Big data-based express transportation cost real-time dynamic accounting method and device

Abstract

The invention provides an express transportation cost real-time dynamic accounting method and device based on big data, and relates to the field of data processing. In the method, a real-time data stream is constructed based on multi-source data, a node toughness state quantity is dynamically updated, and a risk energy propagation trajectory is formed, so that a network risk state can be perceived continuously. The cost is disassembled into three parts of basic freight price, risk premium and emergency deployment compensation, so that the cost structure is associated with the risk strength. And constructing an anti-factual cost channel to compare the real cost deviation and executing dynamic calibration. And a cost probability cloud is constructed for each package and deviates along with risk evolution, and a potential cost rising trend is formed. And abnormal event identification is carried out in combination with a disturbance mode and historical risk energy memory, and a real-time dynamic accounting result is output in linkage with total elements of the cost, so that self-adaptive mapping of the cost and the risk is realized. By implementing the technical scheme provided by the invention, the checking accuracy of the express transportation cost can be improved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, specifically to a method and apparatus for real-time dynamic calculation of express delivery costs based on big data. Background Technology

[0002] As urban logistics continue to expand and the demand for on-demand delivery grows rapidly, express delivery networks are gradually evolving into highly dynamic, multi-node collaborative systems. The cost structure between trunk transportation, regional transshipment, and last-mile delivery is constantly changing, making real-time freight accounting a core capability to ensure logistics timeliness and controllable operating costs.

[0003] Meanwhile, the increasing frequency of extreme weather events globally, the growing complexity of public health emergencies and urban emergency response measures, and the significantly increased disruption to transportation networks caused by uncertainties have led to chain reactions of congestion in short periods of time due to road closures, traffic paralysis, or warehouse shutdowns. This results in drastic fluctuations in transportation costs. In this context, especially when responding to natural disasters or major emergencies, relying solely on routine traffic data and static billing rules is no longer sufficient for accurate calculation of express delivery costs.

[0004] Therefore, there is an urgent need for a method and device for real-time dynamic calculation of express delivery costs based on big data. Summary of the Invention

[0005] This application provides a method and apparatus for real-time dynamic calculation of express delivery costs based on big data, which facilitates the improvement of the accuracy of express delivery cost calculation.

[0006] The first aspect of this application provides a real-time dynamic calculation method for express delivery costs based on big data. The method includes: acquiring road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk line nodes, and terminal nodes to form a real-time data stream; initializing the resilience state quantity of each node and dynamically adjusting it when the real-time data stream is input, causing the resilience state quantity to decay and diffuse towards adjacent nodes to form a risk energy propagation trajectory; and decomposing the express delivery cost into basic freight cost, risk premium cost, and other components based on the resilience state quantity and the risk energy propagation trajectory. This includes the costs of emergency dispatch and compensation; a counterfactual cost channel is constructed as a benchmark cost trajectory, and the deviation between the benchmark cost trajectory and the actual network cost changes is compared. Based on the deviation results, the risk premium cost and the emergency dispatch and compensation cost are adjusted; a cost probability cloud is generated for each package, and the cost probability cloud is shifted according to the changes in the resilience state quantity and the risk energy propagation trajectory to generate a potential cost increase trend; based on the similarity matching between the disturbance pattern and the historical risk energy memory, the abnormal event type, event level, and diffusion direction are identified, and the resilience state quantity, the cost probability cloud, the counterfactual cost channel, and the potential cost increase trend are linked to output the express delivery cost calculation result.

[0007] A second aspect of this application provides a real-time dynamic accounting device for express delivery costs based on big data. The device includes an acquisition module and a processing module. The acquisition module is used to acquire road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk nodes, and terminal nodes, to form a real-time data stream. The processing module is used to initialize the resilience state quantity of each node and dynamically adjust it when the real-time data stream is input, causing the resilience state quantity to decay and diffuse towards adjacent nodes, thus forming a risk energy propagation trajectory. The processing module is also used to decompose the express delivery cost into basic freight cost, based on the resilience state quantity and the risk energy propagation trajectory. The processing module includes risk premium costs and emergency allocation compensation costs. It also constructs a counterfactual cost channel as a baseline cost trajectory, compares the deviation between the baseline cost trajectory and actual network cost changes, and adjusts the risk premium costs and emergency allocation compensation costs based on the deviation results. Furthermore, the processing module generates a cost probability cloud for each package and shifts the cost probability cloud based on changes in the resilience state quantity and the risk energy propagation trajectory, generating a potential cost increase trend. Finally, the processing module identifies abnormal event types, event levels, and propagation directions based on similarity matching between disturbance patterns and historical risk energy memories, and links the resilience state quantity, the cost probability cloud, the counterfactual cost channel, and the potential cost increase trend to output express delivery cost calculation results.

[0008] A third aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, and both the user interface and the network interface are used to communicate with other devices. The processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method described above.

[0009] A fourth aspect of this application provides a non-transitory computer-readable storage medium storing instructions that, when executed, perform the method described above.

[0010] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: By introducing real-time data streams composed of multi-source data, express delivery cost accounting can comprehensively perceive the dynamic changes in the transportation network, avoiding cost judgment biases caused by the distortion of single data points. Through dynamic adjustment of node resilience parameters and their diffusion to adjacent nodes, risk changes can be continuously expressed even in the absence of data, thereby enabling early prediction of risk energy propagation trends. By breaking down transportation costs into basic freight costs, risk premium costs, and emergency allocation compensation costs, the cost composition closely corresponds to actual resource consumption and risk intensity, enhancing the authenticity and rationality of cost accounting results. Furthermore, by constructing a counterfactual cost channel, normal... Using cost changes under operating conditions as a baseline trajectory, cost accounting can identify the intensity of abnormal disturbances in real time and perform dynamic calibration, improving the adaptability of the cost structure to extreme scenarios. By generating a cost probability cloud for packages and shifting its distribution according to risk evolution, cost accounting can reveal potential cost increase trends in advance, achieving proactive risk warnings and scheduling guidance. By identifying abnormal events based on the matching of disturbance patterns and historical risk energy memory, and linking resilience state quantities, cost probability clouds, counterfactual cost channels, and potential cost increase trends to jointly correct cost output, the final accounting result can maintain a dynamic mapping of the actual risk level of the transportation network at any stage. Therefore, this solution can improve the accuracy of cost accounting even under conditions of incomplete data for extreme emergencies. Attached Figure Description

[0011] Figure 1 A flowchart illustrating a real-time dynamic calculation method for express delivery costs based on big data, provided for an embodiment of this application; Figure 2 A schematic diagram of a real-time dynamic calculation device for express delivery costs based on big data, provided for an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0012] Explanation of reference numerals in the attached figures: 21. Acquisition module; 22. Processing module; 31. Processor; 32. Communication bus; 33. User interface; 34. Network interface; 35. Memory. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0014] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0015] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0016] To address the aforementioned technical problems, this application provides a real-time dynamic calculation method for express delivery costs based on big data, referring to... Figure 1 , Figure 1 This is a flowchart illustrating a real-time dynamic calculation method for express delivery costs based on big data, provided in an embodiment of this application. The method is applied to a server and includes steps S110 to S160, as follows:

[0017] S110: Acquire road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk nodes, and terminal nodes to form a real-time data stream.

[0018] Specifically, in this technical solution, the server is used to perform data acquisition, data processing, and cost calculation tasks. It is a computing node deployed in a cloud platform or data center, possessing continuous network connectivity and high-concurrency data processing capabilities. For example, in a nationwide express delivery network, the operator can deploy a set of servers in the cloud to receive data from warehouses in various cities, transport vehicles, road condition platforms, and regulatory departments. The server performs unified cleaning, integration, and analysis of this data, thereby providing computational support for the subsequent dynamic calculation of express delivery costs.

[0019] Warehousing nodes refer to warehousing facilities within the express delivery network that handle the centralized collection, sorting, temporary storage, and transshipment of parcels. These include regional distribution centers, transit centers, consolidation warehouses, or forward warehouses. Trunk line nodes are key nodes located on the main channels of the transportation network, typically large distribution centers, railway or airport freight hubs, or inter-provincial or inter-regional transit stations. They primarily handle large-volume cargo transportation and transfer tasks across cities and regions. Terminal nodes are delivery outlets directly connected to recipients, including express delivery outlets, community delivery stations, and management nodes corresponding to self-service lockers.

[0020] Road traffic status data refers to a dataset characterizing the capacity and traffic conditions of different roads at different times, including information such as whether the road is currently closed, road speed, road congestion level, restricted road sections, and available detour routes. Environmental disturbance status data refers to a dataset reflecting the impact of weather and the natural environment on the transportation network, including rainfall intensity, snowfall level, visibility, road flooding, storm or typhoon warning levels, and geological disaster warnings. Policy control status data refers to constraint information issued by government departments or management agencies related to traffic flow, regional lockdowns, vehicle restrictions, and emergency control, used to characterize the policy-based traffic restrictions in different areas at specific times. Order flow distribution data refers to the temporal and spatial distribution of orders throughout the entire express delivery network, including the time of order creation, order origin location, order destination location, order density in different areas, order peaks and troughs in different time periods, and the proportion of orders flowing along each path.

[0021] Transportation capacity distribution data refers to the distribution of resources available for transportation tasks across different nodes and routes within the entire transportation network. This includes the number of vehicles, their locations, remaining cargo space, driver work status, shift schedules, sorting capacity at each warehouse node, number of delivery personnel at each end node, and remaining vehicle energy. Historical risk record data refers to a collection of risk information related to abnormal events recorded during past operations. This includes statistics on delays during historical natural disasters, the number of packages stranded during severe congestion, cargo damage and compensation records caused by extreme weather, order cancellation rates during traffic accidents or regional lockdowns, records of cross-regional vehicle support during historical emergency dispatches, and corresponding cost consumption. Real-time data streams are continuous data sequences formed by continuously accessing and updating road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity distribution data, and historical risk record data in chronological order on a server. Real-time data streams preserve the relationships between different data types and reflect the overall operational status of the entire transportation network at each time slice along a timeline.

[0022] Furthermore, the server periodically retrieves real-time traffic speeds, road status labels, and passability markers for each road segment through an interface with the map service platform. It also obtains traffic flow data and congestion assessment levels from fixed monitoring points through a traffic monitoring system, and continuously receives trajectory point sequences of moving vehicles from a vehicle positioning system. The server projects these vehicle trajectory points onto specific road segments on the road network topology, calculates the average road traffic speed and effective traffic flow for each road segment within the current time slice, and constructs a road congestion level index by combining this with the planned free-flow speed. For example, the road congestion level index can be defined as:

[0023] in, Indicates time Road sections at all times The road congestion level index has a value range of greater than or equal to 0 and less than or equal to 1. The closer the value is to 1, the higher the degree of congestion. Indicates time Road sections are obtained in real time based on vehicle location trajectory and traffic monitoring data. Average road traffic speed; Indicates road segment The reference road speed under free-flow conditions is obtained from historical low-traffic periods. The server also analyzes the road closure areas and construction control sections provided by the traffic management department, and obtains a set of feasible detour routes from the map service platform. It then uniformly associates road congestion level, road speed, road closure area, and detour route information with the road segments covered by storage nodes, trunk nodes, and terminal nodes, thereby forming structured road traffic status data.

[0024] The server obtains regional wind speed levels, rainfall intensity, snowfall levels, and meteorological disaster warning signals through interfaces connecting to national or local meteorological monitoring platforms. Simultaneously, it collects data on road surface water level, visibility, and surface displacement from environmental monitoring terminals installed around key road sections or nodes. The server spatially aligns various raw environmental observations according to geographic grids or administrative divisions, mapping each storage node, trunk node, and terminal node to its corresponding environmental grid. It integrates wind speed, rainfall, snowfall, water accumulation, and geological disaster warning signals for the node's location to construct environmental disturbance status data. To more intuitively utilize the degree of environmental disturbance in the cost accounting model, the server can internally construct comprehensive environmental disturbance indicators, such as:

[0025] in, Indicates the region In time The comprehensive environmental disturbance index at any given time is limited to a range of values ​​greater than or equal to 0 and less than or equal to 1 or greater than or equal to 0 and less than or equal to 10, depending on the engineering requirements. , , , , Representing regions In time The wind speed level function, rainfall intensity function, snowfall level function, water accumulation degree function, and geological disaster warning level function corresponding to each moment are obtained by piecewise mapping or normalization of the original observations, and their values ​​are limited to the interval greater than or equal to 0 and less than or equal to 1. to The environmental disturbance weight parameter reflects the impact weight of different environmental factors on the transportation network. The parameter value ranges from 0 to 1, and the sum of all weights is 1. The server uses this comprehensive index to mark the current environmental disturbance state of each node and stores the detailed raw observations along with the comprehensive index in the environmental disturbance state data.

[0026] The server periodically acquires structured or semi-structured policy texts through interfaces with information systems of government departments, traffic management departments, and emergency management departments. These texts include rules on vehicle restrictions, checkpoint control areas, epidemic prevention and control lockdown areas, temporary road closures during major events, and special control rules for the transportation of hazardous materials. The server performs structured parsing on the announcements or notices, extracting elements such as the effective time interval, administrative division scope, road scope, vehicle type restrictions, permit requirements, and emergency passage levels, and mapping them to the corresponding warehouse nodes, trunk nodes, and terminal nodes. For nighttime restriction policies, the server discretizes the permitted and prohibited passage periods into time slices, marking the policy accessibility status of each node in different time slices. Finally, the server encodes the aforementioned time interval information, spatial scope information, and control rules into node-level and path-level policy control status data, used to subsequently determine whether a candidate path has legal passage conditions in a given time slice.

[0027] The server receives the order creation time, shipping address, delivery address, cargo weight, volume, and service time requirements for each order from the order system, and resolves the shipping and delivery addresses to corresponding warehouse nodes or end-point nodes. The server combines historical routing rules, current road traffic status data, and transportation capacity distribution data to calculate the set of possible flow paths for orders between warehouse nodes, trunk nodes, and end-point nodes, and counts the number and weight of orders for each region or each path in each time slice. This is to measure the time... The server can define an order density function to determine the order density at any given time, as follows:

[0028] in, Indicates the region In time The order density at any given time; a higher value indicates a greater degree of order accumulation within a unit area. Indicates time Time zone The number of orders counted within the time slice can be defined based on the number of new orders generated during that time slice or the number of orders in transit within that area. Indicates the region The area or a reference value used to measure the size of the region can be an equivalent scale of geographical area, service population, or business coverage. The server simultaneously performs time-series analysis on order density, identifying peak and off-peak periods, and forming the spatial and temporal distribution pattern of order flow, thus constituting order flow distribution data.

[0029] The server receives real-time data from the vehicle-mounted terminal and dispatch system, including the current location of each transport vehicle, the weight and volume of the currently loaded cargo, the vehicle's maximum allowable load capacity, the driver's current task status, and remaining available working hours. It also receives data from the warehouse management system regarding the number of available sorting lines, sorting rates, and staff on-duty status at each warehouse node. Simultaneously, the server obtains the remaining fuel or electricity for each vehicle from the vehicle energy monitoring system. By integrating vehicle quantity, vehicle load capacity, driver available working hours, and warehouse node sorting capacity across regional and route dimensions, the server constructs a model depicting the sufficiency and strain of transportation resources. For example, an internal vehicle capacity utilization rate indicator can be constructed, as follows:

[0030] in, Indicates vehicle In time The capacity utilization rate at any given time is within the range of 0 to 1. Indicates vehicle In time The equivalent load of the cargo already loaded at any given time can be obtained by combining the weight and volume of the cargo and performing an equivalent conversion. Indicates vehicle The maximum allowable load capacity. The server generates capacity resource distribution data for each node and each route in the current time slice by statistically analyzing the distribution of vehicle capacity utilization in different areas, combined with warehouse sorting capacity and drivers' remaining available working hours, for subsequent assessment of whether emergency dispatch needs to be triggered.

[0031] The server retrieves key risk data from historical periods from the claims system, after-sales system, and operational risk control system. This includes the time, location, package type, compensation amount, and liability determination result for each claims event; the number of returned shipments and the distribution of return reasons for each region within a specific time period; the average delay duration and delay ratio for each route at different event stages; and the number of vehicles, resource consumption, and evaluation of handling effectiveness during emergency cross-regional deployments. The server reorganizes the above information according to event, region, and route dimensions, indexing different types of risk events such as natural disasters, traffic accidents, and regional lockdowns, and storing the corresponding delay characteristics, cargo damage characteristics, cost characteristics, and handling effectiveness characteristics in a structured manner. To quickly apply historical risk experience in real-time accounting, the server can construct historical risk intensity indicators, as follows:

[0032] in, Indicates the region Historical risk intensity indicators; Indicates the region Average delay time obtained from historical events; Indicates the region The statistical average or percentile of historical cargo damage compensation amounts; Indicates the region The return rate or probability of returns during historical events; Indicates the region Average resource consumption intensity in historical emergency deployments; This represents a combined function that maps multiple risk characteristics to a single risk intensity index. It can be a weighted sum function, a nonlinear combined function, or a mapping function based on scoring rules. The server uses this index to assign long-term risk labels to different regions and nodes, providing a foundation for subsequent retrieval of risk energy memories.

[0033] The server assigns a unique node number to each warehouse node, trunk node, and terminal node, and assigns a path number to each legal transportation route in the road network topology. Simultaneously, the server divides the timeline into consecutive time slices, each corresponding to a timestamp. For each timestamp, corresponding to the node number and path number, the server extracts road congestion level, road speed, and road closure information from road traffic status data; comprehensive environmental disturbance indicators and original environmental element values ​​from environmental disturbance status data; traffic restriction status and emergency passage level from policy control status data; order density and order volume from order flow distribution data; capacity utilization rate and resource redundancy from transportation resource distribution data; and historical risk intensity and corresponding risk labels from historical risk record data. This multi-source data is then aggregated into a high-dimensional feature vector. The server can formally represent the real-time data stream as follows:

[0034] in, This represents the set of real-time data streams maintained by the server. Indicates time Always targeting node number With path number The multi-source feature vector contains all or part of the fields of road traffic status, environmental disturbance status, policy control status, order flow distribution, transportation capacity resource distribution and historical risk records; This represents a set of node numbers, including all warehouse nodes, trunk nodes, and terminal nodes. This represents the set of route numbers, containing all possible legal transport routes; This represents a set of timestamps arranged in chronological order. Through unified node numbers, path numbers, and timestamp indexes, the server integrates heterogeneous data sources into a continuously evolving real-time data stream, providing a unified data foundation and query entry point for subsequent updates to resilience states, construction of risk energy propagation trajectories, and real-time dynamic calculation of express delivery costs.

[0035] S120. Initialize the resilience state of each node and dynamically adjust it when the real-time data stream is input, so that the resilience state decays and spreads to adjacent nodes to form a risk energy propagation trajectory.

[0036] Specifically, resilience state quantity is a numerical state description used to characterize the strength of a node's ability to maintain normal service and recover when facing external disturbances. The higher the resilience state quantity, the less likely the node is to fail and the easier it is to recover to normal operation in abnormal scenarios such as natural disasters, traffic congestion, and regional lockdowns. Conversely, the lower the resilience state quantity, the more fragile the node is and the more likely it is to experience problems such as sorting backlog, vehicles being unable to enter or exit, and delivery interruptions.

[0037] Risk energy propagation trajectory refers to the path record of risk energy spreading and evolving to other nodes in the transportation network over time. It is used to show the process of risk gradually expanding from the initial affected node to surrounding nodes, and then further propagating from the surrounding nodes to more distant nodes. At each time slice, the server determines whether the risk is spreading from one node to adjacent nodes based on the decay of resilience state quantities and the topological relationship between adjacent nodes, and records the path of this propagation on the time axis and spatial topology.

[0038] Furthermore, the server first establishes a structural attribute profile for each node, normalizing the number and availability of alternative paths available to the node as path redundancy, normalizing the average recovery time and success rate of the node in multiple historical anomalies as historical recovery speed, normalizing the amplification effect of node order volume on external disturbances as order density sensitivity, and normalizing the number of spare vehicles, spare drivers, and warehouse emergency handling capacity of the node under high load conditions as capacity reserve. After normalization, an initialization function for resilience state variables is constructed, for example, setting the node... The initial toughness state parameters are: in, Represents a node The initial toughness state quantity can be set to a range greater than or equal to 0 and less than or equal to 1. The larger the value, the stronger the anti-interference ability in the early stage of external disturbance. Represents a node The path redundancy score is obtained by normalizing factors such as the number of alternative paths around the node, path capacity, and path reliability. Represents a node The historical recovery speed score is obtained by normalizing factors such as the node's average recovery time, recovery success rate, and recovery stability in historical events; Represents a node The order density sensitivity score is used to characterize the strength of the amplification effect of small fluctuations in order volume on node operation pressure. Represents a node The capacity reserve score is obtained by normalizing factors such as the number of spare vehicles, spare driver working hours, and emergency sorting capacity. , , , A weighting parameter is assigned to resilience, which controls the influence of different structural properties on the resilience state quantity. The weighting parameter takes values ​​in the range of 0 to 1, and their sum is set to 1. The server uses this initialization function to map structural properties to initial resilience state quantities, so that each node has a quantifiable level of resistance to disturbances before external disturbances occur significantly.

[0039] At each time slice, the server extracts multi-source perturbation information related to the nodes from the real-time data stream and constructs the node structure. In time The comprehensive disturbance level index for each moment reflects traffic congestion or interruption as the traffic disturbance level, heavy rainfall, heavy snowfall, typhoon or geological disaster warnings as the environmental disturbance level, road closures, traffic restrictions or control measures as the policy disturbance level, sudden increases in order volume or abnormal order density as the order disturbance level, and severe vehicle shortages or insufficient warehousing processing capacity as the capacity disturbance level. This comprehensive index is then used to derive the node disturbance level function. in, Represents a node In time The overall disturbance level at any given time can be set to a range of greater than or equal to 0 and less than or equal to 1 or greater than or equal to 0 and less than or equal to 10. The larger the value, the stronger the external disturbance. , , , , Representing nodes respectively In time The traffic disturbance level, environmental disturbance level, policy disturbance level, order disturbance level, and capacity disturbance level at any given time are obtained by segmenting or normalizing the raw quantities from the corresponding data sources. to The perturbation weight parameter controls the impact of different types of perturbations on the overall perturbation level. The weight parameter ranges from 0 to 1, and their sum is set to 1. In each time slice, the server performs a decay update on the resilience state variables based on the overall perturbation level and the time step, employing a gradual decay strategy. For example, it can be maintained internally.

[0040] in, Represents a node In time The resilience state variables updated at each time step remain within the range of greater than or equal to 0 and less than or equal to 1. Indicates the previous time slice node The toughness state quantity; It represents the time interval between two adjacent time slices, which can be seconds, minutes, or other uniformly selected time scales; This represents the toughness decay sensitivity coefficient, used to control the influence of the overall disturbance level on the decay rate of toughness state quantities. Its value ranges from 0 to 1. The above update relationship reflects that the higher the overall disturbance level and the longer the duration, the more significant the decay of toughness state quantities, thus naturally forming a dynamic adjustment result of "decay according to disturbance level and disturbance duration". The server can also record the toughness change for each time slice as follows:

[0041] in, Represents a node In time The incremental change value of the toughness state quantity at time , when A negative value indicates the attenuation effect, providing a quantitative basis for the subsequent propagation of the attenuation effect.

[0042] The server maintains a topological adjacency matrix between nodes in the network topology, describing which nodes are directly connected by roads or transportation routes. It also assigns a propagation attenuation coefficient to each pair of adjacent nodes, characterizing the degree of attenuation and directional preference of risk transmission from one node to its neighbors. For any given node... When significant toughness degradation occurs within a certain time slice, i.e. When the value is negative and the absolute value is large, the server treats this decay change as a source of decay and distributes it to adjacent nodes according to the topological adjacency relationship. Propagation, and the intensity of propagation can be described internally as follows:

[0043] in, Indicates time Time Node The intensity of risk energy input received from all neighboring nodes; Represents a set of nodes; Represents a node With nodes The topological adjacency indicator between nodes With nodes The value is 1 if there is a direct road connection, and 0 otherwise. Represents a node To the node The propagation attenuation coefficient, which describes the tightness of the risk propagation path between two nodes, is a value that is greater than or equal to 0 and less than or equal to 1. It can be preset according to factors such as path length, road grade, and connection importance. Indicates only if node In time The attenuation effect only propagates outward when resilience decay occurs, and the propagation intensity is directly proportional to the attenuation magnitude. The server, based on... The risk energy input of each node is accumulated on the time axis, and combined with the changes in the resilience state of the node itself, the degree of impact of each node and the affected source node in different time slices are continuously recorded. This generates a risk energy propagation trajectory, which is a sequence composed of a series of node risk energy input intensity distributions in the time dimension. This sequence is used to depict how the risk spreads from the initial affected node to a larger area along the topological adjacency relationship.

[0044] S130. Based on the resilience state quantity and the risk energy propagation trajectory, the express delivery cost is decomposed into basic freight cost, risk premium cost and emergency dispatch compensation cost.

[0045] Specifically, the base freight cost refers to the regular cost required to complete a single express delivery shipment without considering abnormal risks and emergency dispatch factors. It is the most fundamental part of the cost breakdown and generally consists of distance, cargo weight and volume, normal vehicle fuel or electricity consumption, normal driver working hours, and regular toll fees and warehousing handling fees. The risk premium cost refers to the additional cost added to the base freight cost after considering abnormal risk factors such as natural disasters, traffic paralysis, and regional lockdowns, to cover potential delays, damage, returns, and claims. The magnitude of the risk premium cost is closely related to resilience status and risk energy propagation trajectory. When the resilience status of nodes along a route is generally low, and the risk energy propagation trajectory indicates that the route is in a risk front or high-risk concentrated area, the risk premium cost of that route will increase significantly to reflect the additional risk pressure borne by that route. Emergency dispatch compensation cost refers to the additional cost incurred when normal transport capacity and regular routes cannot meet service requirements, in order to maintain business continuity through emergency transport capacity allocation and special guarantee measures. Emergency deployment and compensation costs are closely related to the spread and duration of high-risk areas in the risk energy propagation trajectory. If the risk energy propagation trajectory indicates that a certain area is in a state of severe risk for a long time, and regular vehicles cannot enter or their frequency of entry and exit is greatly reduced, then the platform may need to call vehicles from other areas, temporarily activate backup warehouses, arrange night shifts, or apply for special passes. These actions all bring significantly higher additional costs than regular operations.

[0046] Furthermore, based on historical normal operating data, the server establishes a basic resource consumption model for each candidate path, quantifying path length, standard runtime, standard vehicle fuel or electricity consumption, regular toll fees, and regular warehousing and handling costs as the path's basic resource consumption. This model is then combined with resilience state variables related to the path within the current time slice to determine whether the path is in a stable range. Specifically, the server can analyze the path... In time Modeling the base fare cost for each time slot, for example: in, Representing a path In time The base freight cost at any given time is used to characterize the level of resource consumption required to complete the transportation task along this route under normal operating conditions; Representing a path In time The stability indicator function at any given time takes a value of 1 when all the resilience state quantities of the nodes constituting the path are within the preset stable range. When the resilience state quantity of any critical node falls out of the stable range, the value can be reduced to a reduction value less than 1 or taken as 0 to indicate that the path is no longer priced according to the completely normal operating condition. Representing a path Total driving distance; Representing a path The standard operating time under normal working conditions can be obtained from historical data without anomalies. Representing a path Standard fuel or electricity consumption under normal operating conditions; Representing a path Standard loading, unloading and warehousing operation hours or workload under normal working conditions; , , , These represent the unit distance cost coefficient, unit time cost coefficient, unit energy consumption cost coefficient, and unit operating cost coefficient, respectively, with values ​​preset based on factors such as vehicle type, labor costs, and regional prices; the above relationships are maintained when the toughness state quantities are stable. Approaching 1, the basic freight cost is mainly determined by physical distance and conventional resource consumption. When the resilience state quantity decreases significantly, the stability indication function is reduced or turned off, and the subsequent cost increment is borne by risk premium cost and emergency dispatch compensation cost, thereby achieving a structural separation between the "normal operating condition part" and the "abnormal operating condition part".

[0047] The server identifies the relative position of each path to the risk front in the current time slice based on the risk energy propagation trajectory. Nodes closer to the risk front and located in high-density risk energy areas are marked as high-risk nodes, and the server calculates the cumulative attenuation and duration of resilience state quantities of these high-risk nodes on the time axis. For a path... In time The risk premium cost at any given time can be established using the following function: in, Representing a path In time The risk premium cost at any given moment is used to characterize the additional costs that need to be accrued due to increased risk. , , Representing paths In time The probability of delay, damage, and return at any given moment is estimated by comprehensively considering the degree of resilience state decay of the nodes on the time axis, the duration of decay, and the minimum distance from the node to the risk front. The greater the resilience decay, the longer the duration, and the closer to the risk front, the higher the probability. , , These represent a delay event, a damage event, and a return event in the path, respectively. The average level of loss can be obtained from historical risk record data, including the amount of compensation paid, additional processing costs, and the converted value of brand reputation loss; , , The risk sensitivity coefficient is used to adjust the weight of different types of risks in the cost. Its value is preset by the enterprise based on its sensitivity to timeliness default, goods integrity, and customer returns. The above relationship maps the resilience state quantity decay information and the spatial location of the risk front to the probability of risk occurrence, and then multiplies it with the corresponding expected loss. This makes the risk premium cost actually equivalent to the weighted sum of the expected values ​​of delay, damage, and return losses under the current risk environment, thus naturally reflecting the trend of the overall increase in the probability of the three types of risks when "the resilience is weaker, the closer to the risk front, and the longer the exposure time".

[0048] The server continuously monitors the resilience status and capacity resource distribution data of each node. When the resilience status of a node consistently falls below a preset threshold, and road access status data and policy control status data indicate that its entrance and exit roads are restricted, causing the node's reachability to drop to a point where it cannot meet the current order flow distribution requirements, the system determines that the node has entered an emergency state and triggers cross-regional scheduling and temporary capacity expansion measures based on the resource redundancy of upstream and downstream nodes; for the path In time The emergency deployment and compensation cost corresponding to a given moment can be expressed as follows: in, Representing a path In time Emergency deployment and compensation costs in real time are used to characterize the additional resource consumption required to maintain the transportation capacity of the route or related nodes during a sudden event. This indicates a guaranteed path within the time window. The number of emergency vehicle cross-regional dispatches or the number of vehicles dispatched across regions that are available and triggered can be obtained by statistical analysis of emergency dispatch instructions recorded by the dispatch system. This indicates a path within the time window. The number of temporary passage permits applied for and actually used, or the number of times special passage tickets are used, is used to reflect the frequency of the activation of temporary passage fees; This indicates a path within the time window. The total working hours for nighttime overtime work and related nodes are used to reflect the intensity of nighttime work input; , , These represent the cost coefficients for single emergency vehicle cross-regional dispatch, unit temporary pass or special pass, and unit nighttime operation hours, respectively. The values ​​are determined by a combination of vehicle operating costs, policy fee standards, and labor overtime costs.

[0049] S140. Construct a counterfactual cost channel as a benchmark cost trajectory, compare the deviation between the benchmark cost trajectory and the actual network cost changes, and adjust the risk premium cost and emergency allocation compensation cost based on the deviation results.

[0050] Specifically, the counterfactual cost path is a "hypothetical cost evolution path" constructed in this embodiment to simulate the cost changes that the express delivery network should exhibit under the premise of no abnormal events such as natural disasters, traffic paralysis, or regional lockdowns. The counterfactual cost path does not directly originate from real-time data in the current abnormal scenario, but rather from historical road traffic status data, order flow distribution data, capacity resource distribution data, and the stable performance of node resilience states under normal operating conditions. Through modeling and prediction, it provides a reference result of "how the cost should have changed if there were no abnormal disturbances." The baseline cost trajectory refers to the cost sequence changing over time along the counterfactual cost path, used to characterize the "normal trajectory" that the cost should exhibit over time for a certain path or a certain type of order under the assumption of no abnormal disturbances. The real network cost change refers to the actual cost change over time, calculated comprehensively based on real-time data flow, resilience states, risk energy propagation trajectory, risk premium costs, and emergency dispatch compensation costs under real-world operating conditions.

[0051] Furthermore, when training the baseline cost trajectory model based on node resilience state variables, base freight costs, risk premium costs, and emergency dispatch compensation costs under historical normal operating conditions, the server no longer uses a simple linear mapping. Instead, it constructs a temporal kernel mapping model with time memory and path structure weights. For the reference cost of any path at any time slice, the server selects multi-time feature vectors within a time window from the historical normal operating condition samples and weights them with time kernel weights and path structure weights to form a nonlinear reference cost estimate. For example, a baseline cost trajectory model of the following form can be constructed:

[0052] in, Indicates the path under conditions without abnormal disturbances. In time Reference cost value for each time slot; This represents a nonlinear compression function used to compress the internal linear combination result to a preset cost range. Its output is monotonically increasing and its value range is limited to a finite interval. Indicated by time A set of historical time windows, constructed with the center or the end, is used to extract time slices similar to the current moment from historical normal operating conditions; This represents the time kernel weighting function, used to calculate weights based on historical time points. Reference time Different weights are assigned to the time intervals between them, with values ​​ranging from 0 to 0, and the closer the time interval, the greater the weight. Indicates the path under normal operating conditions In time The feature vector at any given time contains normalized features such as node resilience state quantity, basic freight cost, historical average risk premium cost, and historical average emergency dispatch compensation cost. Representing a path The set of nodes participating in the modeling; This represents the global feature weight vector, used to characterize the impact of stable features related to the overall path on the cost; Indicates nodes on the path The relevant local feature weight vector is used to reflect the different contributions of different nodes to the path cost; Represents nodes The corresponding feature transformation matrix is ​​used to perform linear projection or feature reconstruction on the feature vector to extract features related to the nodes. More relevant feature components; Represents a node In time The toughness state quantity at time step is vectorized and used as an element-wise modulation factor after feature transformation; This represents the element-wise multiplication operation of vectors, used to couple the feature transformation result with the node resilience state variables element-wise.

[0053] The server not only calculates the instantaneous difference between the two, but also introduces time weighting and path-dependent constraints to form a multi-scale bias result. For the path... In time The server first obtains the instantaneous normalized deviation, and then performs smooth aggregation within a local time window using a time kernel and path weights. For example, it can construct...

[0054] in, Representing a path In time The instantaneous normalization deviation at any given moment reflects the degree of deviation of the actual network cost from the benchmark cost; Representing a path In time The real-time network cost is calculated by adding up the base freight cost, the real-time risk premium cost, and the real-time emergency dispatch compensation cost. The reference cost output by the aforementioned baseline cost trajectory model; It is a very small positive constant used to avoid numerical instability when the base cost is close to zero; Representing a path In time The time-smoothing deviation results at any given moment comprehensively characterize the magnitude of the deviation within a local time window and the co-deviation of neighboring paths; This represents the set of time windows used for bias smoothing, the range of which may differ from the time windows of the baseline model; This represents the deviation smoothing kernel function, which is used to assign different time weights to deviations at different historical moments, with higher weights for deviations closer to the current moment. Representing a path The weight of the path's own deviation is used to control the importance of the path's own deviation in the aggregation result; Representation and path A set of adjacent paths that are coupled in network topology or business logic, such as a set of paths that share critical nodes or critical road segments; Representing a path and adjacent paths The deviation propagation weight is used to characterize the effect of adjacent path deviation on the path. The intensity of the impact of the overall deviation.

[0055] The server introduces positive and negative deviation accumulation functions, performs signed integration on deviations over a period of time, and uses exponential or logistic modulation functions to nonlinearly amplify or compress cost weights. This makes risk premium costs and emergency allocation compensation costs more sensitive to long-term deviations and relatively less sensitive to short-term noise deviations. For example, the deviation accumulation and cost adjustment relationship can be constructed in the following form:

[0056] in, Representing a path In time The length of time that has passed is The time-weighted cumulative amount of positive deviation within the time window, only for Integrating the positive portion indicates the cumulative strength of the actual cost consistently exceeding the baseline cost trajectory; Representing a path The time-weighted cumulative amount of negative bias within the same time window, only for Integrating the negative portion represents the cumulative strength of the actual cost falling below or gradually returning to the baseline cost trajectory over a certain period of time. This indicates the length of the deviation accumulation time window, used to control the time range for deviation memory. This represents the deviation accumulation kernel function, which assigns different memory weights to the deviations at different times within the window, with higher weights for times closer to the current time. Indicates the path before adjustment In time Risk premium cost at any given moment; This represents the risk premium cost adjusted for accumulated deviations. The positive deviation enhancement coefficient representing the risk premium cost, when When the risk premium cost increases, it is amplified more rapidly through an exponential function. The negative deviation reduction coefficient representing the risk premium cost, when As the risk premium increases, it decreases through an exponential function. Indicates the path before adjustment In time The cost of emergency deployment and compensation at any time; This represents the emergency allocation compensation cost after adjusting for accumulated deviations. The positive deviation enhancement base coefficient represents the cost of emergency allocation compensation and is used to control the sensitivity of positive deviation to the increase in emergency allocation compensation costs. The negative deviation reduction base coefficient represents the cost of emergency allocation compensation, and is used to control the sensitivity of negative deviation to the reduction of emergency allocation compensation costs. This represents the saturation control parameter for adjusting the emergency allocation compensation cost, ensuring that the enhancement or weakening effect tends to saturate when the deviation is extremely large, thus avoiding unlimited cost amplification or excessive compression.

[0057] S150. Generate a cost probability cloud for each package, and shift the cost probability cloud according to the changes in resilience state quantity and risk energy propagation trajectory to generate a potential cost increase trend.

[0058] Specifically, the cost probability cloud, in this embodiment, is a probabilistic representation of the various cost outcomes and their distribution that may occur during the future transportation of a single package. Instead of using a single, fixed cost value to represent the package's transportation cost, it uses a comprehensive "cost distribution cloud" to represent the possible cost levels and their likelihood under different scenarios, route conditions, and risk levels. The horizontal axis of the cost probability cloud can be understood as a cost range, from lower to higher costs, while the vertical axis can be understood as the probability density of each cost level.

[0059] The potential cost increase trend is based on the cost probability cloud offset. It proactively characterizes the trend of "a relatively high probability of higher costs in the future" before costs have actually risen to their highest level. The potential cost increase trend does not only look at whether costs have increased at a certain moment, but also whether the probability of high cost intervals in the cost probability cloud is continuously increasing, and whether the center of the cost probability cloud is continuously moving towards the high cost side across multiple time slices, thus determining that costs are in an "upward channel".

[0060] Furthermore, the server establishes a two-dimensional index of path sets and time slice sets for each target package. The cost results of the package under different candidate path and time slice combinations are treated as a set of discrete cost samples, and then smoothed into a continuous cost distribution through kernel density mapping. Specifically, for each package... In this regard, let its candidate path set be... The relevant time slice set is , in the path and time slices The cost sample under the combination is The cost sample is composed of the base freight cost, risk premium cost, and emergency dispatch compensation cost, with weights... Used to characterize wrapping in the path Time slice The probability of actual or planned adoption under a combination, bandwidth parameters. Kernel function is used to control the smoothness of the distribution. To expand each cost sample into a locally smooth distribution, the package... In time The corresponding cost probability cloud can be expressed in the form of kernel density superposition as follows:

[0061] in, Indicates time Always targeting packages The cost probability density function, where the cost takes values ​​of The higher the probability density at a given time, the more likely that cost level is to occur. and Packages The set of paths and time slices that may be encountered in business planning; Indicates package via path And in time slice The cost of completing the transportation is calculated by summing the base freight cost, risk premium cost, and emergency dispatch compensation cost for the corresponding route and time slice. This represents the weight of the cost sample, used to characterize the package along the path. Time slice The probability of occurrence under a combination is such that the weights sum to 1 across all path and time slice combinations. This represents the bandwidth parameter used to control the smoothing scale of the cost space. The larger the bandwidth, the smoother the cost probability cloud, but the weaker the detail. The kernel function can be selected, such as a Gaussian kernel or other symmetric kernel functions, to extend discrete cost samples into a continuous distribution.

[0062] The server first calculates the deviation between the expected value of the package cost probability cloud and the base freight cost in the current time slice. Then, it uses a translation transformation to align the probability cloud to near the base freight cost, ensuring that, under the premise of stable resilience states, the cost distribution primarily reflects routine resource consumption rather than abnormal risks. Specifically, for packages... In time Cost probability density Its expected cost can be expressed as:

[0063] The base fare cost for the current time slice is then denoted as... And construct the translation amount: When the resilience state variables of all nodes related to the package path are determined to be within a preset stable range, the server performs a translation transformation: The center of the cost probability cloud is moved from Realign to ,in This represents the aligned cost probability density. Therefore, the first potential cost increase trend indicator can be defined as follows:

[0064] When the toughness state is stable and When the value is close to 0, it indicates that the expected cost of the package is close to the base freight cost, and the cost fluctuation is in a low range. This represents the average cost level under the current cost distribution. This is the base freight cost under the current route and business conditions; The center offset that needs to be corrected; To execute the aligned probability cloud. By continuously performing this center alignment and trend monitoring operation during the resilient and stable phase, the server can ensure that the cost probability cloud closely follows the base freight cost when there is no significant risk, so that the first potential cost increase trend reflects a state of "no significant increase trend" or "only slight natural fluctuations".

[0065] The server will perform asymmetric deformation of the cost probability cloud in the high-cost range based on the decay degree of the node resilience state quantity and the relative position of the package path in the risk energy propagation trajectory. This stretches the high-cost tail and shifts the overall centroid towards the high-cost direction. To this end, the server first performs a transformation on the package in the current time slice. Set a fee threshold This threshold can be determined by a combination of the base freight cost and the historical high-risk scenario cost quantiles. For the original cost random variable... Define the nonlinear transformation as follows:

[0066] in, Indicates time For packages The cost transformation function transforms the original cost. Mapping to the transformed cost ; This is a first-order offset coefficient used to control the strength of linear tension after exceeding the cost threshold; its value is a non-negative number. The larger the value, the stronger the overall high-cost range is pushed towards higher costs; This is a quadratic offset coefficient used to control the degree of nonlinear amplification of costs above the threshold in the tail region; its value is a non-negative number. The larger the value, the more pronounced the accelerated stretching occurs in the extremely high cost range; This ensures that stretching is applied only to cost samples exceeding a threshold, while remaining unchanged for samples below the threshold. The deformed probability density is obtained by performing variable substitution on the cost probability cloud, as follows:

[0067] in, This represents the cost probability density after performing asymmetric stretching. Indicates at a fixed time The inverse mapping of cost by the time transformation function. This represents the local scaling factor of the cost under the inverse mapping, used to ensure that the integral of the probability density after the variable transformation is still 1. The centroid of the transformed cost distribution is then defined as follows:

[0068] A second potential cost increase trend indicator is constructed as follows: When the node resilience state quantity continues to decay and the risk energy propagation trajectory wavefront enters the package path region, the server dynamically adjusts based on the resilience decay intensity and the risk energy accumulation level. and Thus Significantly greater than ,Right now If the value remains positive and continues to increase, the upward trend of the second potential cost clearly reflects the state that "costs are on an upward channel and the tail risk of high costs is rapidly increasing."

[0069] The server needs to perform a shrinking transformation on the previously stretched cost probability cloud during the risk mitigation phase. This will gradually reduce the variance of the cost distribution and the probability of high-cost tails, and reposition the overall centroid closer to the base freight cost area. This avoids making overly pessimistic cost forecasts even after the abnormal event has subsided. To this end, the transformed cost probability density can be targeted... Define a time-varying contraction factor. With the center of regression During the recovery phase of the toughness state, the contraction factor is calculated based on the retreat degree of the nodal toughness state and the risk energy propagation trajectory. and define the transition center fee. To create a smooth transition between the base freight cost and the current expected cost, a shrinkage transformation can be constructed as follows:

[0070] in, For the shrinking transformation function, the current cost is... Map to closer The shrinkage value; This indicates the intensity of contraction; when it is close to 1, the contraction is weak; when it is close to 0, the contraction is strong. This represents the reference center of the current distribution, for example, it can be set to... Or its smoothed version; This represents the target equilibrium center, used for iterative regression between the base freight cost and the current expectation. After performing variable substitution, the shrunken cost probability density is obtained as follows:

[0071] in, This represents the cost probability density after the contraction. It is the inverse mapping of the shrinking transformation. Let be the local scaling factor for the shrinking transformation. Then, the expected cost of the shrunken distribution is calculated as follows:

[0072] And define a third potential cost increase trend indicator. As the resilience state gradually recovers and the risk energy propagation trajectory recedes away from the package path area, the server adjusts the contraction factor. Gradually decrease, center of equilibrium Gradually approaching the base freight cost, prompting Towards To get closer, to make The cost probability cloud gradually decays to near zero, thus smoothly shrinking from a high-risk, high-expectation state to a normal, low-risk state, avoiding excessively exaggerated expectations of cost increases even after abnormal events have subsided.

[0073] S160 identifies the types, levels, and directions of abnormal events based on similarity matching of disturbance patterns and historical risk energy memory, and links them with resilience state quantities, cost probability clouds, counterfactual cost channels, and potential cost increase trends to output express delivery cost calculation results.

[0074] Specifically, at the node or region level, the server maps environmental disturbance state data and policy control state data within the same time slice into a unified disturbance feature vector, and superimposes a memory effect at the time dimension to construct the current disturbance pattern. Let the region or node be... The time is Then the original perturbation feature vector can be constructed as follows:

[0075] in, Represents a region or node In time The perturbation feature vector at time step; , , , These are normalized indices representing the intensity of wind speed disturbance, rainfall disturbance, snowfall disturbance, and water accumulation disturbance, respectively, with values ​​limited to the range of greater than or equal to 0 and less than or equal to 1. This indicator represents the intensity of regional lockdown measures and is used to characterize whether the region is under lockdown and the degree of strictness of the lockdown. Indicates the intensity of traffic restriction or traffic control operations; This indicates a special access control level. To incorporate time memory, the server uses a sliding time window. The perturbation pattern vector is obtained by performing a kernel-weighted integral on the perturbation features:

[0076] in, Represents a region or node In time The perturbation mode vector at time t is used to describe the recent perturbation evolution in this region; Indicates the length of the time window for constructing the perturbation pattern; This represents the time kernel weighting function, used to distinguish the contribution of different historical moments to the current perturbation mode. Generally, it takes a larger value when the time difference is small and gradually decreases when the time difference is large. This represents the perturbation feature transformation matrix, used to extract combined features more relevant to event identification from the original perturbation features. Through the above combined expression, the server connects environmental perturbation state data with policy control state data into a perturbation pattern that reflects temporal evolution and the superposition of multiple factors, providing structured input for subsequent similarity matching with historical risk energy memory.

[0077] The server retrieves multiple typical event templates from the historical risk energy memory database. Each event template contains the disturbance pattern trajectory and risk energy propagation trajectory during the event. Let the set of historical event templates be... any of the templates Contained in a region or node ,time The perturbation mode on and risk energy distribution The server is at the current time. A similarity score is calculated for each template, as follows:

[0078] in, Indicates the current disturbance pattern and historical event template In the time window The overall distance index indicates that the smaller the value, the more similar the two people are. Indicates the set of regions or nodes participating in the comparison; Indicates the length of the time window for similarity matching; This represents the similarity time weighting function, used to emphasize perturbation segments that are closer to the current time. This represents a weighting matrix in the perturbation feature space, used to adjust the influence weights of different perturbation components in the distance metric. The server is based on... The normalized similarity is constructed as follows:

[0079] in, This indicates that the current disturbance pattern belongs to the event template. The similarity probability is greater than or equal to 0 and less than or equal to 1, and the sum of the similarity probabilities on all templates is 1; This represents the similarity sensitivity coefficient, used to control the strength of the nonlinear mapping between distance and similarity.

[0080] The server will amplify the risk effect through three interconnected aspects: updates to resilience state variables, adjustments to cost-probabilistic cloud architecture, and the weighting of risk premium costs and emergency deployment compensation costs. Let the set of nodes be... ,node In time The toughness state quantity is The event matching results are in the node The event association weight is This is used to characterize the influence of the event propagation direction and package path on the node; when the event level is increased or the propagation direction covers the package path, the server adjusts the time step accordingly. Perform accelerated decay updates:

[0081] in, This indicates the node under the influence of the first express delivery cost calculation result. The toughness state quantity in the next time slice; This represents the toughness acceleration factor, which is a non-negative number. The larger the value, the faster the toughness decays under the same level of disturbance. Represents a node In time The strength of the association affected by the current event at any given time. This value is larger when the node is located in the main channel of event propagation and on the package path, and smaller otherwise. This refers to the comprehensive disturbance level index constructed in the aforementioned scheme, used to quantify the intensity of the combined effects of environmental disturbances, traffic disturbances, and policy controls on the current node. For the cost probability cloud, the server increases the high-cost range stretching parameter of the corresponding package along the event-related path, such as the aforementioned... , The structure is designed as an event-driven mechanism, allowing it to follow... Cumulative index of positive deviation This increases, leading to stronger asymmetric stretching in the high-cost direction. Regarding risk premium costs and emergency allocation compensation costs, when the server indicates amplified deviations in the deviation results, it represents the path... Introducing event-driven amplification terms, such as constructing an adjustment relationship for risk premium costs:

[0082] in, This indicates the route under the First Express shipping cost calculation results. In time Risk premium cost at any given moment; This represents the risk premium cost adjusted by the bias mechanism when the results of anomaly identification are not taken into account. This represents the risk premium amplification sensitivity coefficient, used to control the combined amplification of the risk premium by deviation amplification and event severity escalation; A comprehensive indicator representing the correlation between positive deviation and the event can be obtained by combining the positive cumulative amount of the deviation result with the weight of the path in the event propagation direction. Emergency deployment compensation costs can be similarly amplified by introducing an event amplification term, causing the resilience state of the corresponding region to decay more rapidly and the cost probability cloud to shift to the high-cost side more quickly when the event level increases and the propagation direction points towards the package path. Risk premium costs and emergency deployment compensation costs increase synchronously during the deviation amplification phase, thus enabling a rapid response to high-risk situations.

[0083] The server will adjust resilience state variables, cost probability cloud, risk premium costs, and emergency deployment compensation costs in reverse order, so that the cost structure can gradually return to normal as the event mitigates. Let the event mitigation correlation weight be... Used to characterize nodes The degree to which the node is affected by "risk reduction" during the event retreat phase is defined by the upper bound of the target resilience of the node under theoretically stable conditions. Then in the time step Internal execution recovery trend update:

[0084] in, This indicates the node under the influence of the second express delivery cost calculation results. The toughness state quantity in the next time slice; This represents the resilience recovery sensitivity coefficient, which is a non-negative number. The larger the value, the faster the recovery speed under the same mitigation intensity. Represents a node In time It is constantly affected by the positive recovery of the event level decreasing and the direction of diffusion moving away, and takes a larger value in the area through which the event wave recedes; Represents a node The upper limit of target resilience under normal operating conditions. For cost-probabilistic clouds, the server stretches the previously amplified high cost parameters. , Cumulative index based on event mitigation level and negative deviation The cost is gradually reduced, and a contraction factor is introduced to gradually lower the probability of high-cost tails and pull the cost distribution center closer to the base freight cost. For risk premium costs and emergency dispatch compensation costs, the server determines the path when the deviation results indicate deviation convergence. Introducing a decay adjustment term, for example, it can be constructed for risk premium costs:

[0085] in, This indicates the path under the second express delivery cost calculation result. In time Risk premium cost at any given moment; This represents the sensitivity coefficient to the decay of the risk premium; This is a comprehensive indicator representing the degree of negative deviation and event mitigation. The indicator increases as the deviation continuously decreases over a period and the event severity declines, thereby gradually reducing risk premium costs through an exponential factor. Emergency deployment compensation costs can be similarly reduced by introducing a decay term with saturation characteristics. This ensures that during the event mitigation and deviation convergence phases, high levels of emergency costs are no longer maintained, gradually restoring the cost structure to a normal configuration dominated by basic freight costs, supplemented by risk premiums and emergency compensation.

[0086] The second aspect of this application provides a real-time dynamic calculation device for express delivery costs based on big data, referring to... Figure 2 , Figure 2This application provides a schematic diagram of a real-time dynamic calculation device for express delivery costs based on big data. The device is a server, comprising an acquisition module 21 and a processing module 22. The acquisition module 21 acquires road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk nodes, and terminal nodes to form a real-time data stream. The processing module 22 initializes the resilience state quantity of each node and dynamically adjusts it upon input of the real-time data stream, causing the resilience state quantity to decay and diffuse towards adjacent nodes, thus forming a risk energy propagation trajectory. The processing module 22 is also used to calculate the express delivery costs based on the resilience state quantity and the risk energy propagation trajectory. The cost breakdown includes basic freight cost, risk premium cost, and emergency dispatch compensation cost. Processing module 22 is also used to construct a counterfactual cost channel as a benchmark cost trajectory, compare the deviation between the benchmark cost trajectory and the actual network cost changes, and adjust the risk premium cost and emergency dispatch compensation cost based on the deviation results. Processing module 22 is also used to generate a cost probability cloud for each package, and shift the cost probability cloud based on the changes in resilience state quantity and risk energy propagation trajectory to generate a potential cost increase trend. Processing module 22 is also used to identify the type, level, and direction of abnormal events based on the similarity matching of disturbance patterns and historical risk energy memory, and link the resilience state quantity, cost probability cloud, counterfactual cost channel, and potential cost increase trend to output the express delivery cost calculation result.

[0087] This application also provides an electronic device, with reference to... Figure 3 , Figure 3This is a schematic diagram of an electronic device provided in an embodiment of this application. The electronic device may include: at least one processor 31, at least one network interface 34, a user interface 33, a memory 35, and at least one communication bus 32. The communication bus 32 is used to enable communication between these components. The user interface 33 may include a display screen or a camera; optionally, the user interface 33 may also include a standard wired interface or a wireless interface. The network interface 34 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The processor 31 may include one or more processing cores. The processor 31 connects to various parts of the server using various interfaces and lines, and performs various functions of the server and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 35, and by calling data stored in the memory 35. Optionally, the processor 31 may be implemented using at least one hardware form selected from Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). Processor 31 may integrate one or more of the following: a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip, without being integrated into processor 31.

[0088] The memory 35 may include random access memory (RAM) or read-only memory. Optionally, the memory 35 may include a non-transitory computer-readable storage medium. The memory 35 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 35 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 35 may also be at least one storage device located remotely from the aforementioned processor 31. Figure 3 As shown, the memory 35, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for real-time dynamic calculation of express delivery costs based on big data.

[0089] exist Figure 3 In the electronic device shown, the user interface 33 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 31 can be used to call the application stored in the memory 35, which is a real-time dynamic calculation method for express delivery costs based on big data. When executed by one or more processors, the electronic device executes one or more methods as described in the above embodiments.

[0090] This application also provides a non-transitory computer-readable storage medium storing instructions. When executed by one or more processors, these instructions cause an electronic device to perform one or more of the methods described in the above embodiments.

[0091] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for real-time dynamic calculation of express delivery costs based on big data, characterized in that, The method includes: Acquire road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk nodes, and terminal nodes to form a real-time data stream; The resilience state quantity of each node is initialized and dynamically adjusted when the real-time data stream is input, so that the resilience state quantity decays and spreads to adjacent nodes to form a risk energy propagation trajectory. Based on the resilience state quantity and the risk energy propagation trajectory, the express delivery cost is broken down into basic freight cost, risk premium cost, and emergency dispatch compensation cost. Construct a counterfactual cost channel as a benchmark cost trajectory, compare the deviation between the benchmark cost trajectory and the actual network cost changes, and adjust the risk premium cost and the emergency allocation compensation cost based on the deviation results; A cost probability cloud is generated for each package, and the cost probability cloud is shifted according to the changes in the resilience state quantity and the risk energy propagation trajectory to generate a potential cost increase trend. Based on the similarity matching between the disturbance pattern and the historical risk energy memory, the abnormal event type, event level and diffusion direction are identified, and linked with the resilience state quantity, the cost probability cloud, the counterfactual cost channel and the potential cost increase trend, the express delivery cost calculation result is output.

2. The method for real-time dynamic calculation of express delivery costs based on big data as described in claim 1, characterized in that, The acquisition of road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk nodes, and terminal nodes to form a real-time data stream specifically includes: Based on a global map service platform, a traffic condition monitoring system, and a vehicle positioning system, information on road congestion levels, road speeds, road closures, and detour routes is obtained to form the road traffic status data. Based on the meteorological monitoring platform and environmental monitoring terminal, wind speed, rainfall, snowfall, water accumulation and geological disaster early warning information are obtained to form the environmental disturbance state data; Based on traffic restriction announcements, temporary lockdown notices, emergency traffic level change records, and nighttime restriction policy change information, the aforementioned policy control status data is generated. Based on order generation records, order shipping address information, order flow path estimation, and order density trend analysis, the order flow distribution data is formed. The transportation capacity resource distribution data is formed based on vehicle location, vehicle load, driver working status, warehouse sorting capacity, and vehicle energy remaining. The historical risk record data is formed based on compensation history, return frequency records, delay duration statistics, resource allocation records and risk handling results during emergencies; Road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data are structurally integrated based on node numbers and path numbers combined with timestamps to form a real-time data stream with a unified indexing system.

3. The method for real-time dynamic calculation of express delivery costs based on big data as described in claim 1, characterized in that, The process of initializing the resilience state quantity of each node and dynamically adjusting it during the input of the real-time data stream, causing the resilience state quantity to decay and diffuse towards adjacent nodes to form a risk energy propagation trajectory, specifically includes: The resilience state of each node is initialized based on path redundancy, historical recovery speed, order density sensitivity, and capacity reserve, so that the resilience state can characterize the node's anti-interference ability in the early stage of external disturbances. The resilience state quantity is dynamically adjusted based on road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, and transportation capacity resource distribution data. When the external disturbance signal intensifies, the resilience state quantity is attenuated according to the disturbance level and duration, thus obtaining the attenuation effect. The attenuation effect is propagated along the topological adjacency relationship to adjacent nodes, forming the risk energy propagation trajectory.

4. The method for real-time dynamic calculation of express delivery costs based on big data according to claim 1, characterized in that, Based on the resilience state quantity and the risk energy propagation trajectory, the express delivery cost is broken down into basic freight cost, risk premium cost, and emergency dispatch compensation cost, specifically including: The base freight cost is determined based on the resilience state quantity in the stable range to characterize the resource consumption level under normal operating conditions; The risk premium cost is determined based on the degree of attenuation, duration of attenuation, and distance from the risk wave front of the toughness state quantity, in order to characterize the changing trends of the probability of delay, probability of damage, and probability of return. The emergency allocation compensation cost is generated based on the decrease in node accessibility or shortage of transportation resources caused by the decay of the resilience state quantity, which triggers emergency vehicle cross-regional dispatch, activation of temporary toll fees and nighttime operation. This is used to characterize the additional resource consumption to maintain the operation of the transportation link during emergencies.

5. The method for real-time dynamic calculation of express delivery costs based on big data according to claim 1, characterized in that, The construction of a counterfactual cost channel as a baseline cost trajectory, the comparison of the baseline cost trajectory with actual network cost changes, and the adjustment of the risk premium cost and the emergency allocation compensation cost based on the deviation results, specifically include: A benchmark cost trajectory model is trained based on the node resilience state quantity, basic freight cost, risk premium cost, and emergency dispatch compensation cost under historical normal operating conditions, so as to form a counterfactual cost channel based on the benchmark cost trajectory model under the condition of cost change without being affected by abnormal disturbances. In real-time operation, the baseline cost trajectory output by the counterfactual cost channel is compared with the actual network cost changes on the same time axis and the same path set to generate the deviation result; Based on the magnitude, direction, and duration of the deviation, the deviation is decomposed into the risk premium cost and the emergency allocation compensation cost to enhance or weaken the weight, so that the cost structure maintains a dynamic mapping relationship that matches the actual risk level of the network throughout the entire cycle of the abnormal event.

6. The method for real-time dynamic calculation of express delivery costs based on big data according to claim 1, characterized in that, The process of generating a cost probability cloud for each package and shifting the cost probability cloud based on changes in the resilience state quantity and the risk energy propagation trajectory to generate a potential cost increase trend specifically includes: Based on the set of routes and time slices of the package during transportation, the cost results corresponding to the basic freight cost, risk premium cost, and emergency dispatch compensation cost are formed into a continuous distribution to generate the cost probability cloud. When the resilience state quantity is in a stable range, the center position of the cost probability cloud is aligned with the base freight cost to obtain the first potential cost increase trend, so as to keep the cost fluctuation range low. When the resilience state quantity decays and the risk energy propagation trajectory advances to the package path area, the cost probability cloud is asymmetrically stretched and the centroid shifted to the side where the cost meets the preset threshold, resulting in a second potential cost increase trend, so as to reflect the high-risk cost increase state in advance. When the resilience state quantity recovers, the cost probability cloud is reduced to obtain a third potential cost increase trend, in order to avoid the continued overestimation of cost expectations.

7. The method for real-time dynamic calculation of express delivery costs based on big data as described in claim 5, characterized in that, The similarity matching based on perturbation patterns and historical risk energy memory identifies the type, level, and direction of anomalous events, and links these with the resilience state quantity, the cost probability cloud, the counterfactual cost channel, and the potential cost increase trend to output the express delivery cost calculation result, specifically including: Disturbance patterns are extracted by combining and expressing the environmental disturbance state data with the policy control state data; The disturbance pattern is matched with the historical risk energy memory that records the risk energy propagation trajectory to identify the event type, event level and diffusion direction of the abnormal event, and the identification result is obtained. If the identification result indicates that the event level is escalating or the spread direction points to the package path, then the first express delivery cost calculation result is output. The first express delivery cost calculation result is used to accelerate the decay of the resilience state quantity in the corresponding area and enhance the stretching and shift of the cost probability cloud to the side where the cost meets the preset threshold. At the same time, when the deviation result indicates that the deviation is amplified, the risk premium cost and the emergency allocation compensation cost are increased. If the identification result indicates that the event level has decreased or the diffusion direction has deviated from the package path, a second express delivery cost calculation result is output. The second express delivery cost calculation result is used to slow down the decay of the resilience state quantity and shrink the cost probability cloud. At the same time, when the deviation result indicates that the deviation has converged, the risk premium cost and the emergency dispatch compensation cost are reduced.

8. A real-time dynamic calculation device for express delivery costs based on big data, characterized in that, The device is used to execute the real-time dynamic calculation method for express delivery costs based on big data as described in any one of claims 1 to 7. The device includes an acquisition module and a processing module, wherein... The acquisition module is used to acquire road traffic status data, environmental disturbance status data, policy control status data, order flow distribution data, transportation capacity resource distribution data, and historical risk record data covering warehousing nodes, trunk nodes, and terminal nodes, in order to form a real-time data stream; The processing module is used to initialize the resilience state quantity of each node and dynamically adjust it when the real-time data stream is input, so that the resilience state quantity decays and spreads to adjacent nodes to form a risk energy propagation trajectory. The processing module is also used to decompose the express delivery cost into basic freight cost, risk premium cost and emergency dispatch compensation cost based on the resilience state quantity and the risk energy propagation trajectory. The processing module is also used to construct a counterfactual cost channel as a benchmark cost trajectory, compare the deviation between the benchmark cost trajectory and the actual network cost changes, and adjust the risk premium cost and the emergency allocation compensation cost based on the deviation results. The processing module is also used to generate a cost probability cloud for each package, and to shift the cost probability cloud according to the changes in the resilience state quantity and the risk energy propagation trajectory, thereby generating a potential cost increase trend. The processing module is also used to identify the type, level and direction of abnormal events based on the similarity matching between the disturbance pattern and the historical risk energy memory, and to link the resilience state quantity, the cost probability cloud, the counterfactual cost channel and the potential cost increase trend to output the express delivery cost calculation result.

9. An electronic device, characterized in that, The electronic device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions. The user interface and the network interface are both used to communicate with other devices. The processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores instructions that, when executed, perform the method as described in any one of claims 1 to 7.

Images