Supply chain risk assessment system and method based on big data

Abstract

The invention belongs to the technical field of data analysis, and discloses a supply chain risk assessment system and method based on big data, and the system comprises a data collection and processing module which collects multi-dimensional supply chain data from nodes of a supply chain, carries out the format standardization and time alignment of the collected multi-dimensional supply chain data, generates a unified supply chain state data set, and carries out the data collection and processing; extracting multi-dimensional supply chain features of the nodes; the risk index construction module is used for fusing the multi-dimensional supply chain features of the nodes to form a unified risk representation vector, calculating node risk related indexes, and generating comprehensive risk features of the nodes through weighted integration and feature importance evaluation of various features; the risk prediction simulation module is used for constructing a supply chain network according to the supply dependency relationship among the nodes and dynamically simulating risk flow by combining the comprehensive risk characteristics of the nodes; the precision and efficiency of supply chain risk early warning are improved, and the potential impact of systematic risks on enterprise operation is reduced.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and more specifically, to a supply chain risk assessment system and method based on big data. Background Technology

[0002] Existing supply chain risk assessment systems and methods mainly suffer from the following problems:

[0003] With the continuous expansion of the globalized supply chain system, the supply chain structure exhibits characteristics of multi-level, highly coupled, and dynamically changing. Fluctuations or disruptions at any node in the supply chain network can trigger chain reactions, significantly impacting the production, logistics, and capital flow of upstream and downstream nodes, and even the overall stability of the supply chain. Therefore, the quantitative assessment and dynamic transmission modeling of supply chain risks have become important research directions in the field of supply chain management.

[0004] Existing supply chain risk assessment systems and methods typically analyze risk indicators for individual nodes or local supply chain conditions, neglecting the complex dependencies and risk coupling mechanisms between nodes. Traditional methods are mostly based on static risk indicators or single-node risk calculation models, which are insufficient to reflect the transmission and cumulative effects of risks in the supply chain, and cannot reveal potential chain reactions or multi-level risk diffusion processes.

[0005] Furthermore, existing technologies have shortcomings in risk propagation modeling. Traditional risk management methods often focus on single-point risk monitoring, lacking dynamic simulation mechanisms for the transmission, accumulation, and secondary disruptions of overall supply chain risk energy within the network. This makes it difficult to proactively identify and intervene in systemic risks. Especially in complex supply chain networks, where nodes exhibit nonlinear dependencies and heterogeneous response characteristics, existing models often fail to consider historical response characteristics, risk sensitivity, and nonlinear amplification effects, leading to significant biases in risk assessment results and hindering accurate decision-making and dynamic prevention and control.

[0006] In view of this, the present invention proposes a supply chain risk assessment system and method based on big data to solve the above problems. Summary of the Invention

[0007] To overcome the aforementioned shortcomings of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a supply chain risk assessment system based on big data, comprising:

[0008] The data acquisition and processing module collects multi-dimensional supply chain data from nodes in the supply chain, standardizes the format and aligns the time of the collected multi-dimensional supply chain data, generates a unified supply chain status dataset, and extracts multi-dimensional supply chain features from the nodes.

[0009] The risk indicator construction module integrates the multi-dimensional supply chain characteristics of nodes to form a unified risk representation vector, and calculates node risk-related indicators. Through weighted integration of various characteristics and assessment of feature importance, it generates the comprehensive risk characteristics of nodes.

[0010] The risk prediction and simulation module constructs a supply chain network based on the supply dependencies between nodes and dynamically simulates risk flow by combining the comprehensive risk characteristics of the nodes. When the accumulated risk energy of a node exceeds the preset risk energy accumulation threshold, a rupture warning is triggered, and the secondary risks caused by the transmission of rupture energy along the supply chain network are simulated.

[0011] The risk quantification and assessment module, based on the results of risk transmission simulation, quantifies and assesses the risk level of each node and the entire supply chain network, generates a ranked list of risk nodes, and uploads it to the risk assessment terminal.

[0012] Specifically, the multidimensional supply chain data includes supplier operation data, logistics and transportation data, inventory status data, market demand data, and financial credit data.

[0013] Specifically, the method for generating a unified supply chain status dataset and extracting multidimensional supply chain features of nodes includes:

[0014] ETL data processing technology is used to perform format conversion and field mapping of multidimensional supply chain data. Field standardization and unit conversion are completed through a preset database mapping table. Missing values, outliers and duplicate records are handled by group mean filling, box plot method and primary key deduplication technology.

[0015] For data sources with different sampling frequencies, a linear interpolation event resampling method is used for time synchronization to align each node to a unified time base; the standardized and time-synchronized multidimensional supply chain data is integrated with the nodes of the supply chain as the center to construct a unified supply chain status dataset.

[0016] Based on the supply chain status dataset, statistical analysis is performed on the operating status of nodes to extract multidimensional supply chain features of nodes. These features include inventory fluctuation characteristics, order fulfillment rate characteristics, cash flow characteristics, transportation timeliness characteristics, and market fluctuation characteristics. Principal component analysis is used to generate multidimensional supply chain features of supply chain nodes.

[0017] Specifically, the method for forming a unified risk representation vector includes:

[0018] Statistical analysis of multidimensional supply chain operation data and risk events at historical nodes; calculation of the correlation between various features in the multidimensional supply chain and node risk events using Pearson correlation coefficient, and using the absolute value of the correlation as the weight of the contribution of each feature to node risk.

[0019] The multidimensional supply chain characteristics of nodes are subjected to min-max normalization to obtain normalized multidimensional supply chain characteristics; the normalized characteristics are then weighted and combined according to their contribution weights to generate a unified risk representation vector for the nodes.

[0020] Specifically, the method for generating the comprehensive risk characteristics of nodes includes:

[0021] Based on the unified risk representation vector of nodes, a risk assessment model is constructed by combining historical node risk events. The importance of various features in the risk representation vector is analyzed by using the random forest feature importance assessment algorithm. The information gain value of each feature in node risk prediction is calculated, and the information gain value is normalized as the feature importance coefficient.

[0022] Based on the contribution weights of various features to node risk and the feature importance coefficients, a two-layer weighted integration method is adopted to weight and fuse the risk representation vectors to generate the comprehensive risk features of the nodes.

[0023] Specifically, the method for dynamically simulating risk flows includes:

[0024] The correlation analysis of different types of features in the multidimensional supply chain characteristics of nodes is carried out by mutual information method to calculate the supply dependence index between nodes; the supply dependence index between nodes is normalized to obtain the supply dependence matrix between nodes.

[0025] Using the supply dependency matrix as the edge weight matrix, with nodes representing enterprise entities and edge weights representing the supply dependency index between nodes, a weighted directed graph form of supply chain network is constructed; combining the comprehensive risk characteristics of nodes, the risk energy of nodes is defined to characterize the risk accumulation level of nodes at time t.

[0026] Based on the principle of risk energy conservation, a risk energy transmission equation is constructed to dynamically transmit risks in the supply chain network. By iteratively solving the risk energy transmission equation, the temporal transmission and accumulation of risks in the supply chain network are simulated, and the dynamic distribution and flow trajectory of risks in the supply chain network are obtained.

[0027] Specifically, the method for simulating the secondary risks caused by the transmission of fracture energy along the supply chain network includes:

[0028] When the accumulated risk energy of any node in the supply chain network exceeds the preset threshold for accumulated risk energy of a node, a breakup warning is triggered, and the breakup probability of the node is calculated. When the breakup probability of a node is greater than or equal to the preset threshold for the breakup probability of a node, the node is determined to have broken.

[0029] When a node breaks down, the risk energy of the node that exceeds the preset risk energy accumulation threshold is allocated to the neighboring nodes through an energy allocation function according to the supply dependence index between the node and its neighboring nodes, simulating the loss of the node's risk energy during the transmission process in the supply chain network.

[0030] After a neighboring node receives the risk energy from the assigned node, if the accumulated risk energy exceeds the preset risk energy accumulation threshold for the node, a breakage warning for the secondary node is triggered and the breakage probability is calculated. When the breakage probability of the secondary node is greater than or equal to the preset breakage probability threshold for the node, the secondary node is determined to have broken, thereby generating secondary risks. The process is iteratively applied to the neighboring nodes of all broken nodes, and by passing the risk energy layer by layer, the secondary risks caused by the transmission of breakage energy along the supply chain network are simulated.

[0031] Specifically, the method for quantitatively assessing the risk level of each node and the entire supply chain network includes:

[0032] Statistical analysis is performed on the accumulated risk energy of each node in the supply chain network during the risk transmission simulation process to obtain the risk energy index of each node. The risk energy index of each node includes the average accumulated risk energy, the peak accumulated risk energy, and the number of times the node's accumulated risk energy exceeds the preset risk energy accumulation threshold.

[0033] The risk energy index of each node is standardized by Z-score to obtain a quantitative index of the node's risk level; combined with the supply dependence index between nodes, the quantitative risk index of each node is weighted and integrated to obtain the risk energy index of the entire supply chain network.

[0034] Specifically, the method for generating the risk node sorting list and uploading it to the risk assessment terminal includes:

[0035] Based on the risk energy index of each node and the risk energy index of the entire supply chain network, the risk levels of each node in the supply chain network are ranked to generate a risk node ranking list; the risk node ranking list is then uploaded to the risk assessment terminal.

[0036] Big data-based supply chain risk assessment methods include:

[0037] S1. Collect multidimensional supply chain data from nodes in the supply chain, standardize the format and align the time of the collected multidimensional supply chain data to generate a unified supply chain status dataset, and extract multidimensional supply chain features from the nodes.

[0038] S2. Integrate the multi-dimensional supply chain characteristics of nodes to form a unified risk representation vector, and calculate node risk-related indicators. Through weighted integration of various characteristics and assessment of characteristic importance, generate the comprehensive risk characteristics of nodes.

[0039] S3. Construct a supply chain network based on the supply dependencies between nodes, and dynamically simulate risk flow by combining the comprehensive risk characteristics of the nodes; when the accumulated risk energy of a node exceeds the preset risk energy accumulation threshold, trigger a rupture warning and simulate the secondary risks caused by the transmission of rupture energy along the supply chain network.

[0040] S4. Based on the risk transmission simulation results, quantitatively assess the risk level of each node and the entire supply chain network, generate a risk node ranking list, and upload it to the risk assessment terminal.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] By representing the supply chain as a weighted directed graph, with nodes representing enterprise entities and edge weights representing supply dependence indicators, a structured and quantitative description of the entire supply chain network is achieved. This enables the analysis of risk propagation paths and vulnerable links from the perspective of the entire network. Through dynamic simulation of risk flow using risk energy transmission equations, the characteristics of risk evolution over time can be captured, supporting risk evolution prediction across multiple cycles and scenarios. The dynamic distribution and flow trajectory of risks are provided, allowing enterprises to identify potential risk accumulation points and key nodes in advance. The initial value of risk energy is determined by the comprehensive risk characteristics of nodes, and the dynamic changes of risk are calculated by combining the risk energy difference of neighboring nodes and the transmission coefficient, enabling precise quantification of each node's contribution to the overall network risk.

[0043] This system can dynamically simulate the accumulation and breakdown of risks at nodes within a supply chain network, and simulate multi-level secondary risks through iterative transmission, predicting potential cascading breakdown paths and systemic risk points. By using node breakdown probabilities and energy distribution functions, the system can obtain the risk status and risk energy distribution of each node at any given time, enabling quantitative and visualized risk management, facilitating decision-makers' control over key nodes and resource allocation. By identifying breakdown energy transmission paths and high-risk nodes, it can provide early warnings of potential risks in the supply chain system, offering supply chain managers scientific and actionable risk mitigation solutions and decision-making basis. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the supply chain risk assessment system based on big data according to the present invention;

[0045] Figure 2 This is a schematic diagram of the supply chain risk assessment method based on big data according to the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Example 1

[0048] Please see Figure 1 As shown, this embodiment provides a supply chain risk assessment system based on big data, specifically including the following steps:

[0049] The data acquisition and processing module collects multi-dimensional supply chain data from nodes in the supply chain, standardizes the format and aligns the time of the collected multi-dimensional supply chain data, generates a unified supply chain status dataset, and extracts multi-dimensional supply chain features from the nodes.

[0050] The risk indicator construction module integrates the multi-dimensional supply chain characteristics of nodes to form a unified risk representation vector, and calculates node risk-related indicators. Through weighted integration of various characteristics and assessment of feature importance, it generates the comprehensive risk characteristics of nodes.

[0051] The risk prediction and simulation module constructs a supply chain network based on the supply dependencies between nodes and dynamically simulates risk flow by combining the comprehensive risk characteristics of the nodes. When the accumulated risk energy of a node exceeds the preset risk energy accumulation threshold, a rupture warning is triggered, and the secondary risks caused by the transmission of rupture energy along the supply chain network are simulated.

[0052] The risk quantification and assessment module, based on the results of risk transmission simulation, quantifies and assesses the risk level of each node and the entire supply chain network, generates a ranked list of risk nodes, and uploads it to the risk assessment terminal.

[0053] Multidimensional supply chain data includes supplier operation data, logistics and transportation data, inventory status data, market demand data, and financial credit data.

[0054] Supplier operational data includes supplier capacity, production plans, delivery cycles, fulfillment rates, and supply concentration; logistics and transportation data includes transportation routes, transit time, transportation costs, warehousing status, and accessibility of logistics nodes; inventory status data includes raw material inventory, work-in-process inventory, finished goods inventory levels, inventory turnover rate, and records of stockouts and backlogs; market demand data includes order volume, customer demand volatility, sales forecasts, and market price trends; and financial credit data includes the cash flow status, credit rating, debt ratio, and payment cycle of node enterprises.

[0055] Methods for generating a unified supply chain status dataset and extracting multidimensional supply chain features of nodes include:

[0056] ETL data processing technology is used to perform format conversion and field mapping of multidimensional supply chain data. Field standardization and unit conversion are completed through a preset database mapping table. Missing values, outliers and duplicate records are handled by group mean filling, box plot method and primary key deduplication technology.

[0057] For data sources with different sampling frequencies, a linear interpolation event resampling method is used for time synchronization to align each node to a unified time base; the standardized and time-synchronized multidimensional supply chain data is integrated with the nodes of the supply chain as the center to construct a unified supply chain status dataset.

[0058] Based on the supply chain status dataset, statistical analysis is performed on the operating status of nodes to extract multidimensional supply chain features of nodes. These features include inventory fluctuation characteristics, order fulfillment rate characteristics, cash flow characteristics, transportation timeliness characteristics, and market fluctuation characteristics. Principal component analysis is used to generate multidimensional supply chain features of supply chain nodes.

[0059] It should be noted that inventory volatility characteristics are obtained by calculating the mean, standard deviation, and sliding window volatility of node inventory; order fulfillment rate characteristics are obtained by statistically analyzing the proportion of orders completed on time within a preset time window at each node; cash flow characteristics are obtained by calculating the mean cash inflows and outflows and turnover cycle of each node; transportation timeliness characteristics are obtained by analyzing the average duration, delay rate, and volatility of the transportation time series at each node; and market volatility characteristics are obtained by calculating the standard deviation and volatility of the market demand or price series at the node.

[0060] Methods for forming a unified risk representation vector include:

[0061] Statistical analysis of multidimensional supply chain operation data and risk events at historical nodes; calculation of the correlation between various features in the multidimensional supply chain and node risk events using Pearson correlation coefficient, and using the absolute value of the correlation as the weight of the contribution of each feature to node risk.

[0062] The multidimensional supply chain characteristics of nodes are subjected to min-max normalization to obtain normalized multidimensional supply chain characteristics; the normalized characteristics are then weighted and combined according to their contribution weights to generate a unified risk representation vector for the nodes.

[0063] Methods for generating comprehensive risk characteristics of nodes include:

[0064] Based on the unified risk representation vector of nodes, a risk assessment model is constructed by combining historical node risk events (such as stockouts, delays, defaults, financial anomalies, etc.). The importance of various features in the risk representation vector is analyzed by using the random forest feature importance assessment algorithm. The information gain value of each feature in node risk prediction is calculated and the information gain value is normalized as the feature importance coefficient.

[0065] Based on the contribution weights of various features to node risk and the feature importance coefficients, a two-layer weighted integration method is adopted to weight and fuse the risk representation vectors to generate the comprehensive risk features of the nodes.

[0066] Methods for dynamically simulating risk flows include:

[0067] The correlation analysis of different types of features in the multidimensional supply chain characteristics of nodes is carried out by mutual information method to calculate the supply dependence index between nodes; the supply dependence index between nodes is normalized to obtain the supply dependence matrix between nodes.

[0068] Using the supply dependency matrix as the edge weight matrix, with nodes representing enterprise entities and edge weights representing the supply dependency index between nodes, a weighted directed graph form of supply chain network is constructed; combining the comprehensive risk characteristics of nodes, the risk energy of nodes is defined to characterize the risk accumulation level of nodes at time t.

[0069] Based on the principle of risk energy conservation, a risk energy transmission equation is constructed to dynamically transmit risks in the supply chain network. By iteratively solving the risk energy transmission equation, the temporal transmission and accumulation of risks in the supply chain network are simulated, and the dynamic distribution and flow trajectory of risks in the supply chain network are obtained.

[0070] The risk energy transmission equation is: in, β represents the rate of change of risk energy accumulation at node i at time t; j represents the neighboring nodes of node i; N(i) represents the set of neighboring nodes of node i; ij E represents the risk transmission coefficient from node i to its neighboring node j. i (t represents the accumulated risk energy of node i at time t, which is the dynamic value of the node's comprehensive risk characteristics under network propagation. The initial value can be determined by the node's comprehensive risk characteristics; E) j (t) represents the accumulated risk energy of adjacent node j at time t; E j (t)-E i (t) represents the cumulative difference in risk energy between adjacent nodes; λ i This represents the node's own risk attenuation factor, indicating the enterprise's risk buffer or resilience, such as inventory redundancy and capital reserves; i and j represent the node's index; t represents the index at time t.

[0071] This addresses the following technical problems of existing technologies: Existing technologies typically focus only on risk indicators of a single node or the status of a local supply chain, making it difficult to quantify and model the complex dependencies between nodes. Traditional risk assessments often rely on static indicators or single-node risk calculations, making it difficult to consider the transmission and cumulative effects of risks along the supply chain; single-point or local risk assessment methods cannot reveal potential chain reactions or multi-level risk diffusion effects.

[0072] Compared to existing technologies, the advantages are as follows: Representing the supply chain as a weighted directed graph, with nodes representing enterprise entities and edge weights representing supply dependence indicators, enables a structured and quantitative description of the entire supply chain network, allowing analysis of risk propagation paths and vulnerable links from a holistic network perspective; dynamically simulating risk flow through risk energy transmission equations captures the characteristics of risk evolution over time, supporting multi-period and multi-scenario risk evolution prediction; providing dynamic risk distribution and flow trajectories, enabling enterprises to identify potential risk accumulation points and key nodes in advance; the initial value of risk energy is determined by the comprehensive risk characteristics of nodes, and the dynamic changes in risk are calculated by combining the risk energy difference between neighboring nodes and the transmission coefficient, accurately quantifying the contribution of each node to the overall network risk.

[0073] Methods for simulating the secondary risks resulting from the transmission of fracture energy along the supply chain network include:

[0074] When the accumulated risk energy of any node in the supply chain network exceeds the preset threshold for accumulated risk energy of a node, a breakup warning is triggered, and the breakup probability of the node is calculated. When the breakup probability of a node is greater than or equal to the preset threshold for the breakup probability of a node, the node is determined to have broken.

[0075] The probability of node failure is: Among them, P i (t) represents the probability of node i breaking at time t; C i (t) represents the failure probability threshold of the preset node at time t; k represents the proportional adjustment coefficient, used to control the overall linear proportional relationship between risk energy accumulation and failure probability; γ represents the nonlinear amplification coefficient, used to adjust the nonlinear effect of the ratio of risk energy accumulation to the failure probability threshold of the preset node; S i (t) represents the node risk sensitivity amplification term, which reflects the node's sensitivity to risk. It is set by expert experience based on the node's response records in past risk events.

[0076] When a node breaks down, the risk energy of the node that exceeds the preset risk energy accumulation threshold is allocated to the neighboring nodes through an energy allocation function according to the supply dependence index between the node and its neighboring nodes, simulating the loss of the node's risk energy during the transmission process in the supply chain network.

[0077] The energy distribution function is: Where, ΔE i→j ΔE represents the risk energy propagated from node i to adjacent node j. i ω represents the risk energy of node i exceeding the preset risk energy accumulation threshold for nodes; ij η represents the supply dependency index from node i to its neighboring node j; ij The attenuation coefficient represents the rate of risk transmission, indicating the degree to which a risk may be reduced during transmission due to control measures, buffering mechanisms, or material delays; ω ip This represents the supply dependency index from node i to any adjacent node p; p represents an index variable used to traverse all adjacent nodes of node i.

[0078] After a neighboring node receives the risk energy from the assigned node, if the accumulated risk energy exceeds the preset risk energy accumulation threshold for the node, a breakage warning for the secondary node is triggered and the breakage probability is calculated. When the breakage probability of the secondary node is greater than or equal to the preset breakage probability threshold for the node, the secondary node is determined to have broken, thereby generating secondary risks. The process is iteratively applied to the neighboring nodes of all broken nodes, and by passing the risk energy layer by layer, the secondary risks caused by the transmission of breakage energy along the supply chain network are simulated.

[0079] This solution addresses the following technical problems of existing technologies: Traditional supply chain risk management typically focuses only on the risk status of individual nodes, making it difficult to quantify risk accumulation or the probability of node failure, and also unable to accurately simulate the dynamic transmission of risk along the supply chain network. Existing technologies lack dynamic simulation methods for the transmission, accumulation, and triggering of secondary failures of risk energy from failed nodes along the supply chain network. Existing technologies typically do not consider the historical response, sensitivity, and nonlinear risk amplification effects of nodes in the risk transmission model, leading to prediction bias.

[0080] Compared to existing technologies, the advantages include: the ability to dynamically simulate the accumulation and breakdown of node risks in the supply chain network, and to achieve multi-level secondary risk simulation through iterative transmission, predicting potential cascading breakdown paths and systemic risk points. By using node breakdown probabilities and energy distribution functions, the risk status and risk energy distribution of each node at any given time can be obtained, enabling quantitative and visualized risk management, facilitating decision-makers in controlling key nodes and allocating resources. By identifying breakdown energy transmission paths and high-risk nodes, potential risks in the supply chain system can be predicted in advance, providing supply chain managers with scientific and actionable risk mitigation solutions and decision-making basis.

[0081] For example, consider a large electronics manufacturing company whose system includes multiple nodes such as upstream chip supplier A1, battery supplier B1, display module supplier C1, foundry D1, logistics center E1, and downstream distributor F1. Each node continuously uploads multi-dimensional supply chain big data, including production data, inventory data, transportation delays, and raw material price fluctuations.

[0082] Battery supplier B1 experienced a sharp decline in lithium raw material inventory due to international shipping disruptions, leading to a continuous accumulation of production delays. The system calculated, using time-series data, that the accumulated risk energy at node B1 was steadily increasing. When this accumulated risk energy exceeded a preset threshold for that node, a rupture warning was triggered. After node B1 ruptured, the system, according to an energy allocation function, distributed the excess risk energy to adjacent nodes—contractor D1 and display module supplier C1. Contractor D1, after receiving accumulated risk energy, again exceeded its risk energy accumulation threshold, triggering a secondary rupture warning. The system recalculated the rupture probability and found that the contractor's risk energy amplification effect was stronger; its rupture probability exceeded the threshold, and node D1 was also determined to have ruptured. At this point, the risk energy continued to propagate downstream to logistics center E1 and distributor F1, causing a sharp increase in the risk index of the entire supply chain system.

[0083] After completing the risk transmission simulation, the system quantitatively assesses the risk energy indicators of all nodes, calculates the network vulnerability index, and generates a ranking list of risky nodes. The results show that node B1 (battery supplier) is the initial risk source; node D1 (contract manufacturer) is a high-propagation node; and node F1 (distributor) is severely damaged but manageable. The ranking results are automatically uploaded to the risk assessment terminal, providing dynamic risk monitoring and emergency response suggestions for enterprise decision-makers.

[0084] Methods for quantifying the risk level of each node and the entire supply chain network include:

[0085] Statistical analysis is performed on the accumulated risk energy of each node in the supply chain network during the risk transmission simulation process to obtain the risk energy index of each node. The risk energy index of each node includes the average accumulated risk energy, the peak accumulated risk energy, and the number of times the node's accumulated risk energy exceeds the preset risk energy accumulation threshold.

[0086] The risk energy index of each node is standardized by Z-score to obtain a quantitative index of the node's risk level; combined with the supply dependence index between nodes, the quantitative risk index of each node is weighted and integrated to obtain the risk energy index of the entire supply chain network.

[0087] Methods for generating a sorted list of risk nodes and uploading it to the risk assessment terminal include:

[0088] Based on the risk energy index of each node and the risk energy index of the entire supply chain network, the risk levels of each node in the supply chain network are ranked to generate a risk node ranking list; the risk node ranking list is then uploaded to the risk assessment terminal.

[0089] It should be noted that the risk energy accumulation threshold of the preset node is set by the staff. By collecting the risk energy accumulation of different nodes, the average of the risk energy accumulation of multiple nodes is taken as the risk energy accumulation threshold of the preset node; similarly, the failure probability threshold of the preset node is set.

[0090] This embodiment represents the supply chain as a weighted directed graph, with nodes representing enterprise entities and edge weights representing supply dependence indicators, achieving a structured and quantitative description of the entire supply chain network. It can analyze risk propagation paths and vulnerable links from the perspective of the entire network; by dynamically simulating risk flow through risk energy transmission equations, it can capture the characteristics of risk evolution over time and support risk evolution prediction in multiple cycles and scenarios; it provides dynamic distribution and flow trajectory of risks, enabling enterprises to identify potential risk accumulation points and key nodes in advance; the initial value of risk energy is determined by the comprehensive risk characteristics of nodes, and the dynamic changes of risk are calculated by combining the risk energy difference of neighboring nodes and the transmission coefficient, which can accurately quantify the contribution of each node to the overall network risk.

[0091] This system can dynamically simulate the accumulation and breakdown of risks at nodes within a supply chain network, and simulate multi-level secondary risks through iterative transmission, predicting potential cascading breakdown paths and systemic risk points. By using node breakdown probabilities and energy distribution functions, the system can obtain the risk status and risk energy distribution of each node at any given time, enabling quantitative and visualized risk management, facilitating decision-makers' control over key nodes and resource allocation. By identifying breakdown energy transmission paths and high-risk nodes, it can provide early warnings of potential risks in the supply chain system, offering supply chain managers scientific and actionable risk mitigation solutions and decision-making basis.

[0092] Example 2

[0093] Please see Figure 2 As shown, parts not described in detail in this embodiment are described in Embodiment 1. A supply chain risk assessment method based on big data is provided, including:

[0094] S1. Collect multidimensional supply chain data from nodes in the supply chain, standardize the format and align the time of the collected multidimensional supply chain data to generate a unified supply chain status dataset, and extract multidimensional supply chain features from the nodes.

[0095] S2. Integrate the multi-dimensional supply chain characteristics of nodes to form a unified risk representation vector, and calculate node risk-related indicators. Through weighted integration of various characteristics and assessment of characteristic importance, generate the comprehensive risk characteristics of nodes.

[0096] S3. Construct a supply chain network based on the supply dependencies between nodes, and dynamically simulate risk flow by combining the comprehensive risk characteristics of the nodes; when the accumulated risk energy of a node exceeds the preset risk energy accumulation threshold, trigger a rupture warning and simulate the secondary risks caused by the transmission of rupture energy along the supply chain network.

[0097] S4. Based on the risk transmission simulation results, quantitatively assess the risk level of each node and the entire supply chain network, generate a risk node ranking list, and upload it to the risk assessment terminal.

[0098] Since the electronic device described in this embodiment is the electronic device used to implement the big data-based supply chain risk assessment system and method in the embodiments of this application, those skilled in the art can understand the specific implementation methods and various variations of the electronic device in this embodiment based on the big data-based supply chain risk assessment system and method described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. Any electronic device used by those skilled in the art to implement the big data-based supply chain risk assessment system and method in the embodiments of this application falls within the scope of protection of this application.

[0099] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0100] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for users of ordinary technical skills, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A big data based supply chain risk assessment system, characterized in that, The method comprises the following steps: The data acquisition and processing module collects multi-dimensional supply chain data from the nodes of the supply chain, standardizes the format and time alignment of the collected multi-dimensional supply chain data, generates a unified supply chain state data set, and extracts the multi-dimensional supply chain features of the nodes; The risk index construction module fuses the multi-dimensional supply chain features of the nodes to form a unified risk representation vector, calculates the risk-related indexes of the nodes, and generates the comprehensive risk features of the nodes by weighted integration and feature importance evaluation of various features; The risk prediction and simulation module constructs a supply chain network according to the supply dependence relationship between the nodes, and dynamically simulates the risk flow combined with the comprehensive risk features of the nodes; When the risk energy accumulation of the node is greater than the preset risk energy accumulation threshold of the node, a rupture warning is triggered, and the secondary risk caused by the conduction of the rupture energy along the supply chain network is simulated; The risk quantification and evaluation module quantifies and evaluates the risk level of each node and the entire supply chain network based on the simulation results of the risk transmission, and generates a risk node ranking list uploaded to the risk evaluation terminal.

2. The big data based supply chain risk assessment system of claim 1, wherein, The multi-dimensional supply chain data includes supplier operation data, logistics transportation data, inventory status data, market demand data, and financial credit data.

3. The big data based supply chain risk assessment system of claim 2, wherein, The method for generating a unified supply chain state data set and extracting multi-dimensional supply chain features of the nodes comprises the following steps: ETL data processing technology is used for format conversion and field mapping of multi-dimensional supply chain data, field standardization and unit conversion are completed through a preset database mapping table, and missing values, outliers and duplicate records are processed by using grouping mean filling, box plot method and primary key de-duplication technology; For data sources with different sampling frequencies, a linear interpolation event resampling method is used for time synchronization, and each node is aligned to a unified time reference; the multi-dimensional supply chain data after standardization and time synchronization is integrated with the nodes of the supply chain as the center to construct a unified supply chain state data set; Based on the supply chain state data set, the running state of the node is statistically analyzed, and the multi-dimensional supply chain features of the node are extracted, including inventory fluctuation features, order fulfillment rate features, fund flow features, transportation time efficiency features and market fluctuation features; the multi-dimensional supply chain features of the supply chain node are generated by principal component analysis.

4. The big data based supply chain risk assessment system of claim 3, wherein, The method for forming a unified risk representation vector comprises the following steps: Based on the statistical analysis of the multi-dimensional supply chain running data of the historical nodes and the historical node risk events; the correlation degree of each type of feature in the multi-dimensional supply chain features and the node risk event is calculated by Pearson correlation coefficient, and the absolute value of the correlation degree is taken as the contribution weight of each type of feature to the node risk; The multi-dimensional supply chain features of the node are subjected to minimum-maximum normalization processing to obtain normalized multi-dimensional supply chain features; the weighted combination of each type of normalized feature is carried out according to the contribution weight to generate a unified risk representation vector of the node.

5. The big data based supply chain risk assessment system of claim 4, wherein, The method for generating the comprehensive risk features of the node comprises the following steps: Based on the node-unified risk representation vector, a risk assessment model is constructed in combination with historical node risk events; through a random forest feature importance evaluation algorithm, the importance of each type of feature in the risk representation vector is analyzed, the information gain value of each type of feature in node risk prediction is calculated, and the information gain value is normalized as a feature importance coefficient; According to the contribution weight of each type of feature to node risk and the feature importance coefficient, a double-layer weighted integration method is used to weight and integrate the risk representation vector to generate the comprehensive risk feature of the node.

6. The big data based supply chain risk assessment system of claim 5, wherein, The method for simulating the dynamic flow of the risk comprises: Through mutual information method, the correlation of different types of features in the multi-dimensional supply chain features of the node is analyzed, and the supply dependence index between nodes is calculated; the supply dependence index between nodes is normalized to obtain the supply dependence relationship matrix between nodes; The supply dependence relationship matrix is used as an edge weight matrix, and the node represents an enterprise entity and the edge weight represents the supply dependence index between nodes to construct a supply chain network in the form of a weighted directed graph; in combination with the comprehensive risk feature of the node, the risk energy of the node is defined to represent the risk accumulation level of the node at time t; Based on the risk energy conservation principle, a risk energy conduction equation for the dynamic conduction of the risk in the supply chain network is constructed; through iterative solution of the risk energy conduction equation, the time-series conduction and accumulation simulation of the risk in the supply chain network is realized, and the dynamic distribution and flow trajectory of the risk in the supply chain network are obtained.

7. The big data based supply chain risk assessment system of claim 6, wherein, The method for simulating the secondary risk caused by the conduction of the breaking energy along the supply chain network comprises: When the risk energy accumulation of any node in the supply chain network is greater than the preset node risk energy accumulation threshold, a breaking early warning is triggered, and the breaking probability of the node is calculated; when the breaking probability of the node is greater than or equal to the preset node breaking probability threshold, it is determined that the node breaks; When the node breaks, the risk energy of the node that exceeds the preset node risk energy accumulation threshold is distributed to the adjacent nodes according to the proportion of the supply dependence index between the node and the adjacent nodes through an energy distribution function, simulating the loss of the risk energy of the node in the conduction process in the supply chain network; After the adjacent nodes receive the risk energy of the distribution node, if the accumulated risk energy exceeds the preset node risk energy accumulation threshold, a breaking early warning of the secondary node is triggered and the breaking probability is calculated; when the breaking probability of the secondary node is greater than or equal to the preset node breaking probability threshold, it is determined that the secondary node breaks, thereby generating a secondary risk; the adjacent nodes of all the broken nodes are iteratively applied, and the risk energy is transmitted layer by layer to simulate the secondary risk caused by the conduction of the breaking energy along the supply chain network.

8. The big data based supply chain risk assessment system of claim 7, wherein, The method for quantitatively evaluating the risk level of each node and the entire supply chain network comprises: Statistical analysis is performed on the accumulated risk energy of each node in the supply chain network during the risk transmission simulation process to obtain the risk energy index of each node; the risk energy index of each node includes the average risk energy accumulation, the peak risk energy accumulation, and the number of times that the risk energy accumulation of the node exceeds the preset node risk energy accumulation threshold; The risk energy indexes of the nodes are subjected to Z-score standardization processing to obtain quantified indexes of the risk levels of the nodes; the quantified risk indexes of the nodes are subjected to weighted fusion in combination with the supply dependency indexes between the nodes to obtain a risk energy index of the entire supply chain network.

9. The big data based supply chain risk assessment system of claim 8, wherein, The method for generating the risk node ranking list and uploading to the risk assessment terminal comprises: Based on the risk energy index of each node and the risk energy index of the entire supply chain network, the risk levels of the nodes in the supply chain network are ranked to generate a risk node ranking list; the risk node ranking list is uploaded to the risk assessment terminal.

10. The method of supply chain risk assessment based on big data, implemented by the system of supply chain risk assessment based on big data according to any one of claims 1 to 9, characterized in that, Comprise: S1, collecting multi-dimensional supply chain data from nodes of a supply chain, performing format standardization and time alignment on the collected multi-dimensional supply chain data, generating a unified supply chain state data set, and extracting multi-dimensional supply chain features of the nodes; S2, fusing the multi-dimensional supply chain features of the nodes to form a unified risk representation vector, and calculating node risk related indexes, generating comprehensive risk features of the nodes through weighted integration and feature importance evaluation of various features; S3, constructing a supply chain network according to the supply dependency relationship between the nodes, and dynamically simulating risk flow in combination with the comprehensive risk features of the nodes; When the accumulated risk energy of the node is greater than a preset accumulated risk energy threshold of the node, a rupture early warning is triggered, and secondary risks caused by the conduction of rupture energy along the supply chain network are simulated; S4, based on the risk transmission simulation results, quantitatively evaluating the risk levels of each node and the entire supply chain network, and generating a risk node ranking list, which is uploaded to the risk assessment terminal.

Images