Resource intelligent decision-making method based on industrial chain collaborative optimization

Abstract

The invention provides an intelligent resource decision-making method based on industrial chain collaborative optimization, and relates to the technical field of industrial internet management, and the method comprises the steps: integrating resource data, business data, external environment and other multi-source heterogeneous data from all participants, combining a causal conduction rule and an operation constraint rule, and carrying out the collaborative optimization of an industrial chain. According to the method, a dynamic digital model capable of simulating a real industrial chain operation mechanism is generated, concurrent extreme conditions of multiple risks in reality are simulated through a composite interference scene for deduction, and a corresponding multi-dimensional toughness index and a visual toughness profile are generated based on a simulation result; precise positioning of key fragile nodes and fragile conduction paths is realized, deduction of different types of composite interference scenes can simulate influences of various complex factors appearing in an actual industrial chain, so that corresponding resource allocation strategies are generated, a coping basis and a precaution basis are provided for appearing of the actual interference scenes, and the resource allocation efficiency is improved. And the coping capability of the industrial chain network to the emergency scene interference is improved.

Description

Technical Field

[0001] This invention relates to the field of industrial internet management technology, specifically to a resource intelligent decision-making method based on supply chain collaborative optimization. Background Technology

[0002] In today's context of globalization and high specialization, the stable and efficient operation of the industrial chain is crucial to economic security and corporate survival. Because the industrial chain comprises multiple suppliers, manufacturers, logistics providers, and distributors, its complex and interconnected structure makes it extremely vulnerable to internal fluctuations and external shocks. Disruptive events such as raw material delays, equipment failures, logistical disruptions, and sudden changes in market demand can easily trigger a chain reaction through business connections, causing losses throughout the entire industrial chain. The inability to predict or provide adequate forecasts for disruptions is a major obstacle to the efficient operation of the industrial chain. Furthermore, the lack of corporate collaboration and failure to consider the needs of multiple participants in resource allocation within the industrial chain leads to low satisfaction with resource allocation, hindering the stable development of the industrial chain.

[0003] The existing technology has the following technical problems: Problem 1: Existing technical solutions typically simulate pre-defined, isolated single failure points (such as a single supplier's production stoppage), failing to depict the complex scenarios in reality where multiple interfering factors (such as geopolitics, natural disasters, and market fluctuations) intertwine, occur concurrently, and evolve in time and space. The simplistic simulation results in decision-making solutions that cannot cope with real and complex crises, insufficient simulation capabilities for interfering scenarios, and a lack of forward-looking diagnosis of the industrial chain, leading to the inability of the industrial chain network to respond in a timely manner to sudden impacts. The second problem is that most existing models for optimizing the industrial chain are centralized, with the system directly outputting directive solutions for execution. This completely ignores the private information, local interests, and autonomous decision-making rights of each participant. The resulting globally optimal solutions are often difficult to implement because they harm the interests of some participants or are not feasible in certain areas. This makes it impossible to achieve true resource synergy, resulting in low satisfaction among participants and failing to meet the goal of maximizing resource synergy and balance in the industrial chain. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: a resource intelligent decision-making method based on supply chain collaborative optimization, the method comprising: Acquire multi-source heterogeneous data from participants in the industrial chain network, construct an entity relationship network, and embed causal transmission rules and operational constraint rules into the entity relationship network to generate a dynamic digital model; A structured interference element library is constructed, and composite interference scenarios are synthesized and injected into a dynamic digital model. Simulation results are generated through deduction. Based on simulation results, multidimensional resilience indices are calculated to generate resilience profiles, and key vulnerable nodes and vulnerability transmission paths in the industrial chain network are located based on the resilience profiles. An initial resource allocation plan is generated based on key vulnerable nodes and vulnerability transmission paths and released to the participants. Feedback proposals from the participants are received, and the initial resource allocation plan is iteratively optimized based on the feedback proposals to generate a digital collaboration protocol. The digital collaboration protocol is executed and the execution process and results are monitored to generate an evaluation result that includes actual multi-dimensional resilience indicators and actual participant satisfaction. The method is then optimized based on the evaluation result.

[0005] Furthermore, the participants include suppliers, manufacturers, logistics providers, and distributors; the multi-source heterogeneous data includes resource data, business data, and external environment data; the entity relationship network uses participants as nodes and business connections formed between participants based on business data as edges; and the entity relationship network contains multi-source heterogeneous data of each participant.

[0006] Furthermore, the construction of a structured interference element library and the synthesis of composite interference scenarios are injected into a dynamic digital model, and simulation results are generated through deduction, including: The structured interference element library contains N interference elements, each of which contains four dimensions of attribute information: type information, intensity information, spatiotemporal attribute information, and occurrence probability information. Based on preset rules or random sampling, n interference elements are extracted from the structured interference element library. The n interference elements are dynamically combined based on spatiotemporal attribute information to generate a composite interference scenario containing the concurrent effects of the n interference elements. The complex interference scenario is injected into the dynamic digital model. Based on the simulation engine and using the causal transmission rule, the interference propagation path of the concurrent influence of n interference elements in the entity relationship network is deduced. The state data of each node at a specified time within the deduction period is recorded, and the simulation results of the state data of each node in the entity relationship network changing over time are generated.

[0007] Furthermore, the step of generating a resilience profile by calculating multidimensional resilience indices based on simulation results, and locating key vulnerable nodes and vulnerability transmission paths in the industrial chain network based on the resilience profile, includes: Based on the simulation results, time series data of each node in the simulation period are extracted, and multidimensional resilience index is calculated. The resilience dimensions of the multidimensional resilience index include shock resistance index, recovery agility index, adaptive reconfiguration index and impact isolation index. The calculated values ​​of shock resistance, recovery agility, adaptive reconfiguration, and impact isolation indices are normalized and plotted on the same radar chart to generate a toughness profile. By analyzing the profile of the resilience profile, resilience dimensions with index values ​​lower than preset health values ​​are identified. Combined with the interference propagation path, the nodes and edges in the interference propagation path whose state data fluctuation range exceeds the preset fluctuation threshold are traced back to locate key vulnerable nodes and vulnerable transmission paths.

[0008] Furthermore, the initial resource allocation plan is generated based on key vulnerable nodes and vulnerability transmission paths and released to the participants. Feedback proposals from the participants are received, and the initial resource allocation plan is iteratively optimized based on these proposals to generate a digital collaboration protocol, including: Based on key vulnerable nodes and vulnerability transmission paths, and combined with a preset strategy library, a set of candidate enhancement strategies are matched and generated and converted into specific resource scheduling instructions. The resource scheduling instructions are integrated to generate an initial resource allocation plan, which is a resource scheduling report generated for key vulnerable nodes and vulnerability transmission paths. The initial resource allocation plan is structurally encapsulated into a resource allocation strategy matrix, which contains a description of each resource scheduling task. The resource allocation strategy matrix is ​​published to the relevant participants, and feedback proposals from the participants are received. The feedback proposals include suggestions for modifying the resource scheduling report, execution preconditions, and compensation requirements. Based on the interest balancing model, a global feasibility check is performed on all feedback proposals. Feedback proposals that pass the feasibility check are marked as feasible feedback proposals. Feedback proposals with resource conflicts or logical contradictions are identified and marked.

[0009] Furthermore, the iterative optimization of the initial resource allocation plan based on feedback proposals to generate a digital collaboration protocol includes: For feedback proposals that pass the feasibility verification, a compensation plan is generated based on the benefit balancing model; Based on the feedback proposals regarding compensation schemes and feasibility, the initial resource allocation plan is revised, and the descriptions of resource scheduling tasks in the resource allocation strategy matrix are updated. The process iteratively executes the acceptance of M rounds of feedback proposals and the revision of the initial resource allocation plan until the updated resource allocation strategy matrix meets the requirements of the participants or reaches the preset maximum number of iterations. Based on the final consensus of the initial resource allocation plan and compensation scheme, a digital collaboration agreement is generated. Based on the benefit balance model, the predicted multidimensional resilience index and predicted participant satisfaction are obtained after the execution of the digital collaboration agreement.

[0010] Furthermore, the method optimization based on the evaluation results includes optimizing the dynamic digital model, the interest balancing model, and the preset strategy library. Optimizing the dynamic digital model based on the evaluation results includes: After the digital collaboration protocol is executed, collect the actual operation data of the industrial chain network and calculate the difference between the actual operation data and the simulation results at the same point in time. With the goal of minimizing the differences, the intensity coefficients of the causal transmission rules and the threshold parameters of the operational constraint rules in the dynamic digital model are automatically calibrated, and the calibrated intensity coefficients and threshold parameters are updated to the dynamic digital model.

[0011] Furthermore, the benefit balance model is optimized based on the evaluation results, including: Analyze the execution results of the digital collaboration protocol, calculate the actual multidimensional resilience index and the actual participant satisfaction, and calculate the deviation between the actual participant satisfaction and the predicted participant satisfaction. If the deviation is greater than the set satisfaction threshold, the internal parameters used to calculate the compensation scheme in the benefit balance model will be adjusted; otherwise, they will not be adjusted.

[0012] Furthermore, based on the evaluation results, the preset strategy library is optimized, including: Based on the execution results of the digital collaboration protocol, the adopted enhancement strategies are evaluated. The evaluation dimensions include the contribution of the enhancement strategies to the improvement of multi-dimensional resilience indicators, the execution cost, and the acceptance by the participants. Based on the results of the strategy evaluation, update the utility scores and applicable conditions of the candidate enhancement strategies in the preset strategy library.

[0013] Furthermore, a unified representation and measurement standard is established based on the multi-source heterogeneous data, and initial parameters are configured based on the simulation engine, including the simulation time span, simulation step size, and simulation convergence conditions.

[0014] This invention provides a resource-intelligent decision-making method based on supply chain collaborative optimization. It has the following beneficial effects: 1. This invention integrates multi-source heterogeneous data, including resource data, business data, and external environmental data from various participants, and combines causal transmission rules and operational constraint rules to generate a dynamic digital model that can simulate the operating mechanism of the real industrial chain. It simulates extreme situations of multiple concurrent risks in reality through composite interference scenarios, dynamically predicting the propagation path and impact of interference throughout the entire industrial chain network. Based on the simulation results, it generates corresponding multi-dimensional resilience indicators and visualized resilience profiles, achieving precise positioning of key vulnerable nodes and vulnerable transmission paths. This transforms traditional qualitative and empirical judgments into data-driven quantitative and visualized analysis, greatly improving the accuracy and depth of identifying potential risks in the industrial chain. The simulation of different types of composite interference scenarios can simulate the impact of various complex factors in the actual industrial chain, thereby generating corresponding resource allocation strategies. This provides a basis for responding to and preventing actual interference scenarios, increasing the industrial chain network's ability to cope with sudden interference.

[0015] 2. This invention, based on an iterative negotiation mechanism using a benefit-balancing model, generates and executes a digital collaboration protocol. It incorporates feedback proposals from all participants into the iterative negotiation process, designing a fair value distribution and compensation scheme. While protecting the individual interests of each participant, it incentivizes all parties to voluntarily and actively participate in collaboration. Furthermore, it optimizes each step based on the execution results of the digital collaboration protocol, allowing for self-evolution of the method and improving the accuracy of simulation simulations. This ensures that simulation results for complex interference scenarios more closely reflect actual impacts, enabling proactive resource scheduling and collaboration. This forms a sustainable industrial chain resource collaboration ecosystem that can address common risks while respecting and protecting individual interests, achieving collaborative optimization and closed-loop improvement of the overall resilience of the industrial chain network. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of a resource intelligent decision-making method based on supply chain collaborative optimization according to the present invention. Figure 2 This is a data transmission relationship diagram for a resource intelligent decision-making method based on supply chain collaborative optimization according to the present invention. Detailed Implementation

[0017] 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. Specific Implementation Example 1: like Figures 1 to 2As shown, a resource-intelligent decision-making method based on supply chain collaborative optimization is proposed, the method comprising: Step S100: Obtain multi-source heterogeneous data from participants in the industrial chain network, construct an entity relationship network, and embed causal transmission rules and operational constraint rules into the entity relationship network to generate a dynamic digital model; The industry chain network is a structure comprised of multiple participants, including suppliers, manufacturers, logistics providers, and distributors. Multi-source heterogeneous data refers to data sets from different sources, formats, and structures. This data includes resource data, business data, and external environmental data. Resource data includes raw material inventory, semi-finished product inventory, finished product inventory, production line capacity, equipment status, and human resources. Business data includes purchase orders, production work orders, sales orders, and logistics waybills. External environmental data includes raw material futures prices, market demand forecasts, macroeconomic indicators, and traffic information. The acquired multi-source heterogeneous data consists of publicly available data from each participant that does not involve core secrets. Each participant retains at least a portion of its private information, ensuring that its corporate secrets are not leaked while maximizing the acquisition of multi-source heterogeneous data from participants, and simultaneously guaranteeing the security and privacy of data transactions. First, multi-source heterogeneous data is collected in real time or periodically from the participants' internal systems (such as ERP and SCM systems) and external databases (such as publicly available data from government websites and market data platforms) through API interfaces, IoT sensors, or manual input. The collected multi-source heterogeneous data is then cleaned, normalized, and fused to eliminate format and unit differences and form a unified data representation standard. For example, all resource data is converted into standard units, such as tons or pieces. Then, using the participants as nodes and the business connections formed between them through business data as edges (such as order flow and logistics routes), an entity relationship network is constructed. This network contains multi-source heterogeneous data from each participant. Based on this network, a visualized supply chain network topology is created, facilitating the analysis of dependencies and transmission effects among participants. Next, causal transmission rules and operational constraint rules are embedded into the entity relationship network using a rule engine. Specifically, historical data is generated from the entire process of the impact on each participant and their response methods during each actual disruption scenario in the supply chain network. The causal transmission rules and operational constraint rules are then defined based on historical data analysis and domain expert knowledge. The causal transmission rules are logical rules that describe the causal relationships of events in the supply chain network. For example, supplier delays can lead to production interruptions, raw material delays can lead to production start-up delays, and traffic congestion and severe weather can lead to delivery delays. The causal transmission rules are embedded in the dynamic digital model through the rule engine to simulate the dynamics of interference propagation and predict chain reactions. The operational constraint rules are used to define the operational restrictions of the supply chain network, such as the maximum capacity limit and the minimum safety stock standard. The operational constraint rules include constraint conditions (such as inequalities) and threshold parameters for each item, which are derived from the operational policies and physical constraints of the participants. The operational constraint rules are used to ensure that the simulation results conform to the actual operating conditions and improve the realism of the simulation. Finally, the initial parameters of the simulation engine are configured, including the simulation time span, simulation step size, and simulation convergence conditions, to generate a dynamic digital model. This dynamic digital model stores the entity relationship network in the form of a graph database and embeds it into the rule engine and simulation engine to simulate the operation of the industrial chain network. The simulation time span defines the overall time range of the simulation, such as the next 30 days, and is used in step S203 to determine the length of the simulation cycle, ensuring that the simulation time range covers the entire impact period and recovery process of the interfering factors, thereby generating complete time series data. The simulation step size sets the time interval during the simulation process, such as the simulation interval per hour, and is used in step S203 to control the iteration frequency of the simulation engine. The simulation engine applies causal propagation rules and operational constraint rules at each step (e.g., per hour) to update the state data of each node in the entity relationship network, achieving fine-grained dynamic simulation. The simulation convergence condition specifies a threshold for early termination of the simulation, such as a state change rate below 1%. The node state change rate is calculated by comparing the difference in state data corresponding to adjacent step points. In step S203, it is used to monitor the stability of the nodes. When the simulation engine detects that the node state change rate is below the set threshold, it indicates that the node is approaching a stable state, and the simulation is automatically stopped to avoid unnecessary consumption of computational resources and improve simulation efficiency. The parameters of the simulation engine are derived from user settings or default configurations and are used to drive the simulation process and generate time series data. By integrating multi-source heterogeneous data and modeling entity relationship networks, the physical form of the industrial chain network is mapped into a digital twin, which solves the problems of heterogeneity of multi-source data and isolation of participants, realizes unified visualization of the industrial chain status, and provides a data foundation for simulation.

[0019] Step S200: Construct a structured interference element library, synthesize composite interference scenarios and inject them into a dynamic digital model, and generate simulation results through deduction; Step S201: Obtain the data of interference elements in the structured interference element library, as well as the entity relationship network, causal transmission rules, and operational constraint rules in the dynamic digital model; wherein, the structured interference element library contains N interference elements, each of which contains four dimensions of attribute information, namely type information, intensity information, spatiotemporal attribute information, and occurrence probability information; wherein: Type information describes the category of interference factors, including supply-side, production-side, logistics-side, demand-side, policy-side, and environmental-side. For example, supply-side interference involves raw material shortages, while production-side interference involves equipment failures. Type information is used to classify interference sources, facilitating the rapid location of affected nodes in the entity relationship network and ensuring that the simulation can address interference from different sources in a targeted manner. Intensity information quantifies the severity or impact of interference factors, including specific values ​​such as delay days, percentage of capacity loss, and price fluctuation amplitude. For example, the intensity of logistics disruption is expressed as a delivery delay of 24 hours. Intensity information provides measurable input parameters for the simulation, enabling the simulation to accurately reflect the impact of interference and adjust the changes in node states in the dynamic digital model. Node states refer to the changes in the dynamic attributes of each node, represented by state data. Spatiotemporal attribute information is used for... Define the temporal and spatial characteristics of the interference element, including its physical location (such as specific geographical coordinates or region), start time, duration, and spread range (such as radius of influence). For example, an interference element might be specified as occurring in a warehouse in East China, lasting for 48 hours from a specific time. Spatiotemporal attribute information is used to ensure that the interference scenario is aligned with the actual situation in time and space, guiding the simulation engine to apply interference at the correct time and location, and simulating the spread of interference in the industrial chain network. Occurrence probability information is used to represent the likelihood of the interference element occurring, usually a probability value (such as 0.1 representing a 10% probability), such as the probability of natural disasters based on historical data statistics. Occurrence probability information is used to support random sampling in risk assessment and scenario generation, prioritizing high-probability interference, improving the practicality of complex interference scenarios and the real-world reference value of simulation results. Step S202: Extract n interference elements from the structured interference element library according to preset rules or random sampling methods, and dynamically combine the n interference elements based on spatiotemporal attribute information to generate a composite interference scene containing the concurrent effects of n interference elements. Based on spatiotemporal attribute information, n interference elements are dynamically combined, including: First, n interference elements are extracted from the structured interference element library according to preset rules or random sampling methods. The preset rules are based on the occurrence probability information for weighted sampling to give priority to high probability interference, or based on type information to ensure diversity. Then, for the extracted n interference elements, their spatiotemporal attribute information is analyzed, including the start time, duration, physical location, and diffusion range, and time alignment and spatial integration are performed. Among them, time alignment is achieved by adjusting the start time and duration of the interference elements to ensure that the n interference elements occur concurrently or sequentially within the simulation time span. For example, if the start time or duration of two interference elements overlap, they are combined into a simultaneous interference. Spatial integration is achieved by checking the overlap or proximity of physical location and diffusion range, and interference elements with overlapping physical locations and overlapping diffusion ranges are merged. For example, if the supplier delay occurs in location A and the logistics interruption occurs in the adjacent location B, they are combined into a regional composite interference scenario. Finally, a composite interference scenario containing the concurrent effects of n interference elements is generated. The composite interference scenario encapsulates the attributes of all interference elements and defines their interaction in time and space, such as the simultaneous occurrence of supplier delays and logistics interruptions. In step S203, this scenario is injected into the dynamic digital model as an external stimulus to simulate the chain reaction under multiple concurrent interferences. Step S203: Inject the complex interference scenario as an external stimulus into the dynamic digital model. Based on the simulation engine and using the causal transmission rule, deduce the interference propagation path of the concurrent impact of n interference elements in the entity relationship network, and record the state data of each node at a specified time within the deduction period. Generate simulation results of the state data of each node in the entity relationship network changing over time. By simulating the concurrent impact of multiple interference elements, the problem that traditional methods cannot handle complex interference scenarios is solved. It provides behavioral data of the industrial chain network under stress testing, which can be used to assess resilience and identify potential risks. Based on the simulation engine and utilizing causal propagation rules, the interference propagation path of the concurrent influence of n interference elements in the entity relationship network is deduced, including: First, the simulation engine loads the entity relationship network, causal transmission rules, and operational constraint rules from the dynamic digital model, and injects the complex interference scenario as an external stimulus to initialize the state of each node; Then, the simulation engine advances time step by step (e.g., per hour). Within each step, it applies causal propagation rules for deduction. Based on these rules, it checks whether the current node state meets the rule's preconditions. If so, it triggers the rule's conclusion and updates the downstream node state. For example, when the supplier's node state changes to "delay," the manufacturer's node state is automatically updated to "production interruption" according to the causal propagation rules. During the deduction process, the simulation engine records the propagation order of each node's state change, forming an interference propagation path. This path represents a chain that starts from the node initially affected by the interference element and sequentially affects subsequent nodes through business connection edges. Simultaneously, operational constraint rules ensure that the node state conforms to actual limitations (e.g., maximum capacity limit) to prevent the deduction results from deviating from reality. The deduction continues until the simulation time span is reached or the convergence condition is met (e.g., the state change rate is less than 1%). Finally, the interference propagation path is output, which includes the time series data of the affected nodes, the impact time points, and intensity information for subsequent identification of vulnerable links. The status data of each node refers to the dynamic attribute values ​​of each node in the entity relationship network during the simulation process. Specifically, it includes resource data, business data, and derived performance indicators (such as order delivery delay time and inventory turnover rate). It is used to reflect the operating status of the node under disturbance in real time, serve as the basic data for evaluating the node, and reveal the dynamic process of the disturbance impact through the changes in time series data, providing input for calculating multidimensional resilience indicators.

[0020] The simulation results are the output of the state data of each node changing over time during the simulation period. Specifically, they include the time series of the state values ​​of each node at each simulation step, detailed records of the interference propagation path (such as the list of affected nodes and the order of impact), and overall industry chain network performance indicators (such as total capacity loss and average recovery time). The simulation results are generally in the form of structured data tables (such as CSV or database tables, where rows represent time points and columns represent node state variables), visualization charts (such as line graphs showing changes in inventory levels over time), or network topology diagrams (highlighting interference propagation paths and key nodes). This allows users to intuitively analyze the behavior of the industry chain network under interference and is used to generate the resilience profile in step S300 and locate key vulnerable nodes and vulnerable propagation paths. Step S300: Calculate multidimensional resilience indices based on simulation results to generate a resilience profile, and locate key vulnerable nodes and vulnerability transmission paths in the industrial chain network based on the resilience profile; Step S301: Based on the simulation results, extract the time series data of each node during the simulation period. Calculate the multidimensional resilience index using statistical methods or differential equations. The multidimensional resilience index is used to quantitatively and comprehensively evaluate the anti-interference capability of each node in the industrial chain network. It decomposes the complex behavior of nodes under interference into multiple measurable dimensions, thereby overcoming the subjectivity and imprecision of traditional qualitative analysis. This provides an objective, data-driven decision-making basis for accurately locating key vulnerable nodes, vulnerability transmission paths, and formulating resource allocation plans. The resilience dimensions of the multidimensional resilience index include shock resistance indicators, recovery agility indicators, adaptive reconfiguration indicators, and impact isolation indicators. Among them: The resilience index is used to measure the maximum deviation of a node's performance from its baseline operating level under disturbance shocks. By extracting the time series data of the node within the simulation period, the maximum negative deviation of the performance index (such as inventory level, order completion rate) from the baseline value is identified (e.g., the percentage difference between minimum inventory and safety stock). The smaller the value of this index, the stronger the node's ability to resist performance degradation in the early stages of disturbance. It is used to identify weak nodes that are prone to severe performance degradation during disturbance shocks. Recovery agility metrics are used to measure the time efficiency required for a node to recover from its performance low to a baseline level. Based on time series data, the timestamps of the performance low point and the timestamps of the performance reaching or exceeding the baseline value are used to calculate the time interval between the two. The smaller the value of this metric, the faster the node recovers. It is used to evaluate the ability of a node to quickly recover normal operation from the disturbance, which is crucial for reducing overall losses. The adaptive reconfigurability index is used to evaluate the ability of a node to establish a new and stable operating state after the disturbance subsides. It is achieved by comparing the average performance of the node over a period of time after the disturbance subsides (such as the last few simulation steps) with the initial baseline value, or by analyzing the rate of change of the performance level of the new steady state relative to the old steady state through differential equations. The higher the index, the stronger the node's adaptability and robustness. It is used to determine whether the node has the ability to optimize its own structure or process to achieve better operation after the disturbance. The impact isolation index is used to quantify the concentration of node performance loss caused by interference in the entire industry chain network. It analyzes the distribution of performance loss of all nodes on the interference propagation path through statistical methods (such as the Gini coefficient or entropy method). If the loss is highly concentrated in a few nodes, the index value is low, indicating that the impact of interference has not been effectively dispersed. It is used to assess the overall robustness and risk dispersion capability of the industry chain network and identify bottleneck nodes that, once damaged, would cause serious global impact.

[0021] Step S302: Normalize the calculated values ​​of the shock resistance index, recovery agility index, adaptive reconfiguration index and impact isolation index to the range of [0,1], plot them in the same radar chart, and generate a toughness profile; Step S303: Analyze the profile of the resilience profile, identify resilience dimensions whose index values ​​are lower than the preset health value (e.g., 0.5), and in conjunction with the interference propagation path, trace back the nodes and edges in the interference propagation path whose state data fluctuation range exceeds the preset fluctuation threshold (e.g., standard deviation greater than 10%), and locate them as key vulnerable nodes and vulnerable transmission paths; by analyzing the resilience profile, automatically identify resilience dimensions whose values ​​on the radar chart are lower than the preset health value, mark the resilience dimension as weak, and then, based on the reverse tracing algorithm, start from the node with weak resilience dimension, and trace back along the interference propagation path recorded in the simulation results to find the point where the state data fluctuation is earliest and most severe; by calculating the variance or maximum value of the state data of each node on the interference propagation path, mark the nodes that exceed the preset fluctuation threshold as key vulnerable nodes, and mark the connection chain between key vulnerable nodes as vulnerable transmission paths; Step S400: Generate an initial resource allocation plan based on key vulnerable nodes and vulnerable transmission paths and release it to the participants; receive feedback proposals from the participants; iterate and optimize the initial resource allocation plan based on the feedback proposals; and generate a digital collaboration protocol. Step S401: Based on the critical vulnerable nodes and vulnerable transmission paths, and in conjunction with the preset strategy library, a set of candidate enhancement strategies is generated. Each enhancement strategy in the preset strategy library is configured with corresponding applicable conditions. After successful matching, the enhancement strategy is converted into a specific resource scheduling instruction. The resource scheduling instructions are integrated to generate an initial resource allocation plan. The initial resource allocation plan is a resource scheduling report generated for the critical vulnerable nodes and vulnerable transmission paths. The resource scheduling report includes at least the types, quantities, sources, and targets of the resources that need to be scheduled for the critical vulnerable nodes and vulnerable transmission paths of the process. The pre-defined strategy library is a structured knowledge base or database that stores a variety of predefined coping strategies to enhance the resilience of the industry chain network. Each strategy entry is associated with detailed metadata, including strategy identifier, strategy description, applicable conditions, expected effects, utility score, and historical execution case references. The pre-defined strategy library provides strategy templates and decision support for automatically generating initial resource allocation plans. After identifying key vulnerable nodes and vulnerability transmission paths, it can quickly match and recommend applicable enhancement strategies, transforming them into specific resource scheduling instructions. This improves the efficiency and scientific nature of initial resource allocation plan generation and avoids the blindness of ad-hoc decisions. Enhancement strategies are specific countermeasures or solutions proposed to address the vulnerabilities of specific critical vulnerable nodes and vulnerability transmission paths. These include resource scheduling strategies (such as emergency procurement of raw materials from redundant suppliers, sharing inventory among participants, and temporarily allocating human resources or production capacity and equipment), process optimization strategies (such as adjusting production scheduling priorities and activating backup logistics routes), and collaborative strategies (such as establishing information sharing mechanisms and joint procurement). For example, for a manufacturer identified as a critical vulnerable node, the enhancement strategy is to "urgently allocate 100 tons of raw materials from backup supplier B". The applicable conditions are a set of prerequisite rules or constraints that determine whether an enhancement strategy can be activated and applied, including node type constraints (e.g., the strategy only applies to manufacturer nodes), disturbance type matching (e.g., the strategy applies to supply-side disruptions), resource availability requirements (e.g., the sender must have sufficient idle inventory), cost-effectiveness thresholds (e.g., the expected cost of implementing the strategy must be lower than the loss that could be avoided), and time and space constraints (e.g., the strategy must be activated within a specific time window after the disturbance occurs). For example, the applicable conditions for the above-mentioned enhancement strategy's exemplary emergency procurement strategy include "when the primary supplier is delayed for more than 48 hours and the backup supplier's inventory level is greater than 50 tons". Step S402: The initial resource allocation plan is structurally encapsulated into a resource allocation strategy matrix. The resource allocation strategy matrix is ​​in tabular form and contains a description of each resource scheduling task. Each row represents a resource scheduling task, and the columns must include at least the resource identifier, scheduling time, originating party identifier, originating party identifier, scheduling quantity, expected completion time, and a textual description of the collaborative benefits that the resource scheduling task can bring. The resource allocation strategy matrix is ​​published to the participants, and feedback proposals based on private information and local constraints are received from the participants. Feedback proposals include suggestions for modifying the resource scheduling report, execution preconditions, and compensation requirements. Wherein: Private information of participating parties refers to core internal data that each participating party is unwilling to fully disclose due to commercial confidentiality considerations. This includes detailed cost structure data (such as raw material procurement costs, unit production costs, and inventory holding costs), internal operational efficiency indicators (such as Overall Equipment Effectiveness (OEE) and employee labor productivity), precise capacity planning and production scheduling details, undisclosed financial data (such as cash flow status and profit margins), core technical parameters or formulas, and specific customer information or trade secret contract terms. Local constraints are restrictions that each participating party must comply with due to its internal operating policies, physical limitations, or existing contractual obligations. These include physical capacity limits (such as maximum daily output of production lines), warehouse capacity limits, minimum safety stock standards, human resource availability (such as maximum working hours limits), cash flow constraints (such as maximum advance payment limits), contract performance commitments (such as order delivery deadlines for other customers), and internal quality management standards or policy and regulatory requirements. Private information and local constraints are not fully obtained when the initial resource allocation plan is generated. However, after receiving the plan, the participating parties assess the impact of the initial resource allocation plan on their own operations based on the private information. They then propose modification suggestions, execution preconditions, or compensation requirements in their feedback proposals to ensure that the resource scheduling task is feasible in practice and does not harm their own interests.

[0022] Step S403: Based on the benefit balancing model, perform a global feasibility check on all feedback proposals. Feedback proposals that pass the feasibility check are marked as feasible feedback proposals. Feedback proposals with resource conflicts or logical contradictions are identified and marked. Wherein: Resource conflicts are determined by examining the competitive demand of multiple resource scheduling tasks for the same resource pool within the resource allocation strategy matrix. First, all scheduling tasks involving the same resource identifier (such as a specific type of raw material or equipment) are identified. Then, it is verified whether the total number of scheduling tasks within the same time period exceeds the actual available inventory or capacity of that resource at the sending end. If it does, it is marked as a resource conflict; otherwise, there is no resource conflict. For example, if two different enhancement strategies simultaneously require scheduling the same batch of raw materials from the same supplier, and the total demand exceeds its inventory, a conflict is determined to exist. The determination of logical contradictions is achieved by analyzing the logical consistency between the modification suggestions, preconditions, or compensation requirements proposed in the feedback proposals and the initial resource allocation plan or other feedback proposals. First, temporal logic is checked; for example, one feedback proposal requires task A to start only after task B is completed, but another requires task B to execute before task A. Then, causal logic is checked; for example, one feedback proposal denies the existence of a vulnerable transmission path, which is the basis for resource scheduling in the initial resource allocation plan. Finally, constraint violations are checked; for example, a feedback proposal's scheduling scheme violates known operational constraints (such as causing a node's inventory to fall below its safety standard). The benefit balancing model automatically identifies and marks feedback proposals with the aforementioned conflicts or contradictions by traversing the resource allocation strategy matrix and all feedback proposals, applying the above determination methods for consistency checks and resource supply and demand simulations. Marked proposals require further negotiation, revision, or removal to ensure that the final digital collaboration protocol is feasible and consistent globally.

[0023] Step S404: For feedback proposals that pass the feasibility verification, generate a compensation plan based on the benefit balance model: First, estimate the resilience value added that the initial resource allocation plan could create for the entire industry chain network if fully implemented; second, calculate the share of each participant in the resilience value added based on the value distribution rules in cooperative game theory; then, compare this share with the additional costs or risks borne by each participant for executing the resource allocation task, and generate a set of compensation plans for filling the gap, such as transfer payments or future priority rights; wherein: Resilience value enhancement refers to the sum of losses avoided and gains expected to be achieved by fully implementing the initial resource allocation plan for the entire industry chain network, compared to the situation without any intervention. This includes directly avoided financial losses (such as lost sales due to reduced production interruptions and avoided penalties due to stockouts), indirect gains from improved operational efficiency (such as reduced holding costs due to inventory optimization and transportation cost savings due to optimized logistics routes), and intangible strategic value (such as brand reputation maintenance, market share retention, and improved customer satisfaction). Resilience value enhancement provides a quantifiable value basis for the subsequent distribution of benefits and compensation among various participants, enabling the contributions and sacrifices of different participants to be measured and compared on a unified scale. The compensation scheme is a reward mechanism that balances the interests of participating parties and fills the gap between the additional costs or risks they bear in executing resource allocation tasks and their share of the overall resilience value-added. It is generated based on the value distribution rules (such as Shapley values) in cooperative game theory. For example, it is first estimated that executing the initial resource allocation plan creates 1 million yuan of resilience value-added for the entire industry chain network. Then, based on the contribution of each participating party (such as the amount of resources provided and the magnitude of risks borne), it is calculated that the manufacturer should receive 400,000 yuan, the supplier 300,000 yuan, and the logistics provider 300,000 yuan. However, if the supplier incurs an additional 100,000 yuan in costs due to providing emergency supplies, the compensation scheme may include a direct transfer payment of 80,000 yuan to the supplier and a commitment to give it priority or price discounts of 20,000 yuan in future procurements, thereby ensuring that its net income is not less than its share and incentivizing its active participation in collaboration.

[0024] Step S405: Based on the compensation plan and the feedback proposal on feasibility, revise the initial resource allocation plan, update the description of resource scheduling tasks in the resource allocation strategy matrix, and add compensation clauses and execution preconditions based on the compensation plan. Step S406: Iteratively execute steps S402 to S406 for M rounds until the updated resource allocation strategy matrix meets the requirements of the participants or reaches the preset maximum number of iterations. Generate a digital collaboration protocol based on the initial resource allocation plan and compensation scheme that have reached the final consensus. Based on the benefit balance model, predict the multidimensional resilience index and the predicted participant satisfaction obtained after the execution of the digital collaboration protocol.

[0025] Step S500: Execute the digital collaboration protocol and monitor the execution process and results to obtain actual operational data and feedback data from participants during the execution process. Generate evaluation results that include actual multi-dimensional resilience indicators and actual participant satisfaction. Optimize the dynamic digital model, interest balance model, and preset strategy library based on the evaluation results. Step S501: Optimize the dynamic digital model based on the evaluation results, including: After the digital collaboration protocol is executed, the actual operation data of the industrial chain network during the operation cycle is collected, the actual operation data is compared with the simulation results, and the difference between the predicted and actual values ​​of key state variables at the same point in time is calculated. With the goal of minimizing the differences, the strength coefficients of the causal transmission rules and the threshold parameters of the operational constraint rules in the dynamic digital model are automatically calibrated, and the calibrated strength coefficients and threshold parameters are updated to the dynamic digital model.

[0026] Actual operational data refers to the real data collected during the actual operation of a complete operational cycle (e.g., one month) of the industrial chain network after the digital collaboration protocol is executed, corresponding to the state variables used in the simulation. This includes actual resource data of each participant (such as daily records of raw materials, semi-finished products, and finished products inventory, actual output of production lines, actual operating status and downtime of equipment), actual business data (such as actual delivery time of purchase orders, actual completion status of production work orders, actual delivery status and delay records of sales orders, actual trajectory and delivery time of logistics waybills), and actual environmental data obtained from external sources (such as actual price fluctuations in the raw material market and actual sales volume of final products). By comparing with simulation results, the predictive accuracy of the dynamic digital model is evaluated, and the calibration of model parameters is driven, thereby forming a closed-loop learning process and continuously improving the reliability of the method and the scientific nature of decision-making. When comparing actual operational data with simulation results, the main focus is on quantitatively comparing predicted values ​​with actual observed values ​​at the same time points and nodes. First, time alignment is performed, synchronizing the time series of simulation results with the time series of actual operational data (summarized at the same time granularity, such as hourly or daily summaries). Then, variable comparison is conducted, primarily including: comparison of state variable time series curves, for example, plotting the curve of the simulated "finished goods inventory level of Manufacturer A" over time with the actual recorded inventory curve on the same chart to visually observe trend differences; comparison of key event occurrence times, such as comparing the time differences between the simulation and actual occurrence of the event "inventory falling below the safety line"; and comparison of statistical characteristic values, such as calculating the absolute error or root mean square error between the predicted and actual values ​​of "average logistics delay time" throughout the entire simulation period. The entire comparison process is automatically executed through a specialized verification algorithm, outputting a difference analysis report. Key state variables refer to dynamic attributes in an entity relationship network that are sensitive to disturbances and crucial for measuring supply chain performance; these include inventory levels (raw materials, semi-finished products, finished products), order completion rate, capacity utilization rate, and order delivery delay time. The calculation of key state variables is based on collected raw business and resource data. For example, the daily order completion rate of a manufacturer node is calculated as: (Number of completed orders / Number of planned orders for the day) * 100%; the capacity utilization rate is calculated as: (Actual output / Maximum designed capacity) * 100%. In the comparison, the calculated actual values ​​are compared with the corresponding predicted values ​​in the simulation. The strength coefficient of a causal transmission rule is a parameter that quantifies the influence of causal relationships in the rule. For example, in the rule "Supplier delay will cause manufacturer production interruption", the strength coefficient may be specified as the coefficient "0.8" in "For every 1 day of supplier delay, the manufacturer's production will be interrupted for 0.8 days". It determines the degree of attenuation or amplification of interference propagating in the network. The threshold parameters of the operational constraint rules are the specific values ​​that define the operational restriction boundaries; for example, the "maximum production capacity limit" is "1,000 pieces per day", the "minimum safety stock standard" is "200 tons", and the "maximum allowable order delay" is "48 hours". Step S502: Optimize the benefit balance model based on the evaluation results, including: Analyze the execution results of the digital collaboration protocol, calculate the actual multidimensional resilience index and the actual participant satisfaction, compare the actual participant satisfaction with the predicted participant satisfaction, and calculate the deviation between the actual participant satisfaction and the predicted participant satisfaction. If the deviation is greater than the set satisfaction threshold, the internal parameters used to calculate the compensation scheme in the benefit balance model will be adjusted; otherwise, they will not be adjusted.

[0027] The specific results include actual multidimensional resilience indicators and actual participant satisfaction. The actual multidimensional resilience indicators are calculated based on actual operational data, reflecting the actual values ​​of shock resistance, recovery agility, adaptability to restructuring, and impact isolation indicators. They reflect the real resilience improvement effect of the industrial chain network after the implementation of the digital collaboration agreement. Actual participant satisfaction is quantitatively assessed through questionnaires, feedback systems, or by analyzing behavioral data such as participants' willingness to renew agreements and depth of cooperation after the implementation of the agreement. It reflects the subjective satisfaction of each participant with this collaboration. The satisfaction threshold is a pre-set minimum value used to judge whether the satisfaction of the participants is acceptable. It is usually set by the system administrator or the main managers of the alliance or branch of the industry chain network based on historical cooperation experience and management objectives. It is generally a value between 0 and 1, such as 0.7, which is 70% satisfaction. If the actual satisfaction of a participant is lower than 0.7, it is considered that the participant is not satisfied and the model needs to be adjusted. The internal parameters of the bias-adjusted benefit balance model include: First, compare the predicted participant satisfaction of each participant with the actual participant satisfaction and calculate the deviation (absolute or relative value). Then, identify the participants whose deviation exceeds the set satisfaction threshold (e.g., 0.7). Next, analyze the reasons for the deviation, such as whether uneven value distribution or insufficient compensation led to some participants' actual gains being lower than expected. Based on this, with the goal of narrowing the gap between predicted and actual satisfaction, use optimization algorithms (e.g., gradient descent) to fine-tune the internal parameters of the benefit balance model. The adjustment is based on the data patterns of historical execution results, with the goal of making the benefit balance model's next prediction closer to the actual result. The internal parameters specifically include: contribution weighting parameters, used to calculate the contribution ratio of each participant in the overall resilience value-added; cost and risk conversion coefficients, used to uniformly convert the additional costs or risks borne by the participants into a compensable value; and compensation preference coefficients, used to reflect the participants' preference for different forms of compensation. These coefficients affect the composition of the final compensation scheme.

[0028] Step S503: Optimize the preset strategy library based on the evaluation results, including: Based on the execution results of the digital collaboration protocol, the adopted enhancement strategies are evaluated. The evaluation dimensions include three aspects: First, contribution, which is the degree to which the enhancement strategy contributes to the improvement of multi-dimensional resilience indicators. This is calculated by comparing the changes in the actual resilience indicators of nodes or paths directly related to the enhancement strategy before and after the execution of the digital collaboration protocol. Second, execution cost, which is the total resource cost incurred in executing the enhancement strategy. This is obtained directly from the financial systems or cost reports of the participating parties, including direct procurement costs, transportation costs, and labor costs. Third, participant acceptance, which is the participants' subjective acceptance of the enhancement strategy and their willingness to implement it. This is extracted by analyzing the participants' attitudes during the feedback phase and the ratings of relevant items in the satisfaction survey after the protocol's implementation. After the digital collaboration protocol is executed, data from the above three dimensions is automatically collected, and then each enhancement strategy adopted is comprehensively scored through a multi-objective evaluation function. For example, one enhancement strategy may have a high contribution but also a high cost, while another enhancement strategy may have a low cost but poor acceptance. The evaluation function will weigh these dimensions and give an overall evaluation. Based on the strategy evaluation results, update the utility scores and applicable conditions of candidate enhancement strategies in the preset strategy library. First, for enhancement strategies executed this time, update their historical average utility scores according to this strategy evaluation. Second, analyze the success and failure experiences of this enhancement strategy execution. If the enhancement strategy is effective or ineffective under certain specific conditions, refine or modify its applicable conditions accordingly. For example, if it is found that the enhancement strategy of "urgent procurement from supplier B" has low acceptance when "raw material price fluctuations are greater than 10%", the constraint "and raw material price fluctuations are less than 10%" can be added to its applicable conditions. Among them, the utility score of the enhancement strategy is used to quantify the average effectiveness of the enhancement strategy in history, and comprehensively reflects the contribution, implementation cost and acceptance of the enhancement strategy in past applications; the applicable conditions are a set of logical rules that define the environment and prerequisites that must be met for the successful application of the enhancement strategy. The entire process from the generation of the composite interference scenario to the completion of the digital cooperative protocol execution is treated as a new structured case. It includes the composite interference scenario, the enhancement strategies used, the digital cooperative protocol, the execution results, and the evaluation results. These are stored in a pre-set strategy library as reference knowledge for future strategy matching and generation.

[0029] In this embodiment, by integrating multi-source heterogeneous data such as resource data, business data, and external environment from various participants, and combining causal transmission rules and operational constraint rules, a dynamic digital model capable of simulating the real operation mechanism of the industrial chain is generated. Through the simulation of extreme situations where multiple risks occur simultaneously in reality through composite interference scenarios, the simulation results can dynamically deduce the propagation path and impact of interference in the entire industrial chain network. Based on the simulation results, corresponding multi-dimensional resilience indicators and visualized resilience profiles are generated, realizing the accurate positioning of key vulnerable nodes and vulnerable transmission paths. This transforms traditional qualitative and empirical judgments into data-driven quantitative and visualized analysis, greatly improving the accuracy and depth of identifying potential risks in the industrial chain. The simulation of different types of composite interference scenarios can simulate the impact of various complex factors that occur in the actual industrial chain, thereby generating corresponding resource allocation strategies, providing a basis for responding to and preventing actual interference scenarios, and increasing the ability of the industrial chain network to cope with sudden interference scenarios. Based on an iterative negotiation mechanism using a benefit-balancing model, a digital collaboration protocol is generated and executed. Feedback proposals from all participants are incorporated into the iterative negotiation process, resulting in a fair value distribution and compensation scheme. While protecting the individual interests of participants, this approach incentivizes all parties to voluntarily and actively participate in collaboration. Furthermore, the execution results of the digital collaboration protocol are used to optimize each step, leading to self-evolution of the method and improved simulation accuracy. This ensures that simulation results for complex interference scenarios more closely reflect actual impacts, enabling proactive resource scheduling and collaboration. This fosters a sustainable supply chain resource collaboration ecosystem that addresses shared risks while respecting and protecting individual interests, ultimately achieving collaborative optimization and closed-loop enhancement of the overall resilience of the supply chain network. Specific Implementation Example 2: This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code, which, when executed by the one or more processors, can perform a resource-intelligent decision-making method based on supply chain collaborative optimization as described above.

[0031] The method or system according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as ROM or hard disk, may store the resource intelligent decision-making method based on supply chain collaborative optimization provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components in the electronic device shown in this application may be omitted according to actual needs. Specific Implementation Example 3: One embodiment of this application discloses a computer-readable storage medium. The computer-readable storage medium stores computer-readable instructions. When the computer-readable instructions are executed by a processor, a resource intelligent decision-making method based on supply chain collaborative optimization according to an embodiment of this application, as described with reference to the above figures, can be performed. The storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0033] Furthermore, according to embodiments of this application, the processes described in the above-referenced flowcharts can be implemented as computer software programs. For example, this application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be executed by a processor to perform instructions corresponding to the method steps provided in this application, such as a resource intelligent decision-making method based on supply chain collaborative optimization. When this computer program is executed by a central processing unit (CPU), it performs the functions defined in the method of this application.

[0034] A resource-intelligent decision-making method system based on supply chain collaborative optimization includes a processor and a machine-readable storage medium. The machine-readable storage medium and the processor are connected. The machine-readable storage medium is used to store programs, instructions, or code. The processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-mentioned method.

[0035] It should be noted that, in this document, relational terms such as "first" and "second" are used merely 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 a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A resource-intelligent decision-making method based on supply chain collaborative optimization, characterized in that, The method includes: Acquire multi-source heterogeneous data from participants in the industrial chain network, construct an entity relationship network, and embed causal transmission rules and operational constraint rules into the entity relationship network to generate a dynamic digital model; A structured interference element library is constructed, and composite interference scenarios are synthesized and injected into a dynamic digital model. Simulation results are generated through deduction. Based on simulation results, multidimensional resilience indices are calculated to generate resilience profiles, and key vulnerable nodes and vulnerability transmission paths in the industrial chain network are located based on the resilience profiles. An initial resource allocation plan is generated based on key vulnerable nodes and vulnerability transmission paths and released to the participants. Feedback proposals from the participants are received, and the initial resource allocation plan is iteratively optimized based on the feedback proposals to generate a digital collaboration protocol. The digital collaboration protocol is executed and the execution process and results are monitored to generate an evaluation result that includes actual multi-dimensional resilience indicators and actual participant satisfaction. The method is then optimized based on the evaluation result.

2. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 1, characterized in that, The participants include suppliers, manufacturers, logistics providers, and distributors. The multi-source heterogeneous data includes resource data, business data, and external environment data. The entity relationship network uses participants as nodes and business connections formed between participants based on business data as edges. The entity relationship network contains multi-source heterogeneous data of each participant.

3. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 1, characterized in that, The process involves constructing a structured interference element library, synthesizing composite interference scenarios, injecting them into a dynamic digital model, and generating simulation results through deduction, including: The structured interference element library contains N interference elements, each of which contains four dimensions of attribute information: type information, intensity information, spatiotemporal attribute information, and occurrence probability information. n interference elements are extracted from the structured interference element library, and the n interference elements are dynamically combined based on spatiotemporal attribute information to generate a composite interference scenario containing the concurrent effects of the n interference elements. The complex interference scenario is injected into the dynamic digital model. Based on the simulation engine and using the causal transmission rule, the interference propagation path of the concurrent influence of n interference elements in the entity relationship network is deduced. The state data of each node at a specified time within the deduction period is recorded, and the simulation results of the state data of each node in the entity relationship network changing over time are generated.

4. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 3, characterized in that, The process of calculating multidimensional resilience indices based on simulation results to generate a resilience profile, and then using this profile to locate key vulnerable nodes and vulnerability transmission paths in the industrial chain network, includes: Based on the simulation results, time series data of each node in the simulation period are extracted, and multidimensional resilience index is calculated. The resilience dimensions of the multidimensional resilience index include shock resistance index, recovery agility index, adaptive reconfiguration index and impact isolation index. The calculated values ​​of shock resistance, recovery agility, adaptive reconfiguration, and impact isolation indices are normalized and plotted on the same radar chart to generate a toughness profile. By analyzing the profile of the resilience profile, resilience dimensions with index values ​​lower than preset health values ​​are identified. Combined with the interference propagation path, the nodes and edges in the interference propagation path whose state data fluctuation range exceeds the preset fluctuation threshold are traced back to locate key vulnerable nodes and vulnerable transmission paths.

5. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 4, characterized in that, The process involves generating an initial resource allocation plan based on key vulnerable nodes and vulnerability transmission paths, releasing it to participating parties, receiving feedback proposals from participants, iteratively optimizing the initial resource allocation plan based on these feedback proposals, and generating a digital collaboration protocol, including: Based on key vulnerable nodes and vulnerability transmission paths, and combined with a preset strategy library, a set of candidate enhancement strategies are matched and generated and converted into resource scheduling instructions. The resource scheduling instructions are integrated to generate an initial resource allocation plan, which is a resource scheduling report generated for key vulnerable nodes and vulnerability transmission paths. The initial resource allocation plan is structurally encapsulated into a resource allocation strategy matrix, which contains a description of each resource scheduling task. The resource allocation strategy matrix is ​​published to the relevant participants, and feedback proposals from the participants are received. The feedback proposals include suggestions for modifying the resource scheduling report, execution preconditions, and compensation requirements. Based on the interest balancing model, a global feasibility check is performed on all feedback proposals. Feedback proposals that pass the feasibility check are marked as feasible feedback proposals. Feedback proposals with resource conflicts or logical contradictions are identified and marked.

6. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 5, characterized in that, The process of iteratively optimizing the initial resource allocation plan based on feedback proposals to generate a digital collaboration protocol includes: For feedback proposals that pass the feasibility verification, a compensation plan is generated based on the benefit balancing model; Based on the feedback proposals regarding compensation schemes and feasibility, the initial resource allocation plan is revised, and the descriptions of resource scheduling tasks in the resource allocation strategy matrix are updated. The process iteratively executes the acceptance of M rounds of feedback proposals and the revision of the initial resource allocation plan until the updated resource allocation strategy matrix meets the requirements of the participants or reaches the preset maximum number of iterations. Based on the final consensus of the initial resource allocation plan and compensation scheme, a digital collaboration agreement is generated. Based on the benefit balance model, the predicted multidimensional resilience index and predicted participant satisfaction are obtained after the execution of the digital collaboration agreement.

7. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 6, characterized in that, Method optimization based on evaluation results includes optimizing the dynamic numerical model, the interest balancing model, and the preset strategy library. Optimization of the dynamic numerical model based on the evaluation results includes: After the digital collaboration protocol is executed, collect the actual operation data of the industrial chain network and calculate the difference between the actual operation data and the simulation results at the same point in time. With the goal of minimizing the differences, the intensity coefficients of the causal transmission rules and the threshold parameters of the operational constraint rules in the dynamic digital model are automatically calibrated, and the calibrated intensity coefficients and threshold parameters are updated to the dynamic digital model.

8. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 7, characterized in that, Based on the evaluation results, the benefit balance model is optimized, including: Analyze the execution results of the digital collaboration protocol, calculate the actual multidimensional resilience index and the actual participant satisfaction, and calculate the deviation between the actual participant satisfaction and the predicted participant satisfaction. If the deviation is greater than the set satisfaction threshold, the internal parameters used to calculate the compensation scheme in the benefit balance model will be adjusted; otherwise, they will not be adjusted.

9. A resource intelligent decision-making method based on supply chain collaborative optimization according to claim 7, characterized in that, Based on the evaluation results, the preset strategy library is optimized, including: Based on the execution results of the digital collaboration protocol, the adopted enhancement strategies are evaluated. The evaluation dimensions include the contribution of the enhancement strategies to the improvement of multi-dimensional resilience indicators, the execution cost, and the acceptance by the participants. Based on the results of the strategy evaluation, update the utility scores and applicable conditions of the candidate enhancement strategies in the preset strategy library.

10. The resource intelligent decision-making method based on supply chain collaborative optimization according to claim 1, characterized in that, A unified representation and measurement standard is established based on the multi-source heterogeneous data. Initial parameters are configured based on the simulation engine, including the simulation time span, simulation step size, and simulation convergence conditions.

Images