CN-121998326-A - On-line management method and system for queuing transport vehicles

CN121998326ACN 121998326 ACN121998326 ACN 121998326ACN-121998326-A

Abstract

The invention provides an on-line management method and system for queuing transport vehicles, wherein the on-line management method comprises the steps of obtaining a historical clearance information set and vehicle cargo information, carrying out time sequence dependency modeling and clearance efficiency analysis on the preprocessed historical clearance information set based on a time sequence modeling method to obtain a clearance efficiency model, setting a static priority weight set for each queuing vehicle queue, setting a dynamic priority weight set based on a saturation limit decision algorithm according to the clearance efficiency model and the vehicle cargo information set, fusing entropy weights of the static priority weight set and the dynamic priority weight set into a standard priority weight set, carrying out reservation processing time analysis for each queuing vehicle queue based on the standard priority weight set and the clearance efficiency model to obtain a reservation soft time window set, extracting clearance loads of each logistics clearance opening based on the reservation soft time window set, and carrying out service constraint verification and clearance guidance for the transport vehicles based on each clearance load.

Inventors

WANG CHAOHUI
LI LIANG
YANG JIAN

Assignees

运脉云技术有限公司

Dates

Publication Date: 20260508
Application Date: 20260119

Claims (10)

1. An on-line management method for queuing transport vehicles, the method comprising: acquiring a historical clearance information set of each logistics gateway and a vehicle cargo information set of a queuing vehicle queue corresponding to each logistics gateway, wherein the vehicle cargo information set comprises cargo types, customer grades, transportation states and order data of each queuing vehicle; Performing time sequence data preprocessing on the historical clearance information set, and performing time sequence dependency modeling and clearance model training on the preprocessed historical clearance information set based on a time sequence modeling method to obtain a clearance efficiency model; Setting a static priority weight set for each queuing vehicle queue according to the vehicle cargo information set based on a static business rule, and setting a dynamic priority weight set for each queuing vehicle queue according to the gateway clearance efficiency model and the vehicle cargo information set based on a saturation limit decision algorithm; The static priority weight set and the dynamic priority weight set are fused into a standard priority weight set, reservation processing time analysis is carried out on each queuing vehicle queue based on the standard priority weight set and the gateway clearance efficiency model, and a reservation soft time window set is obtained; and extracting clearance loads of all logistics gateways based on the reserved soft time window set, and carrying out business constraint verification and gateway guidance for newly arrived transport vehicles according to all the clearance loads.
2. An on-line management method of queuing transport vehicles according to claim 1, wherein said time-series data preprocessing of said historical clearance information set comprises: performing field mapping and structuring processing on the historical clearance information set to obtain a standard clearance information set; Carrying out missing value identification and abnormal value screening on the standard clearance information set, and carrying out smooth filling on the identified missing value and the missing position generated after the abnormal value screening to obtain a smooth clearance information set; Unifying time axes of the smooth clearance information sets, and performing time sequence arrangement on the smooth clearance information sets after unifying the time axes according to preset time granularity to obtain clearance data sequence sets; and carrying out efficiency statistics and feature labeling on each clearance data in the clearance data sequence set to obtain a clearance feature sequence set, and carrying out feature normalization on the clearance feature sequence set to obtain a standard clearance feature sequence set.
3. The method for on-line management of queuing of transportation vehicles according to claim 2, wherein the performing time-sequence dependent modeling and clearance model training on the preprocessed historical clearance information set based on the time-sequence modeling method to obtain a clearance efficiency model of a gateway comprises: Respectively extracting a static clearance characteristic sequence set, a known clearance characteristic sequence set and an observation clearance characteristic sequence set from the standard clearance characteristic sequence set by using a preset clearance model; Weighting and fusing the static clearance feature sequence set, the known clearance feature sequence set and the observation clearance feature sequence set by using a gating and fusing network of the clearance model to obtain a fused clearance feature sequence set; Carrying out multi-layer causal expansion convolution on the fusion clearance characteristic sequence set to obtain a local clearance time sequence characteristic sequence set; Performing low-rank approximate attention weighting on the local clearance time sequence feature sequence set based on a low-rank projection mechanism to obtain a global clearance time sequence feature sequence set; Taking the static clearance characteristic sequence set as a condition, and embedding the static clearance characteristic sequence set into the global clearance time sequence characteristic sequence set by using a gating mechanism to obtain a standard input characteristic sequence set; and training the clearance model according to the standard input characteristic sequence set and the standard clearance characteristic sequence set to obtain a clearance efficiency model of the gateway.
4. An on-line management method for queuing transport vehicles according to claim 3, wherein said low-rank approximate attention weighting is performed on said local clearance time series feature sequence set based on a low-rank projection mechanism to obtain a global clearance time series feature sequence set, comprising: carrying out layer normalization on each local clearance time sequence characteristic sequence in the local clearance time sequence characteristic sequence set to obtain a standard local clearance time sequence characteristic sequence; Performing attention characteristic projection on the standard local clearance time sequence characteristic sequence to obtain a query vector sequence, a key vector sequence and a value vector sequence; Performing dimension reduction projection on the key vector sequence and the value vector sequence based on a preset low-rank projection matrix to obtain a low-rank key vector sequence and a low-rank value vector sequence; calculating the attention weight between the query vector sequence and the low-rank key vector sequence by using a low-rank approximate attention algorithm to obtain a low-rank approximate attention weight matrix; The low-rank approximate attention weight matrix is utilized to carry out weighted summation on the low-rank value vector sequence, and a global dependency vector sequence is obtained; And carrying out output projection on the global dependency vector sequence, carrying out residual connection on the global dependency vector sequence subjected to output projection and the local clearance time sequence characteristic sequence to obtain a global clearance time sequence characteristic sequence, and collecting all the global clearance time sequence characteristic sequences into a global clearance time sequence characteristic sequence set.
5. An on-line management method for queuing transport vehicles according to claim 3, wherein said training said clearance model based on said standard input feature sequence set and said standard clearance feature sequence set to obtain a gateway clearance efficiency model comprises: extracting a clearance efficiency characteristic sequence set from the standard clearance characteristic sequence set; Sliding screening is carried out on the standard input characteristic sequence set according to a preset time period window, so that a periodic input characteristic sequence set formed by a plurality of periodic input characteristic sequences with the same length is obtained; The clearance efficiency features after the last time step of the input feature sequence of each period in the clearance efficiency feature sequence set are selected one by one to be used as hysteresis efficiency features, and a hysteresis efficiency feature set is obtained; And taking the periodic input characteristic sequence set as input, and taking the hysteresis efficiency characteristic set as a label to carry out supervision training on the clearance model to obtain a clearance efficiency model of the gateway.
6. An on-line management method for queuing transport vehicles according to claim 1, wherein said setting a static priority weight set for each queuing vehicle queue based on said set of vehicle cargo information based on static business rules comprises: Selecting transport vehicles in each queuing vehicle queue as target transport vehicles, taking vehicle cargo information corresponding to the target transport vehicles in the vehicle cargo information set as target vehicle cargo information, and extracting a static service weight item set from the target vehicle cargo information; judging whether absolute service priority items exist in the static service weight item set; if yes, setting static priority weights for the target transport vehicles according to the absolute service priority items; If not, carrying out weight score statistics on each cargo in the target transport vehicle according to the static service weight item set to obtain a weight score set, and setting static priority weights for the target transport vehicle according to the average value of the weight score set.
7. An on-line management method of transportation vehicle queuing as claimed in claim 1, wherein said saturation-limit-based decision algorithm sets a dynamic priority weight set for each queuing vehicle queue in accordance with said gate clearance efficiency model and said vehicle cargo information set, comprising: performing multi-step rolling efficiency analysis on each queuing vehicle queue according to the vehicle cargo information set by using the gateway clearance efficiency model to obtain a queuing clearance efficiency sequence set; Performing waiting time analysis on each queuing vehicle queue based on the queuing clearance efficiency sequence set to obtain a primary waiting time interval sequence set; Extracting the transportation state and order data corresponding to each queuing vehicle queue from the vehicle cargo information set to obtain a transportation state sequence set and an order data sequence set; Constructing the transportation state sequence set, the order data sequence set and the primary waiting time interval sequence set into a vehicle time sensitive characteristic sequence set, and carrying out priority regression decision on the vehicle time sensitive characteristic sequence set to obtain a dynamic priority score set; And carrying out saturation function mapping and change rate threshold limiting on the dynamic priority score set, and carrying out weight normalization operation on the dynamic priority score set subjected to change rate limiting to obtain a dynamic priority weight set.
8. The method of on-line management of transport vehicle queuing as claimed in claim 7, wherein said fusing said static priority weight set and said dynamic priority weight set entropy weights into a standard priority weight set, comprises: carrying out weight alignment on the static priority weight set and the dynamic priority weight set as column vectors based on the transport vehicle IDs corresponding to the queuing vehicle queues to obtain a weight evaluation matrix; performing minimum-maximum normalization operation on the two column vectors of the weight evaluation matrix to obtain a normalized weight matrix; Carrying out information entropy calculation on the two column vectors of the normalized weight matrix to obtain a static weight entropy value and a dynamic weight entropy value; Respectively calculating a static difference coefficient and a dynamic difference coefficient based on the static weight entropy value and the dynamic weight entropy value, and carrying out normalization processing on the static difference coefficient and the dynamic difference coefficient to obtain a static fusion weight and a dynamic fusion weight; and calculating a standard priority weight set based on the static fusion weight, the dynamic fusion weight, the static priority weight set and the dynamic priority weight set.
9. The method for on-line management of transport vehicle queuing as claimed in claim 1, wherein said performing a reservation processing time analysis for each of the queuing vehicle queues based on said set of standard priority weights and said gateway clearance efficiency model to obtain a set of reserved soft time windows comprises: the priority ranking is carried out on each queuing vehicle queue based on the standard priority weight set, and a priority vehicle queue set is obtained; performing multi-step rolling efficiency analysis on the priority vehicle queue set according to the vehicle cargo information set by using the gateway clearance efficiency model to obtain a standard clearance efficiency sequence set; performing waiting time analysis for the priority vehicle queue set based on the standard clearance efficiency sequence set to obtain a standard waiting time interval sequence set; performing elastic time window setting on the standard waiting time interval sequence set based on the standard priority weight set to obtain an elastic window sequence set; and carrying out window combination and window conflict verification on the standard waiting time interval sequence set and the elastic window sequence set to obtain a reserved soft time window set.
10. An on-line management system for queuing transport vehicles, which is characterized by comprising an information acquisition module, a clearance analysis module, a priority setting module, a reservation analysis module and a gateway guide module, wherein: The system comprises an information acquisition module, a data processing module and a data processing module, wherein the information acquisition module acquires a historical clearance information set of each logistics gateway and a vehicle cargo information set of a queuing vehicle queue corresponding to each logistics gateway, and the vehicle cargo information set comprises cargo types, client grades, transportation states and order data of each queuing vehicle; The clearance analysis module is used for preprocessing the time sequence data of the history clearance information set, and carrying out time sequence dependency modeling and clearance model training on the preprocessed history clearance information set based on a time sequence modeling method to obtain a clearance efficiency model; The priority setting module is used for setting a static priority weight set for each queuing vehicle queue according to the vehicle cargo information set based on a static business rule, and setting a dynamic priority weight set for each queuing vehicle queue according to the gateway clearance efficiency model and the vehicle cargo information set based on a saturation limit decision algorithm; The reservation analysis module is used for fusing the static priority weight set and the dynamic priority weight set entropy weight into a standard priority weight set, and carrying out reservation processing time analysis on each queuing vehicle queue based on the standard priority weight set and the gateway clearance efficiency model to obtain a reservation soft time window set; And the gateway guiding module extracts clearance loads of all logistics gateways based on the reserved soft time window set, and performs business constraint verification and gateway guiding for newly arrived transport vehicles according to the clearance loads.

Description

On-line management method and system for queuing transport vehicles Technical Field The invention relates to the technical field of logistics scheduling, in particular to an on-line management method and system for queuing transport vehicles. Background In large logistics hubs, cross-border ports and multi-channel storage parks, transportation vehicles need to finish inspection, clearance, loading and unloading and other operations in sequence at designated logistics gateways, and as logistics flow increases and ageing requirements are improved, problems of vehicle queuing congestion, uneven resource utilization, high-priority goods delay and the like before the gateways are increasingly highlighted, and the realization of intelligent and efficient on-line queuing management becomes a key for improving the throughput capacity and service level of the logistics hubs. However, the existing vehicle queuing management methods still have significant technical limitations when dealing with dynamic, complex actual operation scenarios. Firstly, most systems rely on first-come or static priority based on fixed rules to schedule, and cannot respond to fluctuation of gateway processing efficiency and dynamic change of vehicle states in real time, so that a scheduling strategy is stiff, and overall clearance efficiency is low. And secondly, the accurate prediction capability of the future processing capability of the gateway is lacking, and the estimated queuing waiting time is seriously misaligned, so that the vehicle reservation and scheduling lack a reliable data base, and the schedule is frequently disabled. Finally, the existing scheme is concentrated on queue optimization in a single gateway, lacks an intelligent guiding mechanism for realizing dynamic load balancing among a plurality of parallel gateways, cannot avoid uneven distribution of queue resources from a system level, and limits the upper limit of utilization of overall resources. Disclosure of Invention (One) solving the technical problems Aiming at the defects of the prior art, the invention provides an on-line management method and system for queuing transport vehicles, which have the advantages of time sequence prediction driving decision, multi-stage weight dynamic fusion and global load intelligent guidance, and solve the problems of long average waiting time of vehicles and low overall throughput efficiency of gateway clusters caused by static stiffness of scheduling rules, difficult prediction of processing efficiency and unbalanced multi-gateway loads of vehicles in a logistics gateway queuing management scene. (II) technical scheme In order to achieve the above purpose, the present invention provides the following technical solutions: The invention provides an on-line management method for queuing transport vehicles, which comprises the following steps: acquiring a historical clearance information set of each logistics gateway and a vehicle cargo information set of a queuing vehicle queue corresponding to each logistics gateway, wherein the vehicle cargo information set comprises cargo types, customer grades, transportation states and order data of each queuing vehicle; Performing time sequence data preprocessing on the historical clearance information set, and performing time sequence dependency modeling and clearance model training on the preprocessed historical clearance information set based on a time sequence modeling method to obtain a clearance efficiency model; Setting a static priority weight set for each queuing vehicle queue according to the vehicle cargo information set based on a static business rule, and setting a dynamic priority weight set for each queuing vehicle queue according to the gateway clearance efficiency model and the vehicle cargo information set based on a saturation limit decision algorithm; The static priority weight set and the dynamic priority weight set are fused into a standard priority weight set, reservation processing time analysis is carried out on each queuing vehicle queue based on the standard priority weight set and the gateway clearance efficiency model, and a reservation soft time window set is obtained; and extracting clearance loads of all logistics gateways based on the reserved soft time window set, and carrying out business constraint verification and gateway guidance for newly arrived transport vehicles according to all the clearance loads. According to one preferred embodiment of the present invention, the performing time-series data preprocessing on the historical clearance information set includes: performing field mapping and structuring processing on the historical clearance information set to obtain a standard clearance information set; Carrying out missing value identification and abnormal value screening on the standard clearance information set, and carrying out smooth filling on the identified missing value and the missing position generated after the abnormal value screening to obta