Supply chain logistics planning method based on ant colony algorithm

Abstract

The invention belongs to the technical field of business data processing, and particularly discloses a supply chain logistics planning method based on an ant colony algorithm, which comprises the steps of extracting historical logistics data from a supply chain database, a time sequence analysis method is adopted to process the data to identify a periodic fluctuation mode, and serialized representation of abnormal features is obtained; according to the obtained serialized representation, abnormal features are grouped by adopting a clustering algorithm, similar congestion or delay events are classified into the same category, and a clustering center of a repetitive abnormal mode is determined; if the determined clustering center displays repetition of a specific time period, judging whether the modes reach a preset repetition threshold value or not by calculating the frequency and interval indexes of the clustering center, and obtaining a list of high-risk abnormal modes; the objective of the invention is to solve the problems of periodic congestion and distribution delay in a supply chain in the prior art.

Description

Technical Field

[0001] This invention relates to the field of business data processing technology, specifically to a supply chain logistics planning method based on ant colony algorithm. Background Technology

[0002] Supply chain logistics planning, as a crucial pillar of modern economic operation, directly impacts a company's operational efficiency and market competitiveness. Its research is invaluable for ensuring efficient goods flow, reducing costs, and improving service quality. In complex supply chain networks, logistics planning must address fluctuating market demands and unpredictable external disturbances. Therefore, optimizing resource allocation and enhancing system adaptability are critical issues that urgently need to be addressed. Currently, while many logistics planning methods have achieved path optimization and resource scheduling to some extent, they often struggle to handle recurring anomalies, especially when facing cyclical or seasonal logistics bottlenecks, lacking in-depth analysis of historical problems and the ability to continuously address them. This deficiency leads to the system needing to recalculate and adjust every time a similar problem arises, preventing the accumulation of effective experience and ultimately affecting overall efficiency.

[0003] Against this backdrop, the core technical challenges in supply chain logistics planning lie in how to persistently identify and effectively warn of recurring anomaly patterns. The primary problem is the insufficient long-term retention of anomaly characteristics. Systems often only record fault information briefly, such as congestion or delivery delays at a transportation node. However, this information is gradually lost over time, preventing the system from reacting quickly to similar problems in the future. A deeper challenge lies in the weak ability to identify recurring anomaly patterns. When certain problems repeatedly occur within specific time periods or combinations of nodes, the system struggles to transform these characteristics into long-term effective response strategies and cannot anticipate potential risks. For example, during specific holidays each year, a certain logistics route consistently experiences delays due to surges in demand, but the system cannot adjust resource allocation in advance based on past records to create an early warning mechanism.

[0004] Therefore, how to build a mechanism in supply chain logistics planning that can persistently preserve abnormal characteristics while enabling early detection and resource optimization for recurring problems has become a critical issue that urgently needs to be addressed. Solving this problem will directly affect the stability and responsiveness of the logistics system in complex environments. Summary of the Invention

[0005] This invention provides a supply chain logistics planning method based on ant colony algorithm, aiming to solve the problems of periodic congestion and delivery delays in the existing supply chain.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A supply chain logistics planning method based on ant colony optimization includes: extracting historical logistics data from a supply chain database, including records of transportation node congestion and delivery delay information; processing this data using time series analysis to identify periodic fluctuation patterns and obtain serialized representations of abnormal features; grouping the abnormal features using a clustering algorithm based on the obtained serialized representations, classifying similar congestion or delay events into the same category, and determining cluster centers for recurring abnormal patterns; if the determined cluster centers show recurrence within a specific time period, determining whether these patterns reach a preset recurrence threshold by calculating the frequency and interval indicators of the cluster centers, and obtaining a list of high-risk abnormal patterns; and then, based on the obtained list of high-risk abnormal patterns, from... Historical adjustment records of corresponding nodes are obtained from a pre-established resource allocation template. A time series forecasting model is used to simulate future demand surge scenarios to obtain the predicted distribution of potential logistics bottlenecks. Based on the obtained predicted distribution, a resource scheduling scheme for high-risk modes is generated, integrating the backup paths and backup capacity of the predicted bottleneck nodes into the scheme to determine the optimized path configuration. If the backup capacity in the optimized path configuration exceeds the currently available resources, the scheduling priority of adjacent nodes is adjusted to reallocate transportation tasks and obtain a balanced resource optimization plan. By injecting the obtained balanced resource optimization plan into the supply chain simulation environment, multiple rounds of iterative testing are run to verify its response speed under simulated abnormal scenarios, and the final early warning mechanism deployment parameters are obtained.

[0007] In one aspect of this disclosure, the method of extracting historical logistics data from a supply chain database, including records of transportation node congestion and delivery delay information, and processing this data using time series analysis to identify periodic fluctuation patterns and obtain a serialized representation of abnormal features includes: Based on the above business content and the extracted relevant attributes, the following business solutions are generated, and the technical process design revolves around attributes such as supply chain data, historical logistics, transportation nodes, congestion records, delivery delays, time series, analysis methods, periodic fluctuations, pattern recognition, anomaly features, sequence representation, and data processing. The following steps are based on interrelated attributes, discarding isolated content that cannot form a logical connection, using no more than two machine learning algorithms, and emphasizing the specific implementation of logical progression and technical processing; Historical logistics records are obtained from the supply chain database, including transportation node information, congestion records, and delivery delay data. Data cleaning tools are used to standardize the format of the original records and fill in missing values ​​to obtain a structured logistics dataset. For structured logistics datasets, time series analysis is used to decompose transportation node and delivery delay data, extract periodic fluctuation components, and determine the regular patterns of periodic changes. Based on the regular patterns of periodic changes, combined with congestion records and delivery delay data, the distribution of abnormal features is identified, and significant time points of anomalies are determined by comparing preset thresholds. If the significant time point of the anomaly exceeds the preset threshold range, the anomaly features are serialized and encoded to obtain the sequence representation of the anomaly features. By segmenting and analyzing the sequence representation of abnormal features and combining it with the periodic fluctuation patterns of historical logistics data, the correspondence between abnormal features and periodic fluctuations is determined. Based on the correspondence between abnormal characteristics and cyclical fluctuations, regression analysis is used to predict future possible congestion records and delivery delay trends, thereby obtaining early warning information on potential anomalies. By classifying and storing early warning information of potential anomalies, and combining it with the geographical location data of transportation nodes, a spatiotemporal distribution record of anomaly characteristics is generated, thus completing the closed loop of anomaly monitoring for supply chain data.

[0008] In one aspect of this disclosure, the step of grouping anomalous features using a clustering algorithm based on the obtained serialized representation, classifying similar congestion or delay events into the same category, and determining cluster centers for recurring anomalous patterns includes: Sequence representation data containing anomalous features are obtained from the supply chain database. This data is then preliminarily processed, and redundant information in the sequence representation is removed using data cleaning tools to obtain a standardized anomalous feature dataset. For the standardized abnormal feature dataset, a clustering algorithm is used to group the abnormal features, classify congestion events and delay events according to similarity groups, and determine the cluster center point of each group; Based on the cluster center point of each group, the distribution of repeating patterns is analyzed, and the concentration of abnormal features within each group is compared using a preset threshold to determine whether there are significant repeating patterns. If significant repetitive patterns exist, timestamps are added to the congestion and delay events corresponding to these repetitive patterns to obtain time-stamped abnormal grouping data. Based on the time-stamped abnormal group data, we analyze the changing trends of abnormal features in each group over time. We then use a time window segmentation tool to segment the changing trends and obtain the time distribution segments of the abnormal features. For the time distribution segments of abnormal features, combined with the transportation node information in the supply chain database, the correlation between abnormal groups and node locations is analyzed, and the spatial distribution characteristics of abnormal groups are determined by location mapping tools. Based on the spatial distribution characteristics of abnormal groups, congestion events and delay events within each group are classified and stored. Data visualization tools are used to generate spatiotemporal distribution records of abnormal groups, thus completing a comprehensive archive of abnormal patterns.

[0009] In one aspect of this disclosure, if the identified cluster centers show recurrence over a specific time period, then by calculating the frequency and interval indices of the cluster centers, it is determined whether these patterns reach a preset recurrence threshold, thereby obtaining a list of high-risk abnormal patterns, including: Obtain time period records related to cluster centers from supply chain anomaly data, conduct preliminary screening for repetitive phenomena within specific time periods, extract data fragments that match the repetitive phenomena using timestamp comparison tools, and determine a preliminary set of anomaly patterns. For the initial set of abnormal patterns, the frequency and time interval data corresponding to each cluster center are analyzed. Statistical tools are used to process the distribution of frequency and interval to obtain quantitative indicators of frequency and interval. Based on the quantitative indicators of frequency and interval, a comparison is made with a preset threshold. If the frequency and interval indicators of a certain cluster center exceed the preset threshold, it is marked as a potential high-risk pattern, and a list of potential high-risk patterns is determined. For a list of potentially high-risk patterns, combined with correlation data of specific times and recurring phenomena, time series analysis tools are used to segment the time distribution of the patterns to obtain characteristic segments of the time distribution; Based on the characteristic segments of the time distribution, the triggering conditions of high-risk patterns in different time periods are analyzed. If a certain characteristic segment has a high degree of consistency with the conditions of historical anomaly records, it is classified as a priority anomaly pattern and the set of priority anomaly patterns is determined. For the set of anomaly patterns that are of priority, the results of indicator analysis and comparison are combined, and the anomaly patterns are associated with the corresponding cluster center information and archived through data storage tools to obtain the final list of anomaly patterns.

[0010] In one aspect of this disclosure, the step of obtaining historical adjustment records of corresponding nodes from a pre-established resource allocation template for the obtained list of high-risk abnormal patterns, and simulating future demand surge scenarios using a time series prediction model to obtain the predicted distribution of potential logistics bottlenecks includes: Obtain the corresponding node data related to the high-risk mode from the preset resource configuration file, divide the historical adjustment data into time periods, and extract the historical fragments related to traffic surges through data filtering tools to obtain a preliminary traffic change dataset. Based on the preliminary traffic change dataset, a time series prediction model was used to simulate the scenario of a surge in future logistics traffic. The traffic change trend was broken down using a segmented analysis tool to determine the distribution of potential traffic peak periods. Based on the distribution of potential traffic peak periods, obtain the resource allocation records in the corresponding node data, analyze the matching degree between node resource allocation and traffic peaks through comparison tools, determine whether there are potential resource shortage risks, and derive a resource shortage risk list. For the resource shortage risk list, if the resource allocation record of a certain node shows that it is lower than the preset threshold during peak hours, the resource replenishment priority ranking is generated by the data mapping tool to obtain the priority sequence of resource allocation. Based on the priority sequence of resource allocation, available resource data is extracted from the preset backup resource pool. The node status after resource replenishment is simulated using allocation simulation tools to determine the degree of relief in the bottleneck area. Based on the distribution of the degree of relief in bottleneck areas, if the degree of relief in a certain area does not reach the preset standard, similar scenarios in historical adjustment data are analyzed through data backtracking tools to obtain alternative adjustment plans and derive the final resource optimization allocation plan. Based on the final resource optimization and allocation plan, the optimized resource allocation records are linked and archived with potential bottleneck area information through data storage tools to obtain a complete logistics bottleneck prediction and resource allocation archive.

[0011] In one aspect of this disclosure, generating a resource scheduling scheme for high-risk patterns based on the obtained predicted distribution, integrating the backup paths and backup capacities of the predicted bottleneck nodes into the scheme, and determining the optimized path configuration includes: Bottleneck data is extracted from the predicted distribution data. For bottleneck nodes in high-risk modes, relevant traffic concentration areas are obtained through data filtering tools to determine the distribution range of traffic concentration nodes. Based on the distribution range of traffic-concentrated nodes, a path analysis tool is used to organize alternative paths, obtain feasibility data of alternative paths, and determine whether alternative paths meet the traffic diversion conditions. If the feasibility data of the backup path meets the preset threshold, the backup capacity data is obtained through the capacity matching tool. Based on the matching of the backup capacity data with the traffic concentration node, a preliminary capacity allocation plan is obtained. Based on the initial capacity allocation plan, the backup paths and backup capacity are integrated using a path configuration tool. Traffic simulation is then performed on the integrated path data to determine the optimized path configuration results. Based on the optimized path configuration results, the load changes of the traffic concentration nodes after configuration are analyzed using data comparison tools to obtain the distribution data of load changes and determine whether the load is evenly distributed. If the load distribution data does not meet the balancing standard, the priority of the backup paths is reordered using the path adjustment tool, and the reordered path data is reconfigured to obtain the final configuration scheme. Based on the final configuration scheme, the path configuration results are associated with and stored with the bottleneck data of the high-risk mode through a data archiving tool. The execution basis of the scheduling scheme is generated based on the stored data, and the implementation basis of the scheduling scheme is determined.

[0012] In one aspect of this disclosure, if the spare capacity in the optimized path configuration exceeds the currently available resources, then by adjusting the scheduling priorities of adjacent nodes and reallocating transportation tasks to obtain a balanced resource optimization plan, the following steps are taken: For the standby capacity data in the path configuration, the matching between standby capacity and available resources is analyzed using a capacity comparison tool. If the standby capacity exceeds the available resources, the capacity difference data of the excess portion is obtained to determine the initial range of capacity adjustment. Based on the initial range of capacity adjustment, the scheduling order data of adjacent nodes are extracted using node analysis tools. The nodes are then sorted according to their correlation with each other to obtain a priority list of scheduling order. By using a priority list of scheduling order, transportation tasks are reassigned using a task allocation tool. Data comparison is performed on the matching between task allocation and resource allocation to determine the balance of task allocation. If the balance of task assignment does not reach the preset threshold, the scheduling order of adjacent nodes is rearranged a second time using the node adjustment tool to obtain the adjusted scheduling order data and determine a new task allocation scheme. Based on the new task allocation scheme, resource integration tools are used to verify resource allocation and capacity limitations. Based on the verified data, allocation details of the balancing scheme are generated to obtain intermediate results of resource optimization. By analyzing the intermediate results of resource optimization, data mapping tools are used to analyze the execution of transportation tasks and task assignments, record the data distribution of the execution status, and determine the final resource allocation plan. Based on the final resource allocation plan, the path configuration, scheduling order, and task assignment data are associated and archived using data storage tools. Execution basis is then generated from the archived data to obtain a complete business processing record.

[0013] In one aspect of this disclosure, the method involves injecting the balanced resource optimization plan into a supply chain simulation environment, running multiple rounds of iterative testing to verify its response speed under simulated abnormal scenarios, and obtaining the final deployment parameters of the early warning mechanism, including: By using the simulation system in the supply chain system, loading the data of the balancing plan, initializing the allocation details for resource optimization, and obtaining the initial state of the scenario simulation; Based on the initial state of the scenario simulation, a loop testing tool is used to run the resource optimization allocation scheme multiple times under abnormal scenarios to judge the fluctuation of response efficiency. If the response efficiency is lower than the preset threshold, the specific parameters of the abnormal scenario are recorded to obtain preliminary data for efficiency evaluation. Based on the preliminary data from the efficiency assessment, key points with low response efficiency in abnormal scenarios are extracted using data filtering tools, bottleneck locations in cyclic testing are analyzed, and key directions for testing and verification are determined. Based on the key areas of testing and verification, a simulation system is used to conduct in-depth analysis of the key points of abnormal scenarios, stress tests are performed on resource allocation at bottleneck locations, and detailed feedback on response efficiency is obtained. By providing detailed feedback on response efficiency and leveraging the logical framework of the early warning mechanism, parameters are adjusted to target weaknesses in stress testing, thus determining the optimization range of configuration data. Based on the optimization scope of the configuration data, the results of the cyclic test are matched with the triggering conditions of the early warning mechanism using data integration tools to determine the response strategy under abnormal scenarios and obtain the final deployment parameters.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This invention identifies periodic fluctuations and anomalies through time series analysis, and groups similar events using clustering algorithms to accurately pinpoint high-risk patterns. Subsequently, it uses predictive models to simulate future demand surges, generating targeted resource scheduling solutions and achieving resource balance by dynamically adjusting the priorities of adjacent nodes. Finally, the solution's response is validated in a simulated environment, and the optimized plan is injected into the supply chain simulation environment. Multiple rounds of iterative testing are run to obtain the final deployment parameters of the early warning mechanism, ensuring rapid response in abnormal scenarios. This invention, through the integration of data-driven approaches and intelligent prediction, significantly improves the supply chain's resilience and resource utilization efficiency, providing scientific decision support for logistics management. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a supply chain logistics planning method based on ant colony algorithm according to the present invention. Detailed Implementation

[0017] The present invention will be further described below with reference to embodiments. These embodiments are merely some, not all, of the embodiments described. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the protection scope of the present invention.

[0018] Please see Figure 1 As shown in the figure, this embodiment discloses a supply chain logistics planning method based on ant colony algorithm, which may specifically include: Step 101: Historical logistics data, including transportation node congestion records and delivery delay information, is extracted from the supply chain database. Time series analysis is then used to process this data to identify periodic fluctuation patterns and obtain a serialized representation of abnormal features.

[0019] Based on the aforementioned business content and extracted relevant attributes, the following business solution is generated, focusing on the technical process design around attributes such as supply chain data, historical logistics, transportation nodes, congestion records, delivery delays, time series, analysis methods, cyclical fluctuations, pattern recognition, anomaly features, sequence representation, and data processing. The following steps, based on interrelated attributes, discard isolated content that cannot form a logical connection, employ no more than two machine learning algorithms, emphasizing logical progression and the specific implementation of technical processing.

[0020] Step 1: Obtain historical logistics records from the supply chain database, covering transportation node information, congestion records, and delivery delay data. Use data cleaning tools to standardize the format of the original records and fill in missing values ​​to obtain a structured logistics dataset.

[0021] Step 2: For the structured logistics dataset, time series analysis is used to decompose the transportation node and delivery delay data, extract the periodic fluctuation components, and determine the regular patterns of periodic changes.

[0022] Step 3: Based on the regular patterns of periodic changes, combined with congestion records and delivery delay data, identify the distribution of abnormal features, and determine the significant time points when the abnormality occurred by comparing preset thresholds.

[0023] Step 4: If the significant time point of the anomaly exceeds the preset threshold range, the anomaly features are serialized and encoded to obtain the sequence representation of the anomaly features.

[0024] Step 5: By segmenting and analyzing the sequence representation of abnormal features and combining it with the periodic fluctuation patterns of historical logistics data, determine the correspondence between abnormal features and periodic fluctuations.

[0025] Step Six: Based on the correspondence between abnormal characteristics and periodic fluctuations, regression analysis is used to predict future possible congestion records and delivery delay trends, thereby obtaining early warning information for potential anomalies.

[0026] Step 7: By classifying and storing early warning information of potential anomalies, and combining it with the geographical location data of transportation nodes, a spatiotemporal distribution record of anomaly characteristics is generated, thus completing the closed loop of anomaly monitoring for supply chain data.

[0027] Specifically, historical logistics data is extracted from the supply chain database. Let's assume we obtained logistics records from the system for the past three years, including the number of monthly transportation node congestion events and delivery delay hours. For example, a node recorded 15 congestion events and a total delay of 120 hours in January 2022. Next, time series analysis is used to process this data. First, the congestion events and delay durations are cleaned to remove outliers. For instance, if the number of congestion events suddenly reaches 100 in a certain month, significantly deviating from the mean ± 2 standard deviations, it is replaced with the mean of adjacent months. Subsequently, a fast Fourier transform algorithm is used to perform frequency domain analysis on the data to identify periodic fluctuation patterns. It was found that the number of congestion events has a significant peak every 6 months, corresponding to peak logistics periods during holidays. Spectrum analysis shows a dominant frequency of 0.1667 (i.e., a 6-month cycle). Furthermore, the delay duration sequence was fitted using an autoregressive moving average model with parameters set to (p=1, d=1, q=1). The prediction results showed that the delay duration increased by approximately 30% in specific months, such as December, consistent with historical data. To obtain a serialized representation of anomalous features, outliers were extracted based on the standard deviation method. If the delay duration in a certain month exceeded the mean plus twice the standard deviation (e.g., if the mean is 50 hours and the standard deviation is 10 hours, then the threshold is 70 hours), it was marked as an anomaly, generating a serialized feature vector. For example, [0,0,1,0,1] indicates that the 3rd and 5th months out of 5 months are anomalies. The above process was completed through database SQL queries, FFT and ARIMA analysis implemented using Python scripts, and automated anomaly detection algorithms, forming a complete logical chain from data extraction to anomaly feature identification, ensuring the accuracy of anomaly early warning in supply chain optimization.

[0028] Step 102: Based on the obtained serialized representation, a clustering algorithm is used to group the abnormal features, classify similar congestion or delay events into the same category, and determine the cluster center of the recurring abnormal pattern.

[0029] Step 1: Obtain sequence representation data containing anomalous features from the supply chain database. Perform preliminary processing on this data, using data cleaning tools to remove redundant information from the sequence representations, resulting in a standardized anomalous feature dataset. Step 2: For the standardized anomalous feature dataset, use a clustering algorithm to group the anomalous features. Classify congestion events and delay events according to similarity groups and determine the cluster centroids for each group. Step 3: Based on the cluster centroids of each group, analyze the distribution of recurring patterns. Compare the concentration of anomalous features within each group using a preset threshold to determine if significant recurring patterns exist. Step 4: If significant recurring patterns exist, timestamp the congestion and delay events corresponding to these patterns to obtain time-stamped anomalous grouping data. Step 5: Based on the time-stamped anomalous grouping data, analyze the changing trends of anomalous features within each group over time. Use a time window segmentation tool to segment the changing trends, obtaining time distribution segments of the anomalous features. Step Six: Analyze the correlation between anomalous groups and node locations based on the temporal distribution segments of abnormal features, combined with transportation node information in the supply chain database. Determine the spatial distribution characteristics of the anomalous groups using location mapping tools. Step Seven: Based on the spatial distribution characteristics of the anomalous groups, classify and store congestion and delay events within each group. Generate spatiotemporal distribution records of the anomalous groups using data visualization tools, completing a comprehensive archive of anomalous patterns.

[0030] Specifically, in supply chain management, for the generated serialized representation of abnormal features, the system first uses an automated algorithm to group and analyze these features. Assuming we process logistics anomaly data from the past two years (24 months), the resulting anomaly feature vector is a sequence of length 24, where a value of 1 indicates anomaly and a value of 0 indicates normal. For example, a sequence might be [1,0,1,0,0,1,0,0,1,0,0,0,1,0,1,0,0,1,0,1,0]. The system uses the K-means clustering algorithm to group abnormal events based on feature similarity, setting the number of cluster centers K to 3, representing three possible anomaly patterns. The algorithm calculates the Euclidean distance between each anomaly point and the initial random center, assigning it to the nearest cluster. For example, if an anomaly point is 2.5 distance from the first center, 3.8 from the second, and 1.2 from the third, it will be assigned to the third cluster. Subsequently, the system iteratively updates the center position of each cluster until the change in center position is less than 0.01, taking approximately 5 iterations. The analysis revealed that the anomalies in the first cluster were concentrated at specific time points every four months, involving eight outliers with a center value of [1,0,0,1]. The second cluster contained six outliers, centered at [0,1,0,1], reflecting another periodicity. The third cluster contained ten outliers, centered at [1,0,1,0], potentially related to temporary issues on specific routes. The system further combined the geographical location data of logistics nodes, associating the clustering results with specific routes or warehouses. It was found that the anomalies in the first cluster frequently occurred on a certain inter-provincial route, involving insufficient loading and unloading capacity, forming a complete logical chain from feature grouping to business issues. The entire process was implemented through automated scripts, data was stored in a distributed database, and clustering computation was supported by a cloud computing platform, ensuring efficient processing of large-scale data.

[0031] Step 103: If the identified cluster centers show recurrence within a specific time period, then by calculating the frequency and interval indicators of the cluster centers, it is determined whether these patterns have reached the preset recurrence threshold, and a list of high-risk abnormal patterns is obtained.

[0032] Step 1: Obtain time period records related to the cluster center from the supply chain anomaly data, conduct preliminary screening for repetitive phenomena within a specific time period, extract data fragments that match the repetitive phenomena using timestamp comparison tools, and determine the preliminary set of anomaly patterns.

[0033] Step 2: For the initial set of abnormal patterns, analyze the frequency and time interval data corresponding to each cluster center, and use statistical tools to process the distribution of frequency and interval to obtain quantitative indicators of frequency and interval.

[0034] Step 3: Based on the quantitative indicators of frequency and interval, compare them with the preset thresholds. If the frequency and interval indicators of a certain cluster center exceed the preset thresholds, mark it as a potential high-risk pattern and determine the list of potential high-risk patterns.

[0035] Step 4: For the list of potential high-risk patterns, combine the correlation data of specific times and recurring phenomena, and use time series analysis tools to segment the time distribution of the patterns to obtain the characteristic segments of the time distribution.

[0036] Step 5: Based on the characteristic segments of the time distribution, analyze the triggering conditions of high-risk patterns in different time periods. If a certain characteristic segment has a high degree of consistency with the conditions of historical anomaly records, it is classified as a priority anomaly pattern and the set of priority anomaly patterns is determined.

[0037] Step Six: For the set of anomaly patterns that require priority attention, combine the results of indicator analysis and comparison, and use data storage tools to associate and archive the anomaly patterns with their corresponding cluster center information to obtain the final list of anomaly patterns.

[0038] Specifically, in the field of supply chain management, the system automatically conducts further frequency and interval analysis on identified anomalous pattern clusters to screen out high-risk recurring anomalous patterns. First, the system extracts the anomalous distribution of cluster centers from logistics data over the past 36 months. Assuming a cluster center exhibits an anomalous pattern occurring once every 6 months, specifically in months 6, 12, 18, 24, 30, and 36, totaling 6 anomalous records, the system uses a built-in frequency calculation algorithm to determine the pattern's occurrence frequency as 0.1667 (i.e., once every 6 months). Combined with an interval analysis algorithm, the standard deviation of the time interval between each anomalous event is calculated to be 0.5 months, indicating a highly regular pattern. Next, the system compares the frequency and interval indicators with preset thresholds. Assuming the threshold is a frequency greater than 0.1 and an interval standard deviation less than 1.0, the pattern is identified as a recurring high-risk pattern. Furthermore, through correlation analysis, the system matches this pattern with peak holiday data in logistics operations, finding that anomalous events frequently occur in the second month before specific annual holidays, possibly due to delivery pressure caused by a surge in orders. The system then automatically generates a list of high-risk anomaly patterns, marks the pattern as priority 1, and links it to specific distribution center data, forming a complete logical chain from pattern recognition to business risk assessment. The entire analysis process is driven by automated scripts, and data is stored in a cloud database, ensuring the real-time nature and accuracy of the analysis.

[0039] Step 104: For the obtained list of high-risk abnormal patterns, retrieve the historical adjustment records of the corresponding nodes from the pre-established resource allocation template, and use a time series prediction model to simulate future demand surge scenarios to obtain the predicted distribution of potential logistics bottlenecks.

[0040] Step 1: Obtain the corresponding node data related to the high-risk mode from the preset resource configuration file, divide the historical adjustment data into time periods, and extract historical fragments related to traffic surges through data filtering tools to obtain a preliminary traffic change dataset.

[0041] Step 2: Based on the preliminary traffic change dataset, a time series prediction model is used to simulate the future surge in logistics traffic. The traffic change trend is broken down using a segmented analysis tool to determine the distribution of potential traffic peak periods.

[0042] Step 3: Based on the distribution of potential traffic peak periods, obtain the resource allocation records in the corresponding node data, analyze the matching degree between node resource allocation and traffic peaks through comparison tools, determine whether there are potential resource shortage risks, and derive a resource shortage risk list.

[0043] Step 4: For the resource shortage risk list, if the resource allocation record of a certain node shows that it is lower than the preset threshold during the peak period, the resource replenishment priority ranking is generated by the data mapping tool to obtain the priority sequence of resource allocation.

[0044] Step 5: Based on the priority sequence of resource allocation, extract the available resource data from the preset backup resource pool, and use the allocation simulation tool to simulate the node status after resource replenishment to determine the degree of relief in the bottleneck area.

[0045] Step Six: Based on the distribution of the degree of relief in the bottleneck areas, if the degree of relief in a certain area does not reach the preset standard, similar scenarios in historical adjustment data are analyzed through data backtracking tools to obtain alternative adjustment plans and derive the final resource optimization allocation plan.

[0046] Step 7: Based on the final resource optimization and allocation plan, use data storage tools to link and archive the optimized resource allocation records with potential bottleneck area information to obtain a complete logistics bottleneck prediction and resource allocation archive.

[0047] Specifically, in the field of supply chain management, the system automatically extracts historical adjustment records of corresponding nodes from a pre-established resource allocation template database for a list of identified high-risk abnormal patterns, and combines this with a time series forecasting model to simulate future demand surge scenarios, ultimately generating a predicted distribution of potential logistics bottlenecks.

[0048] For example, the system first retrieves resource adjustment data for a delivery node from the database over the past 48 months. It finds that the node's average order volume growth rate at the end of a specific quarter is 15.3%, and historical adjustment records show that warehouse capacity utilization exceeds 90% during each order surge, leading to delivery delays. Subsequently, the system uses the ARIMA time series forecasting model, based on order volume, inventory turnover rate, and transportation timeliness data from the past 48 months, setting model parameters (p=2, d=1, q=1), to predict potential demand peaks within the next 12 months. The system concludes that order volume may increase by 18.2% and 20.5% in the 3rd and 9th months, respectively. Next, combining the forecast results with historical adjustment records, the system analyzes the potential inadequacies of warehouse capacity and transportation vehicle scheduling under demand surge scenarios, calculating that warehouse capacity may exceed maximum capacity by 12.7% in the 9th month, while the estimated shortage of transportation vehicles is 8 vehicles. Through further bottleneck analysis algorithms, the system identified that the inventory turnover efficiency of this delivery node dropped to 75% of the average level during a specific period. This result was correlated with upstream supplier replenishment cycle data, revealing that the average supplier replenishment cycle was 14 days, potentially hindering timely responses to demand fluctuations. Ultimately, the system generated a predictive distribution report of potential logistics bottlenecks, marking high-risk time points and resource gap data, forming a complete analytical chain from historical data extraction to future risk prediction. The entire process was automated, with data stored on a distributed server.

[0049] Step 105: Based on the obtained predicted distribution, generate a resource scheduling scheme for high-risk modes, integrate the backup paths and backup capacity of the predicted bottleneck nodes into the scheme, and determine the optimized path configuration.

[0050] Step 1: Extract bottleneck data from the predicted distribution data. For bottleneck nodes in high-risk modes, use data filtering tools to obtain relevant traffic concentration areas and determine the distribution range of traffic concentration nodes.

[0051] Step 2: Based on the distribution range of the traffic concentration nodes, use path analysis tools to organize the alternative paths, obtain the feasibility data of the alternative paths, and determine whether the alternative paths meet the traffic diversion conditions.

[0052] Step 3: If the feasibility data of the backup path meets the preset threshold, the backup capacity data is obtained through the capacity matching tool. Based on the matching of the backup capacity data with the traffic concentration node, a preliminary capacity allocation plan is obtained.

[0053] Step 4: Based on the preliminary capacity allocation plan, integrate the backup paths and backup capacity using the path configuration tool, perform traffic simulation on the integrated path data, and determine the optimized path configuration results.

[0054] Step 5: Based on the optimized path configuration results, analyze the load changes of the traffic concentration nodes after configuration using data comparison tools, obtain the load distribution data, and determine whether the load is evenly distributed.

[0055] Step 6: If the load distribution data does not meet the balancing standard, the priority of the backup paths is reordered using the path adjustment tool, and the reordered path data is reconfigured to obtain the final configuration scheme.

[0056] Step 7: Based on the final configuration scheme, use a data archiving tool to associate and store the path configuration results with the bottleneck data of the high-risk mode. Generate the execution basis of the scheduling scheme based on the stored data, and determine the implementation basis of the scheduling scheme.

[0057] Specifically, in the field of supply chain management, the system automatically generates resource scheduling schemes for high-risk patterns based on the predicted distribution of potential bottlenecks, and integrates alternative routes and backup capacity into the optimized configuration to achieve fully automated processing. First, the system extracts high-risk data for a key transit node from the predicted distribution database for the next six months, discovering that transportation demand in the second and fifth months may surge by 17.5% and 19.8%, respectively. Then, the system calls a route optimization algorithm, employing Dijkstra's algorithm based on graph theory, combined with real-time traffic data and historical transportation timeliness records, to calculate three alternative routes from this node to the downstream distribution center. These routes take 5%, 8%, and 12% longer than the main route, respectively, but can alleviate 40% of the traffic pressure on the main route. Next, the system analyzes the backup capacity configuration, extracting available storage data from warehouses surrounding the node, finding that the available backup capacity is 22.6% of the total demand. Using a linear programming algorithm, the system optimizes the backup capacity allocation ratio to 15% for emergency inventory and 7.6% for buffer inventory, ensuring that warehouse utilization does not exceed 85% during peak periods. Simultaneously, the system integrates the route optimization results with backup capacity data to generate a comprehensive scheduling plan. It calculates that activating the second backup route can reduce the main route congestion rate by approximately 18.3%. Furthermore, by correlating the real-time order processing speed of downstream delivery nodes, it finds that their average processing efficiency is 1200 orders per hour, sufficient to handle the diverted traffic. Finally, the system stores the optimized route configuration and capacity allocation plan on the cloud scheduling platform, automatically connecting with the transportation management system to ensure resource pre-allocation is completed before high-risk time points, forming a closed-loop logic from prediction to execution.

[0058] Step 106: If the spare capacity in the optimized path configuration exceeds the currently available resources, the transportation tasks are reallocated by adjusting the scheduling priority of adjacent nodes to obtain a balanced resource optimization plan.

[0059] Step 1: For the spare capacity data in the path configuration, analyze the matching between spare capacity and available resources using a capacity comparison tool. If the spare capacity exceeds the available resources, obtain the capacity difference data of the excess portion to determine the initial range of capacity adjustment.

[0060] Step 2: Based on the initial range of capacity adjustment, use node analysis tools to extract the scheduling order data of adjacent nodes, sort the adjacent nodes according to the correlation between them, and obtain a priority list of scheduling order.

[0061] Step 3: Using the priority list of the scheduling order, the transportation tasks are reassigned using a task allocation tool. Data comparison is performed on the matching between task assignment and resource allocation to determine the balance of task assignment.

[0062] Step 4: If the balance of task assignment does not reach the preset threshold, the scheduling order of adjacent nodes is rearranged a second time using the node adjustment tool to obtain the adjusted scheduling order data and determine a new task allocation scheme.

[0063] Step 5: Based on the new task allocation scheme, use resource integration tools to verify the resource allocation and capacity limits. Generate allocation details of the balancing scheme based on the verified data to obtain intermediate results of resource optimization.

[0064] Step Six: Using the intermediate results of resource optimization, analyze the execution of transportation tasks and task assignments using data mapping tools, record the data distribution of the execution status, and determine the final resource allocation plan.

[0065] Step 7: Based on the final resource allocation plan, use data storage tools to associate and archive the path configuration, scheduling order, and task assignment data, generate execution basis for the archived data, and obtain a complete business processing record.

[0066] Specifically, in the field of supply chain management, when the spare capacity in the optimized route configuration exceeds the currently available resources, the system automatically activates a resource balancing mechanism. This mechanism uses intelligent algorithms to reallocate transportation tasks to optimize resources. First, the system detects that the spare capacity requirement of a core node is 28.5% of the total resources, while the currently available resources are only 21.3%, resulting in a 7.2% gap. Next, the system extracts the available transportation capacity data of adjacent nodes from the resource scheduling database, finding that two auxiliary nodes within 50 kilometers of this node have 15.7% and 12.4% of their capacity idle, respectively. The system then invokes a priority adjustment model based on a greedy algorithm to dynamically increase the scheduling priority of these two nodes, calculating that the transferable transportation task volume is 5.8% of the total gap, with the first node undertaking 3.5% and the second node undertaking 2.3%. Simultaneously, the system analyzes historical transportation data to assess the impact of the adjusted task allocation on the overall logistics network. It finds that this reallocation reduces resource pressure on core nodes to an acceptable level of 23.9%, while the capacity utilization rates of adjacent nodes increase to 78.6% and 81.2% respectively, remaining within their maximum capacity threshold of 85%. To ensure stability during task transfer, the system automatically generates a temporary buffer mechanism, utilizing backup inventory data from nearby small transit stations and reserving 3.1% of total demand for temporary storage to handle potential short-term fluctuations. Finally, the system updates the adjusted resource optimization plan to the central dispatch system and automatically synchronizes it to the task execution modules of each node, ensuring seamless connection of transportation tasks and forming a complete closed loop from resource detection to task reallocation.

[0067] Step 107: By injecting the balanced resource optimization plan into the supply chain simulation environment, multiple rounds of iterative testing are run to verify its response speed under simulated abnormal scenarios, and the final early warning mechanism deployment parameters are obtained.

[0068] Step 1: Load the balancing plan data through the simulation system in the supply chain system, initialize the allocation details for resource optimization, and obtain the initial state of the scenario simulation.

[0069] Step 2: Based on the initial state of the scenario simulation, use a loop testing tool to run the resource optimization allocation scheme multiple times under abnormal scenarios to judge the fluctuation of response efficiency. If the response efficiency is lower than the preset threshold, record the specific parameters of the abnormal scenario to obtain preliminary data for efficiency evaluation.

[0070] Step 3: Based on the preliminary data from the efficiency assessment, use data filtering tools to extract key points with low response efficiency in abnormal scenarios, analyze the bottleneck locations in the loop test, and determine the key directions for test verification.

[0071] Step 4: Based on the key areas of testing and verification, use a simulation system to conduct in-depth analysis of the key points of abnormal scenarios, perform stress tests on resource allocation at bottleneck locations, and obtain detailed feedback on response efficiency.

[0072] Step 5: Through detailed feedback on response efficiency and with the help of the logical framework of the early warning mechanism, adjust parameters to target weak points in the stress test and determine the optimization range of the configuration data.

[0073] Step Six: Based on the optimization range of the configuration data, use data integration tools to match the results of the cyclic test with the triggering conditions of the early warning mechanism, determine the response strategy for abnormal scenarios, and obtain the final deployment parameters.

[0074] Specifically, in the field of supply chain management, after the system injects the optimized resource balancing plan into the simulation environment, it automatically initiates multiple rounds of iterative testing to verify its responsiveness under abnormal scenarios and ultimately generates deployment parameters for the early warning mechanism. First, the system imports the resource optimization plan into the supply chain simulation platform and sets abnormal scenarios in the simulation environment, such as a sudden 18.2% increase in transportation demand at a key node. By running a stochastic simulation model based on the Monte Carlo algorithm, 1000 scenarios with varying degrees of fluctuation are generated, and the system's response time is recorded in each simulation. The average response time is found to be 3.7 seconds, with a peak latency of 5.1 seconds. Subsequently, the system performs in-depth analysis of the simulation data, using time series analysis algorithms to extract key influencing factors on response speed. It finds a positive correlation between the fluctuation amplitude of transportation demand and response time; for every 5% increase in fluctuation, the response time increases by approximately 0.8 seconds. To optimize this metric, the system automatically adjusted the resource pre-allocation ratio in the simulated environment, increasing the reserve resource ratio of critical nodes from the initial 10.5% to 13.8%. After running 500 iterations of testing, the average response time was reduced to 2.9 seconds, and the peak latency decreased to 4.3 seconds, validating the effectiveness of the adjustment. Next, the system combined the simulation results of abnormal scenarios and invoked a decision tree-based classification algorithm to model the correlation between response time and anomaly type, generating early warning threshold parameters. For example, when the response time exceeds 3.5 seconds, a level-two early warning is automatically triggered, notifying the system to adjust the resource scheduling strategy. Finally, the system aggregated all test data to generate the final early warning mechanism deployment parameter table, including a response time threshold of 3.5 seconds, a resource pre-allocation ratio of 13.8%, and an anomaly fluctuation tolerance of 6.4%. These parameters were automatically embedded into the supply chain monitoring system to ensure rapid response to potential risks in a real-world environment, forming a complete logical closed loop from simulation testing to parameter deployment.

[0075] 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 alterations 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 supply chain logistics planning method based on ant colony algorithm, characterized in that, include: By extracting historical logistics data from the supply chain database, including records of congestion at transportation nodes and information on delivery delays, time series analysis is used to process this data to identify periodic fluctuation patterns and obtain a serialized representation of abnormal features. Based on the obtained serialized representation, a clustering algorithm is used to group the abnormal features, classify similar congestion or delay events into the same category, and determine the cluster centers of recurring abnormal patterns. If the identified cluster centers show recurrence within a specific time period, then by calculating the frequency and interval indicators of the cluster centers, it can be determined whether these patterns reach the preset recurrence threshold, and a list of high-risk abnormal patterns can be obtained. For the obtained list of high-risk abnormal patterns, the historical adjustment records of the corresponding nodes are obtained from the pre-established resource allocation template. A time series prediction model is used to simulate future demand surge scenarios and obtain the predicted distribution of potential logistics bottlenecks. Based on the obtained predicted distribution, a resource scheduling scheme for high-risk patterns is generated, and the backup paths and backup capacity of the predicted bottleneck nodes are integrated into the scheme to determine the optimized path configuration. If the spare capacity in the optimized path configuration exceeds the currently available resources, the transportation tasks are reallocated by adjusting the scheduling priority of adjacent nodes to obtain a balanced resource optimization plan. By injecting the balanced resource optimization plan into the supply chain simulation environment, multiple rounds of iterative testing are run to verify its response speed under simulated abnormal scenarios, and the final deployment parameters of the early warning mechanism are obtained.

2. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: The process involves extracting historical logistics data from a supply chain database, including records of transportation node congestion and delivery delay information. Time series analysis is then used to process this data to identify periodic fluctuation patterns and obtain a serialized representation of anomalous features, including: Based on the above business content and the extracted relevant attributes, the following business solutions are generated, and the technical process design revolves around attributes such as supply chain data, historical logistics, transportation nodes, congestion records, delivery delays, time series, analysis methods, periodic fluctuations, pattern recognition, anomaly features, sequence representation, and data processing. The following steps are based on interrelated attributes, discarding isolated content that cannot form a logical connection, using no more than two machine learning algorithms, and emphasizing the specific implementation of logical progression and technical processing; Historical logistics records are obtained from the supply chain database, including transportation node information, congestion records, and delivery delay data. Data cleaning tools are used to standardize the format of the original records and fill in missing values ​​to obtain a structured logistics dataset. For structured logistics datasets, time series analysis is used to decompose transportation node and delivery delay data, extract periodic fluctuation components, and determine the regular patterns of periodic changes. Based on the regular patterns of periodic changes, combined with congestion records and delivery delay data, the distribution of abnormal features is identified, and significant time points of anomalies are determined by comparing preset thresholds. If the significant time point of the anomaly exceeds the preset threshold range, the anomaly features are serialized and encoded to obtain the sequence representation of the anomaly features. By segmenting and analyzing the sequence representation of abnormal features and combining it with the periodic fluctuation patterns of historical logistics data, the correspondence between abnormal features and periodic fluctuations is determined. Based on the correspondence between abnormal characteristics and cyclical fluctuations, regression analysis is used to predict future possible congestion records and delivery delay trends, thereby obtaining early warning information on potential anomalies. By classifying and storing early warning information of potential anomalies, and combining it with the geographical location data of transportation nodes, a spatiotemporal distribution record of anomaly characteristics is generated, thus completing the closed loop of anomaly monitoring for supply chain data.

3. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: The step of grouping anomalous features based on the obtained serialized representation using a clustering algorithm, classifying similar congestion or delay events into the same category, and determining the cluster centers of recurring anomalous patterns includes: Sequence representation data containing anomalous features are obtained from the supply chain database. This data is then preliminarily processed, and redundant information in the sequence representation is removed using data cleaning tools to obtain a standardized anomalous feature dataset. For the standardized abnormal feature dataset, a clustering algorithm is used to group the abnormal features, classify congestion events and delay events according to similarity groups, and determine the cluster center point of each group; Based on the cluster center point of each group, the distribution of repeating patterns is analyzed, and the concentration of abnormal features within each group is compared using a preset threshold to determine whether there are significant repeating patterns. If significant repetitive patterns exist, timestamps are added to the congestion and delay events corresponding to these repetitive patterns to obtain time-stamped abnormal grouping data. Based on the time-stamped abnormal group data, we analyze the changing trends of abnormal features in each group over time. We then use a time window segmentation tool to segment the changing trends and obtain the time distribution segments of the abnormal features. For the time distribution segments of abnormal features, combined with the transportation node information in the supply chain database, the correlation between abnormal groups and node locations is analyzed, and the spatial distribution characteristics of abnormal groups are determined by location mapping tools. Based on the spatial distribution characteristics of abnormal groups, congestion events and delay events within each group are classified and stored. Data visualization tools are used to generate spatiotemporal distribution records of abnormal groups, thus completing a comprehensive archive of abnormal patterns.

4. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: If the identified cluster centers show recurrence within a specific time period, then by calculating the frequency and interval indices of the cluster centers, it is determined whether these patterns reach a preset recurrence threshold, thereby obtaining a list of high-risk abnormal patterns, including: Obtain time period records related to cluster centers from supply chain anomaly data, conduct preliminary screening for repetitive phenomena within specific time periods, extract data fragments that match the repetitive phenomena using timestamp comparison tools, and determine a preliminary set of anomaly patterns. For the initial set of abnormal patterns, the frequency and time interval data corresponding to each cluster center are analyzed, and statistical tools are used to process the distribution of frequency and interval to obtain quantitative indicators of frequency and interval. Based on the quantitative indicators of frequency and interval, a comparison is made with a preset threshold. If the frequency and interval indicators of a certain cluster center exceed the preset threshold, it is marked as a potential high-risk pattern, and a list of potential high-risk patterns is determined. For a list of potentially high-risk patterns, combined with correlation data of specific times and recurring phenomena, time series analysis tools are used to segment the time distribution of the patterns to obtain characteristic segments of the time distribution; Based on the characteristic segments of the time distribution, the triggering conditions of high-risk patterns in different time periods are analyzed. If a certain characteristic segment has a high degree of consistency with the conditions of historical anomaly records, it is classified as a priority anomaly pattern and the set of priority anomaly patterns is determined. For the set of anomaly patterns that are of priority, the results of indicator analysis and comparison are combined, and the anomaly patterns are associated with the corresponding cluster center information and archived through data storage tools to obtain the final list of anomaly patterns.

5. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: For the obtained list of high-risk abnormal patterns, historical adjustment records of corresponding nodes are retrieved from a pre-established resource allocation template. A time series prediction model is used to simulate future demand surge scenarios to obtain the predicted distribution of potential logistics bottlenecks, including: Obtain the corresponding node data related to the high-risk mode from the preset resource configuration file, divide the historical adjustment data into time periods, and extract the historical fragments related to traffic surges through data filtering tools to obtain a preliminary traffic change dataset. For the initial traffic change dataset, a time series prediction model is used to simulate the future surge in logistics traffic. The traffic change trend is broken down using a segmented analysis tool to determine the distribution of potential traffic peak periods. Based on the distribution of potential traffic peak periods, obtain the resource allocation records in the corresponding node data, analyze the matching degree between node resource allocation and traffic peaks through comparison tools, determine whether there are potential resource shortage risks, and derive a resource shortage risk list. For the resource shortage risk list, if the resource allocation record of a certain node shows that it is lower than the preset threshold during peak hours, the resource replenishment priority ranking is generated by the data mapping tool to obtain the priority sequence of resource allocation. Based on the priority sequence of resource allocation, available resource data is extracted from the preset backup resource pool. The node status after resource replenishment is simulated using allocation simulation tools to determine the degree of relief in the bottleneck area. Based on the distribution of the degree of relief in bottleneck areas, if the degree of relief in a certain area does not reach the preset standard, similar scenarios in historical adjustment data are analyzed through data backtracking tools to obtain alternative adjustment plans and derive the final resource optimization allocation plan. Based on the final resource optimization and allocation plan, the optimized resource allocation records are linked and archived with potential bottleneck area information through data storage tools to obtain a complete logistics bottleneck prediction and resource allocation archive.

6. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: Based on the obtained predicted distribution, a resource scheduling scheme for high-risk patterns is generated, integrating the backup paths and backup capacity of the predicted bottleneck nodes into the scheme, and determining the optimized path configuration, including: Bottleneck data is extracted from the predicted distribution data. For bottleneck nodes in high-risk modes, relevant traffic concentration areas are obtained through data filtering tools to determine the distribution range of traffic concentration nodes. Based on the distribution range of traffic-concentrated nodes, a path analysis tool is used to organize alternative paths, obtain feasibility data of alternative paths, and determine whether alternative paths meet the traffic diversion conditions. If the feasibility data of the backup path meets the preset threshold, the backup capacity data is obtained through the capacity matching tool. Based on the matching of the backup capacity data with the traffic concentration node, a preliminary capacity allocation plan is obtained. Based on the initial capacity allocation plan, the backup paths and backup capacity are integrated using a path configuration tool. Traffic simulation is then performed on the integrated path data to determine the optimized path configuration results. Based on the optimized path configuration results, the load changes of the traffic concentration nodes after configuration are analyzed using data comparison tools to obtain the distribution data of load changes and determine whether the load is evenly distributed. If the load distribution data does not meet the balancing standard, the priority of the backup paths is reordered using the path adjustment tool, and the reordered path data is reconfigured to obtain the final configuration scheme. Based on the final configuration scheme, the path configuration results are associated with and stored with the bottleneck data of the high-risk mode through a data archiving tool. The execution basis of the scheduling scheme is generated based on the stored data, and the implementation basis of the scheduling scheme is determined.

7. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: If the spare capacity in the optimized path configuration exceeds the currently available resources, then by adjusting the scheduling priorities of adjacent nodes, transportation tasks are reallocated to obtain a balanced resource optimization plan, including: For the standby capacity data in the path configuration, the matching between standby capacity and available resources is analyzed using a capacity comparison tool. If the standby capacity exceeds the available resources, the capacity difference data of the excess portion is obtained to determine the initial range of capacity adjustment. Based on the initial range of capacity adjustment, the scheduling order data of adjacent nodes are extracted using node analysis tools. The nodes are then sorted according to their correlation with each other to obtain a priority list of scheduling order. By using a priority list of scheduling order, transportation tasks are reassigned using a task allocation tool. Data comparison is performed on the matching between task allocation and resource allocation to determine the balance of task allocation. If the balance of task assignment does not reach the preset threshold, the scheduling order of adjacent nodes is rearranged a second time using the node adjustment tool to obtain the adjusted scheduling order data and determine a new task allocation scheme. Based on the new task allocation scheme, resource integration tools are used to verify resource allocation and capacity limitations. Based on the verified data, allocation details of the balancing scheme are generated to obtain intermediate results of resource optimization. By analyzing the intermediate results of resource optimization, data mapping tools are used to analyze the execution of transportation tasks and task assignments, record the data distribution of the execution status, and determine the final resource allocation plan. Based on the final resource allocation plan, the path configuration, scheduling order, and task assignment data are associated and archived using data storage tools. Execution basis is then generated from the archived data to obtain a complete business processing record.

8. The supply chain logistics planning method based on ant colony algorithm according to claim 1, characterized in that: The process involves injecting the balanced resource optimization plan into a supply chain simulation environment and running multiple rounds of iterative testing to verify its response speed under simulated abnormal scenarios, thereby obtaining the final deployment parameters for the early warning mechanism, including: By using the simulation system in the supply chain system, loading the data of the balancing plan, initializing the allocation details for resource optimization, and obtaining the initial state of the scenario simulation; Based on the initial state of the scenario simulation, a loop testing tool is used to run the resource optimization allocation scheme multiple times under abnormal scenarios to judge the fluctuation of response efficiency. If the response efficiency is lower than the preset threshold, the specific parameters of the abnormal scenario are recorded to obtain preliminary data for efficiency evaluation. Based on the preliminary data from the efficiency assessment, key points with low response efficiency in abnormal scenarios are extracted using data filtering tools, bottleneck locations in cyclic testing are analyzed, and key directions for testing and verification are determined. Based on the key areas of testing and verification, a simulation system is used to conduct in-depth analysis of the key points of abnormal scenarios, stress tests are performed on resource allocation at bottleneck locations, and detailed feedback on response efficiency is obtained. By providing detailed feedback on response efficiency and leveraging the logical framework of the early warning mechanism, parameters are adjusted to target weaknesses in stress testing, thus determining the optimization range of configuration data. Based on the optimization scope of the configuration data, the results of the cyclic test are matched with the triggering conditions of the early warning mechanism using data integration tools to determine the response strategy under abnormal scenarios and obtain the final deployment parameters.

Images