Supply chain resilience path identification method based on fuzzy set qualitative comparative analysis

Abstract

The present application relates to the technical field of supply chain management, and in particular to a supply chain resilience path identification method based on fuzzy set qualitative comparative analysis. The method comprises: acquiring historical operation data of nodes in a supply chain and response data under a sudden incident; on the basis of the similarity of the historical operation data between the nodes, obtaining the degree of operational correlation between the nodes; grouping all the nodes to obtain a plurality of node group clusters; on the basis of the response consistency of the nodes under the sudden incident, screening out a key node group cluster; on the basis of the distribution characteristics of the response data of different nodes, obtaining the response priority of each node; on the basis of the change difference of the response data between each node in the key node group cluster and other nodes in the same batch, obtaining the degree of external interference of each node; and on the basis of the degree of external interference, correcting the response priority to obtain a corrected response priority. The present invention can improve the evaluation efficiency and decision support capability of supply chain resilience.

Description

A Supply Chain Resilience Path Identification Method Based on Fuzzy Set Qualitative Comparison Technical Field

[0001] This invention belongs to the field of supply chain management and data analysis technology, specifically a supply chain resilience path identification method based on fuzzy set qualitative comparison. Background Technology

[0002] As supply chains become increasingly globalized and complex, supply chain resilience has become a core concern for enterprises. Traditional supply chain resilience assessment methods rely heavily on qualitative analysis, lacking systematic quantitative indicators and models, thus failing to fully reflect the supply chain's ability to cope with unforeseen events. A supply chain resilience path identification method based on fuzzy set qualitative comparison has emerged to improve the accuracy and comprehensiveness of supply chain assessment and provide decision support for supply chain management.

[0003] A search revealed a method, system, electronic device, and storage medium for assessing the cascading impacts of a local energy crisis, with publication number CN117273446B, published on June 25, 2024. This patent proposes a method for assessing the cascading impacts of a local energy crisis based on an input-output model, capable of analyzing the cascading effects of the energy crisis from a supply chain perspective and identifying the supply chain transmission paths of the energy crisis. However, this technical solution has the following shortcomings:

[0004] Insufficient qualitative analysis: This method mainly relies on input-output models. Although it can analyze the cascading effects of the energy crisis from a macroeconomic perspective, it lacks specific qualitative analysis of each node in the supply chain and cannot fully reflect the supply chain's ability to cope under different scenarios.

[0005] Limited quantitative indicators: This method mainly proposes a monetization indicator for the energy crisis, which is helpful in evaluating economic benefits, but it fails to cover multi-dimensional indicators of supply chain resilience, such as time response and resource allocation, resulting in an incomplete evaluation result.

[0006] Limited applicability: This method is mainly used to evaluate local energy crises, and its applicability to other types of supply chain crises (such as market fluctuations, natural disasters, etc.) is poor, making it difficult to fully reflect the multidimensional resilience of the supply chain.

[0007] The aforementioned problems indicate that existing methods for assessing the cascading impacts of local energy crises still have certain shortcomings in terms of qualitative analysis, the multidimensionality of quantitative indicators, and the scope of application. Therefore, this invention provides a supply chain resilience path identification method based on fuzzy set qualitative comparison, aiming to optimize the qualitative analysis method for supply chain resilience assessment, introduce multidimensional quantitative indicators, improve the accuracy and comprehensiveness of the assessment, and thus enhance the decision support capabilities of supply chain management. Summary of the Invention

[0008] To address the technical problems of insufficient qualitative analysis, limited quantitative indicators, and narrow applicability in existing supply chain resilience assessment methods, this invention aims to provide a supply chain resilience path identification method based on fuzzy set qualitative comparison. The specific technical solution adopted is as follows:

[0009] This invention proposes a supply chain resilience path identification method based on fuzzy set qualitative comparison, the method comprising:

[0010] Acquire historical operational data and corresponding response data under emergencies for each node in the supply chain; obtain the operational correlation between nodes based on the similarity of historical operational data; group all nodes according to the operational correlation to obtain multiple node clusters; and select key node clusters based on the consistency of responses of nodes in each node cluster under emergencies.

[0011] For any critical node cluster, the response priority of each node is obtained based on the distribution characteristics of the response data of different nodes; the external interference degree of each node in each critical node cluster is obtained based on the difference in the change of response data between each node in the critical node cluster and other nodes in the same batch.

[0012] The response priority is adjusted based on the external interference level of each node in each critical node cluster to obtain the response adjustment priority of each node in each critical node cluster.

[0013] Furthermore, the method for obtaining the operational correlation degree includes:

[0014] Obtain the feature vector of the historical running data of each node, calculate the cosine similarity of the feature vectors between nodes, and use it as the running correlation between nodes.

[0015] Furthermore, the method for obtaining the node group clusters includes:

[0016] Based on the operational correlation between nodes, hierarchical clustering is performed on all nodes to obtain multiple node groups.

[0017] Furthermore, the method for obtaining the key node group clusters includes:

[0018] For any node cluster, count the number of nodes that respond consistently under a sudden event, and use this as the response matching count.

[0019] Based on the difference between the number of response matches and the total number of nodes, the non-critical probability of each node grouping cluster is obtained. The difference feature is negatively correlated with the non-critical probability.

[0020] If the non-critical probability of a node cluster is less than the preset critical threshold, the corresponding node cluster will be designated as a critical node cluster.

[0021] Furthermore, the method for obtaining the response priority includes:

[0022] If the response data of a node exceeds the preset response threshold range, the corresponding node will be marked as an abnormal node.

[0023] For any critical node cluster, the ratio between the number of abnormal nodes in each node and the total number of nodes is obtained, which is used as the response priority for each node.

[0024] Furthermore, the method for obtaining the external interference degree includes:

[0025] For any critical node cluster, the probability of external interference for each node is obtained based on the difference in response data between each node and other nodes in the same batch.

[0026] If the probability of external interference to a node is greater than a preset interference threshold, obtain the intersection of the abnormal responses between that node and all other nodes in the same batch; for each node within the intersection, use the probability of external interference to the corresponding node as the local interference degree of each node; for each node outside the intersection, set the local interference degree of each node to 0.

[0027] The mean of the local interference degree of each node is obtained as the overall interference degree of each node in the key node group cluster.

[0028] Furthermore, the method for obtaining the probability of external interference includes:

[0029] For any critical node group cluster, obtain the correlation coefficient between each node and the abnormal response set of other nodes in the same batch, as the local interference correlation;

[0030] Calculate the mean of the local interference correlation between each node and all nodes in the same batch, and use it as the external interference probability of each node in each key node group cluster.

[0031] Furthermore, the method for obtaining the response correction priority includes:

[0032] Calculate the difference between the positive integer 1 and the overall interference degree of each node in each key node group cluster, and use it as an adjustment factor;

[0033] The response priority of each node in each critical node cluster is weighted according to the adjustment factor to obtain the response correction priority of each node in each critical node cluster.

[0034] Further, the step of weighting the response priority of each node in each key node cluster according to the adjustment factor to obtain the response correction priority of each node in each key node cluster includes:

[0035] The product of the response priority and adjustment factor of each node in each critical node cluster is obtained and used as the response correction priority of each node in each critical node cluster.

[0036] Furthermore, the correlation coefficient is obtained using the Pearson correlation coefficient.

[0037] This invention offers the following advantages: It obtains the operational correlation between nodes based on the similarity of their historical operational data, quantifying the degree of operational correlation between different nodes; it groups all nodes according to this correlation, obtaining multiple node clusters, which helps identify groups of nodes with similar operational characteristics; it filters out key node clusters based on the consistency of responses of nodes in each cluster under sudden events, identifying key node groups requiring focused monitoring; for any key node cluster, it obtains the response priority of each node based on the distribution characteristics of different nodes' response data, allowing for an understanding of the importance of different nodes in the response; it obtains the external interference degree of each node in each key node cluster based on the differences in response data changes between each node in the key node cluster and other nodes in the same batch; and it obtains the response correction priority of each node in each key node cluster, more accurately reflecting the actual importance of each node in the response. This invention improves the efficiency of supply chain resilience assessment and decision support capabilities by determining the optimal response order by obtaining the accurate response priority of each node in the cluster. Attached Figure Description

[0038] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 is an overall flowchart of a supply chain resilience path identification method based on fuzzy set qualitative comparison provided by an embodiment of the present invention.

[0040] Figure 2 is a flowchart of a node operation correlation calculation method provided in an embodiment of the present invention.

[0041] Figure 3 is a flowchart of a key node grouping cluster screening method provided in an embodiment of the present invention.

[0042] Figure 4 is a flowchart of a node response priority acquisition method provided in an embodiment of the present invention.

[0043] Figure 5 is a flowchart of a node external interference degree calculation method provided in an embodiment of the present invention.

[0044] Figure 6 is a flowchart of a response correction priority determination method provided in an embodiment of the present invention. Detailed Implementation

[0045] This invention provides a supply chain resilience path identification method based on fuzzy set qualitative comparison. Its core lies in quantifying the operational correlation between nodes, screening key node clusters, calculating response priorities and external interference levels, and ultimately determining response correction priorities, thereby achieving efficient assessment and optimization of supply chain resilience. The specific implementation of this invention is described in detail below with reference to Figures 1 to 6 and specific embodiments.

[0046] First, as shown in Figure 1, the overall process of this invention includes multiple steps, from data acquisition to final response correction priority determination. In the first step, it is necessary to acquire historical operational data for each node in the supply chain, as well as response data under corresponding emergencies. This data can come from records in the supply chain management system, such as historical operational data like inventory levels, logistics transportation time, and order completion rates, as well as response data during emergencies (such as natural disasters, sudden changes in market demand, etc.). To ensure the comprehensiveness and accuracy of the data, it is recommended to collect at least one year's worth of data and preprocess the data to remove noise and outliers. Preprocessing methods may include data smoothing, missing value imputation, and standardization.

[0047] Next, based on the similarity of historical running data between nodes, the operational correlation between nodes is obtained. The specific implementation of this process is shown in Figure 2. First, the historical running data of each node is transformed into a feature vector. Assuming that the historical running data of a node contains n features, the feature vector of that node can be represented as X = [x1, x2, ..., x...]. n For any two nodes i and j, their eigenvectors are X and X, respectively. i and X j The operational correlation between nodes can be calculated using the cosine similarity formula: S ij =(X i ·X j ) / (|X i |·|X jThe symbol |) represents the vector dot product, and || represents the magnitude of the vector. The cosine similarity ranges from -1 to 1; the closer the value is to 1, the more similar the operational characteristics of the two nodes. By calculating the cosine similarity between all node pairs, an operational association matrix can be constructed for subsequent node grouping operations.

[0048] After obtaining the operational correlation, the next step is to group all nodes to obtain multiple node clusters. As shown in Figure 3, this invention uses a hierarchical clustering algorithm to group nodes. Hierarchical clustering is a bottom-up clustering method. Its basic idea is to treat each node as an independent cluster and then gradually merge the most similar clusters until a stopping condition is met. In this invention, the stopping condition can be set as the average operational correlation of nodes within a cluster reaching a certain threshold T1. Specifically, assuming there are currently m clusters C1, C2, ..., C m For any two clusters C i and C j Its similarity is defined as the average of the operational associations between all pairs of nodes within a cluster: S(C i C j )=(1 / |C i ||C j |)∑(S n ), where r∈C i , t∈C j In each iteration, the two clusters with the highest similarity are selected for merging until the average operational correlation of all clusters is greater than or equal to T1. The resulting clusters of multiple nodes represent a group of nodes with similar operational characteristics.

[0049] After grouping the nodes, it is necessary to further filter out key node clusters. As shown in Figure 4, the selection criterion for key node clusters is the consistency of node responses under sudden events. For any node cluster, the number of nodes with consistent responses under sudden events is counted as the response matching number M. Response consistency can be determined by comparing whether the response data of nodes under sudden events falls within a certain preset range. For example, if the proportion of a node's response data deviating from the normal range is less than a certain threshold T2, then the node's response is considered consistent. Based on the difference between the response matching number M and the total number of nodes N, the non-critical probability P = 1 - (M / N) of each node cluster is calculated. The difference characteristic is negatively correlated with the non-critical probability, that is, the more response matching numbers there are, the lower the non-critical probability. If the non-critical probability of a node cluster is less than the preset critical threshold T3, then the node cluster is marked as a key node cluster. Through this screening process, key node groups that need to be monitored can be identified.

[0050] For each selected key node cluster, the next step is to obtain the response priority of each node based on the distribution characteristics of its response data. As shown in Figure 5, the response priority is calculated as follows: First, a preset response threshold range [T4, T5] is set. If the response data of a node exceeds this range, the node is marked as an anomalous node. For any key node cluster, the number of times each node is marked as an anomalous node, K, is counted, and the ratio R = K / N of this number to the total number of nodes, N, is calculated as the response priority of that node. A higher response priority indicates that the node's response to sudden events is more unstable and requires priority attention.

[0051] After obtaining the response priority, it is also necessary to consider the impact of external interference on the node response. As shown in Figure 6, the calculation method for external interference includes the following steps: First, for any critical node cluster, calculate the variation difference of the response data between each node and other nodes in the same batch to obtain the probability of external interference for each node. The variation difference can be measured by the Pearson correlation coefficient. Assume that the response data sequence of a certain node is Y = [y1, y2, ..., y...]. n The response data sequences of other nodes in the same batch are Z1, Z2, ..., Z. l The local interference correlation between this node and every other node can be calculated using the Pearson correlation coefficient formula: Where cov represents the covariance and σ represents the standard deviation. The mean of the local interference correlation between this node and all other nodes is calculated, which is taken as the external interference probability of this node: E = (1 / l)∑ρ(Y,Z) i If the probability of external interference to a node exceeds the preset interference threshold T6, the intersection of the abnormal responses of that node with all other nodes in the same batch is further analyzed. For each node within the intersection, its probability of external interference is used as the local interference degree; for each node outside the intersection, its local interference degree is set to 0. Finally, the mean of the local interference degree of each node is calculated as the overall interference degree of that node, D = (1 / m)∑d i , where m is the number of nodes in the intersection.

[0052] After obtaining the overall interference level for each node, the next step is to adjust the response priority based on the external interference level to obtain the adjusted response priority. Specifically, first, the difference F = 1 - D between the positive integer 1 and the overall interference level for each node is calculated as an adjustment factor. Then, the adjustment factor is multiplied by the response priority to obtain the adjusted response priority P' = R·F. The adjusted response priority comprehensively considers both the node's response priority and the external interference level, and can more accurately reflect the actual importance of each node in the response.

[0053] In summary, this invention achieves efficient assessment and optimization of supply chain resilience through the aforementioned steps. In practical applications, the method of this invention can be widely applied to various fields such as manufacturing, logistics, and retail. For example, in the manufacturing supply chain, the method of this invention can identify key suppliers and production nodes, thereby enabling the development of targeted risk response strategies; in the logistics industry, the method of this invention can optimize distribution routes and improve logistics efficiency; in the retail industry, the method of this invention can predict changes in market demand and adjust inventory strategies. Furthermore, this invention can be combined with other supply chain management tools, such as ERP systems and SCM systems, to further enhance the intelligence level of the supply chain.

[0054] To verify the effectiveness of this invention, the following experiment was conducted: Supply chain data from a large manufacturing enterprise was selected as the experimental subject. The data included historical operational data of 100 nodes and response data under unforeseen events. Using the method of this invention, 10 key node clusters were successfully identified, and the response correction priority for each node was calculated. Experimental results show that the method of this invention can significantly improve the efficiency of supply chain resilience assessment and decision support capabilities, providing enterprises with a scientific basis for decision-making.

Claims

1. A supply chain resilience path identification method based on fuzzy set qualitative comparison, characterized in that, The method comprises: obtaining historical operation data and corresponding response data of each node in the supply chain; obtaining operation correlation between nodes according to similarity of historical operation data between nodes; grouping all nodes to obtain a plurality of node grouping clusters according to the operation correlation between nodes; and screening out key node grouping clusters according to response consistency of nodes in each node grouping cluster under a burst event; for any key node grouping cluster, obtaining response priority of each node according to response data distribution characteristics of different nodes; and obtaining external interference degree of each node in each key node grouping cluster according to variation difference of response data between each node and other nodes in the same batch. The response priority is corrected according to the external interference degree of each node in each key node grouping cluster to obtain response correction priority of each node in each key node grouping cluster.

2. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 1, characterized in that, The operation correlation acquisition method comprises: obtaining a feature vector of historical operation data of each node, calculating cosine similarity of the feature vectors between nodes as the operation correlation between nodes.

3. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 1, characterized in that, The node grouping cluster acquisition method comprises: hierarchical clustering all nodes according to the operation correlation between nodes to obtain a plurality of node grouping clusters.

4. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 1, characterized in that, The key node grouping cluster acquisition method comprises: for any node grouping cluster, counting the number of nodes responding consistently under a burst event as a response matching number; obtaining non-key possibility of each node grouping cluster according to difference characteristics between the response matching number and the number of all nodes, wherein the difference characteristics and the non-key possibility are negatively correlated; if the non-key possibility of the node grouping cluster is less than a preset key threshold, the corresponding node grouping cluster is taken as a key node grouping cluster.

5. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 1, characterized in that, The response priority acquisition method comprises: if the response data of the node is out of a preset response threshold range, the corresponding node is marked as an abnormal node; for any key node grouping cluster, obtaining a ratio between the number of abnormal nodes and the number of all nodes as the response priority of each node.

6. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 1, characterized in that, The external interference degree acquisition method comprises: for any key node grouping cluster, obtaining external interference possibility of each node according to variation difference of response data between each node and other nodes in the same batch; if the external interference possibility of the node is greater than a preset interference threshold, obtaining an intersection of abnormal responses between the node and all other nodes in the same batch; for each node in the intersection, taking the external interference possibility of the corresponding node as a local interference degree of each node; for each node outside the intersection, setting the local interference degree of each node to 0; obtaining a mean value of the local interference degrees of each node as an overall interference degree of each node in the key node grouping cluster.

7. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 6, characterized in that, The external interference possibility acquisition method comprises: for any key node grouping cluster, obtaining a correlation coefficient of an abnormal response set between each node and other nodes in the same batch as a local interference correlation; calculating a mean value of the local interference correlations between each node and all nodes in the same batch as the external interference possibility of each node in each key node grouping cluster.

8. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 1, characterized in that, The response priority correction method comprises the following steps: calculating the difference between the positive integer 1 and the overall interference degree of each node in each key node group cluster as an adjustment factor; weighting the response priority of each node in each key node group cluster according to the adjustment factor to obtain the response priority correction of each node in each key node group cluster.

9. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 8, characterized in that, The weighting of the response priority of each node in each key node group cluster according to the adjustment factor to obtain the response priority correction of each node in each key node group cluster comprises the following steps: obtaining the product of the response priority of each node in each key node group cluster and the adjustment factor as the response priority correction of each node in each key node group cluster.

10. The supply chain resilience path identification method based on fuzzy set qualitative comparison according to claim 7, wherein, The correlation coefficient is a Pearson correlation coefficient.

Images