Big data-based supply chain procurement data management method and system

Abstract

This invention discloses a supply chain procurement data management method and system based on big data, belonging to the field of data management technology. The method includes collecting resource status data from the supply chain procurement process, forming a node dataset, constructing a SARIMA model, analyzing the trend offset constant and calculating the ADF test statistic based on the unit root coefficient, performing seasonal decomposition on the differenced sequence of the target variable, applying the ADF to verify stationarity, and determining the non-seasonal difference frequency. The method described in this invention obtains a seasonal intensity index by periodically normalizing the predicted seasonal fluctuation values, and combines it with historical service quantile statistics to dynamically map demand forecasts to service levels. It calculates capacity quantiles by statistically analyzing actual arrival volumes within the seasonal phase and superimposes the incoming material defect rate to form an available capacity value, enabling the model to reflect not only the supplier's delivery speed.

Description

Technical Field

[0001] This invention relates to the field of data management technology, and in particular to a supply chain procurement data management method and system based on big data. Background Technology

[0002] Supply chain management is the core of modern enterprises' efforts to improve resource efficiency. The stability of the supply chain has a direct impact on the actual inbound and outbound operations of enterprises, and the resulting operational efficiency directly affects the enterprise's market competitiveness. With the promotion of the concept of big data technology management, simple linear inference models are often used for enterprise procurement data. However, traditional rule-based decision-making and linear models rely too heavily on historical experience data and lack adaptability to dynamic changes. Especially when facing scenarios with seasonality, random fluctuations, and complex supplier collaborations, it is difficult to accurately fit and predict the data, leading to a decline in supply chain collaboration efficiency. Secondly, when allocating resources, traditional technologies rarely consider the interaction of multi-dimensional factors, such as the supplier's fulfillment capabilities, the quality of incoming materials, and the impact of different seasonal phases on demand fluctuations. Therefore, problems such as unreasonable resource allocation or insufficient fulfillment capabilities may occur. Summary of the Invention

[0003] In view of the aforementioned existing problems, the present invention is proposed.

[0004] Therefore, this invention provides a supply chain procurement data management method and system based on big data to address the problem that traditional rule-based decision-making and linear models rely too heavily on historical experience data and lack adaptability to dynamic changes. In particular, when facing scenarios with seasonality, random fluctuations, and complex supplier collaboration, it is difficult to accurately fit and predict the data, resulting in a decline in supply chain collaboration efficiency. Secondly, when allocating resources, traditional technologies rarely consider the interaction of multi-dimensional factors.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a supply chain procurement data management method based on big data, comprising: Resource status data from the supply chain procurement process is collected to form a node dataset. A SARIMA model is constructed, the offset trend constant is analyzed, and the ADF test statistic is calculated based on the unit root coefficient. The target variable difference series is seasonally decomposed using STL decomposition. The stationarity is verified by applying ADF and the non-seasonal difference order is determined. The ACF curve is plotted to determine the non-seasonal autoregression order. The difference order is determined by the seasonal fluctuation value series. The influence of the white noise term on the cyclical trend is analyzed and the corresponding PACF curve is formed. Non-seasonal and seasonal parameters are determined. Seasonal component values ​​are defined to calculate the seasonal intensity index. The data is predicted using the SARIMA model as the arrival intensity value. Data statistics are collected from the time the supplier places an order to the time the goods arrive. Delivery dates are determined based on the unified service quantile. The available material capacity value is calculated. Based on the available material capacity value, the supplier's promised supply range in the seasonal phase is defined. A quantile curve is constructed. Historical utilization rate values ​​are calculated and stability thresholds are defined. The lower limit of supply is determined by combining the arrival intensity value. Supplier allocation constraints are defined and allocation schemes are screened. Define the unit comprehensive monetary consumption value, refer to the supply allocation, calculate the cooperation gain, and determine the gain; By ranking suppliers' cumulative capabilities using the unit comprehensive monetary consumption value, determining the allocation set, and calculating the commitment ratio, risk assessment is performed.

[0006] As a preferred embodiment of the big data-based supply chain procurement data management method of the present invention, the following steps are included: determining the difference order through seasonal fluctuation value sequences, constructing corresponding PACF curves based on the influence of white noise terms on cyclical trends, determining non-seasonal and seasonal parameters, defining seasonal component values ​​to calculate the seasonal intensity index, and using the SARIMA model to predict data as arrival intensity values. Based on the current resource status of nodes in the supply chain procurement process, a SARIMA model is constructed. The historical procurement demand data of the node dataset is used as the target variable, and the ADF algorithm is used to calculate the difference sequence of the target variable. The offset trend constant is analyzed and the ADF test statistic is calculated based on the unit root coefficient. The critical value is determined by using the ADF test table and the stationarity is verified by the test statistic. If the test statistic is less than the critical value, it indicates that the difference sequence of the target variable is stationary. STL decomposition is used to perform seasonal decomposition on the differenced series of the target variable, breaking the series down into long-term trend and seasonal fluctuations. The time series of long-term trend values ​​are used to verify stationarity and determine the non-seasonal differencing order. An ACF curve is plotted based on the differenced series of long-term trend values ​​to verify significance and determine the non-seasonal autoregression order. Non-seasonal noise is detected based on the partial autocorrelation function. The lag order of the PACF test curve that exceeds the confidence interval is selected as the non-seasonal moving average order. For seasonal fluctuation values, the significant long-period lag term is detected by ACF to determine the seasonal autoregressive order. The difference order is determined by the seasonal fluctuation value sequence. The influence of white noise term on the periodic fluctuation and residual on the periodic trend is used to form the corresponding PACF curve. If the significant segment in the PACF curve exceeds the significance level for the first time, the significant lag step is defined as the seasonal moving average order. The non-seasonal autoregression order pa, the non-seasonal difference number, and the non-seasonal moving average order q are used to form non-seasonal parameters. The seasonal autoregression order P, the seasonal difference number D, the seasonal moving average order, and the lag period s extracted from Q are used to form seasonal parameters. The SARIMA model is then used to predict the target variable. The total number of seasonal fluctuation values ​​in each cycle is taken as the seasonal component value, and the seasonal intensity index is calculated by mean normalization. The historical service quantiles are statistically analyzed. If there is "no shortage of goods in the current period and the goods are on time", the quantile is defined as 1, otherwise it is 0. The mean is calculated to form a uniform service quantile, and the SARIMA model prediction output is used as the arrival intensity value.

[0007] As a preferred embodiment of the big data-based supply chain procurement data management method of the present invention, the following steps are included: defining the supplier's committable supply range in a seasonal phase based on the incoming material availability value, constructing a quantile curve, calculating historical utilization rate values ​​and defining a stability threshold, determining the lower bound of supply based on arrival intensity values, defining supplier allocation constraints, and screening allocation schemes, including... Statistics are compiled based on the time from order placement to delivery by the supplier, and the delivery date is determined according to the unified service quantile. At the same time, the actual delivery volume of the supplier in the seasonal phase is calculated to determine the capacity quantile. Combined with the incoming material defect rate, the available capacity value of incoming materials is calculated. Based on the available material supply capacity, the supplier's commit supply range in the seasonal phase is defined, and a quantile curve is constructed. Based on the actual arrival volume data, the historical utilization rate value is statistically calculated and a stability threshold is defined. Combined with the arrival intensity value, the supply lower bound for the current period is calculated based on demand inflow and stability requirements. The total supply allocation of all suppliers in the current period must meet the constraints required for system stability, including that the total supply must not be less than the lower limit of supply, the allocation of each supplier must be within the range of commensurable supply, and supplier allocation schemes that meet the constraints must be selected.

[0008] As a preferred embodiment of the big data-based supply chain procurement data management method of the present invention, the step of defining a unit comprehensive monetary consumption value, referencing supply allocation, and calculating cooperation gain to determine gain includes, The comprehensive monetary consumption value per unit is defined by the overall purchase price, transaction cost per unit, defect rate, and scrap cost. Based on the unit comprehensive monetary consumption, combined with the total allocation of all suppliers in the current period, the amount of shortage penalty stipulated in the contract, and the supply allocation, the cooperation gain is calculated; The gain threshold is based on the sum of the mean and standard deviation of historical cooperative gains. If the cooperative gain is greater than or equal to the gain threshold, the supplier gain is marked as good and the next supplier is introduced. If the cooperative gain is less than the gain threshold, the supplier gain is marked as low and the expansion of suppliers is stopped.

[0009] As a preferred embodiment of the supply chain procurement data management method based on big data described in this invention, the step of determining the allocation set by sorting the cumulative capacity sequence of suppliers based on the unit comprehensive monetary consumption value includes, All suppliers are sorted in ascending order based on the unit comprehensive monetary consumption to form an order of monetary value. Starting from the first supplier, the available capacity value of each supplier is accumulated in turn to calculate the total available capacity of different suppliers, forming a cumulative capacity sequence of different levels. Based on the comparison between the cumulative capacity sequence and the lower bound of supply, the minimum feasible supplier set is selected as the allocation boundary. The allocation rules are defined: if the supplier belongs to the supplier set before the allocation boundary, the allocation amount is the available material capacity value; if the supplier belongs to the allocation boundary, only the remaining amount that makes the total supply reach the lower bound of supply is allocated, and the rest is not allocated. The final allocation of each supplier for the current period.

[0010] As a preferred embodiment of the big data-based supply chain procurement data management method of the present invention, the calculation of the commitment ratio for risk assessment includes, The ratio of the actual allocated quantity to the upper limit of available capacity in the final allocated quantity set is calculated as the commitment percentage. The analysis threshold is based on the sum of the mean and two standard deviations of the supplier's historical commitment percentage. If the actual commitment percentage is greater than or equal to the analysis threshold, it is judged that there is a risk and a risk alert is issued.

[0011] As a preferred embodiment of the big data-based supply chain procurement data management method of the present invention, wherein: the collection of resource status data in the supply chain procurement process forms a node dataset, including, Collect the current resource status of nodes in the supply chain procurement process, including inventory status, logistics status, procurement demand data, timestamps, order volume, and transportation status, and form a node dataset. Define the normal and abnormal range of supply chain behavior, including abnormal order delays, abnormal inventory shortages, and logistics and transportation disruptions.

[0012] Secondly, this invention provides a supply chain procurement data management system based on big data, including: Data acquisition module: Collects resource status data from the supply chain procurement process and generates node datasets; Demand forecasting module: Constructs a SARIMA model, performs seasonal decomposition and sequence data forecasting, and obtains arrival intensity values; Service tiering module: When tracking the order history from supplier placement to delivery, determine the delivery date based on the unified service tiering; Capacity Calculation Module: Calculates the available capacity value of incoming materials based on the incoming material defect rate and defines the range of supplies that suppliers can commit to. Quantile commitment module: Constructs quantile curves and dynamically generates supplier supply ranges; Stability calculation module: Calculates the stability threshold and defines the lower bound of supply based on historical utilization rates; Allocation constraint module: Screens supplier allocation schemes based on the lower limit of supply and performs feasibility assessment; The currency consumption module defines the unit comprehensive currency consumption, which is used for cost calculation and cooperation gain determination, sorts the cumulative capacity sequence of suppliers, and determines the final allocation set. Risk assessment module: Identifies supplier risk status and generates early warnings based on the percentage of commitments made.

[0013] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the supply chain procurement data management method based on big data as described in the first aspect of the present invention.

[0014] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the supply chain procurement data management method based on big data as described in the first aspect of the present invention.

[0015] The beneficial effects of this invention are as follows: by periodically normalizing the predicted seasonal fluctuation values ​​to obtain the seasonal intensity index, and combining it with historical service quantile statistics, a dynamic mapping between demand forecasting and service level is formed. By calculating the capacity quantile by statistically analyzing the actual arrival volume within the seasonal phase, and superimposing the incoming material defect rate to form the available capacity value, the model not only reflects the supplier's delivery speed but also dynamically reflects its quality factors. By dynamically comparing the cumulative capacity sequence with the lower bound of supply, the minimum set of suppliers that meet the stable supply requirements can be automatically identified, avoiding marginal redundancy and resource waste caused by over-expansion. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the 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.

[0017] Figure 1 This is a flowchart illustrating the supply chain procurement data management method based on big data in Example 1.

[0018] Figure 2 This is a schematic diagram of the supply chain procurement data management system based on big data in Example 1. Detailed Implementation

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0020] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0022] Example 1, referring to Figures 1 to 2 This is the first embodiment of the present invention, which provides a supply chain procurement data management method based on big data, including the following steps: S1. Collect resource status data from the supply chain procurement process to form a node dataset, construct a SARIMA model, analyze the offset trend constant and calculate the ADF test statistic based on the unit root coefficient, use STL decomposition to perform seasonal decomposition on the difference series of the target variable, apply ADF to verify stationarity and determine the non-seasonal difference order, plot the ACF curve to determine the non-seasonal autoregression order, determine the difference order through the seasonal fluctuation value series, determine the influence of the white noise term on the cyclical trend and form the corresponding PACF curve, determine the non-seasonal and seasonal parameters, define the seasonal component values ​​to calculate the seasonal intensity index, and use the SARIMA model to predict the arrival intensity value. Preferably, resource status data from the supply chain procurement process is collected to form a node dataset, including: Collect the current resource status of nodes in the supply chain procurement process, including inventory status, logistics status, procurement demand data, timestamps, order volume, and transportation status, and form a node dataset. Define the normal and abnormal ranges of supply chain behavior, including abnormal order delays, abnormal inventory shortages, and logistics disruptions. The time delay threshold is the sum of the mean and twice the standard deviation of the difference between the actual delivery time and the expected delivery time of historical orders. If the current delivery difference is greater than or equal to the time delay threshold, it is judged as an abnormal order delay. The minimum safe inventory level is the sum of the mean and twice the standard deviation of the difference between the inventory quantity and the expected inventory quantity. If the current inventory difference is less than or equal to the minimum safe inventory level, it is judged as an abnormal inventory shortage.

[0023] By collecting multi-source status information such as inventory, logistics, orders, and transportation from nodes in the supply chain procurement process to form a node dataset, the model can comprehensively reflect the dynamic operating status of each link in the time dimension, providing a unified data foundation for anomaly identification. By defining order delays, inventory shortages, and transportation interruptions as unified anomaly domains, the model can achieve cross-dimensional correlation analysis at the node level. When a certain type of anomaly occurs, it can automatically trace its correlation with other nodes, forming a causal chain-like anomaly propagation identification.

[0024] Furthermore, the difference order is determined through the seasonal fluctuation value series; the influence of white noise on the cyclical trend is used to construct the corresponding PACF curve; non-seasonal and seasonal parameters are determined; seasonal component values ​​are defined to calculate the seasonal intensity index; and data prediction using the SARIMA model is used as the arrival intensity value, including... Based on the current resource status of nodes in the supply chain procurement process, a SARIMA model is constructed. Historical procurement demand data from the node dataset is used as the target variable, and the ADF algorithm is used to calculate the difference sequence of the target variable. The offset trend constant is analyzed, and the ADF test statistic is calculated based on the unit root coefficient, expressed as:

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] in, This represents the linear trend term, where T represents the total time. This represents the mean of the time index. and Let these represent the target variable data at times t and t-1, respectively. This represents the mean of the target variable data. Represents the unit root coefficient. and Let represent the difference between the target variables at the current time t and ti, respectively. The coefficients of higher-order lag terms are represented. denoted by the offset trend constant, p represents the lag step size when the ACF exceeds the significance range, and i is the lag step index. The noise term, determined based on historical experience, represents the random fluctuation term after model fitting. This represents the test statistic. express The standard error; The critical value is determined by using the ADF test table and the stationarity is verified by the test statistic. If the test statistic is less than the critical value, it indicates that the difference sequence of the target variable is stationary. Using STL decomposition to perform seasonal decomposition on the differenced series of the target variable, the series is broken down into long-term trend and seasonal fluctuations, represented as follows:

[0031]

[0032]

[0033] in, This represents the long-term trend value at time point t. This represents the seasonal fluctuation value at time point t. Represents the white noise term. Indicates the periodic window size. This indicates the total number of cycles that exist within the available cycles. 's' represents the period number, and 's' represents the period length. A time series was constructed based on long-term trend values, and ADF was applied to verify stationarity and determine the non-seasonal differencing order. An ACF curve was plotted based on the differrated series of long-term trend values ​​for significance verification, and the non-seasonal autoregressive order was determined. The lag order at which the first difference exceeds the 95% confidence interval was selected. Non-seasonal noise was detected using the partial autocorrelation function, and the lag order at which the PACF test curve exceeds the confidence interval was selected as the non-seasonal moving average order, expressed as:

[0034]

[0035]

[0036] in, This represents the sequence after differencing the long-term trend values, where d represents the number of non-seasonal differencing operations. The autocorrelation coefficient represents the lagged value, and h represents the lagged index. This represents the mean of the long-term trend series. represents the non-seasonal noise of the partial autocorrelation function, and pa represents the non-seasonal autoregression order; For seasonal fluctuation values, the significant long-period lag term is detected using the Autoregressive Factor (ACF) method to determine the seasonal autoregressive order. The difference order is then determined using the seasonal fluctuation value series. The influence of white noise on the cyclical fluctuations and the residuals on the cyclical trend are used to construct the corresponding PACF curve. If the significant segment in the PACF curve first exceeds the significance level, the significant lag step size is defined as the seasonal moving average order, expressed as:

[0037]

[0038]

[0039] in, The autocorrelation coefficient represents the lag step of seasonal fluctuations. This represents the mean of seasonal fluctuations. This represents the sequence after seasonal differencing, where D represents the degree of seasonal differencing. This represents the impact of periodic fluctuations and residuals on the periodic trend. This represents the seasonal fluctuation value at a time point lagging by s (period). The non-seasonal autoregression order pa, the non-seasonal difference number, and the non-seasonal moving average order q are used to form non-seasonal parameters. The seasonal autoregression order P, the seasonal difference number D, the seasonal moving average order, and the lag period s extracted from Q are used to form seasonal parameters. The SARIMA model is then used to predict the target variable. The total number of seasonal fluctuation values ​​in each cycle is taken as the seasonal component value, and the seasonal intensity index is calculated by mean normalization. The historical service quantiles are statistically analyzed. If there is "no shortage of goods in the current period and the goods are on time", the quantile is defined as 1, otherwise it is 0. The mean is calculated to form a uniform service quantile, and the SARIMA model prediction output is used as the arrival intensity value.

[0040] By combining the ADF test with SARIMA modeling, the difference sequence of historical procurement demand data of nodes can be stabilized in a statistical sense, ensuring that the time characteristics of the model input are predictable. The significance test of the trend term by the ADF statistic calculated by the unit root coefficient can eliminate spurious trend terms before fitting, thereby improving the stability of the model. By decomposing the difference sequence into long-term trends and seasonal fluctuations using STL, the system can simultaneously characterize both periodicity and aperiodity. This allows the trend term to reflect the true growth trend of procurement, while the seasonal term captures the rhythm of the supply chain cycle. The two are processed separately in the model, which can effectively avoid the misleading influence of seasonal interference on the judgment of long-term trends. By determining the autoregressive and moving average orders for the long-term trend and seasonal components using ACF and PACF respectively, and combining the difference order to precisely control the model response lag, the identification accuracy of high-frequency disturbances and low-frequency cycles is improved, thereby giving the SARIMA output a quantifiable "arrival intensity" attribute. By periodically normalizing the predicted seasonal fluctuation values ​​to obtain the seasonal intensity index, and combining it with historical service quantile statistics, a dynamic mapping between demand forecasts and service levels can be formed. This not only enables the quantification of service capacity based on historical behavior, but also allows for the early identification of potential supply and demand imbalances at the forecasting level.

[0041] S2: Statistical analysis is performed on the data from the supplier's order to the delivery. The delivery date is determined based on the unified service quantile. The available material capacity value is calculated. Based on the available material capacity value, the supplier's promised supply range in the seasonal phase is defined. The quantile curve is constructed. The historical utilization rate value is calculated and the stability threshold is defined. The supply lower bound is determined in combination with the arrival intensity value. Supplier allocation constraints are defined and allocation schemes are screened. Preferably, based on the available material supply capacity, the supplier's committable supply range for the seasonal phase is defined, a quantile curve is constructed, historical utilization rates are calculated and a stability threshold is defined, the supply lower bound is determined by combining the arrival intensity value, supplier allocation constraints are defined, and allocation schemes are screened, including... The system compiles statistics on the time from order placement to delivery from suppliers, determines the delivery date based on a unified service quantile, and calculates the supplier's actual delivery volume for each seasonal phase. Combining this with the incoming material defect rate, the system calculates the available incoming material capacity value, expressed as follows:

[0042]

[0043]

[0044]

[0045]

[0046] in, This indicates a unified service segmentation. Indicator functions representing Boolean events, The seasonal intensity index represents the seasonal phase m. The seasonal component level represents the position of the seasonal component within the m-seasonal phase (period). This indicates the supplier's capability quantile in period i during seasonal phase m. This represents the historical quantile threshold based on the unified service quantile. This represents the actual amount of goods received by supplier i during period t. This represents the set of all periods in history where the seasonal phase is m. This indicates the available material capacity. This indicates the defect rate of incoming materials. This indicates the supplier's delivery date in period i, at seasonal phase m. This represents the time from order placement to delivery for supplier i at index k. Based on the available material supply capacity, the supplier's commit supply range for each seasonal phase is defined, and a quantile curve is constructed, represented as follows:

[0047] in, The range of supply capacity that supplier i can commit to during seasonal phase m is represented by the quantile curve. Based on actual arrival data, historical utilization rates are statistically calculated and a stability threshold is defined. Combined with arrival intensity values, the lower bound of supply for the current period, derived from demand inflows and stability requirements, is calculated and expressed as follows:

[0048]

[0049] in, This represents the historical utilization rate value. Indicates the intensity value reached. Indicates the total number of suppliers. The stability threshold is determined by selecting the maximum value among historical utilization rates that satisfies both no stockouts and successful on-time delivery. Indicates the lower bound of supply; The total supply allocation of all suppliers in the current period must meet the constraints required for system stability, including that the total supply must not be less than the lower limit of supply, the allocation of each supplier must be within the range of commensurable supply, and supplier allocation schemes that meet the constraints must be selected.

[0050] By combining the order-to-delivery time of suppliers with a unified service quantile, delivery reliability can be quantified over time. Historical service performance is transformed into calculable quantile thresholds, thereby eliminating subjective judgments caused by individual fluctuations in the supply chain. Capacity quantiles are calculated by statistically analyzing actual arrival volumes within seasonal phases, and the incoming material defect rate is overlaid to form an available capacity value. This allows the model to reflect not only the supplier's delivery speed but also its quality factors dynamically, achieving a transformation from "delivery capacity" to "effective supply capacity." This capacity value defines the committable supply range and constructs quantile curves, allowing each supplier to... The supply range under seasonal phases has verifiable boundaries, thereby enabling predictable and constrained supply fulfillment. Furthermore, by combining historical utilization values ​​and stability thresholds with dynamic calculations, constraints are formed with actual supply and demand matching as the core. This allows allocation decisions to automatically exclude combinations of oversupply or excessive risk while ensuring system stability. Through the combined use of the above methods, multi-dimensional integrated quantification of delivery reliability, quality correction capability, and supply and demand balance control is achieved. This transforms the allocation scheme from static planning into a dynamic decision-making mechanism based on facts and service quantification, significantly improving the adaptability and risk defense capability of the supply chain response.

[0051] S3 defines the unit comprehensive monetary consumption value, refers to the supply allocation, calculates the cooperation gain, and determines the gain; Preferably, a unit comprehensive monetary consumption value is defined, and the cooperative gain is calculated and determined by referring to the supply allocation, including... The comprehensive unit monetary consumption value, defined by the combined purchase price, transaction cost per unit, defect rate, and scrap cost, is expressed as follows:

[0052] in, This represents the unit total monetary consumption of supplier i. This represents the contract unit price for supplier i. Indicates the unit cost of transaction. This represents the unit cost of rework or scrap, used to quantify the direct economic loss caused by defective products; Based on the unit comprehensive monetary consumption, combined with the total allocation of all suppliers in the current period, the amount of shortage penalties stipulated in the contract, and the supply allocation, the cooperation gain is calculated and expressed as:

[0053] in, Indicates the gain from cooperation. This indicates the amount of supply allocated by supplier i during period t. This indicates the amount of the stockout penalty stipulated in the contract, used to quantify the economic losses when demand cannot be met; The gain threshold is based on the sum of the mean and standard deviation of historical cooperative gains. If the cooperative gain is greater than or equal to the gain threshold, the supplier gain is marked as good and the next supplier is introduced. If the cooperative gain is less than the gain threshold, the supplier gain is marked as low (indicating that the marginal cost of introducing the supplier has exceeded the economic benefits it contributes), and the expansion of suppliers is stopped.

[0054] By defining the comprehensive monetary consumption value per unit based on the integrated purchase price, transaction cost per unit, defect rate, and scrap cost, the cost of each supplier reflects not only the contract price but also their actual economic burden in coordination, quality, and reprocessing. This transforms traditional static pricing into a dynamic and measurable economic efficiency indicator. Furthermore, by introducing the calculation of cooperation gains, the model performs a difference analysis between the revenue corresponding to the supply allocation and the stockout penalty. This allows the model to simultaneously assess the trade-off between increased revenue and rising costs resulting from increased supply, shifting from "cost minimization" to "gain maximization." It couples financial, quality, and operational data within the same logical layer, forming a dynamically updated economic constraint model that prevents cost spillovers caused by inefficient expansion and automatically captures high-return combinations, achieving an optimal economic closed loop for supply chain collaborative decision-making.

[0055] S4: Sort suppliers’ cumulative capacity sequence by unit comprehensive monetary consumption value, determine the allocation set, calculate the commitment ratio and make risk assessment. Preferably, the allocation set is determined by ranking the cumulative capacity sequence of suppliers based on their unit comprehensive monetary consumption value, including: All suppliers are sorted in ascending order based on the unit comprehensive monetary consumption to form an order of monetary value. Starting from the first supplier, the available capacity value of each supplier is accumulated in turn to calculate the total available capacity of different suppliers, forming a cumulative capacity sequence of different levels. Based on the comparison between the cumulative capacity sequence and the lower bound of supply, the minimum feasible supplier set is selected as the allocation boundary. The allocation rules are defined: if the supplier belongs to the supplier set before the allocation boundary, the allocation amount is the available material capacity value; if the supplier belongs to the allocation boundary, only the remaining amount that makes the total supply reach the lower bound of supply is allocated, and the rest is not allocated. The final allocation of each supplier for the current period.

[0056] By ranking suppliers based on their unit comprehensive monetary consumption and accumulating available capacity values ​​layer by layer, the allocation process is transformed from price-oriented to an optimal selection mechanism under the dual constraints of "cost-capacity." This ensures that the supplier combination with the lowest unit cost and highest supply efficiency is prioritized. Through dynamic comparison of the accumulated capacity sequence with the lower bound of supply, the minimum set of suppliers that meets the requirements for stable supply can be automatically identified, avoiding marginal redundancy and resource waste caused by overexpansion. This achieves a complementary closed loop where ranking logic ensures economic efficiency, capacity accumulation ensures supply stability, and boundary allocation rules ensure minimum feasibility. The supplier selection and allocation no longer rely on subjective judgment but are based on quantifiable thresholds within the system, forming a dynamic supply decision-making structure with self-constraint and self-termination characteristics.

[0057] Furthermore, the risk assessment is based on the percentage of commitments made, including: The ratio of the actual allocated quantity to the upper limit of available capacity in the final allocated quantity set is calculated as the commitment percentage. The analysis threshold is based on the sum of the mean and two standard deviations of the supplier's historical commitment percentage. If the actual commitment percentage is greater than or equal to the analysis threshold, it is judged that there is a risk and a risk alert is issued.

[0058] By introducing this dynamic early warning mechanism between supply chain allocation results and historical performance data, risk warnings can be triggered immediately after allocation decisions are made, providing early intervention signals for subsequent quality reviews, capacity coordination, and delivery date adjustments.

[0059] This embodiment also provides a supply chain procurement data management system based on big data, including, Data acquisition module: Collects resource status data from the supply chain procurement process and generates node datasets; Demand forecasting module: Constructs a SARIMA model, performs seasonal decomposition and sequence data forecasting, and obtains arrival intensity values; Service tiering module: When tracking the order history from supplier placement to delivery, determine the delivery date based on the unified service tiering; Capacity Calculation Module: Calculates the available capacity value of incoming materials based on the incoming material defect rate and defines the range of supplies that suppliers can commit to. Quantile commitment module: Constructs quantile curves and dynamically generates supplier supply ranges; Stability calculation module: Calculates the stability threshold and defines the lower bound of supply based on historical utilization rates; Allocation constraint module: Screens supplier allocation schemes based on the lower limit of supply and performs feasibility assessment; The currency consumption module defines the unit comprehensive currency consumption, which is used for cost calculation and cooperation gain determination, sorts the cumulative capacity sequence of suppliers, and determines the final allocation set. Risk assessment module: Identifies supplier risk status and generates early warnings based on the percentage of commitments made.

[0060] This embodiment also provides a computer device applicable to the supply chain procurement data management method based on big data, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the supply chain procurement data management method based on big data as proposed in the above embodiment.

[0061] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0062] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the supply chain procurement data management method based on big data as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0063] In summary, this invention obtains a seasonal intensity index by periodically normalizing the predicted seasonal fluctuation values ​​and combining it with historical service quantile statistics to create a dynamic mapping between demand forecasting and service levels. By calculating capacity quantiles based on actual arrival volume within the seasonal phase and superimposing the incoming material defect rate to form an available capacity value, the model not only reflects the supplier's delivery speed but also dynamically reflects its quality factors. Through dynamic comparison of the cumulative capacity sequence with the lower bound of supply, it can automatically identify the minimum set of suppliers that meet the stable supply requirements, avoiding marginal redundancy and resource waste caused by overexpansion.

[0064] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A big data based supply chain procurement data management method, characterized by, include: Resource status data from the supply chain procurement process is collected to form a node dataset. A SARIMA model is constructed, the offset trend constant is analyzed, and the ADF test statistic is calculated based on the unit root coefficient. The difference series of the target variable is seasonally decomposed. The stationarity is verified by applying ADF and the degree of non-seasonal differencing is determined. The ACF curve is plotted to determine the order of non-seasonal autoregression. The difference order is determined by the seasonal fluctuation value series. The influence of white noise on the cyclical trend is analyzed and the corresponding PACF curve is formed. Non-seasonal and seasonal parameters are determined. Seasonal component values ​​are defined to calculate the seasonal intensity index. The data is predicted using the SARIMA model as the arrival intensity value. Data statistics are collected from the time the supplier places an order to the time the goods arrive. Delivery dates are determined based on the unified service quantile. The available material capacity value is calculated. Based on the available material capacity value, the supplier's promised supply range in the seasonal phase is defined. A quantile curve is constructed. Historical utilization rate values ​​are calculated and stability thresholds are defined. The lower limit of supply is determined by combining the arrival intensity value. Supplier allocation constraints are defined and allocation schemes are screened. Define the unit comprehensive monetary consumption value, refer to the supply allocation, calculate the cooperation gain, and determine the gain; By ranking suppliers' cumulative capabilities using the unit comprehensive monetary consumption value, determining the allocation set, and calculating the commitment ratio, risk assessment is performed.

2. The supply chain procurement data management method based on big data as described in claim 1, characterized in that: The process involves determining the difference order through seasonal fluctuation value sequences, constructing corresponding PACF curves based on the influence of white noise terms on cyclical trends, identifying non-seasonal and seasonal parameters, defining seasonal component values ​​to calculate the seasonal intensity index, and using the SARIMA model to predict data as the arrival intensity value. include, Based on the current resource status of nodes in the supply chain procurement process, a SARIMA model is constructed. The historical procurement demand data of the node dataset is used as the target variable, and the ADF algorithm is used to analyze the offset trend constant and calculate the ADF test statistic based on the unit root coefficient. The ADF validation was used to determine the non-seasonal difference order and the seasonal autoregression order. The PACF test curve was selected, and the lag order exceeding the confidence interval was used as the non-seasonal moving average order. The significant lag step was defined as the seasonal moving average order. These were used to form non-seasonal and seasonal parameters, respectively, and the SARIMA model was used to predict the target variable. The historical service quantiles are statistically analyzed. If there is "no shortage of goods in the current period and the goods are on time", it is identified as quantile 1; otherwise, it is 0. The mean is calculated to form a uniform service quantile, and the SARIMA model prediction output is used as the arrival intensity value.

3. The supply chain procurement data management method based on big data as described in claim 2, characterized in that: The process involves defining the supplier's committable supply range based on incoming material availability, constructing quantile curves, calculating historical utilization rates and defining a stability threshold, combining this with arrival intensity values ​​to determine the lower bound of supply, defining supplier allocation constraints, and screening allocation schemes. Statistics are compiled based on the time from order placement to delivery by the supplier, and the delivery date is determined according to the unified service quantile. At the same time, the actual delivery volume of the supplier in the seasonal phase is calculated to determine the capacity quantile. Combined with the incoming material defect rate, the available capacity value of incoming materials is calculated. Based on the available material supply capacity, the supplier's commit supply range in the seasonal phase is defined, and a quantile curve is constructed. Based on the actual arrival volume data, the historical utilization rate value is statistically calculated and a stability threshold is defined. Combined with the arrival intensity value, the supply lower bound for the current period is calculated based on demand inflow and stability requirements. The allocation to each supplier must be within the committed supply range, and supplier allocation schemes that meet the constraints must be selected.

4. The supply chain procurement data management method based on big data as described in claim 3, characterized in that: The defined unit of comprehensive monetary consumption value, with reference to the supply allocation, is used to calculate the cooperative gain and determine the gain, including... The comprehensive monetary consumption value per unit is defined by the overall purchase price, transaction cost per unit, defect rate, and scrap cost. Based on the unit comprehensive monetary consumption, combined with the total allocation of all suppliers in the current period, the amount of shortage penalty stipulated in the contract, and the supply allocation, the cooperation gain is calculated; The gain threshold is based on the sum of the mean and standard deviation of historical cooperative gains. If the cooperative gain is greater than or equal to the gain threshold, the supplier gain is marked as good and the next supplier is introduced. If the cooperative gain is less than the gain threshold, the supplier gain is marked as low and the expansion of suppliers is stopped.

5. The supply chain procurement data management method based on big data as described in claim 4, characterized in that: The step of sorting suppliers' cumulative capabilities by unit comprehensive monetary consumption value to determine the allocation set includes, All suppliers are sorted in ascending order based on the unit comprehensive monetary consumption to form an order of monetary value. Starting from the first supplier, the available capacity value of each supplier is accumulated in turn to calculate the total available capacity of different suppliers, forming a cumulative capacity sequence of different levels. Based on the comparison between the cumulative capacity sequence and the lower bound of supply, the minimum feasible supplier set is selected as the allocation boundary. The allocation rules are defined: if the supplier belongs to the supplier set before the allocation boundary, the allocation amount is the available material capacity value; if the supplier belongs to the allocation boundary, only the remaining amount that makes the total supply reach the lower bound of supply is allocated, and the rest is not allocated. The final allocation of each supplier for the current period.

6. The supply chain procurement data management method based on big data as described in claim 5, characterized in that: The calculation of the commitment ratio is used for risk assessment, including, The ratio of the actual allocated quantity to the upper limit of available capacity in the final allocated quantity set is calculated as the commitment percentage. The analysis threshold is based on the sum of the mean and two standard deviations of the supplier's historical commitment percentage. If the actual commitment percentage is greater than or equal to the analysis threshold, it is judged that there is a risk and a risk alert is issued.

7. The supply chain procurement data management method based on big data as described in claim 1, characterized in that: The resource status data collected in the supply chain procurement process is used to form a node dataset, including: Collect the current resource status of nodes in the supply chain procurement process, including inventory status, logistics status, procurement demand data, timestamps, order volume, and transportation status, and form a node dataset. Define the normal and abnormal range of supply chain behavior, including abnormal order delays, abnormal inventory shortages, and logistics and transportation disruptions.

8. A supply chain procurement data management system based on big data, based on the supply chain procurement data management method based on big data as described in any one of claims 1 to 7, characterized in that: include, Data acquisition module: Collects resource status data from the supply chain procurement process and generates node datasets; Demand forecasting module: Constructs a SARIMA model, performs seasonal decomposition and sequence data forecasting, and obtains arrival intensity values; Service tiering module: When tracking the order history from supplier placement to delivery, determine the delivery date based on the unified service tiering; Capacity Calculation Module: Calculates the available capacity value of incoming materials based on the incoming material defect rate and defines the range of supplies that suppliers can commit to. Quantile commitment module: Constructs quantile curves and dynamically generates supplier supply ranges; Stability calculation module: Calculates the stability threshold and defines the lower bound of supply based on historical utilization rates; Allocation constraint module: Screens supplier allocation schemes based on the lower limit of supply and performs feasibility assessment; The currency consumption module defines the unit comprehensive currency consumption, which is used for cost calculation and cooperation gain determination, sorts the cumulative capacity sequence of suppliers, and determines the final allocation set. Risk assessment module: Identifies supplier risk status and generates early warnings based on the percentage of commitments made.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the supply chain procurement data management method based on big data as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the big data-based supply chain procurement data management method as described in any one of claims 1 to 7.

Images