Search

CN-122022657-A - Cross-border logistics path optimization method and system based on multi-mode logistics information

CN122022657ACN 122022657 ACN122022657 ACN 122022657ACN-122022657-A

Abstract

The invention relates to the technical field of cross-border logistics, in particular to a cross-border logistics path optimization method and system based on multi-mode logistics information Feature coding and fusion are carried out on the multi-modal evidence package, and a feasible constraint set is constructed according to cost parameters of the generated transportation section e and the node v Determining a feasible edge set of the order; constructing a multi-objective optimization model, and solving an optimal path for minimizing an objective function; the method and the system realize the dynamic optimization adjustment of the cross-border logistics path and promote the adaptability of path planning to complex and variable scenes of the cross-border logistics by updating the multi-mode evidence package in real time and quantitatively calculating the abnormal strength in the transportation execution process.

Inventors

  • SUN TAO

Assignees

  • 青岛酒店管理职业技术学院

Dates

Publication Date
20260512
Application Date
20260211

Claims (10)

  1. 1. The cross-border logistics path optimization method based on multi-mode logistics information is characterized by comprising the following steps of: Collecting multi-mode logistics information associated with orders to form a multi-mode evidence package And based on multi-modal evidence packages Generating a consistency metric ; For multi-modal evidence packages Feature coding and fusion are carried out to obtain a unified characterization vector And according to Generating cost parameters of the transportation section e and the node v; Based on compliance rules, forbidden operation constraints, cut-off time windows and consistency metrics Constructing a feasible constraint set Determining a feasible edge set of the order; constructing a multi-objective optimization model by adopting edge selection variables Characterizing the selection condition of the transportation section e, and meeting the network connection and constraint set Solving an optimal path for minimizing an objective function; performing interpretable analysis on the optimal path, and outputting an interpretable result comprising cost decomposition, risk decomposition and time-efficient confidence indexes, wherein the time-efficient confidence indexes comprise the quantile of total time delay And conditional risk value ; Abnormal strength based on multi-modal information during transportation execution Triggering event-driven local re-planning, fixing the executed path segments and re-solving on the rest of the sub-networks to obtain updated paths.
  2. 2. The cross-border logistics path optimization method based on multi-modal logistics information as claimed in claim 1, wherein the multi-modal evidence package is based on Generating a consistency metric The method comprises the steps of carrying out unified alignment processing of space-time dimension on the acquired multi-mode logistics information, and constructing an order with time t as an index Multi-modal evidence package as a subject The multi-modal logistics information comprises a structured field, document text information, image information, space-time track information and Internet of things sensing information, and is based on the multi-modal evidence package Consistency constraint of Chinese document field and observation field, and order is calculated and generated Consistency metric at time t The method is used for quantifying the matching degree and the data credibility of the multi-source multi-mode information, and the expression of the consistency measurement is as follows: , Wherein, the The function is activated for Sigmoid, For consistency weights of the kth cargo core attribute field, A document field value for the kth cargo core attribute, The function is indicated for the event of a conflict, For the observed field value of the kth cargo core attribute, m is the index of the multimodal information conflict event, Negative correction weights for mth multimodal information conflict event.
  3. 3. The cross-border logistics path optimization method based on multi-modal logistics information as claimed in claim 1, wherein the multi-modal evidence package Performing feature encoding and fusion, including Original multi-mode information with different types and different dimensions is converted into high-dimensional feature vectors with uniform dimensions to obtain the feature vectors of each mode coding For the following Respectively calculating modal quality indexes The modal quality index Including at least the rate of absence, time lag and noise intensity of the corresponding modal information, based on Self-adaptive fusion weights of all modes are obtained through soft maximization function calculation Will be And corresponding adaptive fusion weights Weighted summation is carried out to obtain a unified characterization vector The unified characterization vector calculation formula is as follows: , , Wherein, the 、 Are all the modal indexes of the multi-modal evidence package, Is the bias term for the mth modality, For the adaptive fusion weight of the mth modality, And the weight matrix is the m-th modal quality index.
  4. 4. The cross-border logistics path optimization method based on multi-modal logistics information as claimed in claim 1, wherein the method is based on the following steps Generating cost parameters of a transportation segment E and a node V comprises collecting real-time running states of a space-time transportation network G= (V, E), wherein the space-time transportation network G= (V, E) is composed of a cross-border node set V and a transportation segment set E, and uniformly characterizing vectors Consistency metric Performing feature association with real-time running states, realizing association modeling of multiple features and cost parameters through a nonlinear mapping model, and finally performing probability distribution modeling on various cost parameters of a transportation section e and a node v respectively, characterizing the cost parameters into a parameterized form of a non-negative distribution family, and generating an order Strongly associated dynamic cost parameters.
  5. 5. The cross-border logistics path optimization method based on multi-modal logistics information as claimed in claim 1, wherein the construction of the feasible constraint set And determining a feasible edge set of the order, comprising respectively converting the compliance rules, the forbidden operation constraints and the cut time windows into constraint expressions, and constructing a feasible constraint set based on the constraint expressions For consistency measurement And feasible constraint set Performing association adaptation, and determining a feasible constraint set through a preset threshold value Dynamically adjusting, introducing a comprehensive feasibility indication function, carrying out segment-by-segment feasibility quantization judgment on each transport segment E in a time-space transport network G= (V, E), and screening out all feasible constraint sets All the transportation sections of all the constraint conditions in the system are collected through a collection and integration algorithm to form an order Dedicated order feasible edge sets.
  6. 6. The cross-border logistics path optimization method based on multi-modal logistics information according to claim 1, wherein the constructing a multi-objective optimization model comprises defining edges to select binary decision variables, constructing objective functions of the multi-objective optimization model based on cost parameter layering, converting three optimization targets of cost, risk and robust aging into a single-objective optimization form by adopting a weighted summation form, and integrating a feasible constraint set The method is communicated with a network for constraint, and is converted into formal constraint conditions for adapting to a multi-objective optimization model to form a complete multi-objective optimization model, wherein the expression of the multi-objective optimization model is as follows: , Wherein, the For the integrated objective function of the multi-objective optimization model, 、 And The weight coefficients of the cost term, the risk term and the robust aging term are respectively, As a cost term in the objective function, As a risk item in the objective function, Is a robust aging term in the objective function.
  7. 7. The cross-border logistics path optimization method based on multi-modal logistics information as claimed in claim 6, wherein the solving the optimal path for minimizing the objective function comprises pre-checking constraint conditions of the multi-objective optimization model, eliminating invalid constraints, and solving the multi-objective optimization model based on an improved genetic algorithm to obtain decision variables for minimizing the comprehensive objective function J Combining, will The transport segments of the network node are spliced according to the time-space sequence, and the optimal path from the starting node to the destination node is obtained through analysis The network connectivity constraint expression of the multi-objective optimization model is as follows: , , Wherein, the Binary decision variables are selected for the edges, which are used to characterize the selection of the transport segment e, For the transport segment set of the egress node v, For the set of transport segments flowing into node v, For a set of order-feasible nodes, For the order-taking-up node, For the order destination node.
  8. 8. The cross-border logistics path optimization method based on multi-modal logistics information as claimed in claim 1, wherein the performing the interpretable analysis on the optimal path includes performing the analysis on the optimal path Carrying out layering analysis, quantitatively calculating the cost of each transportation section and node and the total cost of the path based on the cost parameters to obtain a cost decomposition result, quantitatively calculating the single type and the comprehensive risk probability of each transportation section according to the comprehensive risk probability parameters, summarizing to obtain the total risk level of the path to form a risk decomposition result, and finally calculating the aging robustness of the optimal path to generate an aging confidence index, wherein the calculation formula of the conditional risk value in the aging confidence index is as follows: , Wherein, the In order to age the confidence level, As an auxiliary variable, a control signal is provided, The method is used for the mathematical expectation operator, And the random variable is the total time delay of the optimal path.
  9. 9. The cross-border logistics path optimization method of claim 1, wherein the fixing of the executed path segments and the re-solving on the remaining sub-networks comprises: Optimal path In the execution process, synchronously updating the multi-mode evidence package Updated multi-modal evidence package Quantitatively calculating abnormal intensity at t time ; Presetting an abnormal intensity threshold Abnormal intensity to be calculated in real time And threshold value Comparing when When the event-driven local re-planning is triggered, when When the optimal path is maintained Continuing execution; After triggering local re-planning, fixing the transport segments which are completed to be executed in the optimal path, removing the executed nodes and the transport segments, and constructing a residual sub-network comprising unexecuted nodes and unexecuted transport segments based on an order feasible edge set; And re-solving on the residual sub-network, and combining the multi-mode information updated in real time with the cost parameter to obtain an updating path adapting to the residual transport link.
  10. 10. A cross-border logistics path optimization system based on multi-modal logistics information, performed in the method of claim 1, comprising: the data acquisition module is configured to acquire multi-modal logistics information related to the order and form a multi-modal evidence package And based on multi-modal evidence packages Generating a consistency metric ; A code fusion module configured to package multi-modal evidence Feature coding and fusion are carried out to obtain a unified characterization vector And according to Generating cost parameters of the transportation section e and the node v; a feasible edge set module configured to intercept the time window and the consistency metric based on the compliance rules, the forbidden operation constraint Constructing a feasible constraint set Determining a feasible edge set of the order; a model module configured to construct a multi-objective optimization model using edge selection variables Characterizing the selection condition of the transportation section e, and meeting the network connection and constraint set Solving an optimal path for minimizing an objective function; A decomposition module configured to perform an interpretable resolution of the optimal path, and output an interpretable result including cost decomposition, risk decomposition, and an age confidence index including a quantile of a total delay And conditional risk value ; An output module configured to output abnormal intensity based on the multi-modal information during transportation execution Triggering event-driven local re-planning, fixing the executed path segments and re-solving on the rest of the sub-networks to obtain updated paths.

Description

Cross-border logistics path optimization method and system based on multi-mode logistics information Technical Field The invention relates to the technical field of cross-border logistics, in particular to a cross-border logistics path optimization method and system based on multi-mode logistics information. Background The cross-border logistics is used as a core link of international trade, the rationality of path planning directly determines the efficiency, cost and risk control level of logistics transportation, and the characteristics of wide node distribution, various transportation modes, complex supervision rules, multi-source information fragmentation and the like exist in a cross-border logistics scene, so that the path optimization needs to comprehensively consider multi-dimensional targets such as compliance, timeliness, economy, risk and the like. In the prior art, the cross-border logistics path optimization method is mostly developed based on space-time transportation networks, an optimization model is built by collecting logistics basic information and an optimal path is solved, although basic path planning requirements can be realized, the method has obvious defects in the aspects of multi-mode logistics information utilization and fusion, the method is difficult to adapt to the actual scene of cross-border logistics full-link multi-source information interaction, comprehensive and accurate data source support cannot be provided for path optimization, and the matching degree of an optimization result and the actual transportation requirements is low. In addition, the cross-border logistics path optimization method has the inherent defects of static planning, and the interpretability and the floor property of an optimization result are insufficient, namely, on one hand, the path planning is mostly one-time static solution, the path planning cannot respond to sudden abnormal events such as track deviation, sensing abnormality, port congestion and the like in real time in the transportation execution process, and a dynamic re-planning mechanism is lacked, so that the original planning path loses optimality and even feasibility in an abnormal scene, on the other hand, the optimization result is mostly only combined by a path node and a transportation section, the indexes such as cost, risk and timeliness are not subjected to refined decomposition, the marginal contribution of each transportation section to an optimization target is not quantized, a 'black box' decision result is formed, logistics operators cannot accurately identify a high cost point and a risk point in the path, and the cooperation of the path adjustment and the operation cannot be completed, the actual floor value of the path optimization result is greatly reduced, and the cross-border logistics path optimization method and the system based on multi-mode logistics information are needed at present. Disclosure of Invention The invention provides a cross-border logistics path optimization method and system based on multi-mode logistics information, which are used for solving the problems of insufficient multi-mode information utilization and poor interpretability and grounding property of an optimization result in the existing cross-border logistics path optimization method. In a first aspect, the cross-border logistics path optimization method based on multi-mode logistics information provided by the invention adopts the following technical scheme: A cross-border logistics path optimization method based on multi-mode logistics information comprises the following steps: Collecting multi-mode logistics information associated with orders to form a multi-mode evidence package And based on multi-modal evidence packagesGenerating a consistency metric; For multi-modal evidence packagesFeature coding and fusion are carried out to obtain a unified characterization vectorAnd according toGenerating cost parameters of the transportation section e and the node v; Based on compliance rules, forbidden operation constraints, cut-off time windows and consistency metrics Constructing a feasible constraint setDetermining a feasible edge set of the order; constructing a multi-objective optimization model by adopting edge selection variables Characterizing the selection condition of the transportation section e, and meeting the network connection and constraint setSolving an optimal path for minimizing an objective function; performing interpretable analysis on the optimal path, and outputting an interpretable result comprising cost decomposition, risk decomposition and time-efficient confidence indexes, wherein the time-efficient confidence indexes comprise the quantile of total time delay And conditional risk value; Abnormal strength based on multi-modal information during transportation executionTriggering event-driven local re-planning, fixing the executed path segments and re-solving on the rest of the sub-networks to obtain updated paths. Fur