fed to an artificial intelligence algorithm that learns actual shipping logistics (loading time, transit time, unloading time, port, delays, etc.) and performs trend analysis. Users can then invoke these trends to make more accurate supply chain forecasts and provide more precise recommendations to their customers.

[0005] For example, the invention patent with announcement number CN116090656A discloses an AI supply chain analysis method and algorithm model, which includes: an AI analysis data extraction functional component to acquire supply chain operation data; the AI ​​analysis data extraction functional component to transmit product flow of users at the supply chain node end to an AI analysis parameter learning functional component via data packets, and to transmit product statistics to an AI supply chain analysis functional component via data packets; the AI ​​analysis parameter learning functional component to extract data packets, execute a parameter learning algorithm to obtain product flow rate, and transmits it to the AI ​​supply chain analysis functional component via data packets; and the AI ​​supply chain analysis functional component to calculate the predicted flow position of products in the supply chain using an AI supply chain simulation algorithm. Summary of the Invention

[0006] To address the problems in existing technologies where enterprise supply chain analysis is fragmented across various business KPIs, lacks high-dimensional strategic index analysis, focuses on internal aspects without a global supply chain perspective, lacks automated external data acquisition, analysis, and feedback mechanisms, and suffers from a lack of traceability from the supply chain to the industrial chain, making it impossible to predict the impact of changes in the external environment in advance, this invention provides an AI analysis method using a global supply chain index model. The technical solution is as follows: On the one hand, an AI analysis method for a global supply chain index model is provided. This method includes: S01, based on the SCOR model for supply chain operations, constructing a full-scale indicator library from preset core dimensions, including planning, procurement, production, distribution, and returns, and selecting core indicators through indicator importance scoring to form a core indicator set; S02, collecting business performance data from historical adjustment cycles, mining the correlation rules between indicators, dynamically adjusting indicator weights and establishing a benchmark year, and recording the original data and corrected indicator weights of the core indicators in the benchmark year; S03, collecting internal and external supply chain data for the target year, standardizing it using a dynamic parameter adjustment mechanism, and calculating the global supply chain comprehensive index and sub-indices of each core dimension based on the corrected indicator weights; S04, based on the data of the target year and the benchmark year, combined with the corrected indicator weights, conducting supply chain index diagnosis and optimization analysis, identifying weak indicators, and formulating optimization plans.

[0007] On the other hand, an AI analysis device for a global supply chain index model is provided. This device is applied to the AI ​​analysis method of the global supply chain index model and includes: a core indicator construction module, a rule weight adjustment module, an index sub-item calculation module, and a diagnostic solution formulation module. The core indicator construction module is used to build a full-scale indicator library based on the SCOR supply chain operation reference model from preset core dimensions. The rule weight adjustment module is used to collect business performance data from historical adjustment cycles, mine the correlation rules between indicators, dynamically adjust indicator weights, establish a base year, and record the original data and corrected indicator weights of the core indicators in the base year. The index sub-item calculation module is used to collect internal and external supply chain data for the target year, perform standardization processing using a dynamic parameter adjustment mechanism, and calculate the global supply chain comprehensive index and the sub-indices of each core dimension based on the corrected indicator weights. The diagnostic solution formulation module is used to perform supply chain index diagnosis and optimization analysis based on the data of the target year and the base year, combined with the corrected indicator weights, to identify weak indicators and formulate optimization solutions.

[0008] Beneficial effects The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: 1. Based on the SCOR model, a comprehensive indicator library is constructed by combining a global supply chain index matrix with an artificial intelligence model. Core indicators are scientifically selected, and weights are dynamically adjusted. Internal and external data are standardized to calculate comprehensive and sub-indices for each dimension. This allows for supply chain diagnosis and optimization through vertical comparison, external benchmarking, and weakness identification. Simultaneously, five core indices—resilience, connectivity, innovation, facilitation, and security—are accurately measured. Association rules are used to uncover the causes of weaknesses and formulate priority optimization plans, tracking and evaluating their effectiveness. This achieves the scientific quantification and dynamic monitoring of supply chain indices, forming a closed-loop management system from index analysis, weakness identification, plan formulation to effect evaluation. It provides enterprises with objective data benchmarks and targeted guidance, effectively enhancing supply chain resilience and competitiveness.

[0009] 2. By constructing a positive-negative judgment matrix and dynamically adjusting the resource input ratio, combined with information entropy and entropy weight calculations, and adjusting the number of historical periods based on the calibration deviation between entropy weight and benchmark weight, the indicator weights and entropy weights are weighted and geometrically calculated for comprehensive scoring, and core indicators are selected to form a standardized indicator set. This ensures the consistency of the judgment matrix, optimizes the efficiency of expert resource utilization, accurately quantifies the degree of indicator variation, and balances sample stability and sensitivity, making the selection of core indicators more scientific, objective, and accurate. This lays a solid and reliable indicator foundation for the subsequent accurate calculation and dynamic weighting of the global supply chain index.

[0010] 3. By collecting historical business performance data and discretizing it to construct a transaction dataset, a frequent pattern tree is built based on minimum support and confidence to form association rules. The lift factor is calculated to screen effective rules, and the comprehensive association impact coefficient of core indicators is measured. The original weights and association impact coefficients are weighted and merged to obtain the corrected dynamic weights, and a deviation verification mechanism ensures the rationality of the weights. This achieves dynamic correction of indicator weights and close alignment with business association logic, enabling the index calculation weight system to dynamically adapt to changes in business development. This significantly improves the scientific nature, accuracy, and business relevance of the weights, making the index calculation results more realistically reflect the actual operating conditions of the supply chain.

[0011] 4. By dynamically processing the contribution directions of internal and external supply chain performance indicators for the target year, standardization operations are carried out separately for positive and negative indicators. The corresponding standardization methods are adapted according to the data distribution characteristics (normal / skewed), identifying and processing outliers outside the three-standard-deviation interval until their impact is reduced to below 3%. This unifies the standardization dimensions of indicators with different contribution directions, accurately adapts to data distribution characteristics, effectively suppresses the adverse effects of outliers, and ensures the scientific, accurate, and consistent nature of data processing. This provides a high-quality, comparable data foundation for subsequent global supply chain index calculations.

[0012] 5. After data standardization, based on the global supply chain index matrix and combined with the corrected dynamic weights, a weighted geometric calculation is performed on the global supply chain composite index and the sub-indices of each core dimension. The index results for the target year are output, and the standardized parameters and weights are fully recorded for subsequent comparisons. This achieves overall quantification of supply chain performance levels and tiered, refined evaluation of core dimensions, ensuring that the index calculation results accurately align with operational logic. It also provides traceable and comparable core data for subsequent supply chain index analysis, guaranteeing the continuity and comparability of index analysis.

[0013] 6. By comparing the standardized values ​​of each core indicator and the comprehensive index in the target year with the base period, the rate of change of the indicators, the total change of the comprehensive index, and the contribution of each indicator are obtained. Indicators with significant impact are screened according to thresholds, and those that improve or worsen are ranked and compiled into a list for inclusion in the longitudinal trend analysis report. This accurately quantifies the actual contribution of each core indicator to the change of the supply chain comprehensive index, identifies key indicators driving performance changes, and provides clear quantitative data support for the longitudinal evolution trend of supply chain performance, offering precise evidence for enterprises to trace the core drivers of performance changes.

[0014] 7. By establishing an external benchmarking mechanism, industry benchmark data is standardized to the same standards. Combined with corrected dynamic weights, the gap between the company and industry benchmarks is quantified, and external optimization directions are identified. A weakness identification mechanism is established to decompose strategic goals into core indicators and use dynamic weights to calculate obstacles and screen weakness indicators. Based on these indicators and association rules, core influencing factors are traced and identified. Candidate optimization measures are generated and ranked using weighted geometric calculations based on resource constraints, implementation difficulty, and expected benefits. A quantifiable implementation plan is then developed, and its effects are included in the next round of tracking and evaluation. This achieves precise guidance from both external benchmarking and internal strategic orientation, forming a complete scientific management loop from weakness identification and solution development to implementation and effect evaluation, significantly improving the pertinence, systematicness, and effectiveness of supply chain optimization work. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 Flowchart of the AI ​​analysis method for the global supply chain index model provided in this application embodiment; Figure 2 A flowchart illustrating the core indicator selection process for the AI ​​analysis method of the global supply chain index model provided in this application embodiment. Detailed Implementation

[0017] The following provides explanations of some terms used in this application. It should be noted that these explanations are for the convenience of those skilled in the art and do not constitute a limitation on the scope of protection claimed in this application.

[0018] References to "one embodiment," "in some examples," or "some embodiments" as described in the embodiments of this application mean that one or more embodiments of this specification include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in some examples," "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] like Figure 1 The diagram shows a flowchart of the AI ​​analysis method for the global supply chain index model provided in this application. This method is based on the Supply Chain Operations Reference Model (SCOR) proposed by the Supply Chain Council, combined with a global supply chain index matrix and a large-scale artificial intelligence model. The method includes the following steps: S01, based on the SCOR model, using algorithms such as AHP / ANP / TOPSIS to construct a full-scale indicator library from preset core dimensions. These core dimensions include planning, procurement, production, distribution, and returns. Core indicators are selected through indicator importance scoring to form a core indicator set; S02, collecting business performance data from historical adjustment cycles, mining the correlation rules between indicators, dynamically adjusting indicator weights, and establishing a benchmark year. The original data and corrected indicator weights of the core indicators in the benchmark year are recorded; S03, collecting internal and external supply chain data for the target year, using a dynamic parameter adjustment mechanism for standardization, and calculating the global supply chain comprehensive index and sub-indices for each core dimension based on the corrected indicator weights; S04, based on the target year and benchmark year data, combined with the corrected indicator weights, conducting supply chain index diagnosis and optimization analysis, identifying bottleneck indicators, and formulating optimization plans.

[0021] The core indicators include the Resilience Index, Connectivity Index, Innovation Index, Facilitation Index, and Security Index, which correspond to five key assessment dimensions: supply chain disruption response and recovery capabilities, node collaboration and information sharing levels, technological and process innovation performance, positive impact on the economic and social environment, and compliance risk control and information security levels. This provides a multi-dimensional, comprehensive, and three-dimensional quantitative characterization of global supply chain capabilities, fundamentally addressing the technical problem of traditional supply chain assessments being limited by their single dimension and inability to fully cover core capabilities such as resilience, collaboration, innovation, sustainability, and security. Specifically, the Resilience Index measures the supply chain's ability to respond to and recover from disruptions, based on supply chain operations. Referring to the SCOR model, a full-scale indicator library was constructed from five core dimensions: planning, procurement, production, distribution, and returns. Indicators related to supply chain disruption response and recovery capabilities were selected. Based on indicator importance scoring, delivery reliability, inventory turnover, and emergency response time were chosen as core indicators of the resilience index. Historical adjustment cycle business performance data was collected to mine the correlation rules among delivery reliability, inventory turnover, and emergency response time. Based on these correlation rules, the correlation impact coefficients of each indicator were calculated, and the initially determined calibrated weights were dynamically adjusted to obtain the corrected values ​​for delivery reliability, inventory turnover, and emergency response time. Indicator weights: Data on delivery reliability, inventory turnover, and emergency response time for the target year are collected and standardized using a dynamic parameter adjustment mechanism. The weighted geometric calculation is performed on delivery reliability, inventory turnover, and emergency response time indicators. The connectivity index measures the level of collaboration and information sharing among nodes in the supply chain. Based on the SCOR model for supply chain operations, a full-scale indicator library constructed from five core dimensions (planning, procurement, production, delivery, and returns) is used to select indicators related to the level of collaboration and information sharing among nodes in the supply chain. Information integration, on-time order delivery rate, and supplier collaboration are selected based on indicator importance scoring. The core indicators of the connectivity index are: collecting business performance data from historical adjustment cycles, mining the correlation rules among the three indicators of information integration, on-time order delivery rate, and supplier collaboration, calculating the correlation influence coefficient of each indicator based on the correlation rules, and dynamically correcting the initially determined calibrated weights to obtain the corrected indicator weights of information integration, on-time order delivery rate, and supplier collaboration; collecting information integration, on-time order delivery rate, and supplier collaboration indicator data for the target year, standardizing them using a dynamic parameter adjustment mechanism, and obtaining the weighted geometric calculation from the information integration, on-time order delivery rate, and supplier collaboration indicators;The innovation index measures the supply chain's performance in technology application and process innovation. Based on the SCOR model for supply chain operations, a comprehensive indicator library is constructed from five core dimensions: planning, procurement, production, distribution, and returns. Indicators related to supply chain technology application and process innovation performance are selected from a full-scale indicator library. Through indicator importance scoring, R&D investment ratio, new technology adoption rate, and process optimization frequency are chosen as the core indicators of the innovation index. Historical adjustment cycle business performance data is collected to mine the correlation rules among the three indicators: R&D investment ratio, new technology adoption rate, and process optimization frequency. Based on these correlation rules, the correlation impact coefficient of each indicator is calculated, and the initially determined calibrated weights are dynamically adjusted to obtain the R&D investment ratio. The revised weights of three indicators—new technology adoption rate, process optimization frequency, and R&D investment ratio, new technology adoption rate, and process optimization frequency—are calculated. Data on these indicators for the target year are collected and standardized using a dynamic parameter adjustment mechanism. The weighted geometric calculation of these three indicators yields the final result. The promotion index measures the positive impact of the supply chain on the economy, society, and environment. Based on the SCOR model for supply chain operations, a full indicator library constructed from five core dimensions—planning, procurement, production, distribution, and returns—is used to select indicators related to the positive impact of the supply chain on the economy, society, and environment. Local sourcing rate, carbon emission intensity, and employee satisfaction are selected through indicator importance scoring. The degree of local procurement rate is used as the core indicator of the promotion index. Historical adjustment cycle business performance data is collected to mine the correlation rules among the three indicators: local procurement rate, carbon emission intensity, and employee satisfaction. Based on these correlation rules, the correlation impact coefficient of each indicator is calculated, and the initially determined calibrated weights are dynamically adjusted to obtain the adjusted weights of the three indicators. Local procurement rate, carbon emission intensity, and employee satisfaction data for the target year are collected and standardized using a dynamic parameter adjustment mechanism. The weighted geometric calculation is then performed on the local procurement rate, carbon emission intensity, and employee satisfaction indicators. The security index measures the level of the supply chain in terms of compliance, risk control, and information security. Based on the SCOR model for supply chain operations, a full-scale indicator library was constructed from five core dimensions: planning, procurement, production, distribution, and returns. Indicators related to supply chain compliance risk control and information security levels were selected. Through indicator importance scoring, the number of compliance incidents, risk identification coverage, and data breaches were chosen as the core indicators of the security index. Business performance data from historical adjustment cycles was collected to mine the correlation rules among the three indicators. Based on these correlation rules, the correlation impact coefficients of each indicator were calculated, and the initially determined calibrated weights were dynamically adjusted to obtain the corrected indicator weights for the three indicators.Data on the number of compliance events, risk identification coverage, and data breach frequency for the target year are collected and standardized using a dynamic parameter adjustment mechanism. The resulting index is calculated using a weighted geometric calculation of these metrics. The five specialized indices are independent yet complementary, forming a complete and comprehensive global supply chain capability assessment index system across five dimensions: basic supply chain operations, cross-node collaboration, innovation-driven development, economic, social, and environmental value empowerment, and compliance and risk security assurance. This provides detailed, precise, and quantifiable core data for subsequent vertical comparisons of supply chain indices, external benchmarking, weakness identification, and optimization plan development, significantly enhancing the assessment depth, coverage breadth, and practical application value of the global supply chain index AI analysis method.

[0022] In this embodiment, the present invention constructs a comprehensive indicator library based on five core dimensions: planning, procurement, production, distribution, and returns. It then forms a standardized indicator set based on indicator importance scoring, establishing a unified and standardized indicator system foundation for supply chain index analysis. Furthermore, by collecting historical business performance data to mine correlation rules between indicators, dynamically adjusting indicator weights, and establishing a benchmark year, it achieves dynamic adaptation of indicator weights and standardized anchoring of benchmark data. This effectively solves the problems of fixed indicators and rigid weights in traditional supply chain index analysis. Subsequently, it collects internal and external supply chain data for the target year, completes data standardization processing using a dynamic parameter adjustment mechanism, and accurately calculates the full range of indicators based on dynamic weights. The Global Supply Chain Comprehensive Index and its core dimension sub-indices leverage a large-scale artificial intelligence model to achieve intelligent and efficient data processing and index calculation, significantly improving the accuracy and efficiency of index calculation. Finally, based on the current index and historical base period and other years' data, combined with the revised indicator weights, supply chain index diagnosis and optimization analysis are conducted. This enables the accurate identification of weak indicators in supply chain operations and the development of targeted optimization solutions. Ultimately, it achieves intelligent analysis, accurate diagnosis, and scientific optimization of the entire global supply chain process in a multi-dimensional and dynamic manner, significantly improving the scientific nature, foresight, and decision-making efficiency of global supply chain management, and contributing to a comprehensive improvement in the overall operational efficiency and risk resistance of the supply chain.

[0023] like Figure 2The diagram shows the core indicator selection flowchart of the AI ​​analysis method for the global supply chain index model provided in this application embodiment. The selection of core indicators through indicator importance scoring specifically includes: constructing a positive-negative judgment matrix based on full indicator data within each historical adjustment cycle using a 1-9 scale method. Scale 1 indicates that two indicators are equally important; scales 3, 5, 7, and 9 indicate that one indicator is slightly more important, significantly more important, strongly more important, and extremely more important than the other, respectively; 2, 4, 6, and 8 represent the median of the above adjacent judgments; and the reciprocal indicates the opposite comparison relationship. The maximum eigenvalue and corresponding eigenvector of the positive-negative judgment matrix are calculated, and it is determined whether the consistency ratio of the positive-negative judgment matrix reaches a preset consistency ratio threshold. If not, the eigenvector is normalized to obtain the weight of each full indicator. If so, the resource input ratio of the positive-negative judgment matrix is ​​dynamically adjusted based on the deviation between the consistency ratio threshold and the consistency ratio of the positive-negative judgment matrix, and the current consistency ratio is calculated relative to the preset threshold. The difference is recorded as the exceedance range. If the exceedance range is larger, resource input is increased proportionally, such as increasing the number of iterations in expert scoring, increasing the number of experts, or increasing the correction range for matrix elements. If the exceedance range is close to a threshold, resource input is reduced accordingly until the consistency ratio of the reconstructed judgment matrix meets the requirement of being less than the threshold. This optimizes the utilization efficiency of expert resources while ensuring the quality of the judgment matrix. A reconstructed positive-negative judgment matrix is ​​then constructed until the consistency ratio of the reconstructed positive-negative judgment matrix is ​​acceptable. The consistency ratio is the ratio of the consistency index to the preset average random consistency index. The consistency index is the deviation between the largest eigenvalue of the positive-negative judgment matrix and the order of the judgment matrix, and its proportion to the deviation between the order of the judgment matrix and 1. The raw data of all indicators in each historical adjustment cycle are standardized to obtain a standardized matrix. Based on the proportion of each full indicator to the sum of the standardized indicator values ​​in each historical adjustment cycle, the information entropy of each full indicator is calculated. The specific calculation formula is as follows: In the formula, i is the historical adjustment cycle number, i=1,2,3,...,m, and j is the total indicator number, j=1,2,3,...,n. Let j be the information entropy of the total index. This refers to the proportion of each total indicator to the sum of standardized indicator values ​​for each historical adjustment cycle. ; Let j be the sum of all standardized values ​​of the j-th total indicator over m historical adjustment periods. Let j be the standardized index value of the j-th full indicator in the i-th historical adjustment period, and calculate the entropy weight of each full indicator based on information entropy. The specific calculation formula is as follows: In the formula, Let the entropy weight of the j-th total index be denoted as . Information entropy redundancy reflects the degree of variation in the total data of the j-th index. The higher the degree of variation, the greater the information entropy redundancy and the greater the entropy weight. This represents the sum of the information entropy redundancy of all n full indicators, ensuring that the sum of the entropy weights of all indicators is 1. It determines whether the calibration deviation of the entropy weight of each full indicator from the preset benchmark weight exceeds the preset weight calibration deviation threshold. The preset benchmark weight is the average entropy weight based on historical adjustment cycles or the weights obtained through the analytic hierarchy process. If so, the number of historical adjustment cycles is dynamically adjusted based on the deviation between the preset weight calibration deviation threshold and the calibration deviation. The difference ratio between the current calibration deviation and the preset threshold is calculated. If the calibration deviation exceeds the threshold by a larger margin, the number of historical adjustment cycles included in the entropy weight calculation is increased proportionally by a preset step size to enhance sample stability and reduce the interference of abnormal cycles on the entropy weight. If the calibration deviation gradually converges... If the value falls within the threshold range, the number of cycles is reduced accordingly to the baseline level to avoid excessive smoothing that could lead to a decrease in sensitivity. During the adjustment process, dynamic upper and lower limits are set for the number of cycles to ensure a balance between computational efficiency and data sufficiency. This continues until the calibration deviation between the recalculated entropy weight and the preset baseline weight meets the threshold requirement. The entropy weight is then recalculated until the calibration deviation meets the requirement. Otherwise, the weights of each full indicator and the entropy weight are directly weighted geometrically to obtain a comprehensive score for the importance of each full indicator. Full indicators with a comprehensive score for importance greater than or equal to the preset core threshold are designated as core indicators, and full indicators with a comprehensive score for importance lower than the preset core threshold are removed to form a core indicator set.

[0024] In this embodiment, the present invention constructs a positive-negative judgment matrix using the 1-9 scaling method and calculates the consistency ratio by combining the maximum eigenvalue and eigenvector. Simultaneously, it dynamically adjusts resource inputs such as the number of expert scoring iterations, the number of experts, and the correction range of matrix elements based on the consistency ratio exceeding the standard. This effectively solves the problems of poor consistency of the judgment matrix and unreasonable utilization of expert resources in the traditional analytic hierarchy process (AHP), maximizing the efficiency of expert resource utilization while ensuring the scientific rigor of the judgment matrix. By standardizing the raw data of all indicators and calculating the indicator entropy weights based on the information entropy formula, objective weighting can be achieved based on the degree of variation of the data itself, avoiding the one-sidedness of single subjective weighting. Furthermore, by dynamically adjusting the number of historical adjustment cycles, the present invention adapts the calibration bias between the entropy weights and the benchmark weights. The method enhances sample stability and reduces abnormal periodic interference while avoiding the decrease in indicator sensitivity caused by excessive smoothing. Furthermore, it balances computational efficiency and data sufficiency by setting upper and lower limits for periodic adjustments. Finally, it performs weighted geometric calculations by combining the weights obtained from the analytic hierarchy process (AHP) and the entropy weights obtained from the information entropy method to obtain a comprehensive score for indicator importance. Core indicators are then selected based on their scores to form a standardized indicator set. This achieves a deep integration and weighting of subjective expert experience judgment and objective data information value. It can accurately and scientifically select core indicators with key value for global supply chain index analysis from the full set of indicators, significantly improving the rationality, relevance, and reliability of the indicator system. This provides a precise and standardized indicator foundation for subsequent supply chain data processing, index calculation, and diagnostic optimization.

[0025] Furthermore, the dynamic adjustment of indicator weights specifically includes: collecting business performance data within a preset historical adjustment period, including the actual values, achievement rates, and deviation values ​​of each core indicator within the preset historical adjustment period; discretizing the business performance data to construct a transaction dataset with the period as the transaction and the indicator performance status as the item, where each transaction corresponds to a historical adjustment period, and the items in the transaction correspond to the performance status of each core indicator. Discretization is used to transform quantitative indicators into an itemset format suitable for association rule mining; constructing a frequent pattern tree based on preset minimum support and minimum confidence; selecting transaction datasets with support greater than or equal to minimum support and confidence greater than or equal to minimum confidence to form association rules. Support is the proportion of transactions containing both the antecedent and consequent of the rule to the total number of transactions, reflecting the universality of the rule; confidence is the proportion of transactions containing both the antecedent and consequent of the rule to the number of transactions containing only the antecedent, reflecting the reliability of the rule; and calculating... Calculate the lift of each association rule, predefine a lift threshold, and remove association rules with lift less than the threshold. The remaining association rules are recorded as valid association rules. Valid association rules with a significant impact on the index are retained. Based on the valid association rules, calculate the association impact coefficient of each core indicator, dynamically adjust the weights of each core indicator, calculate the lift of each valid association rule, normalize the lift of each rule to obtain the weight of each rule, and use the normalized lift as the weight to allocate the impact to the indicators involved in the rule. Calculate the impact value obtained by each indicator from being an antecedent and consequent, and sum the two to obtain the comprehensive association impact coefficient of the indicator. Weight and fuse the original weights of each core indicator with the corresponding comprehensive association impact coefficient to obtain the corrected indicator weights. If the deviation between the corrected indicator weights and the historical dynamic weights is greater than the weight change tolerance threshold, the weighting operation is re-executed until the deviation meets the requirements.

[0026] In this embodiment, the present invention collects and discretizes business performance data such as the actual values, achievement rates, and deviation values ​​of each core indicator within a preset historical adjustment period, constructing a transaction dataset with the period as the transaction and the indicator performance status as the item. This can transform quantitative supply chain performance indicators into an itemset format suitable for association rule mining, effectively solving the technical problem that traditional quantitative data cannot be directly used for intrinsic association mining between indicators, and laying a standardized data foundation for accurately identifying indicator linkage relationships. Based on minimum support and minimum confidence, a frequent pattern tree is constructed and corresponding rules are filtered, enabling the extraction of universal and reliable indicator association rules from massive historical transaction data, eliminating scattered and unrepresentative invalid associations, and ensuring the effectiveness and practicality of association rule mining. By calculating the lift of association rules and setting a threshold to eliminate weak association rules with insufficient lift, only effective association rules that significantly affect the supply chain index are retained, avoiding interference from irrelevant or inefficient rules on weight correction, further enhancing the reference value and guiding significance of the rules. Based on the effective association rules, the association influence coefficient of each core indicator is calculated, and the normalized lift is then used to calculate the association influence coefficient of each core indicator. The weighting of indicators is assigned as a rule weight, and the comprehensive influence coefficient is obtained by considering the influence value of the comprehensive indicators as both antecedents and consequents of the rules. The original weights of each core indicator are then weighted and integrated with the corresponding comprehensive influence coefficients to obtain the corrected dynamic weights. This approach fully aligns with the actual operational rules of mutual influence and linkage among indicators in the global supply chain, realizing the transformation of indicator weights from static fixed to dynamic adaptive adjustment. This fundamentally solves the shortcomings of traditional weight settings, which are rigid and cannot reflect the true strength and degree of correlation between indicators. At the same time, by setting a tolerance threshold for weight changes, the deviation between the corrected dynamic weights and historical dynamic weights is verified. If the deviation exceeds the threshold, the weighting operation is re-executed, which can effectively avoid abnormal weight fluctuations and ensure the stability and rationality of the weight correction process. Overall, this approach achieves dynamic, scientific, and precise correction of the weights of core indicators in the global supply chain, significantly improving the adaptability of indicator weights to actual supply chain operation scenarios. This provides a reliable and dynamically adaptable weight basis for the accurate calculation of the global supply chain comprehensive index and the sub-indices of each core dimension, as well as for the diagnosis and optimization of the supply chain index.

[0027] Furthermore, the dynamic parameter adjustment mechanism specifically includes: dynamic processing of the contribution direction of supply chain performance indicators based on internal and external supply chain data for the target year: data where the supply chain performance indicator reaches the preset supply chain performance threshold is recorded as a positive indicator and standardized in its original direction; data where the negative supply chain performance indicator does not reach the preset supply chain performance threshold is recorded as a negative indicator, reversed, and then standardized; detection of the distribution characteristics of internal and external supply chain data after dynamic processing of contribution direction, if any data conforms to a normal distribution, the proportion of the difference between the original value of the indicator and the mean of the indicator to the standard deviation of the indicator is used as the indicator standard. Standardized values; if any data is skewed, the difference between the original value of the indicator and the minimum value of the indicator is taken as the proportion of the difference between the maximum and minimum values ​​of the indicator as the standardized value of the indicator; data whose standardized values ​​are outside three standard deviations are recorded as outliers. If outliers exist, the mean of the standardized parameters is dynamically accumulated based on a preset adjustment factor (twice the change in the mean) to obtain the adjusted mean of the standardized parameters, and the standardization process is repeated until the proportion of the impact of outliers on the standardization result is lower than the preset impact proportion threshold, which is less than or equal to three percent; if there are no outliers, no additional processing is performed.

[0028] In this embodiment, the present invention dynamically processes the contribution direction of supply chain performance indicators in the target year's internal and external supply chain data. Performance data reaching a preset threshold is recorded as positive indicators and standardized in its original direction; performance data not reaching the preset threshold is recorded as negative indicators and undergoes reverse transformation before standardization. This unifies the quantitative caliber of indicators with different contribution directions, effectively solving the technical problems of mixed positive and negative indicators, inconsistent dimensions, and inconsistent evaluation logic in traditional standardization processes, ensuring the benchmark fairness and comparability of subsequent index calculations. By detecting the distribution characteristics of internal and external data after dynamic contribution direction processing, standardization is applied to normally distributed data based on the indicator mean and standard deviation, and to skewed data based on indicator extreme values. This achieves adaptive matching between the standardization algorithm and the actual data distribution characteristics, avoiding the use of a single standard. The standardization model addresses distortion and bias issues arising under different data distributions, significantly improving the accuracy and adaptability of indicator standardization. By identifying outliers within a three-standard-deviation interval and dynamically accumulating and iterating the standardization process using a preset adjustment factor to correct the mean of the standardization parameters until the impact of outliers on the standardization results falls below a preset threshold of three percent, the excessive interference of extreme outliers on the standardization results can be effectively reduced without disrupting the overall data patterns, thus enhancing the anti-interference capability and robustness of data processing. Overall, this dynamic parameter adjustment mechanism achieves adaptive, high-precision, and highly robust standardization processing of heterogeneous data from multiple sources within and outside the global supply chain. This provides a real, standardized, and reliable data foundation for the accurate calculation of the subsequent global supply chain comprehensive index and its core dimension sub-indices, further improving the accuracy, stability, and credibility of the index results.

[0029] Furthermore, the calculation of the global supply chain composite index and the sub-indices of each core dimension specifically includes: after standardization, based on the global supply chain index matrix, a weighted geometric calculation is performed using the corrected indicator weights to obtain the global supply chain composite index, which is used to quantitatively assess the supply chain performance level of the target year. This involves multiplying the corrected indicator weights of each indicator with their standardized values ​​and then summing the results. For each core dimension, a weighted geometric calculation is performed using the corrected indicator weights of each core indicator under that dimension and their corresponding standardized values ​​to obtain the sub-indices of each core dimension. The sub-indices of each core dimension and the global supply chain composite index for the target year are output, and the standardized parameters and weights are recorded for subsequent comparisons.

[0030] In this embodiment, based on the prior standardization of dynamic parameters and dynamic correction of indicator weights for internal and external supply chain data, and using the global supply chain index matrix, a weighted geometric calculation is performed on the corrected dynamic weights of each core indicator and their standardized values, and the results are accumulated to obtain the global supply chain composite index. This provides a unified, intuitive, and quantitative comprehensive assessment of the overall operational performance of the global supply chain in a target year, accurately reflecting the overall operational efficiency and comprehensive performance level of the global supply chain. Simultaneously, for the five core dimensions of planning, procurement, production, distribution, and returns under the SCOR model, a weighted geometric calculation is performed separately on the corrected dynamic weights of each core indicator within that dimension and their corresponding standardized values ​​to obtain the sub-indices for each core dimension. This allows for the independent decomposition and analysis of each key operational link of the global supply chain. This refined and quantitative assessment not only grasps the overall state of the supply chain from a macro perspective but also accurately identifies the operational level of each link from a detailed dimension. By outputting the sub-indices of each core dimension of the target year and the global supply chain comprehensive index, and simultaneously recording the standardized parameters and corrected dynamic weights used throughout the process, it provides clear and quantifiable results for global supply chain performance evaluation. It also enables the retention and filing of key evaluation parameters and weights, providing traceable, verifiable, and reproducible data for subsequent horizontal comparisons, vertical traceability, index diagnostic analysis, and optimization scheme formulation of supply chain indices across different years. This effectively ensures the accuracy, comparability, and stability of the supply chain index calculation results, further enhancing the evaluation accuracy, scientific rigor, and practical application value of the global supply chain index AI analysis method.

[0031] Furthermore, the supply chain index diagnosis and optimization analysis includes a longitudinal comparison mechanism, specifically: comparing the standardized values ​​of each core indicator in the target year with the corresponding core indicator values ​​in the benchmark year to calculate the rate of change of each core indicator; recording the difference between the composite index of the target year and the composite index of the benchmark year as the total change in the composite index; recording the proportion of the weighted result of the rate of change of each core indicator and the corrected indicator weights to the total change in the composite index as the contribution of each core indicator to the change in the composite index; recording core indicators whose absolute contribution value and absolute value of the rate of change both exceed the preset thresholds for the rate of change and the preset threshold for the contribution value as significant impact indicators; comparing the rate of change of each significant impact indicator with 0, marking significant impact indicators with a rate of change greater than 0 as improvement indicators, and marking significant impact indicators with a rate of change less than 0 as deterioration indicators; sorting improvement indicators by contribution value from largest to smallest, and sorting deterioration indicators by absolute contribution value from largest to smallest, forming a list of significant improvement indicators and a list of significant deterioration indicators, which are included in the longitudinal trend analysis report.

[0032] In this embodiment, the present invention compares the standardized values ​​of each core indicator in the target year with the corresponding indicator values ​​in the base year and calculates the rate of change of the indicators. This allows for the precise capture of the fluctuation range and trend of individual core indicators in the longitudinal time dimension, clearly presenting the development and changes of each indicator from the base period to the target year, laying a precise foundation for individual indicator changes in subsequent analysis. By calculating the difference between the composite index in the target year and the base year, the total change in the composite index can be obtained, which can intuitively quantify the overall change level of the global supply chain performance in the longitudinal time dimension, clarifying whether the overall development trend of the supply chain is improving or declining. By weighting the rate of change of each core indicator with the corrected dynamic weights and then calculating its proportion in the total change of the composite index, the contribution of each indicator to the change of the composite index can be obtained. This can accurately define the degree of influence of individual core indicators on the overall performance change of the supply chain, solving the technical problem that it is difficult to distinguish the contribution weight of each indicator and locate the core influencing factors in traditional longitudinal comparisons. By setting preset thresholds for the rate of change of indicators and preset thresholds for contribution, significant indicators whose absolute values ​​of contribution and absolute values ​​of the rate of change of indicators both exceed the thresholds are selected. The impact indicators effectively eliminate interference from indicators with minor fluctuations and minimal impact on overall performance, accurately identifying core indicators that play a crucial role in changes to the comprehensive supply chain index. By categorizing significant impact indicators into improvement indicators and deterioration indicators based on their rate of change compared to 0, and then ranking them by contribution (or absolute value of contribution) to form corresponding lists and incorporating them into the longitudinal trend analysis report, this not only clearly presents the advantageous and disadvantageous indicators in supply chain operations but also provides clear and targeted basis for the formulation of subsequent optimization plans. This facilitates targeted strengthening of the advantages of improvement indicators and rectification of the disadvantages of deterioration indicators. Overall, this longitudinal comparison mechanism achieves a full-dimensional longitudinal source-tracing analysis of global supply chain performance, from individual indicators to the comprehensive index, and from local fluctuations to overall changes. It can clearly outline the longitudinal development trend of supply chain performance and accurately locate the key core indicators affecting performance changes, providing strong support for the scientific and accurate diagnosis of the supply chain index. This significantly improves the targeting and practicality of supply chain optimization analysis, helping enterprises or related entities to fully grasp the longitudinal development patterns of the supply chain and make more forward-looking and targeted management decisions.

[0033] Furthermore, the supply chain index diagnosis and optimization analysis also includes external benchmarking mechanisms and weakness identification mechanisms. The external benchmarking mechanism includes: acquiring publicly available or third-party supply chain benchmark data from the industry; processing the benchmark data using the same standardization method as the target year to obtain the industry benchmark index and its sub-indices; comparing the company's target year index with the industry benchmark index to obtain the gap values ​​for each dimension and each core indicator; based on the gap values ​​and combined with the corrected indicator weights, identifying the direction for external benchmarking optimization; firstly, calculating the difference between the sub-indices of each core dimension and the industry benchmark sub-indices to identify weak dimensions below the industry benchmark; then calculating the standardized values ​​of each core indicator and the industry benchmark indicator values. The difference and percentage of the gap are analyzed, and the contribution of each indicator to the overall gap between the company and the industry benchmark is analyzed in combination with the corrected indicator weights. Indicators whose contribution exceeds the preset contribution threshold are identified as external benchmarking optimization directions. The shortcoming identification mechanism includes: decomposing the company's strategic goals into target values ​​for each core dimension and each core indicator, and calculating the gap between the current core indicator value and the target value; performing dynamic weight calculation by combining the corrected indicator weights with the gap values ​​of each core indicator to obtain the degree of obstruction of each core indicator to the achievement of strategic goals; sorting each core indicator from largest to smallest according to the degree of obstruction, selecting the top preset number of core indicators as shortcoming indicators, and formulating optimization plans based on the shortcoming indicators.

[0034] By acquiring publicly available or third-party supply chain benchmark data through external benchmarking mechanisms, and using standardized methods consistent with the target year to uniformly process the benchmark data, industry benchmark indices and sub-indices are obtained. This ensures that the company's own index is compared with the industry benchmark index under the same quantitative caliber, eliminating benchmarking bias caused by inconsistent data processing standards at its root. By comparing the company's target year index with the industry benchmark index, the gap values ​​of each dimension and core indicator are obtained. Combined with adjusted dynamic weights, weak dimensions below the industry benchmark are first identified, and then the contribution of each core indicator to the overall gap between the company and the industry benchmark is accurately calculated. Indicators with a contribution exceeding a preset threshold are identified as external benchmarking optimization directions. This allows for precise positioning of the core gap between the company's supply chain and the benchmark level from an industry competition perspective, effectively avoiding the problems of generalized comparison and vague focus in traditional benchmarking analysis, and significantly improving the accuracy and directionality of external benchmarking analysis. At the same time, through the weakness positioning mechanism, the company's strategic goals are decomposed into specific target values ​​for each core dimension and core indicator, and the current... The gap between core indicator values ​​and target values ​​is weighted using modified dynamic weights to calculate the degree of obstruction of each core indicator to the achievement of strategic goals. Indicators are ranked from highest to lowest obstruction level, and the top indicators are selected as bottleneck indicators. Optimization plans are then developed based on these bottleneck indicators. This approach objectively and accurately identifies key bottleneck indicators that constrain supply chain performance improvement and hinder the achievement of strategic goals from an internal strategic implementation perspective. It overcomes the shortcomings of traditional bottleneck identification methods, which rely on subjective judgment, lack data support, and fail to consider weighting. This achieves a scientific, quantitative, and precise identification of bottlenecks. Overall, this dual mechanism identifies competitive gaps through external industry benchmarking and developmental weaknesses through internal strategic benchmarking. This combination of internal and external approaches builds a comprehensive, multi-layered supply chain diagnostic system. Dynamic weighting allows for precise targeting of optimization directions and bottleneck indicators, providing data-driven and targeted decision-making support for supply chain optimization plans. This significantly improves the scientific rigor, practicality, and implementability of supply chain index diagnosis and optimization analysis, helping companies efficiently narrow industry gaps, steadily achieve strategic goals, and comprehensively enhance the overall competitiveness and strategic adaptability of the global supply chain.

[0035] Specifically, the optimization plan based on the bottleneck indicators includes: analyzing influencing factors based on the bottleneck indicator list output by the bottleneck identification mechanism and the results of association rule mining between indicators; using bottleneck indicators as consequents of association rules, searching for association rules in the transaction dataset that meet preset conditions in terms of support, confidence, and lift, identifying antecedent indicators strongly correlated with bottleneck indicators, and tracing the root causes of the identified antecedent indicators in conjunction with business logic to determine the core influencing factors leading to the bottleneck; generating multiple candidate optimization measures for each bottleneck indicator based on the influencing factor analysis results, including but not limited to process improvement, technology upgrades, and resource allocation adjustments; and constructing a analytic hierarchy process (AHP) judgment matrix or using a weighted scoring method to weight the performance of each candidate optimization measure under each decision criterion, using enterprise resource constraints, implementation difficulty, and expected benefits as decision criteria. The calculation process yields a comprehensive priority score for each candidate optimization measure, which is then ranked from highest to lowest. Resource constraints include funding, human resources, and time investment. Implementation difficulty is assessed based on factors such as technological maturity and organizational synergy. Expected benefits are evaluated based on factors such as index improvement potential and cost savings. For the top-ranked candidate optimization measures, corresponding implementation plans are developed and integrated into an overall optimization scheme. The implementation plan includes quantifiable target values, responsible departments, timelines, and monitoring indicators, clarifying the dependencies and synergistic effects between measures. The execution effect of the overall optimization scheme is tracked and evaluated in the next index analysis cycle. By comparing the changes in the index of the next cycle with the current cycle index, the implementation effect of the optimization measures is assessed. The index analysis cycle refers to a fixed time period for calculating the global supply chain index, including quarterly, semi-annual, or annual periods.

[0036] In this embodiment, a list of shortcomings based on the shortcomings identification mechanism is used, combined with the results of association rule mining between indicators, to conduct influencing factor analysis. Shortcoming indicators are used as consequents of association rules. Valid association rules that meet preset conditions in terms of support, confidence, and lift are selected from the transaction dataset. This accurately identifies antecedent indicators strongly correlated with the shortcomings indicators, and, combined with business logic, completes root cause tracing. This objectively and accurately locates the core influencing factors that form the shortcomings indicators, fundamentally avoiding the problems of relying on subjective experience and vague / one-sided cause identification in traditional optimization scheme formulation. It provides a reliable and targeted basis for the generation of subsequent optimization measures. For each shortcomings indicator, candidate optimization measures covering multiple dimensions such as process improvement, technology upgrade, and resource allocation adjustment are generated based on the influencing factor analysis results, enriching the optional paths and applicable scenarios of the optimization scheme and ensuring the measures are targeted and feasible. Using enterprise resource constraints, implementation difficulty, and expected benefits as core decision-making criteria, a weighted geometric calculation is performed on each candidate optimization measure using a analytic hierarchy process (AHP) judgment matrix or a weighted scoring method to obtain a comprehensive priority score and rank them. By considering the constraints of resource input (funding, manpower, time, etc.), implementation difficulties (technology maturity, organizational coordination), and expected benefits (index improvement potential, cost savings, etc.), this approach achieves a scientific and quantitative ranking of optimization measures. It effectively addresses the technical problems of traditional measure selection, such as blindness, unreasonable resource allocation, and low cost-effectiveness. For the top-ranked candidate optimization measures, implementation plans are developed, including quantifiable target values, responsible departments, timelines, and monitoring indicators. The dependencies and synergies between measures are clarified and integrated into a holistic optimization plan. The effectiveness of this overall optimization plan is tracked and evaluated within fixed quarterly, semi-annual, or annual index analysis cycles. The implementation effect of optimization measures is quantitatively assessed by comparing index changes between the current and next cycles. This constructs a full-process supply chain optimization system encompassing weakness identification, cause tracing, measure generation, priority ranking, plan implementation, and closed-loop effect evaluation. This not only significantly improves the scientific rigor, feasibility, and implementability of optimization plans but also enables the quantification, traceability, and iterative nature of supply chain optimization effects, continuously driving steady improvement in global supply chain performance and the self-improvement of index analysis methods.

[0037] Furthermore, the global supply chain index model AI analysis device is characterized by comprising: a core indicator construction module, a rule weight adjustment module, an index component calculation module, and a diagnostic solution formulation module; wherein, the core indicator construction module is used to construct a full-scale indicator library from preset core dimensions based on the supply chain operation reference SCOR model; the rule weight adjustment module is used to collect business performance data from historical adjustment cycles, mine the correlation rules between indicators, dynamically adjust the indicator weights and establish a benchmark year, and record the original data and corrected indicator weights of the core indicators in the benchmark year; the index component calculation module is used to collect internal and external supply chain data for the target year, perform standardization processing using a dynamic parameter adjustment mechanism, and calculate the global supply chain comprehensive index and the component indices of each core dimension in combination with the corrected indicator weights; the diagnostic solution formulation module is used to conduct supply chain index diagnosis and optimization analysis based on the data of the target year and the benchmark year, combined with the corrected indicator weights, to identify weak indicators and formulate optimization solutions.

[0038] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)). Where there is no conflict, the solutions in the above embodiments can be combined.

[0039] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0040] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0041] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0042] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0043] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope and intent of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and variations.

Claims

1. An AI analysis method for a global supply chain index model, characterized in that, Includes the following steps: S01, based on the SCOR model for supply chain operations, construct a full-scale indicator library from preset core dimensions, including planning, procurement, production, distribution and returns. Core indicators are selected through indicator importance scoring to form a core indicator set; S02, collect business performance data from historical adjustment cycles, mine the correlation rules between indicators, dynamically adjust indicator weights and establish a base year, and record the original data and corrected indicator weights of the core indicators in the base year. S03: Collect internal and external supply chain data for the target year, standardize the data using a dynamic parameter adjustment mechanism, and calculate the global supply chain composite index and sub-indices for each core dimension by combining the corrected indicator weights. S04, based on data from the target year and the baseline year, combined with the revised indicator weights, conducts supply chain index diagnosis and optimization analysis to identify weak indicators and formulate optimization plans.

2. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The selection of core indicators through indicator importance scoring specifically includes: Construct a positive and negative judgment matrix based on the full index data within each historical adjustment cycle; Calculate the largest eigenvalue and its corresponding eigenvector of the positive-reverse judgment matrix, and determine whether the consistency ratio of the positive-reverse judgment matrix reaches a preset consistency ratio threshold. If not, normalize the eigenvector to obtain the weights of each total index. If yes, dynamically adjust the resource input ratio of the positive-reverse judgment matrix based on the deviation between the consistency ratio threshold and the consistency ratio of the positive-reverse judgment matrix. The consistency ratio is the ratio of the consistency index to the preset average random consistency index. The consistency index is the deviation between the largest eigenvalue of the positive-reverse judgment matrix and the order of the judgment matrix, which accounts for the proportion of the deviation between the order of the judgment matrix and 1. The raw data of all indicators in each historical adjustment cycle are standardized to obtain a standardized matrix. Based on the proportion of each full indicator to the sum of standardized indicator values ​​in each historical adjustment cycle, the information entropy of each full indicator is calculated, and the entropy weight of each full indicator is calculated based on the information entropy. Determine whether the calibration deviation between the entropy weight of each full indicator and the preset benchmark weight exceeds the preset weight calibration deviation threshold. If so, dynamically adjust the number of historical adjustment cycles based on the deviation between the preset weight calibration deviation threshold and the calibration deviation, and recalculate the entropy weight until the calibration deviation meets the requirements. Otherwise, directly perform weighted geometric calculation on the weight and entropy weight of each full indicator to obtain the comprehensive score of the importance of each full indicator. All indicators whose overall importance score is greater than or equal to the preset core threshold are designated as core indicators, and all indicators whose overall importance score is lower than the preset core threshold are removed to form a core indicator set.

3. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The dynamically adjusted indicator weights specifically include: Collect business performance data within a preset historical adjustment period. The business performance data includes the actual values, achievement rates, and deviation values ​​of each core indicator within the preset historical adjustment period. Discretize the business performance data to construct a transaction dataset with the period as the transaction and the indicator performance status as the item. A frequent pattern tree is constructed based on the preset minimum support and minimum confidence. Transaction datasets with support greater than or equal to the minimum support and confidence greater than or equal to the minimum confidence are selected to form association rules. The support is the proportion of the number of transactions containing both the antecedent and consequent of the rule to the total number of transactions. The confidence is the proportion of the number of transactions containing both the antecedent and consequent of the rule to the number of transactions containing only the antecedent. Calculate the lift of each association rule, remove association rules whose lift is less than the lift threshold, and record the remaining association rules as valid association rules. Calculate the association influence coefficient of each core indicator based on the valid association rules, and dynamically adjust the weights of each core indicator to obtain the adjusted indicator weights. If the deviation between the corrected indicator weight and the historical dynamic weight exceeds the weight change tolerance threshold, the weighting operation will be re-executed until the deviation meets the requirements.

4. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The dynamic parameter adjustment mechanism specifically includes: The contribution direction of supply chain performance indicators based on internal and external supply chain data for the target year is dynamically processed: data that reach the preset supply chain performance threshold are recorded as positive indicators and standardized in the original direction; data that do not reach the preset supply chain performance threshold are recorded as negative indicators, reverse-converted and then standardized. The detection contribution direction is dynamically processed to detect the distribution characteristics of internal and external supply chain data. If any data conforms to a normal distribution, the proportion of the difference between the original value of the indicator and the mean of the indicator to the standard deviation of the indicator is used as the standardized value of the indicator. If any data is skewed, the proportion of the difference between the original value of the indicator and the minimum value of the indicator to the difference between the maximum value and the minimum value of the indicator is used as the standardized value of the indicator. Data whose standardized values ​​fall outside three standard deviations are considered outliers. If outliers exist, the mean of the standardized parameters is dynamically processed based on a preset adjustment factor to obtain the adjusted mean of the standardized parameters. The standardization process is then repeated until the proportion of the outlier's influence on the standardization result is lower than the preset influence proportion threshold. If there are no outliers, no additional processing is performed.

5. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The calculation of the global supply chain composite index and its core dimension sub-indices specifically includes: After standardization, a weighted geometric calculation is performed based on the global supply chain index matrix and the corrected indicator weights to obtain the global supply chain composite index, which is used to quantitatively assess the supply chain performance level in the target year. For each core dimension, the weighted geometric calculation is performed using the corrected weights of the core indicators under that dimension and their corresponding standardized values ​​to obtain the sub-indices of each core dimension. Output the sub-indices of each core dimension and the global supply chain composite index for the target year, and record the standardized parameters and weights for subsequent comparisons.

6. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The supply chain index diagnosis and optimization analysis includes a longitudinal comparison mechanism, specifically: The standardized values ​​of each core indicator in the target year are compared with the corresponding core indicator values ​​in the base year, and the rate of change of each core indicator is calculated. The difference between the composite index of the target year and the composite index of the base year is recorded as the total change in the composite index. The weighted average of the change rate of each core indicator and the adjusted indicator weights, as a percentage of the total change in the composite index, is recorded as the contribution of each core indicator to the change in the composite index. Core indicators whose absolute value of contribution and absolute value of indicator change rate both exceed the preset indicator change rate threshold and preset contribution threshold are defined as significant impact indicators. The rate of change of each significant impact indicator is compared with 0. Significant impact indicators with a rate of change greater than 0 are marked as improvement indicators, and those with a rate of change less than 0 are marked as deterioration indicators. Improvement indicators are sorted from largest to smallest in terms of contribution, and deterioration indicators are sorted from largest to smallest in terms of absolute value of contribution, forming a list of significant improvement indicators and a list of significant deterioration indicators.

7. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The supply chain index diagnosis and optimization analysis also includes external benchmarking mechanisms and weakness identification mechanisms: The external benchmarking mechanism includes: acquiring publicly available or third-party supply chain benchmark data; processing the benchmark data using the same standardization method as the target year to obtain the industry benchmark index and sub-indices; comparing the company's target year index with the industry benchmark index to obtain the gap values ​​for each dimension and each core indicator; and identifying the external benchmarking optimization direction based on the gap values ​​and the corrected indicator weights. The weakness identification mechanism includes: decomposing the enterprise's strategic goals into target values ​​for each core dimension and each core indicator, calculating the gap between the current core indicator value and the target value; performing dynamic weight calculation by combining the corrected indicator weights with the gap values ​​of each core indicator to obtain the degree of obstruction of each core indicator to the achievement of the strategic goals; sorting each core indicator from largest to smallest according to the degree of obstruction, selecting the top-preset number of core indicators as weakness indicators, and formulating optimization plans based on the weakness indicators.

8. The AI ​​analysis method for the global supply chain index model as described in claim 7, characterized in that: The optimization plan based on the weakest link indicator specifically includes: Based on the list of shortcomings output by the shortcomings location mechanism, and combined with the results of the association rule mining between the indicators, the influencing factors are analyzed: the shortcomings indicators are used as the consequents of the association rules, and the association rules that meet the preset conditions of support, confidence and lift are searched from the transaction dataset. The antecedent indicators that are strongly correlated with the shortcomings indicators are identified, and the root causes of the identified antecedent indicators are traced in combination with the business logic to determine the core influencing factors that lead to the formation of shortcomings. For each bottleneck indicator, multiple candidate optimization measures are generated based on the analysis of influencing factors. These candidate optimization measures include, but are not limited to, process improvement, technology upgrade, and resource allocation adjustment. Using enterprise resource constraints, implementation difficulty, and expected benefits as decision-making criteria, a weighted geometric calculation is performed on the performance of each candidate optimization measure under each decision-making criterion to obtain the comprehensive priority score of each candidate optimization measure, and the measures are ranked from high to low according to the comprehensive priority score. For the candidate optimization measures that rank first by a predetermined number, corresponding implementation plans are formulated and integrated into an overall optimization scheme. The implementation plan includes quantifiable target values, responsible departments, time nodes, and monitoring indicators. The execution effect of the overall optimization scheme is included in the next round of index analysis cycle for tracking and evaluation.

9. The AI ​​analysis method for the global supply chain index model as described in claim 1, characterized in that: The core indicators include resilience index, connectivity index, innovation index, facilitation index, and security index, among which: The resilience index is used to measure the supply chain’s ability to cope with and recover from disruptions. It is obtained by weighted geometric calculation of delivery reliability, inventory turnover and emergency response time indicators. The connectivity index is used to measure the level of collaboration and information sharing among nodes in the supply chain. It is obtained by weighted geometric calculation of information integration degree, order on-time delivery rate, and supplier collaboration indicators. The innovation index is used to measure the performance of the supply chain in terms of technology application and process innovation. It is obtained by weighted geometric calculation of the R&D investment ratio, new technology adoption rate and process optimization frequency. The aforementioned facilitation index is used to measure the positive impact of the supply chain on the economy, society, and environment. It is obtained by weighted geometric calculation of local sourcing rate, carbon emission intensity, and employee satisfaction indicators. The security index is used to measure the level of the supply chain in terms of compliance, risk control and information security. It is obtained by weighted geometric calculation of the number of compliance events, risk identification coverage and data breach frequency.

10. An apparatus for applying the AI ​​analysis method of a global supply chain index model as described in any one of claims 1-9, characterized in that, include: The module includes a core indicator construction module, a rule weight adjustment module, an index sub-item calculation module, and a diagnostic solution formulation module. The core indicator construction module is used to build a full indicator library from preset core dimensions based on the SCOR supply chain operation reference model. The rule weight adjustment module is used to collect business performance data from historical adjustment cycles, mine the correlation rules between indicators, dynamically adjust the indicator weights and establish a base year, and record the original data and corrected indicator weights of the core indicators in the base year. The index sub-item calculation module is used to collect internal and external supply chain data for the target year, and uses a dynamic parameter adjustment mechanism for standardization processing. It then calculates the global supply chain comprehensive index and sub-indices for each core dimension by combining the corrected index weights. The diagnostic scheme formulation module is used to conduct supply chain index diagnosis and optimization analysis based on target year and base year data, combined with the corrected indicator weights, to identify weak indicators and formulate optimization schemes.

Images