A dynamic evaluation method and system for electronic information industry chain supply risk

Abstract

The application discloses an electronic information industry chain supply risk dynamic evaluation method and system, and the method comprises the following steps: constructing a dynamic evolution electronic information industry chain knowledge graph, and establishing an edge weight dynamic evolution model according to multi-source heterogeneous risk data, so that the system significantly enhances the capturing ability of sudden and implicit chain breakage risks; real-time parameters are extracted from the knowledge graph, three-dimensional orthogonal characteristics including node endogenous vulnerability, link topological dependency and external environmental impact are calculated and solved, a multi-dimensional supply risk dynamic evaluation index system is established, high-risk signals are output, the high-risk signals are taken as initial disturbances and injected into the knowledge graph, vertical shortage transmission caused by inventory depletion and horizontal overload crowding caused by order transfer are simulated, so that the system can accurately estimate the specific time delay of risk transmission to the terminal, and scientific decision-making basis in a golden window period is provided for enterprises and management departments.

Description

Technical Field

[0001] This invention relates to the field of supply chain risk management and assessment technology, specifically to a dynamic assessment method and system for the risk of supply disruption in the electronic information industry chain. Background Technology

[0002] The electronic information industry is a strategic, fundamental, and pioneering industry in the modern economy. It is characterized by a long industrial chain and rapid technological iteration, with close interdependence between upstream and downstream enterprises. Once the core nodes in the industrial chain, such as high-end chips, semiconductor equipment, or key materials, are cut off from supply, it is very easy to trigger a cascading effect, causing the downstream complete machine manufacturing and application links to come to a standstill, resulting in huge economic losses.

[0003] Existing methods for assessing supply chain risks have the following drawbacks:

[0004] Traditional assessment methods often evaluate the risk resistance of individual enterprises in isolation, lacking in-depth modeling of the complex topology of the electronic information industry chain and the transmission mechanism between nodes, making it difficult to quantify the dynamic impact of a supply disruption in a certain link on the entire industry chain.

[0005] Existing assessment methods suffer from data lag and lack the ability to dynamically capture and integrate real-time intelligence, such as information on logistical disruptions. This makes it impossible to track and provide dynamic early warnings of supply disruption risks in real time, and it is difficult to meet the needs for rapid decision-making and precise policy implementation in extreme environments. Summary of the Invention

[0006] This invention proposes a dynamic assessment method and system for the risk of supply disruption in the electronic information industry chain, in order to solve the technical problems mentioned in the background.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A dynamic assessment method for supply chain disruption risks in the electronic information industry chain, comprising the following steps:

[0008] S1. Collect multi-source heterogeneous risk data and perform structured preprocessing on the multi-source heterogeneous risk data to obtain processed multi-source heterogeneous risk data.

[0009] S2. Construct a dynamically evolving knowledge graph of the electronic information industry chain, and establish a dynamic evolution model of edge weights based on the multi-source heterogeneous risk data. The model is used to correct the supply topology strength between each level in the knowledge graph.

[0010] S3. Extract real-time parameters from the knowledge graph, calculate and solve three-dimensional orthogonal features including node intrinsic vulnerability, link topology dependency and external environmental impact, establish a multi-dimensional dynamic assessment index system for supply disruption risk, and output high-risk signals.

[0011] S4. The high-risk signal is injected into the knowledge graph as an initial disturbance to simulate vertical shortage transmission and horizontal overload squeeze, and the simulation is repeated iteratively to output the risk index curve.

[0012] S5. Trigger an early warning based on the peak value and growth rate of the risk index curve, use the knowledge graph path to traverse and search for alternative suppliers, and calculate the optimal safety stock.

[0013] Preferably, step S1 includes the following steps:

[0014] S11. Obtain the raw data stream through a distributed crawler cluster;

[0015] S12. Natural language processing technology is used to extract the measurement values ​​related to the delivery cycle of specific materials in the original data stream, and the observation values ​​generated by the fluctuation of attention are fused with the predicted values ​​based on historical trends through the Kalman filter algorithm, so as to output a real-time risk matrix that has been aligned in time and space.

[0016] Preferably, step S2 includes the following steps:

[0017] S21. Divide the electronic information industry chain from bottom to top into five vertical dimensions: basic materials layer, process equipment layer, semiconductor components layer, functional module layer and terminal assembly layer. Establish a four-element heterogeneous spatiotemporal graph architecture to represent the knowledge graph, which includes a set of nodes, a set of directed edges, attribute tensors and a time axis dimension.

[0018] The quaternary heterogeneous spatiotemporal graph architecture is as follows:

[0019]

[0020] in, It represents a set of nodes possessing multiple semantic features. For the set of directed edges describing interactions, For multidimensional attribute tensors mounted on nodes and edges, Endow the graph with dynamic evolutionary properties in the time dimension;

[0021] S22. Establish a dynamic evolution model of edge weights based on the real-time risk matrix. The model uses nonlinear coupling calculation to correct the supply topology strength between each level in the knowledge graph in real time.

[0022] The dynamic evolution model of the edge weights is as follows:

[0023]

[0024] in, Characterizing downstream nodes For upstream suppliers The percentage of order quotas This parameter indicates that the real-time risk matrix is ​​extracted in real time. This indicates the abnormal incremental components in the material delivery cycle. Scoring for material substitutability.

[0025] Preferably, step S3 includes the following steps:

[0026] S31. Extract real-time parameters from the knowledge graph using a graph database query language, and quantify and solve the three-dimensional orthogonal features, which include node intrinsic vulnerability, link topology dependency, and external environmental impact.

[0027] S32. A weighted regression formula based on the domestic substitution rate, real-time inventory level ratio and corporate cash flow health is introduced to calculate the intrinsic vulnerability of the node, and the result is normalized by the logistic function.

[0028] S33. Calculate the link topology dependency by characterizing the degree of deadlock between the downstream and the upstream through edge weights and a penalty coefficient for single-source supply relationships;

[0029] S34. Calculate the external environmental impact by accumulating and mapping the event severity score with the node's sensitivity matrix to the disaster;

[0030] S35. Generate and output high-risk signals based on three-dimensional orthogonal features.

[0031] Preferably, step S4 includes the following steps:

[0032] S41. The high-risk signal is injected into the knowledge graph as an initial disturbance to simulate the vertical shortage transmission caused by inventory depletion and the horizontal overload squeeze caused by order transfer.

[0033] S42. Set a fixed step size to simulate discrete events, and combine the Monte Carlo algorithm to inject Gaussian noise into the initial parameters, and repeat the iterative deduction.

[0034] S43. By calculating the output gap and market share weight at the terminal unit level, output a supply chain risk index curve that reflects the degree of damage to the entire supply chain.

[0035] Preferably, step S5 includes the following steps:

[0036] S51. Automatically trigger a four-level circuit breaker warning mechanism based on the peak value and growth rate of the risk index curve, and use the path traversal capability of the knowledge graph to search for alternative suppliers with functional equivalence, capacity margin and geopolitical security in the same level of nodes.

[0037] S52. Based on the predicted delivery date fluctuations, the dynamic safety stock level is calculated, and the final assessment conclusions, failure path diagrams, and action recommendations are packaged into a structured risk report and sent to the decision-making terminal.

[0038] Preferably, the formula for calculating the intrinsic vulnerability of the node is as follows:

[0039]

[0040] in, Indicates the domestic substitution rate, This indicates the safety stock limit (usually 30). Current inventory This indicates an assessment of the company's debts during the period of supply disruption and production stoppage. This indicates the assessment of a company's working capital during the period of supply disruption and production stoppage. , as well as These are the weighting coefficients;

[0041] The formula for calculating the link topology dependency is as follows:

[0042]

[0043] in, The edge weight represents the current time step. supplier to downstream manufacturers Dynamic supply pressure, For enterprises All upstream suppliers, It is a Boolean logic function. This is the weight amplification factor. The whole represents the time. upstream suppliers For downstream manufacturers Dynamic supply pressure;

[0044] The formula for calculating the impact of the external environment is:

[0045]

[0046] in, The score indicates the severity of the event. Represents the node sensitivity matrix, representing the node For event types Exposure level.

[0047] Preferably, the path traversal capability of the knowledge graph is used to search for alternative suppliers with functional equivalence, capacity margin, and geopolitical security among nodes at the same level through a sorting function.

[0048] The sorting function is as follows:

[0049]

[0050] in, The overall risk hedging score representing alternative supply nodes, This refers to the potential supplier entity nodes located in the graph. To map the domestic substitution rate attribute from potential supplier entity nodes, The hidden switching costs involved in switching from the original supplier to the alternative include technical certification time, mold development costs, etc. This represents the estimated delivery time in days under the new supply route. , and The weighting coefficients preset by the system;

[0051] The formula for calculating the dynamic safety stock level is as follows:

[0052]

[0053] in, This indicates the average daily consumption. For delivery cycle, For safety reasons, Used to quantify the uncertainty of downstream orders or production consumption.

[0054] A dynamic assessment system for supply chain disruption risks in the electronics and information industry chain, the system comprising:

[0055] The risk perception module is used to collect multi-source heterogeneous risk data and perform structured preprocessing on the multi-source heterogeneous risk data to obtain processed multi-source heterogeneous risk data.

[0056] The knowledge graph construction module is used to construct a dynamically evolving knowledge graph of the electronic information industry chain. Based on the multi-source heterogeneous risk data, a dynamic evolution model of edge weights is established. The model is used to correct the supply topology strength between different levels in the knowledge graph.

[0057] The quantitative indicator calculation module is used to extract real-time parameters from the knowledge graph, calculate and solve three-dimensional orthogonal features including node intrinsic vulnerability, link topology dependency and external environmental impact, establish a multi-dimensional dynamic assessment indicator system for supply disruption risk, and output high-risk signals.

[0058] The cascaded failure evolution module is used to inject the high-risk signal as an initial disturbance into the knowledge graph, simulate the vertical shortage transmission and the horizontal overload squeeze, repeatedly iterate and deduce, and output the risk index curve.

[0059] The strategy early warning output module is used to trigger an early warning based on the peak value and growth rate of the risk index curve, and to use the knowledge graph path traversal to search for alternative suppliers and calculate the optimal safety stock.

[0060] As can be seen from the above technical solution, this invention provides a dynamic assessment method for the risk of supply chain disruption in the electronic information industry chain. Compared with the prior art, this invention has the following advantages: By constructing a dynamically evolving knowledge graph of the electronic information industry chain and establishing a dynamic evolution model of edge weights based on multi-source heterogeneous risk data, the system significantly enhances its ability to capture sudden and hidden supply chain disruption risks; by injecting high-risk signals as initial perturbations into the knowledge graph, simulating the vertical shortage transmission caused by inventory depletion and the horizontal overload squeeze caused by order transfer, the system can accurately predict the specific time delay of risk transmission to the terminal, providing enterprises and management departments with a scientific decision-making basis within the golden window period; in addition, the automated logic of strategy generation greatly shortens the sourcing and verification cycle of alternative solutions, effectively improving the resilience of the electronic information industry chain under extreme pressure. Attached Figure Description

[0061] Figure 1 This is a flowchart illustrating a dynamic assessment method for supply chain disruption risks in the electronic information industry chain according to the present invention.

[0062] Figure 2 This is a structural block diagram of a dynamic assessment system for the risk of supply disruption in the electronic information industry chain, as described in this invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, but not all embodiments.

[0064] like Figure 1 As shown in this embodiment, a dynamic assessment method for the risk of supply disruption in the electronic information industry chain includes the following steps:

[0065] S1. Collect multi-source heterogeneous risk data and perform structured preprocessing on the multi-source heterogeneous risk data to obtain processed multi-source heterogeneous risk data.

[0066] S2. Construct a dynamically evolving knowledge graph of the electronic information industry chain, and establish a dynamic evolution model of edge weights based on the multi-source heterogeneous risk data. The model is used to correct the supply topology strength between each level in the knowledge graph.

[0067] S3. Extract real-time parameters from the knowledge graph, calculate and solve three-dimensional orthogonal features including node intrinsic vulnerability, link topology dependency and external environmental impact, establish a multi-dimensional dynamic assessment index system for supply disruption risk, and output high-risk signals.

[0068] S4. The high-risk signal is injected into the knowledge graph as an initial disturbance to simulate vertical shortage transmission and horizontal overload squeeze, and the simulation is repeated iteratively to output the risk index curve.

[0069] S5. Trigger an early warning based on the peak value and growth rate of the risk index curve, use the knowledge graph path to traverse and search for alternative suppliers, and calculate the optimal safety stock.

[0070] S1 includes the following steps:

[0071] S11. Obtain the raw data stream through a distributed crawler cluster;

[0072] S12. Utilize natural language processing technology to extract measurement values ​​related to specific material delivery cycles from the raw data stream, and use the Kalman filter algorithm to fuse the observation values ​​generated by fluctuations in attention with the predicted values ​​based on historical trends, thereby outputting a real-time risk matrix that has been spatiotemporally aligned.

[0073] The Kalman filter algorithm predicts the delivery date distribution at the current time using the covariance matrix of the previous time step, and calculates the Kalman gain and corrects the posterior estimate of the material delivery date based on the observation update term. The Kalman filter algorithm includes the state transition equation of the Kalman filter and the Kalman gain formula.

[0074] Due to the significant differences in the frequency of the collected data, direct fusion would lead to severe "time-domain holes." This embodiment employs a state prediction method based on Kalman filtering to estimate missing values. Addressing the mismatch between "fast-frequency news (minute-level)" and "slow-frequency statistical data (monthly customs data)" in the electronics and information industry, a state transition equation based on Kalman filtering is introduced. The state transition equation of the Kalman filter is as follows:

[0075]

[0076] in, This represents the current moment's supply chain risk characteristic vector. It includes multiple dimensions, such as core chip inventory levels and node capacity load. Φ represents the state transition matrix, which describes the natural evolution of supply chain risk over time without external interference, such as the rate at which inventory is naturally depleted over time. This represents process noise. It signifies unpredictable, minor fluctuations in the supply chain, such as small-scale delays in logistics. This equation is used to determine the noise level based on the previous time step (t). 1) Given the known risk state, predict the state of the industrial chain at the current time t;

[0077] The Kalman gain formula is as follows:

[0078]

[0079] in, This indicates the uncertainty of the forecast. This represents observed noise, or "impurities" reflecting the level of news attention. If the news source being collected is not authoritative enough, its noise will be significant. It will increase. This determines how the system finds a balance between "trusting previous predictive models" and "trusting the currently captured news data";

[0080] The formula for the optimal risk state value is as follows:

[0081]

[0082] in, The final determined optimal risk state value at the current moment will be used for the next step, S2, to construct the map. These are real-time observations. For example, the sentiment score for supply disruption calculated from news reports using an NLP module. "Surprise level" or "residual" represents the gap between reality (news) and the ideal (model prediction).

[0083] S2 includes the following steps:

[0084] S21. Divide the electronic information industry chain from bottom to top into five vertical dimensions: basic materials layer, process equipment layer, semiconductor components layer, functional module layer and terminal assembly layer. Establish a four-element heterogeneous spatiotemporal graph architecture to represent the knowledge graph, which includes a set of nodes, a set of directed edges, attribute tensors and a time axis dimension.

[0085] The quaternary heterogeneous spatiotemporal graph architecture is as follows:

[0086]

[0087] in, It represents a set of nodes possessing multiple semantic features. For the set of directed edges describing interactions, For multidimensional attribute tensors mounted on nodes and edges, Endow the graph with dynamic evolutionary properties in the time dimension;

[0088] S22. Establish a dynamic evolution model of edge weights based on the real-time risk matrix. The model uses nonlinear coupling calculation to correct the supply topology strength between each level in the knowledge graph in real time.

[0089] The dynamic evolution model of the edge weights is as follows:

[0090]

[0091] in, Characterizing downstream nodes For upstream suppliers The percentage of order quotas Indicates the first in the industrial chain Each node The dynamic comprehensive risk index at any given time is extracted from the real-time risk matrix. This indicates the abnormal incremental components in the material delivery cycle. Scoring for material substitutability.

[0092] The five vertical dimensions are as follows:

[0093] In five-level heterogeneous nodes and their multidimensional attributes ( In the definition of risk penetration, the L1 level, the key electronic-grade chemicals and basic materials layer, covers high-purity electronic specialty gases, semiconductor-grade photoresists, high-purity chemical reagents, and advanced packaging substrate materials. The core indicator slots for this level are reserved for domestic substitution rate and the geographical concentration index of supply sources.

[0094] The L2 level precision semiconductor manufacturing equipment and key subsystems extend upwards, encompassing lithography systems, plasma etching machines, thin film deposition devices, and testing and sorting systems. Its core indicator slots record the equipment's mean time between failures (MTBF), the overseas dependence on key components, and the average acquisition cycle for maintenance spare parts.

[0095] The L3 layer, which serves as the core transmission hub for risk fluctuations in the industrial chain, focuses on logic processors, memory chips, power management ICs, and analog signal link chips. Its indicator slots include advanced process node scores, yield volatility, and dynamic inventory safety turnover rate.

[0096] The L4-level highly integrated functional modules and precision circuit component layer, which involve smart camera modules, high-speed multilayer PCB boards and RF front-end integrated modules, include bill of materials complexity weight, self-sufficiency rate of key components and universality score in different complete product lines in its core indicator slots.

[0097] Located at the end of the industry chain, the L5 level intelligent complete system and terminal application scenario layer includes AI computing power cluster servers, 5G / 6G core network base station equipment and new energy vehicle central domain controllers. Its indicator slots are equipped with terminal market share, the upper limit of potential production capacity loss caused by specific supply disruptions, and the risk exposure of downstream contract default penalties.

[0098] S3 includes the following steps:

[0099] S31. Extract real-time parameters from the knowledge graph using a graph database query language, and quantify and solve the three-dimensional orthogonal features, which include node intrinsic vulnerability, link topology dependency, and external environmental impact.

[0100] S32. A weighted regression formula based on the domestic substitution rate, real-time inventory level ratio and corporate cash flow health is introduced to calculate the intrinsic vulnerability of the node, and the result is normalized by the logistic function.

[0101] S33. Calculate the link topology dependency by characterizing the degree of deadlock between the downstream and the upstream through edge weights and a penalty coefficient for single-source supply relationships;

[0102] S34. Calculate the external environmental impact by accumulating and mapping the event severity score with the node's sensitivity matrix to the disaster;

[0103] Specifically as follows:

[0104] Dimension 1: Intrinsic Vulnerability of Nodes This indicator focuses on the self-healing ability and resilience of a single enterprise entity or specific material link when facing supply disruptions. The system integrates multi-dimensional data extracted from the attributes of the graph nodes, including a technology resilience sub-indicator mapped from the domestic substitution rate attribute of L1 / L2 level nodes, used to quantify the ability of the local supply chain to provide functionally equivalent substitutes. It also includes an operational resilience sub-indicator mapped from the actual inventory days attribute synchronized from the enterprise's ERP interface to the graph, which identifies potential risks by comparing it with a preset safety stock threshold.

[0105] Dimension Two: Link Topology Dependency This indicator measures the tightness of connections between upstream and downstream suppliers, i.e., the difficulty and cost barriers to changing suppliers. Its data comes directly from the dynamic weights of relationship edges in the S2 graph, including supply share dependence reflecting the downstream's proportion of upstream procurement, and single-source risk automatically generated by calculating node in-degree. It also covers the implicit costs of switching during the recertification cycle.

[0106] Dimension 3: Impact of the External Environment This indicator measures the targeted impact intensity of natural disasters on nodes. Its data originates from the risk event stream output by S1, involving the constraint strength mapped from the risk matrix and the geopolitical sensitivity calculated based on the node's geographical location attributes. To mitigate cross-border link risks, the system also introduces a logistics disruption probability mapped from the shipping data congestion index. This data mapping implementation logic ensures that when the basic data in the graph is updated, the indicator system in S3 can acquire the latest parameters within milliseconds and achieve real-time synchronization between data and indicators.

[0107] S35. Generate and output high-risk signals based on three-dimensional orthogonal features.

[0108] S4 includes the following steps:

[0109] S41. The high-risk signal is injected into the knowledge graph as an initial disturbance to simulate the vertical shortage transmission caused by inventory depletion and the horizontal overload squeeze caused by order transfer.

[0110] S42. Set a fixed step size to simulate discrete events, and combine the Monte Carlo algorithm to inject Gaussian noise into the initial parameters, and repeat the iterative deduction.

[0111] S43. By calculating the output gap and market share weight at the terminal unit level, output a supply chain risk index curve that reflects the degree of damage to the entire supply chain.

[0112] Specifically, production load ( () refers to the current time node The total number of orders accepted is typically set during the initialization phase to the average capacity load of that node under normal production conditions. The corresponding maximum output capacity is... This represents the node's maximum production capacity without altering its physical infrastructure, and its calculation formula is: ,in, The redundancy coefficient reflects a company's flexibility in handling sudden surges in orders. In the precision electronics manufacturing industry, production lines are highly specialized and pursue Just-in-Time (JIT) production. The value is in the range of (0.05, 0.15).

[0113] In addition, critical material buffer inventory ( The setting is directly mapped to the real-time inventory days.

[0114] Based on this, the initial disturbance signal injection mechanism involves setting the nodes with extreme risk characteristics identified in step S3 as the initial failure sources for the simulation, i.e., high-risk signals. In this process, the system does not simply perform binary state switching, but injects negative pulse signals into the topology network to simulate the phenomenon of specific nodes encountering extreme external shocks, such as the instantaneous zeroing of production capacity when the core production line encounters a sudden disaster.

[0115] The injection of this disturbance signal also involves nonlinear adjustments to the node attributes within the potential disaster area, aiming to induce dynamic imbalances across the entire supply chain. For example, in a simulation of a disruption in the supply of electronic-grade neon, the disturbance signal manifests as a sharp decrease in the conduction weights of all relevant supply chain links at a specific geographical coordinate. The core significance of this step lies in testing the cascading failure threshold of the entire supply chain under multiple coupling pressures through artificially created disturbance points, thereby quantitatively assessing the depth of systemic risk collapse caused by the failure of a single critical link. In this implementation case, the L1-level neon supply node and its directly associated L3-level wafer fab are set as excitation points to trigger subsequent iterative calculations.

[0116] This invention also proposes a two-way coupling transmission mechanism tailored to the characteristics of the electronics and information industry chain. Through mathematical modeling, it accurately depicts the dual impacts of material shortages and market imbalances. This mechanism is primarily composed of two interwoven paths: vertical transmission of supply shortages and horizontal overload squeezes. Together, these constitute the core driving force behind the dynamic evolution of risk.

[0117] In a shortage-driven model based on inventory deficits, the system simulates the vertical penetration of supply disruption risk from upstream to downstream along the supply chain links. When upstream nodes... When it fails, the substantial interruption of its output will directly affect downstream nodes. A supply shortage has occurred. Regarding the construction of the inventory depletion kinetics equation, downstream nodes... At any moment Real-time buffer inventory follows The evolutionary logic, in which For a moment The actual supply from upstream nodes to downstream. Regarding the tiered failure threshold determination process, when inventory falls below the safety stock threshold, the node enters a Vulnerable State and triggers hedging procurement logic; once inventory reaches zero and supply is interrupted, the node enters a shortage failure state, and its production activities will be forcibly stopped.

[0118] The overload-driven mode based on elastic overflow in the horizontal transmission path simulates the secondary disasters caused when market demand is forced to shift to other peer nodes after the failure of a core supply node. In the order transfer and load redistribution algorithm, the load originally borne by the failed node will be redistributed among peer competing sets according to the principle of remaining capacity priority, and its increase is... This is affected by the link dependency coefficient. In this process, the overload failure judgment logic is used to identify whether the additional load received by the alternative node exceeds its production ceiling, i.e., when... At this point, the node will crash due to overload. This transmission mechanism analysis simulates the bank run effect commonly seen in the electronics industry, where the shutdown of upstream core plants can quickly overwhelm alternative plants that were originally operating at high capacity through demand surges.

[0119] Given the high degree of stochasticity in the supply chain environment, the system must perform large-scale stochastic simulations to obtain the probability distribution of risks. In the execution steps of the time-series iterative process, the system sets the simulation step size to 1 day and performs global state updates, failure scan determination, and topology collapse recording cyclically over a total period of 90 days.

[0120] To ensure the scientific rigor of the evaluation conclusions, the probabilistic convergence analysis method utilizes a graphics processing unit (GPU) to accelerate computation, repeatedly executing 10,000 Monte Carlo simulations. During this process, the system injects noisy random terms following a Gaussian distribution into the initial parameters, enabling it to simulate the worst-case scenario and average impact under different probability fluctuations, thus making the simulation results closer to the volatility characteristics of the real market.

[0121] S5 includes the following steps:

[0122] S51. Automatically trigger a four-level circuit breaker warning mechanism based on the peak value and growth rate of the risk index curve, and use the path traversal capability of the knowledge graph to search for alternative suppliers with functional equivalence, capacity margin and geopolitical security in the same level of nodes.

[0123] S52. Based on the predicted delivery date fluctuations, the dynamic safety stock level is calculated, and the final assessment conclusions, failure path diagrams, and action recommendations are packaged into a structured risk report and sent to the decision-making terminal.

[0124] In formulating the early warning classification standards, this embodiment constructs a response system with four levels. The system automatically triggers the corresponding early warning process based on the peak value of the risk index curve obtained from S4 simulation and the time when the peak value is reached.

[0125] Specifically, the triggering conditions for Level IV alerts are set as follows: This manifests as slight fluctuations in localized supply, but existing safety stock is sufficient to fully cover demand and there is no risk of end-user production stoppages. At this stage, the system's response is to send email alerts to purchasing managers and relevant business partners, suggesting continuous monitoring of dynamic changes at specific logistics nodes. Level III alert is triggered when the conditions are met. A level 10% warning indicates an anticipated capacity shortfall of less than 10%, necessitating the use of reserve inventory to maintain production. In this case, the system will automatically trigger an ERP system inventory lock command, halting spot shipments to non-core customers or non-high-value product lines. Level II is a severe warning, triggered when the risk index reaches [a certain threshold]. A Level I circuit breaker warning is triggered when a substantial supply disruption of critical materials is anticipated, potentially leading to delivery delays of over two weeks for mainstream models. At this level, the system initiates the Certificate of Replacement Material (PCN) process and automatically activates a pre-defined list of alternative suppliers. The trigger conditions for a Level I circuit breaker warning are as follows: This signifies that the industrial chain is facing systemic paralysis and key nodes are completely ineffective. At this time, the system will take emergency response actions, including recommending the establishment of a crisis management team and forcibly adjusting the priority of capacity allocation.

[0126] In addition to static threshold determination, the system enhances its sensitivity to emergencies by introducing dynamic circuit breaker logic. The system not only monitors the current risk index value but also tracks it in real time. The first derivative is the growth rate. If the system detects... Exceeding the preset surge threshold, even if currently Even if the value has not yet reached the red line for the classification, the system will still forcibly raise the warning level.

[0127] This step fully utilizes the knowledge graph constructed by S2. The provided topology depth allows for the use of specific graph algorithms to search for optimal solutions based on different risk types.

[0128] Preferably, the formula for calculating the intrinsic vulnerability of the node is as follows:

[0129]

[0130] in, Indicates the domestic substitution rate, This indicates the safety stock limit (usually 30). Current inventory This indicates an assessment of the company's debts during the period of supply disruption and production stoppage. This indicates the assessment of a company's working capital during the period of supply disruption and production stoppage;

[0131] The formula for calculating the link topology dependency is as follows:

[0132]

[0133] in, The edge weight represents the current time step. supplier to downstream manufacturers Dynamic supply pressure, For enterprises All upstream suppliers, It is a Boolean logic function. This is the weighting amplification factor;

[0134] The formula for calculating the impact of the external environment is:

[0135]

[0136] in, The score indicates the severity of the event. Represents the node sensitivity matrix, representing the node For event types Exposure level.

[0137] The path traversal capability of the knowledge graph is used to search for alternative suppliers with functional equivalence, capacity margin, and geopolitical security among nodes at the same level through a sorting function.

[0138] The sorting function is as follows:

[0139]

[0140] in, The overall risk hedging score representing alternative supply nodes, This refers to the potential supplier entity nodes located in the graph. To map the domestic substitution rate attribute from potential supplier entity nodes, The hidden switching costs involved in switching from the original supplier to the alternative include technical certification time, mold development costs, etc. This represents the estimated delivery time in days under the new supply route. , and The weighting coefficients preset by the system;

[0141] The formula for calculating the dynamic safety stock level is as follows:

[0142]

[0143] in, This indicates the average daily consumption. For delivery cycle, For safety reasons, Used to quantify the uncertainty of downstream orders or production consumption.

[0144] like Figure 2 As shown, a dynamic assessment system for the risk of supply disruption in the electronic information industry chain is provided, the system comprising:

[0145] The risk perception module is used to collect multi-source heterogeneous risk data and perform structured preprocessing on the multi-source heterogeneous risk data to obtain processed multi-source heterogeneous risk data.

[0146] The knowledge graph construction module is used to construct a dynamically evolving knowledge graph of the electronic information industry chain. Based on the multi-source heterogeneous risk data, a dynamic evolution model of edge weights is established. The model is used to correct the supply topology strength between different levels in the knowledge graph.

[0147] The quantitative indicator calculation module is used to extract real-time parameters from the knowledge graph, calculate and solve three-dimensional orthogonal features including node intrinsic vulnerability, link topology dependency and external environmental impact, establish a multi-dimensional dynamic assessment indicator system for supply disruption risk, and output high-risk signals.

[0148] The cascaded failure evolution module is used to inject the high-risk signal as an initial disturbance into the knowledge graph, simulate the vertical shortage transmission and the horizontal overload squeeze, repeatedly iterate and deduce, and output the risk index curve.

[0149] The strategy early warning output module is used to trigger an early warning based on the peak value and growth rate of the risk index curve, and to use the knowledge graph path traversal to search for alternative suppliers and calculate the optimal safety stock.

[0150] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., a solid-state disk (SSD)).

[0151] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0152] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0153] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic assessment method for supply chain disruption risks in the electronic information industry chain, characterized in that, Includes the following steps: S1. Collect multi-source heterogeneous risk data and perform structured preprocessing on the multi-source heterogeneous risk data to obtain processed multi-source heterogeneous risk data. S1 includes the following steps: S11. Obtain the raw data stream through a distributed crawler cluster; S12. Utilize natural language processing technology to extract measurement values ​​related to specific material delivery cycles from the raw data stream, and use the Kalman filter algorithm to fuse the observation values ​​generated by fluctuations in attention with the predicted values ​​based on historical trends, thereby outputting a real-time risk matrix that has been spatiotemporally aligned. S2. Construct a dynamically evolving knowledge graph of the electronic information industry chain, and establish a dynamic evolution model of edge weights based on the multi-source heterogeneous risk data. The model is used to correct the supply topology strength between each level in the knowledge graph. S2 includes the following steps: S21. Divide the electronic information industry chain from bottom to top into five vertical dimensions: basic materials layer, process equipment layer, semiconductor components layer, functional module layer and terminal assembly layer. Establish a four-element heterogeneous spatiotemporal graph architecture to represent the knowledge graph, which includes a set of nodes, a set of directed edges, attribute tensors and a time axis dimension. S22. Establish a dynamic evolution model of edge weights based on the real-time risk matrix. The model uses nonlinear coupling calculation to correct the supply topology strength between each level in the knowledge graph in real time. S3. Extract real-time parameters from the knowledge graph, calculate and solve three-dimensional orthogonal features including node intrinsic vulnerability, link topology dependency and external environmental impact, establish a multi-dimensional dynamic assessment index system for supply disruption risk, and output high-risk signals. S4. The high-risk signal is injected into the knowledge graph as an initial disturbance to simulate vertical shortage transmission and horizontal overload squeeze, and the simulation is repeated iteratively to output the risk index curve. S5. Trigger an early warning based on the peak value and growth rate of the risk index curve, use the knowledge graph path to traverse and search for alternative suppliers, and calculate the optimal safety stock. The quaternary heterogeneous spatiotemporal graph architecture is as follows: in, A set of nodes possessing multiple semantic features. For the set of directed edges describing interactions, For multidimensional attribute tensors mounted on nodes and edges, This represents the dynamic evolution attribute of the time axis dimension of the graph; The dynamic evolution model of the edge weights is as follows: in, For downstream nodes For upstream suppliers The percentage of order quotas As the first in the industrial chain Each node A dynamic comprehensive risk index at any given time. To record abnormal incremental components in the material delivery cycle, For material substitutability scoring, , , as well as These are the weighting coefficients; The formula for calculating the intrinsic vulnerability of a node is as follows: in, To indicate the rate of domestic substitution, As a safety stock limit, Current inventory To assess the company's debts during the period of supply disruption and production stoppage, To assess a company's working capital during a supply disruption or production stoppage, , as well as As a weighting factor; The formula for calculating the link topology dependency is as follows: in, The edge weight represents the current time step. supplier to downstream manufacturers Dynamic supply pressure, For enterprises All upstream suppliers, It is a Boolean logic function. This is the weight amplification factor. The whole represents the time. upstream suppliers For downstream manufacturers Dynamic supply pressure; The formula for calculating the impact of the external environment is: in, Score the severity of the incident. This is the node sensitivity matrix, representing the node... For event types Exposure level.

2. The dynamic assessment method for supply chain disruption risk in the electronic information industry chain according to claim 1, characterized in that: S3 includes the following steps: S31. Extract real-time parameters from the knowledge graph using a graph database query language, and quantify and solve the three-dimensional orthogonal features, which include node intrinsic vulnerability, link topology dependency, and external environmental impact. S32. A weighted regression formula based on the domestic substitution rate, real-time inventory level ratio and corporate cash flow health is introduced to calculate the intrinsic vulnerability of the node, and the result is normalized by the logistic function. S33. Calculate the link topology dependency by characterizing the degree of deadlock between the downstream and the upstream through edge weights and a penalty coefficient for single-source supply relationships; S34. Calculate the external environmental impact by accumulating and mapping the event severity score with the node pair or disaster sensitivity matrix; S35. Generate and output high-risk signals based on three-dimensional orthogonal features.

3. The dynamic assessment method for supply chain disruption risk in the electronic information industry chain according to claim 2, characterized in that: S4 includes the following steps: S41. The high-risk signal is injected into the knowledge graph as an initial disturbance to simulate the vertical shortage transmission caused by inventory depletion and the horizontal overload squeeze caused by order transfer. S42. Set a fixed step size to simulate discrete events, and combine the Monte Carlo algorithm to inject Gaussian noise into the initial parameters, and repeat the iterative deduction. S43. By calculating the output gap and market share weight at the terminal unit level, output a supply chain risk index curve that reflects the degree of damage to the entire supply chain.

4. The dynamic assessment method for supply chain disruption risk in the electronic information industry chain according to claim 3, characterized in that: S5 includes the following steps: S51. The circuit breaker warning mechanism is automatically triggered based on the peak value and growth rate of the risk index curve, and the path traversal capability of the knowledge graph is used to search for alternative suppliers with functional equivalence, capacity margin and geopolitical security in the same level of nodes. S52. Based on the predicted delivery date fluctuations, the dynamic safety stock level is calculated, and the final assessment conclusions, failure path diagrams, and action recommendations are packaged into a structured risk report and sent to the decision-making terminal.

5. The dynamic assessment method for supply chain disruption risk in the electronic information industry chain according to claim 4, characterized in that: The path traversal capability of the knowledge graph is used to search for alternative suppliers with functional equivalence, capacity margin, and geopolitical security among nodes at the same level through a sorting function. The sorting function is as follows: in, The overall risk hedging score for alternative supply nodes, These are potential supplier entity nodes located in the graph. To map the domestic substitution rate attribute from potential supplier entity nodes, The implicit switching costs involved in switching from the original supplier to the alternative supplier. For the estimated delivery time in days under the new supply route, , and The weighting coefficients preset by the system; The formula for calculating the dynamic safety stock level is as follows: in, This represents the average daily consumption. For delivery cycle, For safety reasons, This is the uncertainty coefficient, used to quantify the uncertainty of downstream orders or production consumption.

6. A dynamic assessment system for the risk of supply disruption in the electronic information industry chain, comprising executing the dynamic assessment method for the risk of supply disruption in the electronic information industry chain as described in any one of claims 1-5, characterized in that, The system includes: The risk perception module is used to collect multi-source heterogeneous risk data and perform structured preprocessing on the multi-source heterogeneous risk data to obtain processed multi-source heterogeneous risk data. The knowledge graph construction module is used to construct a dynamically evolving knowledge graph of the electronic information industry chain. Based on the multi-source heterogeneous risk data, a dynamic evolution model of edge weights is established. The model is used to correct the supply topology strength between different levels in the knowledge graph. The quantitative indicator calculation module is used to extract real-time parameters from the knowledge graph, calculate and solve three-dimensional orthogonal features including node intrinsic vulnerability, link topology dependency and external environmental impact, establish a multi-dimensional dynamic assessment indicator system for supply disruption risk, and output high-risk signals. The cascaded failure evolution module is used to inject the high-risk signal as an initial disturbance into the knowledge graph, simulate the vertical shortage transmission and the horizontal overload squeeze, repeatedly iterate and deduce, and output the risk index curve. The strategy early warning output module is used to trigger an early warning based on the peak value and growth rate of the risk index curve, and to use the knowledge graph path traversal to search for alternative suppliers and calculate the optimal safety stock.

Images