A logistics supply chain security management system and method

Abstract

This invention relates to the field of aviation logistics management technology, specifically a logistics supply chain security management system and method. The system includes an aviation network vulnerability analysis module, which calculates the number of times each shortest path passes through nodes and edges based on the node and flight route edge information of global cargo airports, and counts the number of connected routes to establish an aviation logistics network structural vulnerability map. In this invention, by quantifying the number of path passes and the number of connected routes for the node and flight route edge information of global cargo airports, a network structural vulnerability map can be formed, enabling intuitive identification and risk focus of key nodes and high-frequency segments in the air transport network. Combined with the precise recording of residual vectors and residual decay half-lives of logistics assets, the current pollution potential value can be dynamically calculated, establishing a pollution risk baseline covering ULDs, racking positions, and transport vehicles, making the asset risk status quantifiable and traceable.

Description

Technical Field

[0001] This invention relates to the field of aviation logistics management technology, and in particular to a logistics supply chain security management system and method. Background Technology

[0002] The field of air logistics management technology involves the management of the entire process of achieving efficient, safe, and timely flow of goods globally using air transport.

[0003] Current technologies in air logistics management typically focus on timeliness and cost planning for routes and nodes, making it difficult to identify potential risk clusters in a timely manner, leading to delayed responses to emergencies or node anomalies. Regarding asset risk management, the lack of real-time dynamic monitoring based on residue characteristics and decay patterns easily creates hidden contamination risks in the transportation of highly sensitive goods. In the route selection stage, ranking is often based solely on time or cost, without fully considering the combined impact of cargo characteristics and route risks, potentially selecting routes that appear efficient but have high potential risks. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a logistics supply chain security management system and method.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a logistics supply chain security management system comprising:

[0006] The aviation network vulnerability analysis module calculates the number of times each shortest path passes through nodes and edges based on the node and edge information of global cargo airports and flight routes, and counts the number of connected routes to build a vulnerability map of the aviation logistics network structure.

[0007] The logistics asset risk assessment module, based on each logistics asset in the vulnerability map of the aviation logistics network structure, establishes a record containing residue vectors and residue decay half-lives for each ULD, rack position, and transport vehicle, obtains a digital archive of logistics assets, calculates the current pollution potential value based on the residue vectors and residue decay half-lives in the digital archive of logistics assets, and generates a pollution risk baseline for logistics assets.

[0008] The route safety cost calculation module, based on air waybill information and cargo sensitivity matrix, calls the air logistics network structure vulnerability map and the logistics asset pollution risk baseline to calculate the route risk score, and then calculates the route risk score with transportation time and monetary cost to generate the comprehensive air logistics route safety cost.

[0009] The route optimization and execution module performs route search based on the comprehensive safety cost of different alternative routes for air logistics routes and selects the route with the lowest comprehensive safety cost for air logistics routes to obtain a preferred resilient logistics route scheme. Based on the preferred resilient logistics route scheme, it integrates to obtain logistics route execution instructions.

[0010] Preferably, the steps for obtaining the vulnerability map of the air logistics network structure are as follows:

[0011] Based on global cargo airport node information and flight route edge information, the airport IATA three-letter code and airport geographical coordinates are parsed. The duplicate edges are removed by route number and take-off and landing airports while retaining directionality. Node index and adjacency table are generated and isolated nodes and duplicate edges are verified to obtain the air logistics network topology.

[0012] Based on the topology of the aviation logistics network, a shortest path search is performed for each pair of origin and destination, and the number of nodes and edges traversed along the path is accumulated one by one. At the same time, the number of incoming edges and outgoing edges of each node are counted and summed to obtain the number of connected routes, forming a shortest path traversal statistics table and a node connected route number statistics table.

[0013] Based on the shortest path traversal statistics table and the node connection route number statistics table, the nodes and edges are counted, sorted, and divided into quantile intervals, and the count level and the number of connected routes are labeled. These are then overlaid onto the geographic base map, and the visualization style and layer order of the nodes and edges are configured to generate an air logistics network structure vulnerability map.

[0014] Preferably, the steps for obtaining the digital archives of the logistics assets are as follows:

[0015] Based on the vulnerability map of the aviation logistics network structure, the unique identifiers and geographical locations of each ULD, rack position, and transport vehicle are located one by one, and the timestamp of the most recent contact with the cargo is read. The residual sampling entries reported by each asset are analyzed, and the naming conventions and units of measurement of the residual entries are standardized. The values ​​of the same type of residuals within the same asset are arranged in chronological order and the corresponding residual decay half-life metadata is attached to form an asset-level entry set for each ULD, rack position, and transport vehicle, resulting in a record containing residual vectors and residual decay half-life.

[0016] Based on the records containing residual vectors and residual decay half-lives, all entries for the same asset are merged according to the unique identifiers of ULD, shelf location, and transport vehicle, and duplicate sampling entries are removed. The asset category label and historical operation trajectory summary of each asset are completed, and the field integrity and timestamp continuity are verified. The set of verified asset-level entries is written into the archive entries with a unified structure and a searchable index is generated to obtain the digital archive of logistics assets.

[0017] Preferably, the steps for obtaining the pollution risk baseline of the logistics assets are as follows:

[0018] Based on the digital archives of logistics assets, the residual vector and residual decay half-life of each asset are read, and the residual entries are converted by time decay according to the timestamp in the archive and aggregated and weighted with similar residuals. The asset category label and historical operation trajectory summary are extracted as risk correction factors, and the aggregated residual entries are labeled with threshold segments. The current pollution potential value is output in units of assets and summarized into a hierarchical list according to asset category to generate the pollution risk baseline of logistics assets.

[0019] Preferably, the steps for obtaining the path risk score are as follows:

[0020] Based on the sensitivity matrix of air waybill information and cargo, the cargo name code, processing code and temperature control level in the air waybill are read one by one. The corresponding row and column cells of the sensitivity matrix are matched in turn and the sensitivity values ​​are extracted. The extracted sensitivity values ​​are uniformly converted into the same unit of measurement and then concatenated into a sensitivity measurement vector in the order of the fields. The average calculation is performed on all components of the sensitivity measurement vector to obtain the cargo sensitivity measurement result.

[0021] Based on the cargo sensitivity quantification results, the planned flight segment sequence in the air waybill is analyzed, and the node count and edge count corresponding to each flight segment are retrieved in the air logistics network structure vulnerability map. The node count and edge count are normalized to the flight segment structure vulnerability value. At the same time, the ULD, rack space and transport vehicle associated with the flight segment are matched in the logistics asset pollution risk baseline, and the corresponding asset pollution potential value is extracted. The asset pollution potential value and structural vulnerability value of each flight segment are combined with the planned transport time and transit stay time recorded in the air waybill to form the flight segment planned transport stay time, resulting in a list of path exposure elements including flight segment asset pollution potential value, flight segment structural vulnerability value and flight segment planned transport stay time.

[0022] Calculate the path risk score based on the list of path exposure elements.

[0023] Preferably, the steps for obtaining the comprehensive safety cost of the air logistics route are as follows:

[0024] Based on the route candidate set, the flight segment list is extracted sequentially, the departure time and arrival time of each flight segment are extracted and the time difference is calculated. The transit customs waiting time is added to obtain the flight segment transportation time. At the same time, the take-off and landing fees, fuel surcharges and ground handling fees of each flight segment are extracted from the cost details and summed. The total transportation time of all flight segments and the total monetary cost are summarized by route to generate the route time cost summary.

[0025] Based on the summary of the route time costs, the minimum transportation time and minimum monetary cost of the alternative routes are calculated. The transportation time ratio is obtained by dividing the sum of the transportation time of each route by the minimum transportation time of all alternative routes. The monetary cost ratio is obtained by dividing the sum of the monetary costs of each route by the minimum monetary cost of all alternative routes. The two ratios are then integrated with the route risk score by route to obtain the route risk time cost ratio table.

[0026] Calculate the comprehensive safety cost of air logistics routes based on the aforementioned route risk-time cost ratio table.

[0027] Preferably, the step of obtaining the preferred resilient logistics route scheme is as follows:

[0028] Based on the comprehensive safety cost of the air logistics route for different alternative routes, the route identifier, segment sequence, departure time, arrival time, and comprehensive safety cost of the air logistics route are read. Routes with missing comprehensive safety costs of the air logistics route are removed, and the routes are sorted in ascending order of comprehensive safety costs of the air logistics route and labeled as parallel groups to generate a sorted list of comprehensive safety costs of the air logistics route.

[0029] Based on the comprehensive safety cost ranking list of air logistics routes, priority is given to selecting routes from the group with the lowest comprehensive safety cost, in order of minimum segment sequence length, minimum total planned transport dwell time, and earliest arrival time. This process forms a unique route while retaining the segment sequence, transit nodes, and time parameters, thus obtaining the optimal resilient logistics route scheme.

[0030] Preferably, the step of obtaining the logistics path execution instruction is as follows:

[0031] Based on the optimized resilient logistics route scheme, the system integrates flight segment sequence, transit nodes, planned departure and arrival times, ULD allocation list, rack space pre-occupancy list, transport vehicle scheduling instructions, temperature control settings, customs release nodes, abnormal contact list, and fee authorization number to generate logistics route execution instructions.

[0032] This invention also provides a logistics supply chain security management method, comprising the following steps:

[0033] Based on the node and edge information of global cargo airports and their flight routes, the number of times each shortest path passes through nodes and edges is calculated, and the number of connected routes is counted to establish a vulnerability map of the air logistics network structure.

[0034] Based on each logistics asset in the vulnerability map of the aviation logistics network structure, a record containing residue vectors and residue decay half-life is established for each ULD, rack position and transport vehicle to obtain a digital archive of logistics assets. Based on the residue vectors and residue decay half-life in the digital archive of logistics assets, the current pollution potential value is calculated to generate a pollution risk baseline for logistics assets.

[0035] Based on air waybill information and cargo sensitivity matrix, the vulnerability map of the air logistics network structure and the pollution risk baseline of the logistics assets are called to calculate the path risk score. Then, the path risk score is calculated with transportation time and monetary cost to generate the comprehensive safety cost of air logistics path.

[0036] Based on the comprehensive safety cost of different alternative air logistics routes, a route search is performed and the route with the lowest comprehensive safety cost is selected to obtain a preferred resilient logistics route scheme. Based on the preferred resilient logistics route scheme, logistics route execution instructions are integrated to obtain the logistics route execution instructions.

[0037] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0038] In this invention, by quantifying the number of times a route passes through global cargo airport nodes and flight routes, and the number of connecting routes, a network structure vulnerability map can be formed. This enables intuitive identification and risk focus of key nodes and high-frequency segments in the air transport network. By combining the accurate recording of residual vectors and residual decay half-lives of logistics assets, the current pollution potential can be dynamically calculated, establishing a pollution risk baseline covering ULDs, racking positions, and transport vehicles, making asset risk status quantifiable and traceable. By introducing a cargo sensitivity matrix into the route assessment and jointly calculating the route risk score with transportation time and monetary costs, a comprehensive safety cost with multi-dimensional trade-offs can be generated, taking into account cargo characteristics, operational efficiency, and economic costs. By screening multiple alternative routes based on the comprehensive safety cost, the selected transportation plan can be ensured to balance risk resistance and economic rationality. Furthermore, when outputting execution instructions, operational elements such as flight segment times, asset allocation, temperature control, and customs procedures are integrated, improving the executability and feasibility of transportation organization. In summary, by quantifying route network characteristics, dynamically calculating asset risk status, integrating cargo characteristics and multi-factor costs, and outputting execution information in a unified manner, we have achieved proactive risk prevention and control, scientific route decision-making, and precise execution, thereby enhancing the safety and resilience of the entire air logistics chain. Attached Figure Description

[0039] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0041] Please see Figure 1 The present invention provides a technical solution: a logistics supply chain security management system comprising:

[0042] The aviation network vulnerability analysis module calculates the number of times each shortest path passes through nodes and edges based on the node and edge information of global cargo airports and flight routes, and counts the number of connected routes to build a vulnerability map of the aviation logistics network structure.

[0043] The logistics asset risk assessment module, based on each logistics asset in the vulnerability map of the air logistics network structure, establishes a record containing residue vectors and residue decay half-lives for each ULD, rack position and transport vehicle, obtains a digital archive of logistics assets, calculates the current pollution potential value based on the residue vectors and residue decay half-lives in the digital archive of logistics assets, and generates a pollution risk baseline for logistics assets.

[0044] The route safety cost calculation module, based on air waybill information and cargo sensitivity matrix, calls the air logistics network structure vulnerability map and logistics asset pollution risk baseline to calculate the route risk score, and then calculates the route risk score with transportation time and monetary cost to generate the comprehensive air logistics route safety cost.

[0045] The route optimization and execution module searches for routes based on the comprehensive safety cost of different alternative air logistics routes and selects the route with the lowest comprehensive safety cost to obtain the preferred resilient logistics route solution. Based on the preferred resilient logistics route solution, it integrates to obtain the logistics route execution instructions.

[0046] The steps for obtaining a vulnerability map of the air logistics network structure are as follows:

[0047] Based on global cargo airport node information and flight route edge information, the airport IATA three-letter code and airport geographical coordinates are parsed. The duplicate edges are removed by route number and take-off and landing airports while retaining directionality. Node index and adjacency table are generated and isolated nodes and duplicate edges are verified to obtain the air logistics network topology.

[0048] Based on the topology of the air logistics network, the shortest path is searched for each pair of origin and destination, and the number of nodes and edges traversed along the path is accumulated one by one. At the same time, the number of incoming edges and outgoing edges of each node are counted and summed to obtain the number of connected routes, forming a shortest path traversal statistics table and a node connected route number statistics table.

[0049] Based on the shortest path traversal statistics table and the node connection route number statistics table, the nodes and edges are counted, sorted, and divided into quantile intervals, and the count level and the number of connected routes are labeled. The data is then overlaid onto the geographic base map, and the visualization style and layer order of the nodes and edges are configured to generate an air logistics network structure vulnerability map.

[0050] Specifically, based on global cargo airport node information and flight route information, the data parsing process is first initiated. For each airport node information, string matching and verification functions are called to extract and verify the airport's IATA three-letter code from the original data row, ensuring it is in three uppercase letter format. Simultaneously, the airport's geographic coordinates are extracted, and all non-decimal coordinate formats, such as degrees, minutes, and seconds, are uniformly converted to decimal latitude and longitude values ​​under the WGS84 standard for subsequent geographic information system processing. Next, the flight route information is processed, creating a data structure for each route record consisting of a route number, departure airport, and other relevant information. A unique composite key composed of the IATA three-letter code and the landing airport's IATA three-letter code is used to traverse and deduplicate all flight routes using a hash set data structure to eliminate duplicate flight route data. This process preserves the directionality of the flight routes; that is, a flight route from XXXX (PVG) to XXX (LAX) and a flight route from XXX to XXXX are treated as two independent edges. After data cleaning, the network data structure is constructed. The system traverses all deduplicated airport nodes and assigns a consecutive integer starting from 0 as a node index to each unique airport's IATA three-letter code, establishing an IATA data structure. A bidirectional mapping table from ATA three-letter codes to integer indices is used. Then, an adjacency table is initialized. This table is either an array or a hash table, where the keys are the integer indices of nodes, and the values ​​are lists storing the indices of all directly reachable destination nodes for that node. After populating the adjacency table, network integrity checks are performed. The first step is to check for orphaned nodes. This is done by checking the adjacency table and constructing a reverse adjacency table to count the in-degree and out-degree of each node. If a node's total degree (in-degree plus out-degree) is 0, it is marked as an orphaned node and processed according to a preset rule. This rule is defined as: if the orphaned node is not found in external authoritative data... If a node is marked as an active cargo airport in the list published by the International Air Transport Association (IATA), the node is retained and a warning log is generated for manual verification. Otherwise, it is removed from the node index and adjacency list. The second step is to check for duplicate edges. The adjacency list is traversed again to check if there are duplicate node indexes in the neighbor list of each node. If duplicates are found, one is retained and the others are removed to ensure that there is at most one direct edge in the same direction between any two nodes in the network. After the above parsing, deduplication, construction and verification steps, an air logistics network topology structure is finally formed, consisting of a node index table and an adjacency list that has passed integrity verification.

[0051] Based on the air logistics network topology, two data structures are first initialized for statistics: one is a shortest path traversal count table at the node level, whose structure is a hash table with node indices as keys and count values ​​as values, initially all values ​​are 0; the other is a shortest path traversal count table at the edge level, whose structure is an ordered pair of nodes representing edges (e.g., (starting index, ...). A hash table is created with the destination index as the key and the count value as the value, initially set to 0. Then, for each airport node in the network, the system uses it as the starting point and runs a breadth-first search (BFS) algorithm to calculate the shortest path from that starting point to all other reachable nodes in the network. The shortest path is defined by minimizing the number of flight segments traversed. During the BFS process, in addition to using a queue for node traversal, a predecessor node record table is maintained in parallel to backtrack and reconstruct the complete path after reaching the destination. For a total of N nodes in the network, this process is repeated N times, completing the shortest path calculation for all node pairs. After calculating the shortest path between each pair of starting and ending points, the system backtracks along the path, incrementing the count of the shortest path traversal in the corresponding node dimension of all intermediate nodes (excluding the starting and ending points) by 1. Simultaneously, for each edge constituting the path (i.e., each consecutive node pair), the count is incremented in the corresponding edge dimension. The count in the shortest path traversal statistics table is incremented by 1. If multiple shortest paths of the same length exist between any pair of nodes, the path selection is performed according to a preset rule, such as selecting the path first discovered by the breadth-first search algorithm. Other shortest paths with the same length are not processed. After completing the shortest path traversal and count accumulation for all node pairs, the system calculates the number of connected routes for each node. Using the adjacency list in the existing air logistics network topology, the length of each node's neighbor list is directly read to obtain the number of outgoing edges (out-degree) of that node. At the same time, by traversing the entire adjacency list, the number of times each node appears as a neighbor is accumulated to obtain the number of incoming edges (in-degree) of that node. The number of incoming edges and outgoing edges of each node are added together, and the sum is the number of connected routes for that node. Finally, the traversal counts of all nodes are integrated to form a complete shortest path traversal statistics table, and the number of connected routes for all nodes is summarized to form a node connected route count statistics table.

[0052] Based on the shortest path traversal statistics table and the node connection route count statistics table, the system first quantifies and classifies the vulnerability of nodes and edges. For each node, the system extracts its count value from the shortest path traversal statistics table and sorts all node count values ​​in descending order. Then, it divides the network into intervals according to a preset node vulnerability quantile threshold. This threshold is set based on the industry-accepted Pareto principle, identifying the few key nodes that have the greatest impact on the network. Specifically, the top 10% of nodes by count value are classified as "highly vulnerable," nodes between 10% and 40% are classified as "medium vulnerable," and the remaining 60% of nodes are classified as "low vulnerable." Each node is labeled with its corresponding count level. Similarly, for edges, the system extracts their count values ​​from the shortest path traversal statistics table, sorts them in descending order, and applies a preset edge vulnerability quantile threshold for classification. For example, the top 5% of edges are classified as "highly vulnerable," 5% to 25% as "medium vulnerable," and the rest as "low vulnerable," and each is labeled with its count level. Next, the system uses a node connection route count statistics table to sort nodes in descending order by the number of connected routes, and labels their hub level according to another set of preset node centrality quantile thresholds. For example, the top 5% of nodes by connection number are labeled "core hub," and 5% to 20% are labeled "trunk hub." The remaining airports are labeled "feeder airports." After completing all classification and labeling, the system enters the visualization generation stage. First, a standard Geographic Information System (GIS) base map is loaded, such as a tile map based on OpenStreetMap data. Then, the geographic coordinates of each airport node are obtained from the air logistics network topology, and the corresponding latitude and longitude positions are plotted as point symbols on the map. Next, based on the adjacency list information, arcs representing flight routes are drawn between the map points. These arcs are generated using a great circle flight route algorithm to realistically reflect flight trajectories on the Earth's surface. Finally, the visualization styles of nodes and edges are configured, with the node color adjusted according to its count level. The rendering process is structured in layers, with "high vulnerability" rendered in red, "medium vulnerability" in yellow, and "low vulnerability" in green. The size of a node is proportional to the number of routes it connects to, with "core hubs" being the largest and "feeder airports" the smallest. Similarly, the color of an edge is determined by its count level, and the thickness of an edge is proportional to the absolute value of the number of times it is traversed by its shortest path. Finally, the rendering order of layers is set to ensure that the geographic base map is at the bottom, followed by the route edge layer, and the airport node layer is at the top. Edges and nodes with high vulnerability levels are given higher layer priority to prevent them from being obscured by other elements. This generates a vulnerability map of the air logistics network structure that integrates geographic information, topological relationships, and quantitative vulnerability indicators.

[0053] The steps for obtaining digital archives of logistics assets are as follows:

[0054] Based on the vulnerability map of the air logistics network structure, the unique identifiers and geographical locations of each ULD, racking location, and transport vehicle are located one by one, and the timestamp of the most recent contact with the cargo is read. The residual sampling entries reported by each asset are analyzed and the naming conventions and units of measurement of residual entries are standardized. The values ​​of the same type of residuals within the same asset are arranged in chronological order and the corresponding residual decay half-life metadata is attached to form an asset-level entry set for each ULD, racking location, and transport vehicle, resulting in a record containing residual vectors and residual decay half-life.

[0055] Based on records containing residual vectors and residual decay half-lives, all entries for the same asset are merged according to the unique identifiers of ULD, shelf location, and transport vehicle, and duplicate sampling entries are removed. The asset category label and historical operation trajectory summary for each asset are completed, and the field integrity and timestamp continuity are verified. The set of verified asset-level entries is written into archive entries with a unified structure and a searchable index is generated to obtain a digital archive of logistics assets.

[0056] Specifically, based on the vulnerability map of the air logistics network structure, the system first initiates an asset data collection and preprocessing process. Through interface with the IoT platform, it retrieves the unique electronic identifier for each managed logistics asset, such as the IATA standard code of the ULD, the RFID tag ID of the shelf location, or the VIN code of the transport vehicle. Combined with positioning devices installed on the asset, such as GPS modules or Bluetooth beacons, it obtains the current or last recorded geographical location. It also reads the exact timestamp of the asset's most recent physical contact with goods from the asset management system's operation logs, such as loading, unloading, or storage operations. After obtaining this basic information, the system begins processing residual data reported by on-site detection equipment or manual sampling. This data may initially vary in format, so the system calls a pre-built standardized residual dictionary to parse the reported entries. This dictionary contains multiple aliases, chemical formulas, and a standardized name for hundreds of common logistics pollutants. For example, "industrial dust," "PM2.5 particles," and "particulate matter" are uniformly parsed as the standard name "Particulate Matter." Simultaneously, the system checks the unit of measurement for each residual item. Using a built-in conversion factor table, all non-standard units, such as "ppm" and "ounces per cubic foot," are uniformly converted to international standard units. For example, the concentration unit is standardized to "milligrams per cubic meter," and the surface residue unit is standardized to "micrograms per square centimeter." After standardizing the naming and units, the system aggregates all residue records belonging to the same asset and sorts them in ascending order according to the sampling timestamp to form a time series. Then, the system queries a pre-set "material properties database," which stores the physicochemical properties of various residues, including their properties under specific environmental conditions. The system uses the natural decay half-life of the residue to match and extract the corresponding half-life value based on the standard name of the residue. This half-life is then appended as a metadata field to each residue record. After a series of operations such as location, parsing, standardization, and metadata appending, a set of asset-level entries is finally formed for each independent ULD, shelf location, and transport vehicle. Each record in this set contains the asset identifier, sampling time, standardized residue type, quantified value, standard unit of measurement, and the key residue decay half-life, resulting in a record containing the residue vector and the residue decay half-life.

[0057] Based on records containing residue vectors and residue decay half-lives, the system initiates the asset profile construction process. First, it performs data aggregation, grouping scattered asset-level entries using each asset's unique identifier, such as a ULD code, as the key. Records belonging to the same asset are merged into a single list. During this merging process, the system performs deduplication, comparing the asset's unique identifier, residue type, and sampling timestamp. If identical entries are found, they are considered duplicates and only one is retained. Next, to enrich the profile content, the system connects to the enterprise's core asset management database and transportation management system, providing each... Each asset is supplemented with key static and dynamic information. Asset category tags, such as "Cold Chain ULD," "Ambient Temperature Shelving," and "Dangerous Goods Transport Vehicle," are directly retrieved from the asset management database based on the asset's unique identifier and populated. Historical operation trajectory summaries are dynamically generated by querying the asset's operation records over a past period (e.g., 90 days) in the transportation management system. These summaries contain a series of key node information, such as the three-letter sequence of airport codes, the flight number, and the cargo type code, described in a concise text format, for example, "Recent 30-day trajectory: PVG-LAX-ORD, previously transported cargo type: AVI," After completing the information supplementation, the system performs strict integrity and continuity checks on the archive entries under construction. The integrity check checks whether each entry contains all predefined necessary fields such as asset unique identifier, asset category, timestamp, and residue vector. Any record with missing fields will be marked as incomplete and moved to the pending queue. The timestamp continuity check analyzes the time series of each asset's operation log and sampling records, and sets a preset time interval threshold. This threshold is dynamically set according to the asset type and its normal usage frequency. For example, for high-frequency ULDs, the threshold is 24 hours. If the timestamp interval of two consecutive records exceeds this threshold, it is considered that there may be missing data and is marked. Only the set of asset-level entries that pass both integrity and continuity checks will be written into an archive entry with a unified JSON data structure, and an inverted index will be built for key fields such as asset unique identifier, asset category, and geographical location area to obtain the logistics asset digital archive.

[0058] The steps for obtaining the baseline for pollution risk of logistics assets are as follows:

[0059] Based on the digital archives of logistics assets, the residual vector and residual decay half-life of each asset are read, and the residual entries are converted by time decay according to the timestamp in the archive and aggregated and weighted with similar residuals. The asset category label and historical operation trajectory summary are extracted as risk correction factors, and the aggregated residual entries are labeled with threshold segments. The current pollution potential value is output in units of assets and summarized into a hierarchical list according to asset category to generate the pollution risk baseline of logistics assets.

[0060] Specifically, based on the digital archives of logistics assets, the system initiates a risk baseline calculation process. First, it retrieves all residual vectors and their corresponding residual decay half-lives for each asset from the archives. Then, it applies a time decay model to each residual record for numerical conversion, using the exponential decay formula. ,in, It is the calculated estimated concentration of the current residue. It is the original sampling concentration recorded in the archive. This is the current system timestamp. It is the historical timestamp of the sampling. The decay half-life of the residue is retrieved from the archives. Using this formula, all historical sampling data is uniformly converted to the current time point. Then, for the same type of residue on the same asset after decay conversion, the system performs aggregated weighted calculations, directly adding their estimated concentration values ​​to form the total pollution potential value of that type of residue. Subsequently, the system extracts the asset category label and historical operation trajectory summary of the asset, using this as the basis for risk correction. The system queries a preset "Risk Correction Factor Mapping Table," which was established by industry safety experts based on historical data analysis. This table defines the risk correction coefficients corresponding to different asset characteristics. For example, due to its enclosed and low-temperature environment, the risk correction factor for biological contaminants in "Cold Chain ULD" is set to 1.2, while that for "Ordinary Shelf Space" it is 0.9. If the historical operation trajectory summary contains "high-risk areas" (such as flights that have previously transported live animals, AVI), then its corresponding risk correction factor is 1.5. The system multiplies all applicable correction factors found and multiplies by the previously aggregated value. The system obtains the total pollution potential value of the residues and then calculates the corrected residue risk value. The system then segments and labels the corrected risk values, using the following threshold setting method: for all similar residues (e.g., volatile organic compounds) within a specific asset category (e.g., all ULDs), the system calculates their mean and standard deviation. Residues below the mean are classified as "low risk," those between the mean and the mean plus one standard deviation are classified as "medium risk," and those above the mean plus one standard deviation are classified as "high risk." Finally, the system weights and sums the final risk values ​​of all different types of residues on an asset according to a preset "pollutant hazard level weight table" to obtain the final single quantitative indicator for that asset, namely the current pollution potential value. For example, the weight for highly toxic chemical residues is 0.7, and the weight for common allergens is 0.2. The current pollution potential values ​​and risk levels of all assets are then categorized and summarized by asset category (ULD, shelf space, transport vehicle) to form a clearly hierarchical list, generating a pollution risk baseline for logistics assets.

[0061] The steps to obtain the path risk score are as follows:

[0062] Based on the sensitivity matrix of air waybill information and cargo, the cargo name code, processing code and temperature control level in the air waybill are read one by one. The corresponding row and column cells of the sensitivity matrix are matched in turn and the sensitivity values ​​are extracted. The extracted sensitivity values ​​are uniformly converted into the same unit of measurement and then concatenated into a sensitivity measurement vector in the order of the fields. The average calculation is performed on all components of the sensitivity measurement vector to obtain the cargo sensitivity measurement result.

[0063] Based on the cargo sensitivity quantification results, the planned flight segment sequence in the air waybill is analyzed, and the node count and edge count corresponding to each flight segment are retrieved in the air logistics network structure vulnerability map. The node count and edge count are normalized to the flight segment structure vulnerability value. At the same time, the ULD, rack space and transport vehicle associated with the flight segment are matched in the logistics asset pollution risk baseline, and the corresponding asset pollution potential value is extracted. The asset pollution potential value and structural vulnerability value of each flight segment are combined with the planned transport time and transit stay time recorded in the air waybill to form the flight segment planned transport stay time, resulting in a list of path exposure elements including flight segment asset pollution potential value, flight segment structural vulnerability value and flight segment planned transport stay time.

[0064] Based on the list of exposed elements along the route, the route risk score is calculated using the following formula:

[0065] ;

[0066] in, The path risk score, The number of flight segments in the route exposure element list. This refers to the cargo sensitivity scalar in the cargo sensitivity scalarization results. For the first Planned transport stopover time for each segment The characteristic time constant, For the first The potential pollution value of assets for each flight segment. For the first Structural vulnerability values ​​for each flight segment Risk type weighting coefficients for asset pollution potential. Risk type weighting coefficients for structural vulnerability values. This represents the risk synergy gain coefficient.

[0067] Specifically, based on air waybill information and a cargo sensitivity matrix, the system first initiates a cargo sensitivity assessment process. For each received electronic air waybill, it calls a data parser to read key fields one by one. These include the IATA standard three-digit cargo name code, such as "502" for pharmaceuticals and "001" for live animals. It also reads special handling codes, such as "COL" indicating refrigeration or "FRG" for fragile items, and specific temperature control levels, such as "CRT" representing a controlled room temperature of 15°C to 25°C or "FRO" representing freezing requirements below -20°C. After reading the information, the system calls a pre-built cargo sensitivity matrix for matching. This matrix is ​​a multi-dimensional table, and its construction process is as follows: First, logistics safety and cargo management experts jointly define a set of sensitivity dimensions, including "temperature sensitivity," "humidity sensitivity," etc. The system evaluates the sensitivity of goods based on their product name code, handling code, and temperature control level. Then, for each combination of product name code and special handling code, the expert group uses the Delphi method to anonymously score the sensitivity in each dimension, ranging from 1 to 10. For example, for a medicine with product name code "502" and temperature control level "CRT", its "temperature sensitivity" might be 8, "contamination sensitivity" 9, and "vibration sensitivity" 6. After scoring, the average of all expert scores is taken and entered into the corresponding cells of the matrix, forming a numerical matrix with product code as rows and sensitivity dimensions as columns. The system searches for the corresponding row in the matrix based on the product name code, handling code, and temperature control level read from the waybill, and extracts the values ​​for that row under all sensitivity dimensions, thus obtaining a list of sensitivity values, such as [8, [5, 6, 4, 9, 7] Next, the system performs a unified quantization transformation on these sensitivity values ​​of different dimensions, converting them all to a standardized scale of 0 to 100 through linear mapping. The conversion formula is as follows: ,in The original ratings from 1 to 10 are converted to [77.8, 44.4, 55.6, 33.3, 88.9, 66.7]. The system then concatenates these standardized values ​​according to a predetermined field order (temperature, humidity, vibration, pressure, pollution, and timeliness) to form a multidimensional sensitivity measurement vector. Finally, the average of all components in this vector is calculated as (77.8 + 44.4 + 55.6 + 33.3 + 88.9 + 66.7) / 6, resulting in a single scalar value, which is the sensitivity measurement result for the goods.

[0068] Based on the cargo sensitivity quantification results, the system continues to extract and integrate route risk elements. First, the system analyzes the detailed planned flight segment sequence in the air waybill. For example, a transportation plan from XX (PVG) via XXX (LAX) to XXX (ORD) will be decomposed into two independent segments: the first segment is PVG to LAX, and the second segment is LAX to ORD. For each segment, the system retrieves data from the previously generated air logistics network structure vulnerability map. Specifically, for the first segment PVG to LAX, the system retrieves the "shortest path traversal count" of the origin (PVG) and destination (LAX) nodes, as well as the "shortest path traversal count" of the edge connecting the two points. These counts (e.g., PVG node count is 12500 times, PVG-LAX edge count is 8900 times) are then normalized using the min-max normalization method. The calculation formula is... ,in This is the original count value. and These are the maximum and minimum counts of all nodes or edges in the network. Using this method, the original counts are converted into segment structural vulnerability values ​​between 0 and 1. For example, the vulnerability values ​​for nodes and edges are 0.85 and 0.72 respectively. These two values ​​are then averaged to obtain the overall structural vulnerability value for the segment, 0.785. Simultaneously, the system queries the logistics asset contamination risk baseline. Based on the unique asset identifier bound to the waybill, such as the ULD number "AKE12345CX" used in the first segment, it matches and extracts the "current contamination potential value" of that ULD from the baseline, for example, a value of 35.2 in the range of 0 to 100. This process is also applied to assets involved at the segment's endpoint (transfer station), such as the rack space "C-07-04" reserved at LAX airport. The system extracts the current pollution potential of the ground transport vehicle license plate “CA-T5678” and averages it as the asset pollution potential of the transit segment. Next, the system reads the planned flight time of the first segment from the flight schedule of the air waybill, for example, 12.5 hours, and reads the planned transit stay time at LAX airport from the ground operation plan, for example, 4 hours. The two times are added together to get the total planned transport stay time of the first segment as 16.5 hours. Finally, the system merges the three core elements of the asset pollution potential, structural vulnerability value and planned transport stay time of the segment into a single record. This operation is repeated for all segments to obtain a list of path exposure elements that includes the segment asset pollution potential, segment structural vulnerability value and segment planned transport stay time.

[0069] formula: The advantage of the formula lies in its use of an exponential saturation model. It can scientifically reflect that risk increases with exposure intensity, but it will not increase indefinitely; eventually, it will tend to a level determined by the sensitivity of the goods themselves. The upper limit set aligns with the objective law that risks have a maximum tolerance limit in reality. Secondly, it addresses two completely different sources of risk: pollution risk from the micro-level of logistics assets. and structural risks from the macro level of logistics networks By integrating the data through a weighted sum, the overall risk exposure of the path is measured more comprehensively, while also introducing a risk synergy gain term. This means that when a heavily polluted asset is transferred through a highly vulnerable hub, the risk it generates is far greater than the simple sum of the risks of the two independent assets. This synergistic gain design makes the risk assessment results closer to the risk amplification effect of "1+1>2" in complex systems in the real world, thus enabling the identification of potential high-risk paths that traditional linear models cannot detect.

[0070] The number of segments in the route exposure element list is a positive integer representing the number of basic transport units into which the entire transport route is divided. This parameter is obtained directly from the route exposure element list generated in the previous step through counting. For example, a route from XX (PVG) via XXX (LAX) to XXX (ORD) is decomposed into two segments: PVG-LAX and LAX-ORD. .

[0071] The cargo sensitivity scalar in the cargo sensitivity quantification result is a value ranging from 0 to 100. It quantifies the sensitivity of the cargo itself to adverse external conditions. This value is calculated by the aforementioned "Cargo Sensitivity Quantification Result" step and is the final result after standardizing and averaging the sensitivity scores of the cargo in multiple dimensions such as temperature, humidity, vibration, and pollution. For example, a batch of vaccines that requires strict temperature control might have a calculated cargo sensitivity scalar. .

[0072] For the first The planned transit dwell time for each segment, in hours, represents the cargo's transit time on the [number]th [segment]. The total duration of exposure to potentially risky environments within each flight segment, including the sum of airtime and ground transfer or waiting time, is extracted directly from the route exposure factor list. For example, for the PVG-LAX segment, with a planned flight time of 12.5 hours and a planned transfer time of 4 hours at LAX, the planned transport stop time for the first segment is... Hour.

[0073] The characteristic time constant, expressed in hours, plays a role in time scale normalization in the formula. It places the risk exposure levels of flight segments of different durations on a comparable benchmark. Its value is set based on statistical analysis of historical logistics risk event data. By collecting data on various cargo damage, pollution, and delay risks that occurred over the past five years, the average time interval from the occurrence of the risk factor to the event is calculated, and the 75th percentile of this statistical distribution is used as the characteristic time constant. For example, if statistical analysis shows that 75% of pollution events occur within 24 hours of exposure, then the characteristic time constant is set to... Hour.

[0074] For the first The asset pollution potential value for each flight segment is a normalized dimensionless value ranging from 0 to 1. It quantifies the pollution potential value of the assets in the first flight segment. In each flight segment, the potential risk level of residual contaminants carried by logistics assets (such as ULDs, racking positions, and transport vehicles) that come into contact with cargo is derived from the "Current Contamination Potential" (range 0-100) in the logistics asset contamination risk baseline and normalized using the following formula: For example, if the ULD used in the first leg of the journey has a current pollution potential of 35.2, then... .

[0075] For the first The structural vulnerability value for each flight segment is a normalized dimensionless value ranging from 0 to 1. It quantifies the structural vulnerability of the first flight segment. The importance and irreplaceability of each flight segment within the entire air logistics network reflects the severity of the impact on the entire network should that segment be disrupted. This value is obtained by standardizing the "shortest path traversal count" of the nodes and edges involved in the segment using the maximum and minimum methods described in the preceding steps. For example, for the busy PVG-LAX segment, its calculated structural vulnerability value is... .

[0076] The risk type weighting coefficient for asset contamination potential is a dimensionless value ranging from 0 to 1. It reflects the decision-maker's level of attention to the risks caused by asset contamination. This coefficient was determined using the Analytic Hierarchy Process (AHP). Five supply chain security experts were invited to conduct pairwise comparisons between "asset contamination risk" and "network structure risk" to construct a judgment matrix. For example, if the experts unanimously agreed that asset contamination risk was "slightly more important" than network structure risk, they would assign it a rating of 2. The weight vector was obtained by calculating the largest eigenvalue and its corresponding eigenvector of the judgment matrix and performing a consistency check. .

[0077] The risk type weighting coefficient for structural vulnerability is a dimensionless value ranging from 0 to 1. It reflects the degree of importance that decision-makers attach to the risks caused by network structural vulnerabilities and satisfies the following conditions: In determining back, The value of is also determined accordingly. In the above example, .

[0078] The risk synergy gain coefficient is a dimensionless normal number used to quantify and amplify the coupled enhancement effect when asset contamination risk and structural vulnerability risk coexist. The value of this coefficient is obtained by performing nonlinear regression fitting on historical risk data, collecting data including cargo damage results, etc. Value and Historical transportation records are used to construct a regression model to fit the relationship between actual risk and predicted risk, thereby calibrating the data. The value of , for example, when obtaining the optimal fitting effect through fitting analysis of thousands of historical data, The value is 0.5.

[0079] Calculation process:

[0080] Taking a route like PVG-LAX-ORD as an example, this route contains two segments ( ).

[0081] The parameter values ​​are set as follows:

[0082] Cargo Sensitivity Measurement Scalar: ;

[0083] Characteristic time constant: ;

[0084] Risk type weighting coefficient: , ;

[0085] Risk synergistic gain coefficient: ;

[0086] Segment 1 (PVG-LAX):

[0087] Planned transit stop time: Hour;

[0088] Potential for asset pollution: ;

[0089] Structural fragility value: ;

[0090] Segment 2 (LAX-ORD):

[0091] Planned transit stop time: Hour;

[0092] Potential for asset pollution: ;

[0093] Structural fragility value: ;

[0094] Calculate the risk contribution value of the first leg. :

[0095] ;

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] ;

[0101] ;

[0102] ;

[0103] Calculate the risk contribution value of the second leg. :

[0104] ;

[0105] ;

[0106] ;

[0107] ;

[0108] ;

[0109] ;

[0110] ;

[0111] ;

[0112] Calculate the total path risk score :

[0113] ;

[0114] The result indicates that the route from PVG through LAX to ORD has a comprehensive path risk score of 43.76. This is a quantitative risk assessment result, which can be used to make horizontal comparisons among multiple alternative routes. The lower the score, the safer the route. In addition, this score can be compared with preset risk thresholds. For example, if the system's preset "low risk" threshold is 50 and the "medium risk" threshold is 100, then the result of 43.76 indicates that the route is in the low-risk range and is acceptable.

[0115] The steps to obtain the comprehensive safety cost of air logistics routes are as follows:

[0116] Based on the route candidate set, the flight segment list is extracted sequentially, the departure time and arrival time of each flight segment are extracted and the time difference is calculated. The transit customs waiting time is added to obtain the flight segment transportation time. At the same time, the take-off and landing fees, fuel surcharges and ground handling fees of each flight segment are extracted from the cost details and summed. The total transportation time of all flight segments and the total monetary cost are summarized by route to generate the route time cost summary.

[0117] Based on the summary of route time costs, the minimum transportation time and minimum monetary cost of the alternative routes are calculated. The transportation time ratio is obtained by dividing the sum of the transportation time of each route by the minimum transportation time of all alternative routes. The monetary cost ratio is obtained by dividing the sum of the monetary costs of each route by the minimum monetary cost of all alternative routes. The two ratios are then integrated with the route risk score by route to obtain the route risk time cost ratio table.

[0118] Based on the route risk-time cost ratio table, the comprehensive safety cost of air logistics routes is calculated using the following formula:

[0119] ;

[0120] in, Considering the overall safety costs of air logistics routes, To normalize the path risk score, The ratio of transportation times is the ratio of the total transportation time to the minimum transportation time of all alternative routes. Let be the ratio of monetary costs, which is the ratio of the sum of monetary costs to the minimum monetary cost of all alternative paths. The decision weight coefficient for risk score. The decision weighting coefficient for the ratio of transportation time. The decision weighting coefficients for the monetary cost ratio, the three satisfy the following conditions: , The penalty aversion index is used to control sensitivity to high-value items.

[0121] Specifically, based on the route candidate set and the segment list, the system initiates a precise calculation process for time and monetary costs. For each candidate route in the set, the system parses its segment list one by one. For each segment, the system extracts the scheduled departure and arrival times of the planned flights for that segment from the flight schedule data interface published by the carrier. Both times use Coordinated Universal Time (UTC) to avoid time zone conversion errors. By calculating the difference between the two timestamps, the pure air travel time for that segment is obtained. Then, the system calculates the transfer time. The current segment is not the starting segment of the route. The system then retrieves the estimated transit time at the departure airport for this segment from the operation plan. This time includes the time for cargo to be unloaded from the previous segment's flight, wait in the warehouse or on the tarmac, and load onto the current segment's flight. A crucial customs waiting time is also superimposed. This time is not a fixed value but is obtained by querying a dynamically updated "Airport Customs Efficiency Database." This database is based on historical data and varies depending on the airport (e.g., PVG, LAX), cargo type (e.g., general cargo, live animals, pharmaceuticals), and time period. For example, statistical models are established for weekdays (daytime, nighttime, and holidays) to predict the average time required for customs clearance and transshipment. For instance, for a shipment of general cargo transshipped through LAX, the average customs waiting time during weekdays is predicted to be 2.5 hours. The total transit time is calculated by adding the flight time, ground handling time, and customs waiting time. Simultaneously, the system extracts detailed cost information related to this segment from the integrated financial system or carrier quotation database. This includes landing and takeoff fees charged by the departure and arrival airports, which are related to the aircraft type and weight. For example, at a large hub airport... The takeoff and landing costs are approximately US$8,000, plus a fuel surcharge charged by the carrier based on international oil prices and flight distance fluctuations, such as approximately US$2.5 per kilogram of cargo on the China-US route, and a ground handling fee charged by the airport ground service agent based on cargo weight or volume, such as US$0.3 per kilogram. The system sums these three main costs to obtain the monetary cost of the flight segment. After calculating the transportation time and monetary cost of all segments within the route, the system summarizes the costs by route, adding up the total transportation time and total monetary cost of all segments under the same route to generate a route time cost summary.

[0122] Based on the summarized route time costs, the system begins standardizing various cost indicators. First, the system iterates through all candidate routes in the summary, extracting the total transportation time and total monetary cost for each route, forming two independent lists. Then, the system performs a minimum value lookup operation in each of these lists to obtain the minimum transportation time and minimum monetary cost among all candidate routes. These two minimum values ​​will serve as the benchmark for subsequent ratio calculations. For example, if the total transportation times for three candidate routes are 28.5 hours, 32.0 hours, and 29.5 hours, the minimum transportation time is 28.5 hours, and the total monetary costs are $4500, $4200, and $4800, the minimum monetary cost is $4200. After determining the benchmark values, the system iterates through each candidate route again, calculating its cost ratio. For transportation time, the system divides the total transportation time of that route by the previously calculated minimum transportation time for all candidate routes to obtain the transportation time for that route. The ratio, a dimensionless number greater than or equal to 1, reflects the difference in time efficiency between the current path and the optimal choice. For example, a path with a transportation time of 32.0 hours has a transportation time ratio of 32.0 divided by 28.5, approximately equal to 1.12. Similarly, the system divides the sum of the monetary costs of the current path by the minimum monetary cost of all alternative paths to obtain the monetary cost ratio. For example, a path with a monetary cost of $4,500 has a monetary cost ratio of $4,500 divided by $4,200, approximately equal to 1.07. After calculating the ratios for the two paths, the system integrates these two newly calculated ratios with the path risk score of the current path. The path risk score is a comprehensive risk score calculated in the previous step based on cargo sensitivity, asset contamination, and network vulnerability. The system uses the unique identifier of the path as the primary key to integrate the path risk score, transportation time ratio, and monetary cost ratio of the current path into a single record. The same operation is performed on all alternative paths to finally obtain the path risk time cost ratio table.

[0123] formula: The advantage of the formula lies in its use of the generalized Minkowski distance to quantify the comprehensive cost of an air logistics route that includes three conflicting objectives: risk, time, and money. Firstly, this is achieved by introducing decision weight coefficients ( , , The model allows decision-makers to flexibly adjust the importance of each cost dimension based on different business scenarios and strategic priorities (e.g., for high-value pharmaceutical transportation, security risk has the highest weight; for ordinary e-commerce parcels, monetary cost has the highest weight). Secondly, the penalty aversion index... The introduction of, when When, the formula degenerates into a simple weighted average, when At that time, the formula will impose a heavier penalty on cost terms with higher values, for example when At this time, its form is similar to Euclidean distance, which means that if a path performs particularly poorly in any dimension of risk, time, or monetary cost (i.e., the ratio is significantly high), its final overall safety cost will be high. It will be disproportionately amplified, simulating the real-world aversion of decision-makers to the "weakest link" effect, effectively filtering out paths with extreme flaws, thus obtaining a more robust and balanced optimal solution.

[0124] The normalized path risk score is a dimensionless ratio with a value greater than or equal to 1. It is obtained by standardizing the original path risk scores to ensure comparability with other ratio parameters. The steps to obtain it are as follows: First, obtain the path risk scores of all candidate paths from the aforementioned steps. And find the minimum value among them. Then, using each path's own value divided by this minimum value The calculation formula is: For example, given three alternative paths A, B, and C, their path risk scores, calculated in the previous step, are... The values ​​are 43.76, 60.50, and 55.20 respectively. For path B, its normalized path risk score is... .

[0125] The transportation time ratio is a dimensionless ratio with a value greater than or equal to 1. It is obtained by dividing the total transportation time of a specific route by the minimum total transportation time among all alternative routes. This parameter is directly extracted from the route risk time cost ratio table generated in the previous step. For example, if the total transportation times of alternative routes A, B, and C are 32.0 hours, 28.5 hours, and 29.5 hours respectively, then the minimum transportation time is 28.5 hours. For route B, its transportation time ratio is 28.5 hours, so its transportation time ratio is... .

[0126] The currency cost ratio is a dimensionless ratio with a value greater than or equal to 1. It is obtained by dividing the total currency cost of a specific path by the minimum total currency cost among all alternative paths. This parameter is also extracted from the path risk-time cost ratio table. For example, if the total currency costs of alternative paths A, B, and C are $4,500, $4,900, and $4,200 respectively, then the minimum currency cost is $4,200. For path B, its currency cost is $4,900, and its currency cost ratio is... .

[0127] , , These are decision weight coefficients for risk, time, and cost, respectively. They are three dimensionless positive numbers that sum to 1. Their setting reflects the strategic preferences of the decision-making process and is typically determined using the Analytic Hierarchy Process (AHP). For example, for a batch of high-value electronic products with extremely high requirements for timeliness and security, the decision-making team performs pairwise comparisons: they consider "risk" to be extremely important than "cost" (score 9), "risk" to be equally important than "time" (score 1), and "time" to be very important than "cost" (score 7). They construct a judgment matrix and calculate the normalized eigenvector corresponding to its largest eigenvalue, then perform a consistency test with a value less than 0.1. After passing this test, a set of weight coefficients is obtained, for example... , , .

[0128] The penalty aversion index, a real number greater than or equal to 1, is used to adjust for sensitivity to extremely high values. This value is set by the decision-maker based on their risk preference strategy. If the decision-maker cannot tolerate a significant excess of any cost and tends to choose a path that balances performance across all aspects, a higher value will be selected. Values ​​(e.g.) If decision-makers are more focused on the overall weighted average cost and less sensitive to the peculiarities of individual items, they will choose the lower cost. Values ​​(e.g.) In standard applications, it is usually taken as... This represents a moderate level of risk aversion. In this example, we set... .

[0129] Calculation process:

[0130] Taking alternative path B as an example, substitute the obtained parameter values ​​into the calculation:

[0131] Normalized path risk score: ;

[0132] Transportation time ratio: ;

[0133] Monetary cost ratio: ;

[0134] Decision weight coefficient: , , ;

[0135] Punishment Aversion Index: ;

[0136] Calculate the comprehensive safety cost of air logistics route B. :

[0137] ;

[0138] ;

[0139] ;

[0140] ;

[0141] ;

[0142] ;

[0143] The result indicates that the comprehensive safety cost of air logistics route B is 1.202. This value is a relative, dimensionless cost measure; it does not represent a specific monetary amount or time unit. Its core significance lies in providing a unified comparison benchmark for different alternative routes. The system will calculate the respective costs for all alternative routes. By comparing these K values, we can conclude which path has the lowest overall safety cost. For example, if path A has the lowest overall safety cost... Path C ,So The path with the smallest value is path C, which is the optimal choice in the current decision-making scenario.

[0144] The steps to obtain the preferred resilient logistics route are as follows:

[0145] Based on the comprehensive safety cost of different alternative routes for air logistics, the system reads the route identifier, flight segment sequence, departure time, arrival time, and comprehensive safety cost of air logistics routes. It then removes routes that lack comprehensive safety costs for air logistics routes, sorts them in ascending order by comprehensive safety cost, marks parallel groups, and generates a sorted list of comprehensive safety costs for air logistics routes.

[0146] Based on the comprehensive safety cost ranking list of air logistics routes, priority is given to selecting routes from the group with the lowest comprehensive safety cost, in order of minimum segment sequence length, minimum total planned transport dwell time, and earliest arrival time. This process forms a unique route while retaining the segment sequence, transit nodes, and time parameters, thus obtaining the optimal resilient logistics route scheme.

[0147] Specifically, based on the comprehensive safety cost of different alternative air logistics routes, the system initiates a route sorting and filtering process. First, the system traverses the dataset of all alternative routes, and for each route, it reads its structured information one by one, including a unique route identifier, such as the string "PVG-LAX-ORD-01" generated by combining the origin, destination, and key transfer points, and the complete flight segment sequence, such as a sequence containing "PVG-LAX", The system retrieves an ordered list of routes ("LAX-ORD"), along with the total planned departure time, total planned arrival time, and the comprehensive safety cost of the air logistics route calculated in the previous stage. During the reading process, the system performs a data validity check, verifying that the comprehensive safety cost field for each route is a valid floating-point number. Any route with a null, non-numerical, or invalid cost value is removed from the current candidate set, and its route identifier is recorded in a log called "Calculation Abnormal Route Log" for subsequent analysis. After verification and removal, the system performs a sorting operation on the remaining valid routes. This sorting operation uses the comprehensive safety cost of the air logistics route as the unique and primary sorting key, and arranges them in ascending order, i.e., the route with the lowest cost value. The routes are listed first. To handle cases where the overall safety costs of air logistics routes are exactly the same, the system will execute a parallel group labeling process after sorting. It will traverse the sorted list, compare the overall safety costs of adjacent air logistics routes, and set a very small numerical tolerance, such as 0.0001. If the absolute value of the cost difference between two routes is less than this tolerance, they are considered parallel. The system will assign the same parallel group number to these two routes and all subsequent routes whose cost values ​​are within this tolerance range. For example, if the costs of the first three routes after sorting are 1.150, 1.150 and 1.202 respectively, the first two routes will be labeled as "Parallel Group 1" and the third route will be labeled as "Parallel Group 2", generating a sorted list of overall safety costs for air logistics routes.

[0148] Based on the comprehensive safety cost ranking list of air logistics routes, the system executes a multi-level deterministic screening procedure to select the unique optimal route. This procedure first locates the lowest-cost tie group in the comprehensive safety cost ranking list, specifically the route subset numbered 1, and uses this subset as the current unique candidate pool. All subsequent screening operations will be performed only within this candidate pool. The first level of screening, the primary winning rule, compares the segment sequence lengths of each route in the candidate pool. The system calculates the number of segments contained in each route; for example, if route A contains 2 segments and route B contains 3 segments, the system will select the route with the fewest segments. If all routes have the same number of segments, or if, after screening, multiple routes still have the same number of segments, these tied routes proceed to the next level of screening. The second level of screening, the secondary winning rule, compares the total planned transit time of the remaining candidate routes. This total time is the total planned transit time (i.e., the sum of flight time and transfer time) of all segments on that route. The system extracts or recalculates the sum from the total path time cost and selects the path with the smallest sum. If more than one path remains tied after this round of screening, a third layer of screening is initiated, namely the final winning rule. This rule compares the final planned arrival times of each candidate path. The system directly reads the arrival time timestamps from the path records and selects the path with the earliest timestamp. Since arrival times are usually accurate to the minute or second, the possibility of ties after this layer of screening is extremely low. Through such a strict screening logic designed according to the priority order of "cost-simplicity-timeliness-punctuality", the system can deterministically select a unique path from multiple options with the same cost. Subsequently, the system extracts and solidifies all the key information of the final winning path, including its complete flight segment sequence, the IATA three-letter code list of all transit nodes, and detailed time parameters, such as the takeoff and landing times and transit times of each flight segment, to obtain the optimal resilient logistics path solution.

[0149] The steps for obtaining logistics route execution instructions are as follows:

[0150] Based on the optimized resilient logistics route scheme, the system integrates flight segment sequence, transit nodes, planned departure and arrival times, ULD allocation list, rack space pre-occupancy list, transport vehicle scheduling instructions, temperature control settings, customs release nodes, abnormal contact list, and fee authorization number to generate logistics route execution instructions.

[0151] Specifically, based on the optimized resilient logistics route plan, the system initiates an automated process for generating and integrating execution instructions. First, the system uses the core route information from the optimized resilient logistics route plan, including the flight segment sequence, transit nodes, planned departure time, and planned arrival time, as the skeleton of the instructions. Then, based on the origin, transit points, and destination of the route, the system initiates interface calls to various related operational support systems to obtain and integrate specific execution resources and information. For the allocation of ULDs, the system sends a booking request to the ULD management system, which includes the cargo type, volume, weight, and the first flight segment. Given the departure airport and time, the system will select units based on a preset "ULD selection strategy," such as prioritizing units with the most recent cleaning records and meeting cargo temperature control requirements. The ULD management system will lock a specific unit, such as "AKE54321LH," and return its unique identifier, forming a ULD allocation list. For the pre-reservation of storage resources, for each transit node in the route, the system sends a pre-reservation instruction to the corresponding warehouse management system (WMS). The instruction includes the estimated arrival time, dwell time, and cargo characteristics. The WMS will then allocate a specific shelf location, such as "LAX-WH2-". The system adds "RACK-B07" to the shelf space reservation list. For ground transportation, the system sends a last-mile delivery dispatch instruction to the Transportation Management System (TMS), including the final destination address and required delivery time window. The TMS allocates a vehicle and driver and returns a dispatch confirmation number, generating a transportation vehicle dispatch instruction. Simultaneously, based on the temperature control requirements in the original waybill, such as "maintain between 2°C and 8°C," the system converts them into explicit operational instructions for cold chain ULDs or temperature-controlled warehouses, i.e., temperature control settings. Furthermore, the system automatically generates a customs node list based on cross-border nodes along the route, specifying the... Each airport requiring customs clearance or transit needs to submit the necessary documents and the estimated release time. In addition, the system will match and extract the contact information of the corresponding operations, customer service, and emergency handling personnel for each key node (take-off station, transit station) in the enterprise address book database, forming an abnormal contact list. Finally, through the integration with the financial system, the system generates a unique fee authorization number for this transportation business, which is used for the settlement of all related fees. All the above lists, instructions, settings, numbers, and other information are integrated into a standardized data structure to generate logistics route execution instructions.

Claims

1. A logistics supply chain security management system, characterized in that, The system includes: The aviation network vulnerability analysis module calculates the number of times each shortest path passes through nodes and edges based on the node and edge information of global cargo airports and flight routes, and counts the number of connected routes to build a vulnerability map of the aviation logistics network structure. The logistics asset risk assessment module, based on each logistics asset in the vulnerability map of the aviation logistics network structure, establishes a record containing residue vectors and residue decay half-lives for each ULD, rack position, and transport vehicle, obtains a digital archive of logistics assets, calculates the current pollution potential value based on the residue vectors and residue decay half-lives in the digital archive of logistics assets, and generates a pollution risk baseline for logistics assets. The route safety cost calculation module, based on air waybill information and cargo sensitivity matrix, calls the air logistics network structure vulnerability map and the logistics asset pollution risk baseline to calculate the route risk score, and then calculates the route risk score with transportation time and monetary cost to generate the comprehensive air logistics route safety cost. The route optimization and execution module, based on the comprehensive safety cost of different alternative air logistics routes, performs route search and selects the route with the lowest comprehensive safety cost, obtaining a preferred resilient logistics route solution. Based on the preferred resilient logistics route solution, it integrates to obtain logistics route execution instructions. Based on global cargo airport node information and flight route edge information, the airport IATA three-letter code and airport geographical coordinates are parsed. The duplicate edges are removed by route number and take-off and landing airports while retaining directionality. Node index and adjacency table are generated and isolated nodes and duplicate edges are verified to obtain the air logistics network topology. Based on the topology of the aviation logistics network, a shortest path search is performed for each pair of origin and destination, and the number of nodes and edges traversed along the path is accumulated one by one. At the same time, the number of incoming edges and outgoing edges of each node are counted and summed to obtain the number of connected routes, forming a shortest path traversal statistics table and a node connected route number statistics table. Based on the shortest path traversal statistics table and the node connection route number statistics table, the nodes and edges are counted, sorted, and divided into quantile intervals, and the count level and the number of connected routes are labeled. The data is then overlaid onto the geographic base map, and the visualization style and layer order of the nodes and edges are configured to generate an air logistics network structure vulnerability map. Based on the digital archives of logistics assets, the residual vector and residual decay half-life of each asset are read, and the residual entries are converted by time decay according to the timestamp in the archive and aggregated and weighted with similar residuals. The asset category label and historical operation trajectory summary are extracted as risk correction factors, and the aggregated residual entries are labeled with threshold segments. The current pollution potential value is output in units of assets and summarized into a hierarchical list according to asset category to generate the pollution risk baseline of logistics assets.

2. The logistics supply chain security management system according to claim 1, characterized in that, The steps for obtaining the digital archives of the logistics assets are as follows: Based on the vulnerability map of the aviation logistics network structure, the unique identifiers and geographical locations of each ULD, rack position, and transport vehicle are located one by one, and the timestamp of the most recent contact with the cargo is read. The residual sampling entries reported by each asset are analyzed, and the naming conventions and units of measurement of the residual entries are standardized. The values ​​of the same type of residuals within the same asset are arranged in chronological order and the corresponding residual decay half-life metadata is attached to form an asset-level entry set for each ULD, rack position, and transport vehicle, resulting in a record containing residual vectors and residual decay half-life. Based on the records containing residual vectors and residual decay half-lives, all entries for the same asset are merged according to the unique identifiers of ULD, shelf location, and transport vehicle, and duplicate sampling entries are removed. The asset category label and historical operation trajectory summary of each asset are completed, and the field integrity and timestamp continuity are verified. The set of verified asset-level entries is written into the archive entries with a unified structure and a searchable index is generated to obtain the digital archive of logistics assets.

3. The logistics supply chain security management system according to claim 1, characterized in that, The steps for obtaining the path risk score are as follows: Based on the sensitivity matrix of air waybill information and cargo, the cargo name code, processing code and temperature control level in the air waybill are read one by one. The corresponding row and column cells of the sensitivity matrix are matched in turn and the sensitivity values ​​are extracted. The extracted sensitivity values ​​are uniformly converted into the same unit of measurement and then concatenated into a sensitivity measurement vector in the order of the fields. The average calculation is performed on all components of the sensitivity measurement vector to obtain the cargo sensitivity measurement result. Based on the cargo sensitivity quantification results, the planned flight segment sequence in the air waybill is analyzed, and the node count and edge count corresponding to each flight segment are retrieved in the air logistics network structure vulnerability map. The node count and edge count are normalized to the flight segment structure vulnerability value. At the same time, the ULD, rack space and transport vehicle associated with the flight segment are matched in the logistics asset pollution risk baseline, and the corresponding asset pollution potential value is extracted. The asset pollution potential value and structural vulnerability value of each flight segment are combined with the planned transport time and transit stay time recorded in the air waybill to form the flight segment planned transport stay time, resulting in a list of path exposure elements including flight segment asset pollution potential value, flight segment structural vulnerability value and flight segment planned transport stay time. Calculate the path risk score based on the list of path exposure elements.

4. The logistics supply chain security management system according to claim 1, characterized in that, The steps for obtaining the comprehensive safety cost of the air logistics route are as follows: Based on the route candidate set, the flight segment list is extracted sequentially, the departure time and arrival time of each flight segment are extracted and the time difference is calculated. The transit customs waiting time is added to obtain the flight segment transportation time. At the same time, the take-off and landing fees, fuel surcharges and ground handling fees of each flight segment are extracted from the cost details and summed. The total transportation time of all flight segments and the total monetary cost are summarized by route to generate the route time cost summary. Based on the summary of the route time costs, the minimum transportation time and minimum monetary cost of the alternative routes are calculated. The transportation time ratio is obtained by dividing the sum of the transportation time of each route by the minimum transportation time of all alternative routes. The monetary cost ratio is obtained by dividing the sum of the monetary costs of each route by the minimum monetary cost of all alternative routes. The two ratios are then integrated with the route risk score by route to obtain the route risk time cost ratio table. Calculate the comprehensive safety cost of air logistics routes based on the aforementioned route risk-time cost ratio table.

5. The logistics supply chain security management system according to claim 1, characterized in that, The steps for obtaining the preferred resilient logistics route are as follows: Based on the comprehensive safety cost of the air logistics route for different alternative routes, the route identifier, segment sequence, departure time, arrival time, and comprehensive safety cost of the air logistics route are read. Routes with missing comprehensive safety costs of the air logistics route are removed, and the routes are sorted in ascending order of comprehensive safety costs of the air logistics route and labeled as parallel groups to generate a sorted list of comprehensive safety costs of the air logistics route. Based on the comprehensive safety cost ranking list of air logistics routes, priority is given to selecting routes from the group with the lowest comprehensive safety cost, in order of minimum segment sequence length, minimum total planned transport dwell time, and earliest arrival time. This process forms a unique route while retaining the segment sequence, transit nodes, and time parameters, thus obtaining the optimal resilient logistics route scheme.

6. The logistics supply chain security management system according to claim 1, characterized in that, The steps for obtaining the logistics route execution instructions are as follows: Based on the optimized resilient logistics route scheme, the system integrates flight segment sequence, transit nodes, planned departure and arrival times, ULD allocation list, rack space pre-occupancy list, transport vehicle scheduling instructions, temperature control settings, customs release nodes, abnormal contact list, and fee authorization number to generate logistics route execution instructions.

7. The logistics supply chain security management method of the logistics supply chain security management system according to any one of claims 1-6, characterized in that, Includes the following steps: Based on the node and edge information of global cargo airports and their flight routes, the number of times each shortest path passes through nodes and edges is calculated, and the number of connected routes is counted to establish a vulnerability map of the air logistics network structure. Based on each logistics asset in the vulnerability map of the aviation logistics network structure, a record containing residue vectors and residue decay half-life is established for each ULD, rack position and transport vehicle to obtain a digital archive of logistics assets. Based on the residue vectors and residue decay half-life in the digital archive of logistics assets, the current pollution potential value is calculated to generate a pollution risk baseline for logistics assets. Based on air waybill information and cargo sensitivity matrix, the vulnerability map of the air logistics network structure and the pollution risk baseline of the logistics assets are called to calculate the path risk score. Then, the path risk score is calculated with transportation time and monetary cost to generate the comprehensive safety cost of air logistics path. Based on the comprehensive safety cost of different alternative air logistics routes, a route search is performed and the route with the lowest comprehensive safety cost is selected to obtain a preferred resilient logistics route scheme. Based on the preferred resilient logistics route scheme, logistics route execution instructions are integrated to obtain the logistics route execution instructions.

Images